ecology of coral reefs results of a Workshop on Coral Reef Ecology held by the American Society of Zoologists, Philadelphia, Pennsylvania, December 1983

[From the U.S. Government Printing Office, www.gpo.gov]

SYMPOSIA SERIES FOR UNDERSEA RESEARCH NOAA'S UNDERSEA RESEARCH PROGRAM
VOL. 3 NO. 1, 1985
Coasaal �~ne
introreat:cn

: The Ecology

of Coral Reefs

U. S. DEPARTMENT OF COMMERCE NOAA
COASTAL SEREVICES CENTER
2234 SOUTH HOBSON AVENUE ---,
CHARLEStON, SC 29405-24137 '

~i~i
- '

US DEPARTMENT
OF COMMERCE -
Nationai Oceanic and
Atmospheric Administration

-I '-a.--~j

SYMPOSIA SERIES FOR UNDERSEA RESEARCH
NOAA'S UNDERSEA RESEARCH PROGRAM, VOL. 3 NO. 1, 1985

The Ecology

of Coral Reefs

Results of a Workshop on Coral Reef
Ecology held by the American Society
of Zoologists, Philadelphia, Pennsylvania,
December 1983

Edited by
Marjorie L. Reaka
Department of Zoology
University of Maryland
College Park, Maryland 20742
Washington, D.C.
September 1985

U.S. DEPARTMENT OF COMMERCE
Malcolm Baldrige, Secretary
National Oceanic and Atmospheric Administration
Anthony J. Calio, Deputy Administrator
Oceanic and Atmospheric Research
Ned A. Ostenso, Assistant Administrator
Office of Undersea Research
Elliott Finkle, Director

Symposia Series for Undersea Research

The National Oceanic and Atmospheric Administration's (NOAA) Office Of
Undersea Research provides manned undersea facilities and other research support
for investigations of aquatic environments in the areas of biological, geological,
and ecological research. There are currently five national undersea research
facilities which operate under cooperative agreements between universities and
NOAA. These facilities are located at: the West Indies Laboratory of Fairleigh
Dickinson University, the University of North Carolina at Wilmington, the
University of Connecticut at Avery Point, the University of Southern California,
and the University of Hawaii.

NOAA's Office of Undersea Research provides facilities for scientists to
conduct research supporting NOAA's mission objectives in the areas of: marine
fisheries, marine pollution, sea floor properties and processes, and ocean
services.

Since its inception, NOAA also has encouraged and supported the use of
submersibles to perform in-situ underwater observations and data gathering.
Many shallow-water submersible missions have been supported, including the use
of the Johnson Sea-Link, the Nekton-Ganmma, and the Mermaid, as well as deep-water
missions using the Alvin. The goals of the submersible programs are to support
the research requirements of NOAA's major program elements and its Sea GrantW
College system.

This Symposia Series for Undersea Research has been developed specifically
to provide a pu Ishi~ng me~Tim for national symposia whose contents have addressed
topics related to undersea research activities. Additional information concerning
this series and other activities of NOAA's Office of Undersea Research may be
obtained by contacting:

Director
NOAA's Office of Undersea Research
Mail Stop R/SE2
6010 Executive Boulevard, Roomi 805
Rockville, MD 20852

Photo Credits: Color photographs are from Bruce Nyden, Marjorie Reaka, Phillip
Lobel, or the Office of Undersea Research.

TABLE OF CONTENTS

Chapter I. INTRODUCTION. Marjorie L. Reaka ............................ 1

Chapter II. RESULTS OF A WORKSHOP ON CORAL REEF ECOLOGY: SYNTHESES
OF DISCUSSION SECTIONS ......................................7

Overview: The Dynamics of Recruitment in Coral Reef
Organisms. William N. McFarland ..........................9

Growth and Life History Patterns of Coral Reef Organisms:
A Discussion Group Summary and Overview. Ronald H.
Karlson ................................................. 17

Overview: Coral Reef Community Structure and Function.
Mark A. Hixon ... .......................................... 27

Some Thoughts on the Past, Present, and Future of Studies on
Coral Reef Community Metabolism. S. V. Smith ............. 33

Chapter III. CONTRIBUTED PAPERS .......................................... 35

Recruitment of Young Coral Reef Fishes From the Plankton.
William N. McFarland and John C. Ogden .................... 37

Pelagic Duration, Dispersal, and the Distribution of Indo-
Pacific Coral-Reef Fishes. Edward B. Brothers and
Ronald E. Thresher ........................................ 53

Diurnal Periodicity of Spawning Activity by the Hamlet Fish,
Hypoplectrus guttavarius (Serranidae). Phillip S.
Lobel and Steve Neudecker ................................. 71

Patterns of Reproduction in Small Jamaican Brittle Stars:
Fission and Brooding Predominate. Ronald H. Emson,
Philip V. Mladenov and lain C. Wilkie ..................... 87

Reef Zooplankton Collected Along a Depth Gradient at
Discovery Bay, Jamaica. Sharon L. Ohlhorst ............... 101

Temporal Patterns of Zooplankton Migration. Sharon L.
Ohlhorst and W. David Liddell ............................. 117

Field Analysis of the Dominance Hierarchy of the Bicolor
Damselfish Stegastes partitus (Poey) (Pisces: Poma-
centridae). Yvonne Sa3ovy .............................. 129

A Comparison of Fore and Back Reef Populations of Diadema
antillarum Philippi and Eupomacentrus planifrons
Cuvier at St. Croix, U.S. Virgin Islands. C. Lance
Robinson and Ann Houston Williams ......................... 139

Carbonate Sediment Production by the Rock-Boring Urchin
Echinometra lucunter and Associated Endolithic
Infauna at Black Rock, Little Bahama Bank. Charles
M. Hoskin and John K. Reed ................................ 151

Highly Productive Eastern Caribbean Reefs: Synergistic
Effects of Biological, Chemical, Physical, and
Geological Factors. Walter H. Adey and Robert S.
Steneck ................................................... 163

Studies on the Bio-Optics of Coral Reefs. Phillip Dustan ... 189

Preliminary Studies of Denitrification on a Coral Reef.
Sybil P. Seitzinger and Christopher F. D'Elia ............. 199

CHAPTER I: INTRODUCTION

Marjorie L. Reaka
Department of Zoology
The Universit of Maryland
College Park, Maryland 20742

This volume presents the results of a workshop on the ecology of coral
reefs that was held at the American Society of Zoologists meetings in Philadel-
phia, December 1983. The workshop, which included four symposium presentations,
four discussion sections, and presentations of undersea research facilities by
NOAA's Office of Undersea Research, was sponsored by the Division of Ecology of
the ASZ. The four symposium talks and discussion sections addressed the topics
of growth and life history patterns, recruitment, community organization, and
community metabolism in coral reef systems. The workshop was unusually well
attended (attendance in the Ecology Division of the ASZ was 30% higher than in
previous years and two additional contributed paper sessions in the Division
of Ecology were devoted specifically to coral reef ecology), and vigorous group
discussions continued well into the night. The objective of the workshop was
to focus attention on what we do and do not know about the ecology of coral
reefs, particularly along depth gradients, and to assess the most important
directions for future research on coral reefs.

A previous volume in this series (Reaka, 1983), containing symposium
presentations and contributed papers in each of the four topical areas, was
made available at the Coral Reef Workshop. The present volume presents the
main conclusions derived from the discussion sections (Chapter II) and additional
contributed papers on the ecology of coral reefs (Chapter III).

In Chapter II, the discussion leaders summarize the topics discussed and
the conclusions reached in each discussion section. William McFarland points
out that astonishingly little is known about recruitment in coral reef species.
He identifies some of the reasons for our lack of knowledge in this area, not
the least of which is the prevalence of planktonic larvae in reef organisms.
The production of hundreds to millions of tiny dispersing planktotrophic larvae
distinguishes marine animals from those in every other major environment.
Furthermore, the incidence of species with long-lived planktotrophic larvae is
inversely related to latitude (Thorson, 1950; Mileikovsky, 1971), so that, in
contrast to marine communities from high latitudes, the vast majority of coral
reef species produce planktonic dispersing young. This phenomenon has profound
implications for the population dynamics, community organization, and even
evolutionary patterns of marine species compared to those in terrestrial or
fresh water environments. In marine species thousands of recruits may occur in
local areas in some years but not in others; planktonic larvae may settle in
their parent population, may be carried to distant habitable sites, or even may
be swept out to sea without any favorable substrate having been encountered.
McFarland discusses several mechanisms that enhance the predictability of local
recruitment on coral reefs, including the timing and dynamics of reproduction,
currents or eddies that may return larvae to their parent populations, and
behavioral patterns of the larvae and settling juveniles themselves. These

mechanisms, as well as fine scale observations of the larvae of species from
different reef habitats, deserve extensive research before we will understand
this critically important aspect of the biology of coral reef organisms.

Ronald Karlson summarizes the conclusions of his discussion section on a
related topic, life history patterns. An emerging body of research demonstrates
that the observed growth rates, form, and longevity of corals along depth
gradients result from complex interactions between physical (e.g., light and
wave disturbance) and biotic variables. Recent research shows that sublethal
phenomena such as injuries due to physical and biotic agents have significant
implications for growth and reproduction in many sessile reef organisms. These
processes can vary significantly in different habitats. This discussion group
also addressed the important topic of ecological and evolutionary patterns in
solitary vs. colonial organisms on coral reefs. The importance of fission,
fusion, and fragmentation (and thus of colony size) in both solitary and colonial
species received particular attention as a productive avenue for future research.

Mark Hixon lead the discussion section on community structure and function.
Coral reef ecologists have been in the forefront of the debate on whether
communities are organized by stable predictable processes or whether they
merely represent independent populations in shifting states of disequilibria.
One of the reasons that disequilibrium and stochastic processes have been so
apparent in coral reef communities may be because dispersing, feeding larvae
engender unpredictable recruitment. However, recent research also has demon-
strated the importance of irregular catastrophic perturbations, such as storms
(e.g., Woodley, et al., 1981), population outbreaks (e.g., Birkeland, 1982), or
diseases (e.g., Lessios, et al., 1984) in the organization of coral reef communi-
ties. One of these events-, th-e mass mortality of Caribbean sea urchin populations,W
was ongoing at the time of the workshop and was a major topic of discussion.
Although the causes and mechanisms of this outbreak still are not well under-
stood, Mixon identifies the heuristic value of this "natural experiment" which
has removed a major herbivore over large local and geographical areas of reefs.
This discussion lead to an evaluation of a more deterministic process, herbivory
by urchins vs. fishes, on reefs; this discussion centered on how human activities
such as fishing in certain well studied localities can affect our interpretations
of the past and present organization of reef communities. Although terrestrial
and fresh water ecologists have been concerned about the effects of human inter-
vention on natural communities for some time, the pervasiveness of this problem
in coral reef communities has attracted attention only recently. Hixon argues
that one of the most critical needs for future research is for experimental
manipulations and long term comparative studies.

The discussion section on community metabolism was led by Stephen Smith.
Smith shows how this field has developed from a focus on oxygen and carbon flux
to studies of the physical, chemical, and biological variables that control
metabolism of the reef ecosystem. The discussion addressed how the quantity and
quality of available nutrients affect productivity, how nutrient flux between sub-
communities can affect nutrient balance of the overall reef ecosystem, and how new
technological approaches can improve our ability to measure community metabolism.
Smith stresses that a holistic approach, including comparisons with other ecosys-
tems, will represent important avenues for future research in this area.

Chapter III includes contributed papers whose topics span those covered
Win the four discussion groups. McFarland and Ogden provide a comprehensive
review of recruitment in coral reef fishes and evaluate the various hypotheses
advanced to account for the presence of planktonic larvae. They conclude that
both reduced predation on young and enhanced dispersal favor the development of
long-lived larvae which migrate offshore from local reefs. Due to the hetero-
geneity in patterns of larval development, however, these authors note that the
relative importance of these selective factors must vary among different species
of reef fishes. Brothers and Thresher evaluate the relationship between larval
durations and the extent of the geographic distributions in 115 species of ree'
fishes. Their study yields the new and interesting result that, if larval
durations are < about 45 days, the size of the geographic range (usually an
indication of dispersal patterns) is not correlated with length of larval life.
Larvae which have pelagic phases longer than 45 days, however, all have broad
distributions and can transgress oceanic barriers at least occasionally. This
work, then, has important implications for patterns of colonization and biogeog-
raphy. Lobel and Neudecker analyze another aspect of reproductive biology of
reef fishes that can influence dispersal of larvae on reefs: the timing and
location of spawning. They describe and compare courtship in several species
of hamlets. These fishes spawn at dusk over specific high reef structures,
which may reduce predation and facilitate dispersal of eggs. The time and
duration of spawning is influenced by depth and/or lunar phase as well as
simulated attacks by predators. Emson, Mladenov, and Wilkie provide information
on a different aspect of reproductive biology in a reef invertebrate. Various
modes of asexual reproduction, including fission, are well known in a number of
. relatively sessile invertebrate phyla. However, these authors demonstrate that
fission represents the major method of reproduction in many small species of
motile brittle stars associated with algae and sea grass in lagoonal habitats.
Many of these ophiuroids also apparently produce dispersing planktonic larvae,
an unexpected phenomenon because small size usually is associated with brooding.
These results again have implications for patterns of recruitment and maintenance
of local populations.

Both meroplankton (the larvae of benthic reef organisms) and holoplankton
(organisms that spend their entire life in the plankton) exhibit diel vertical
migrations that are likely to influence their tendency to disperse in currents
and their susceptibility to predation. The long term study reported by Ohlhorst
also suggests that other factors, such as preceding rain and lunar effects,I
can influence abundance of zooplankton above the reef. Ohlhorst further demon-
strates that numbers of zooplankton decline significantly with increased depth,
and that even in the shallow areas these plankton can provide relatively little
of the metabolic requirements for sessile reef organisms. Using frequent
collections throughout the day and night, Ohlhorst and Liddell provide the
first precise documentation of a predawn surge in abundance of plankton over
the reef that complements the postsunset peak of emergence known for these
organisms. A swarming holoplanktonic copepod usually W~as the most common
zooplankter in both Jamaica and St. Croix, althoukgh other taxa, including
meroplankters, became prominent during the hours around sunset. These papers
therefore are related to the topics of dispersal and reproductive biology
discussed above but also bear on community organization and energy flow among
different components of the reef community.

The following three papers in the volume deal with damselfishes and/or
urchins on reefs. The territorial activities of damselfishes are known to be I
significant for the structure and productivity of reef communities (e.g.,
Brawley and Adey, 1977; Hixon, 1983; and others). Sadovy studied the detailed
dominance relations among individuals of bicolored damselfish in the field.
Her results showed a linear, size-dependent dominance hierarchy in these fish
and revealed some interesting similarities and differences between field and
otherwise comparable laboratory studies. Robinson and Williams compare population
characteristics of threespot damselfish and urchins (Diadema) in shallow back
reef and fore reef environments. They found that urchins are larger, occur in
lower densities, and are found higher in coral branches on the back reef than
on the fore reef; in contrast, damselfish live at higher densities, have larger
lawns, and exhibit less aggression toward urchins on the fore reef than on the
back reef. Urchins may remain small and close to the substrate on the shallow
fore reef because high wave action inhibits their grazing; thus interference
between urchins and damselfish is less strong when mediated by disturbance on
the fore reef than in more benign conditions on the back reef. Examining
another species of urchin with a different but very significant impact on the
structure of reefs, Hoskin and Reed use carefully controlled experimental
methods to estimate the rates of bioerosion of carbonate substrates by urchins
(Echinometra) and other burrowing infauna. These urchins excavate approximately
8.9 g substrate/m2/day (9 tons/year for the population), while the remaining
infauna produces about 5 g sediment/m2/day (6.5 tons/year). This study illus-
trates the tremendous amount of sediment that is produced and transported in
benthic reef communities. Scouring and burial by this sediment must represent
one of the most important agents influencing the structure and dynamics of reef
communities, yet the significance of these processes for community organization
are almost unstudied. In addition, this study provides new data on the sources
and rates of degradation of reef substrates, a significant aspect of long term
morphogenesis and diagenesis of reefs. This paper therefore provides a link
between the papers on community organization and those on community metabolism.

Adey and Steneck show how reef productivity is related to substrate
complexity and surface area, light and water flow, and abundance of turf algae.
They demonstrate that all of these factors can be understood in the context of
geological development of the reefs. Using upstream/downstream flow respirometry
techniques on reefs in three different stages of geological development in St.
Croix, these authors provide good evidence for some of the highest rates of
primary productivity reported for any marine environment. Dustan explores a
different method of assessing primary productivity. He provides initial data
on the spectral intensities of light impinging on, absorbed, and reflected by
different substrates. Dustan suggests that much can be learned about the photo-
biology of reef organisms using these nondestructive bio-optical sampling
techniques and indicates that we may be able to estimate primary productivity
over vast tracts of previously inaccessible reefs if optical signals from these
reefs can be detected from orbiting spacecraft. Although we are in the initial
stages of interpreting these data, such new technological approaches offer the
promise of measuring and understanding productivity on a global scale. Seitzinger
and D'Elia present data on another new and potentially important aspect of
community metabolism on reefs. While nitrogen fixation on reefs has received
well deserved attention, Seitzinger and D'Elia show that some reef habitats, in
particular dead coral heads and sediments, may be the sites of extensive denitri-

4
4

ERRATA
Several typographical errors were inadvertently included in our publication
of The Ecology of Deep and Shallow Coral Reefs [Symp. Ser. Undersea Res., Vol.
1(1T 71983], anwe wouTlike totake this opportunity to correct them.

For Chornesky, E. A., and S. L. Williams, "Distribution of Sweeper Tentacles
on Montastrea cavernosa, p. 61, par. 3, line 15 should read "or after contact
and recognition;" p. 63, caption for figure 1 should read "For M. annularis,
change in number of polyps."

For Wolf, N. G., E. B. Bermingham, and M. L. Reaka, "Relationships Between
Fishes and Mobile Benthic Invertebrates on Coral Reefs," p. 69, the address of
E. B. Bermingham should read "Department of Molecular and Population Genetics;"
p. 72, line 3 should read "Although fewer planktivores were recorded on C than
on A reefs."

For Hay, M. E., and T. Goertemiller, "Between-Habitat Differences in
Herbivore Impact on Caribbean Coral Reefs," p. 100, par. 4, line 3 should read
"and 12.8% versus 1% on Lighthouse (fig. 1);" p. 101, par. 2, lines 10-11 should
read "or motility in sea urchin sperm (Norris and Fenical 1982, Paul and Fenical
1983). The polyphenolic compounds produced by Turbinaria."

For Szmant-Froelich, A., "Functional Aspects of Nutrient Cycling on Coral
Reefs," p. 136, Table 1, DIN should be TDN.

fication, whereby NO2 and N03 are lost from the system as atmospheric N2. Such
studies make important contributions to Smith's goal of understanding the
overall balance of nutrient flow within and between major components of reef
communities.

Many people have contributed to the successful completion of this volume.
Thanks are especially extended to the authors for their fine contributions and
cooperation throughout the editing process. In addition, the reviewers of
these papers are gratefully acknowledged for contributing their time and
criticisms, all of which improved the volume. Completion of this symposium
volume would have been impossible without the expert typing and editorial
assistance of Ms. Marcia Collie, Staff Assistant from NOAA's Office of Undersea
Research; her cooperation, advice, and assistance through all the stages of
publication are acknowledged with heartfelt thanks. Similarly, publication was
made possible through NOAA's Office of Undersea Research, and the support of
Mr. William Busch and Mr. Elliott Finkle throughout this endeavor are sincerely
appreciated. Lastly, I thank my husband, Stephen, and our two little blonde
boys, Alexei and Erik, for their sunshine and enthusiasm.

LITERATURE CITED

Birkeland, C. 1982. "Terrestrial Runoff as a Cause of Outbreaks of Acanthaster
planci (Echinodermata: Asteroidea)." Mar. Biol. 69:175-185.

Brawley, S. H., and W. H. Adey. 1977. "Territorial Behavior of Threespot
Damselfish (Eupomacentrus planifrons) Increases Reef Algal Biomass and
Productivity." Environ. Biol. Fish. 2:45-51.

Hixon, M. A. 1983. "Fish Grazing and Community Structure of Reef Corals and
Algae: A Synthesis of Recent Studies," pp. 79-87. In: M. L. Reaka (ed.),
The EcologY of Deep and Shallow Coral Reefs. Symp. Ser. Undersea Res., Vol.
T7). NOAA Undersea Res. Progr., Rockville, Md.

Lessios, H. A., D. R. Robertson, and J. D. Cubit. 1984. "Spread of Diadema
Mass Mortality Through the Caribbean." Science 226:335-337.

Mileikovsky, S. A. 1971. "Types of Larval Development in Marine Bottom
Invertebrates, Their Distribution and Ecological Significance: A Re-
Evaluation." Mar. Biol. 10:390-404.

Reaka, M. L., ed. 1983. The Ecology of Deep and Shallow Coral Reefs. Symp.
Ser. Undersea Res., VolI(1). NOAUnde-rsea Res. Progr., Rockville, Md.
149 pp.

Thorson, G. 1950. "Reproductive and Larval Ecology of Marine Bottom
Invertebrates." Biol. Rev. 25:1-45.

Woodley, J. D., E. A. Chornesky, P. A. Clifford, J. B. C. Jackson, L. S. Kaufman,
N. Knowlton, J. C. Lang, M. P. Pearson, J. W. Porter, M. C. Rooney, K. W.
Rylaarsdam, V. J. Tunnicliffe, C. M. Wahle, J. L. Wulff, A. S. G. Curtis, M.
D. Dallmeyer, B. P. Jupp, M. A. R. Koehl, J. Neigel, and E. M. Sides. 1981.
"Hurricane Allen's Impact on Jamaican Coral Reefs." Science 214:749-755.

CHAPTER II: RESULTS OF A WORKSHOP ON CORAL REEF ECOLOGY: SYNTHESES

ion OF DISCUSSION SECTIONS

les

d I
1 1 Figure 1

Coral reefs have been tremendously important in the history of human
civilizations. Reefs support fisheries and provide protected embayments that
are essential for food and commerce in coastal cultures. It is critical that
we understand the biological dynamics responsible for the formation and
degradation of these systems in order to preserve and manage their resources.
NOAA's underwater habitat, HYDROLAB (fig. 1), has (continued on next page)

Figure 2 Figure 3
7

iN"

Figure 5

Figure 4

allowed scientists to conduct experimental field studies on reproduction,
recruitment, population dynamics, and community organization of coral reef
communities that would have been impossible without the extended diving time
made possible through saturation techniques. Here divers study the factors
that influence recruitment of fishes and invertebrates (figs. 2, 3, 4), many
of them important fisheries and food stocks, on artificial reefs where large
predators and grazers are excluded or allowed access. These facilities also
have allowed extended study of the life histories, growth, and species inter-
actions of the major structural agents on reefs, corals and sponges (fig. 5).
The AEGIR, NOAA's mobile underwater habitat that will replace HYDROLAB in 1986,
also is shown here (fig. 6).

Figure 6
8

discussion section that spawning is a variable act within and between species
a nd an activity that involves considerable behavioral plasticity. If the
function of casting eggs into the plankton is to rapidly remove them from reef
predators, then doing so at dusk (as many fish do) would seem safest. The fact
that many fish do not spawn at dusk may relate to the presence of favorable
currents that flush the eggs rapidly offshore and out of the reach of abundant
small reef predators. Unfortunately, few papers deal simultaneously with
spawning and local currents (McFarland and Ogden, 1985). As a result, it is
rather difficult to resolve the differences that exist between species (see
Robertson, 1983, for spawning variation in several acanthurids).

Although recruitment from the plankton occurs at the end of the larval
stage, knowing how and when potential propagules are introduced into the plankton
is essential for understanding recruitment because the spawning act initiates
the recruitment process. The difficulties identified above-could be partly
resolved by an integrated study in a local area of the seasonal, monthly, and
diel spawnings of many different species of coral reef fishes. Evaluating the
function~s) of differential spawning amongst the various species will require
that the observations be tightly coupled with analysis of local currents (for
example, Robertson, 1983). If this is accomplished in severa~l different areas,
it should be possible to sort out the trade-offs that different species invoke
in order to "safely" deliver young to the plankton.

Once spawned, the eggs of coral reef fishes usually hatch within a day
(Thresher, 1984). Although most coral reef fish eggs and larvae are flushed
offshore, in some circumstances the eggs and larvae are entrained within lagoons
(apparently by local currents). Because eggs and very young larvae cannot swim
* ~~and orient actively (Leis and Rennis, 1983), it is generally assumed that, when
at sea, larvae are widely dispersed (through diffusional and other processes).
As pointed out by Lobel during the discussion, similarly sized larvae of the
same species often are taken in the same plankton hauls. In addition, newly
recruited fish larvae often appear on reef sites in small groups of 3 or more
fish (e.g., pempherids, high hats, grunts, etc.), but may also recruit as
individuals, as noted by McFarland at the discussion. The question arose,
therefore, as to how valid is the assumption of wide dispersion of individuals
at sea? There was general agreement that little is really known about the
distribution and dispersion of the eggs and larvae of individual species at sea
(but see Leis and Goldman, 1984, for a beginning). Shapiro's paper in volume I
of these proceedings develops a model which suggests that the dispersion of
eggs resulting from a single spawning act will retain reasonable cohesiveness
(after 24 hours of drift in a current; this varies depending on assumed
conditions, but could yield as many as I post-.hatchling in each adjacent m3 of
water). If newly hatched larvae can swim actively and are mutually attracted,
it is at least possible that in the first few days they might form cohesive
aggregations. Shapiro's thesis specifically suggests the possibility that
young recruiting reef fishes may, in some instances, be kin. Tests of this
hypothesis can involve electrophoretic analysis of recent recruits that occur
in cohesive groups. A more immediate test involves the aging of all recruits
in a group by use of the otolith aging technique (Brothers and Thresher, 1985;
Victor, 1983). If spawned simultaneously and dispersed as a group, the
cohesiveness of which is retained by active behaviors by individuals in the

OVERVIEW: THE DYNAMICS OF RECRUITMENT IN CORAL REEF ORGANISMS

William N. McFarland
Section of Ecology and Systemnatics
Division of Biological Sciences
Cornell University
Ithaca, New York 14853

In her introduction to volume 1 of these proceedings, Marjorie Reaka
states, "The second set of papers addresses the issue about which we probably
know the least in coral reef ecology: what factors control recruitment?" One
must conclude from the three contributed papers in volume I of these proceedings
(Lobel and Robinson, 1983; Shapiro, 1983; Victor, 1983) and the papers in the
present volume (Brothers and Thresher, 1985; Lobel and Neudecker, 1985; McFarland
and Ogden, 1985) that data on reproduction, larval development, and recruitment
are sparse. This is astonishing when we consider that almost all coral reef
fishes, and a great majority of tropical invertebrates too, cast their eggs and
larvae into the plankton where they must survive and grow before they can settle
again to the reef as juveniles.

The discussion held on December 27, 1983, at the ASZ meeting in Philadelphia,
dwelled on two aspects of recruitment, what we actually know and what we need
to know to better understand the dynamics of reef recruitment. About 20 par-
ticipants were involved in the discussion. Here I will summarize the more
important trends of the discussion.

Given the ubiquity of a planktonic larval phase in so many coral reef0
organisms, why is so little known about recruitment? The answer is multiple.
The eggs and larvae of most reef organisms are small, translucent, and difficult
to identify to species. They are extremely hard to maintain and study in
captivity because they are fragile and nutritional requirements often are
either unknown or difficult to adequately provide. In addition, only over the
last decade have newer techniques (e.g., otolith aging, increased use of SCUBA
and saturation diving for observations, improved methods for measuring currents,
etc.) become accessible to reef biologists. The paper by McFarland and Ogden
(this proceeding) addresses what is known about why so many fishes cast their
eggs and larvae into the plankton. They conclude that there is yet no definitive
answer; that the function of a pelagic larval phase for one reef organism may
not apply to another species; that in most species multiple functions prevail;
and that avoiding predation and dispersal appear to be the most important
factors that favor long-lived planktonic larvae. In addition, the times of
spawning in most reef species are poorly known. In recent years, however,
considerable effort has been placed on in situ observations of where and when
reef fish spawn. Lobel and Neudecker's-pape-r(this volume) describes spawning
in a hamlet (Hypoplectrus guttavaius) and also reviews important papers on
reproduction in other specTies of ishes. Although many reef species spawn at
specific *spawning sites and during a rush toward the surface, the time of
spawning is variable, often occurring in the late afternoon and dusk (Lobel and
Neudecker, 1985). However, spawning also often occurs during midday or midafter-
noon (Colin, 1982; Robertson, 1983). It was the general conclusion in the

. ~group, then each recruit in the group should be of the same daily age. This
specific experiment has not yet been performed.

As tantalizing as the possibility of kinship in recruits may be, a larger
related question prevails. What do larval fishes do when at sea? It is
axiomatic that they feed and grow, but beyond that the discussion group agreed
that the behaviors of larval reef fishes remain virtually unknown. Quantitative
data on the behaviors of a variety of species of reef fishes are crucial to
understanding the recruitment process. Answers will not be easy nor rapidly
forthcoming, for they must involve careful, stratified plankton sampling to
determine vertical distributions (at different times of the day), and also mus-c
include specific experiments in the laboratory and on shipboard to assess the
responses of fish larvae to various stimuli. The investigations are greatly
complicated by larval growth. Because most reef fish larvae spend from I to 6
weeks and in some instances up to 15 weeks at sea (Brothers and Thresher, this
volume), behaviors must change as the larvae grow. It is critical, therefore,
to also evaluate the ontogeny of larval behaviors. The central importance of
such studies to the overall process of tropical fish recruitment, however,
should motivate investigators to overcome the difficulties that the study of
larval fish behavior presents. Especially in this area will creative and
imaginative investigations help, for they will provide short cuts to what
otherwise would be a prolonged undertaking. The lead provided by researchers
studying larval fish behaviors in temperate waters, fortunately, can be of
great assistance in getting started (e.g., Blaxter and Hunter, 1982; Lasker,
1981; on clupeoid fishes).

Since the classic demonstrations of Johannes Schmidt (1922) that the
leptocephalus larvae of the European eel can drift at sea for not only months
but in some instances for more than a year, the duration of larval life of
fishes has provoked constant interest. Knowing how long a larva can exist at
sea has been particularly important in explaining long-distance transport of
reef fishes, e.g., the presence of Indo-Pacific reef species in the eastern
Pacific (Brothers and Thresher, this proceeding). The problem has been frustra-
ting because of the lack of methods to age larvae. The "invention" of the
daily otolith aging technique (Panella, 1971; Brothers, et al., 1976) and the
"discovery" of a transitional check mark in the otolith increments that approxi-
mate settlement of the larvae from the plankton to the reef substrate (Brothers
and McFarland, 1981; Brothers, et al., 1983; Victor, 1983) have revolutionized
the aging of larval fishes and, as a consequence, estimating the duration of
larval life of reef fishes. Brothers and Thresher's paper (this proceeding)
examines 115 species of Pacific reef fishes. Interestingly, over 80% of these
fishes have fairly short larval lives (< 45 days). All fishes with pelagic
larval durations in excess of 45 days are widely distributed (as expected). We
recognized that the use of otolith aging is an innovation that will assi-st in
providing answers to many questions concerning recruitment. Already we have a
better understanding of the lengths of larval life for a considerable number of
species. In some species the length of larval life can be restricted to only a
few days (e.g., French grunts, as noted by McFarland during the discussion) or
can encompass a highly variable number of days (from a short number to many
days and may show a strongly skewed distribution toward fewer days; Victor
* noted during the discussion that this is the case for some wrasses). Given the

novelty of the otolith aging techn~ique, a general warning was sounded that care
should be taken with generalizations. Nevertheless, it was agreed that few
studies on fish recruitment could effectively proceed without the use of this
revealing technique.

The paper by Lo-bel and Robinson (1983) re~viewed attempts to relate larval
drift to mesoscale disturbances in the near surface waters, e.g., local current
gy res .For many years various authors have invoked changes in local currents
or even steady current conditions to explain how spawning, larval duration,
and settlement are integrated to sustain local populations of fishes (Emery,
1972; Johannes, 1978, 1980; Sale, 1970). It is only over thie last decade,
however, that oceanographers have begun to develop instrumentation that allows
ready determination of local current eddies, shears, etc., on a sufficiently
small scale to be useful in biological investigations (see, for example,
Robinson, 1983). Techniques include radio-tracked drogues, expendable rapid
acting bathythermographs, current meters, and satellite infrared imagery. As
deployed in Hawaii, the initial data appear to support the concept that larvae
spawned at a particular site can drift in gyres (at certain times of the year
only) and could be returned close to their point of departure in times coin-
cident with the general length of larval life. We agreed that it was crucial
that local current regimes in various areas of the tropics must be carefully
measured at different times of the year and during all phases of the lunar
month if the potential drift paths of eggs and larvae are to be ascertained.
However, as indicated, daily changes in larval behaviors and the ability to
swim and orient can place fishes in different local current regimes. It is
possible, therefore, that active directed behaviors (such as vertical movements)~
could increase the chance(s) that fishes will be dispersed to suitable reefW
habitats for colonization. Thus, it is critical that subsurface currents as
well as surface currents be measured.

As crucial as the measurements of currents are, they must be coupled to
ongoing investigations of spawning, egg and larval distribution, and larval
behaviors if they are to result in meaningful conclusions for understanding
recruitment dynamics in the tropics.

As indicated by McFarland and Ogden (this proceeding) data on the actual
settlement of reef fishes are sparse. Observations of settlement on a given
day usually record the presence of recruits on a specific reef site, from which
they were absent the previous day. Daily observations of this sort are useful
in establishing recruitment cycles, but they reveal little about actual settle-
ment behaviors of local reef fish. Do most reef fish larvae settle at night,
or is there die] variation? And if so, are settlement behaviors species specific
or plastic? What specific factors trigger settlement, and what attributes of a
reef are attractive to each species? Answers to these questions will not yield
to casual observation. As an example, during the discussion I indicated that
newly recruited French grunts repeatedly established themselves on the same
small coral heads and never recruited on closely adjacent heads that had similar
configurations. Choice of a site might, thus, be resolved not only by appropriate
substrate requirements but also by microscale differences in currents, as noted
by Lobel. To know exactly what factors control site selection by new recruits
will require carefully controlled manipulative experiments for a broad spectrum
of reef species. Answers are critical before generalizations can be made: if W

settlement is to a large degree passive and succeeds only because currents
carry larvae over reefs, then the process can be considered somewhat stochastic;
if, however, searching and testing behaviors, such as vertical movements in the
water column (as observed in many invertebrate reef plankters; see McFarland
and Ogden, this volume, for specific references), are widely utilized by larvae
to identify suitable substrates on which to settle, then recruitment can be
considered as an active determined process. However, even if fish larvae do
actively 'search' for specific substrate as they settle from the plankton, or
soon after, their success in establishing themselves can be negatively affected
by the presence of previously settled juveniles (Shulman, et al., 1983). As a
result, even if the act of settlement is deterministic forea-E'-larva, the
overall success of individual recruits is infused with a large element of
chance (e.g., being in a current that favors transport to native habitat,
arriving in a region of native habitat at the appropriate time of day to avoid
predation, actually finding an appropriate and 'empty' settlement site within
the native habitat). In a very real sense, therefore, settlement can be viewed
as partly determined and partly stochastic. Because we remain largely ignorant
of the details of the settlement process, and because recruitment is central to
understanding coral reef fish community structure (see Helfman, 1978, for review),
generalizations about the proximal and ultimate causes of the high species diver-
sity of fishes associated with coral reefs remain, at best, first approximations.

Although innumerable problems require solution to understand recruitment
dynamics in coral reef organisms, it was generally concluded that rapid progress
would result by focusing attention on four general areas: (1) the timing and
dynamics of reproduction: where, when, and how often; (2) investigation of the
fine scale horizontal, vertical, and diel distribution of larvae over a time
frame that spans reproduction through recruitment; (3) coincident fine scale
measures of local current regimes (at varied depths); and (4) descriptive and
experimental investigations of the behaviors of larval coral reef organisms.
Furthermore, it is not enough to study only one species, although individual
researchers will be hard pressed to examine more than one at a time.

Because of the necessary broad scope of recruitment studies, the research
will involve planktologists, fish biologists, invertebrate biologists,
behaviorists, and physical oceanographers. As a result, interdisciplinary
collaborations will be a requirement in understanding and solving tropical
recruitment processes.

LITERATURE CITED

Blaxter, J. H. S., and J. R. Hunter. 1982. "The Biology of the Clupeoid
Fishes." Adv. Mar. Biol. 20:1-223.

Brothers, E. B., C. P. Matthews, and R. Lasker. 1976. "Daily Growth Increments
in Otoliths From Larval and Adult Fishes." Fish. Bull. 74:1-8.

Brothers, E. B., and W. N. McFarland. 1981. "Correlations Between Otolith
Microstructure, Growth, and Life History Transitions in Newly Recruited
French Grunts (Haemulon flavolineatum (Desmarest), Haemulidae)." In: R.
Lasker and J. Blaxter (eds.), Early Life History of Fish, II. Rapp. P-V.
Reun. Cons. Int. Explor. Mer.:-7T37.

Brothers, E. B., D. McB. Williams, and P. F. Sale. 1983. "Length of Larval
Life in Twelve Families of Fishes at 'One Tree Lagoon,' Great Barrier Reef,
Australia." Mar. Biol. 76:319-324.

Brothers, E. B., and R. E. Thresher. 1985. "Pelagic Duration, Dispersal, and
the Distribution of Indo-Pacific Coral-Reef Fishes," pp. 53-69. In: M. L.
Reaka (ed.), The Ecology of Coral Reefs, Symp. Ser. Undersea Res., Vol. 3(1).
NOAA Undersea Res. Progr., Rockvil-eMd.

Colin, P. L. 1982. "Aspects of the Spawning of Western Atlantic Reef Fishes,"
pp. 69-78. In: G. R. Huntsman, W. R. Nicholson, and W. W. Fox (eds.), The
Biological Basis for Reef Fishery Management. NOAA Tech. Memo NMFS-SEFC-T,
Nat. Mar. FTis. Serv., Beaufort, NC.

Emery, A. R. 1972. "Eddy Formation From an Oceanic Island: Ecological
Effects." Carib. J. Sci. 12:121-128.

Helfman, G. S. 1978. "Patterns of Community Structure in Fishes: Summary and
Overview." Environ. Biol. Fish. 3:129-148.

Johannes, R. E. 1980. "Using Knowledge of the Reproductive Behavior of Reef
and Lagoon Fishes to Improve Fishing Yields." In: J. E. Bardach, J. J.
Magnuson, R. C. May, and J. M. Reinhart (eds.),-lFish Behavior and Its Use in
the Capture and Culture of Fishes. Proc. 5th Int.enter Living Auat. Resource
Managemt., Minila, Philippines. 512 pp.

Lasker, R., ed. 1981. Marine Fish Larvae. Univ. Wash. Press, Seattle. 131 pp@

Leis, J. M., and B. Goldman. 1984. "A Preliminary Distributional Study of Fish
Larvae Near a Ribbon Reef in the Great Barrier Reef." Coral Reefs, 2:197-203.

Leis, J. M., and D. S. Rennis. 1983. The Larvae of Indo-Pacific Coral Reef
Fishes. Univ. Hawaii Press, Honolulu'. 26 pp.

Lobel, P. S., and S. Neudecker. 1985. "Diurnal Periodicity of Spawning Activity
by the Hamlet Fish, Hypoplectrus guttavarius (Serranidae)," pp. 71-86. In:
M. L. Reaka (ed.), The Ecology of Coral Reefs, Symp. Ser. Undersea Res.,
Vol. 3(1). NOAA Uned'ersea Res. Fogr., Rockville, Md.

Lobel, P. S., and A. R. Robinson. 1983. "Reef Fishes at Sea: Ocean Currents
and the Advection of Larvae," pp. 29-38. In: M. L. Reaka (ed.), The Ecology
of Deep and Shallow Coral Reefs, Symp. Ser. Undersea Res., Vol. 1(TT. NOAA
Undersea Res. Progr., Rockville, Md.

McFarland, W. N., and J. C. Ogden. 1985. "Recruitment of Young Coral Reef
Fishes From the Plankton," pp. 37-51. In: M. L. Reaka (ed.), The Ecology of
Coral Reefs, Symp. Ser. Undersea Res., Vol. 3(1). NOAA Undersea Res. Progr.,
RockviTTe, Md.

Panella, G. 1971. "Fish Otoliths: Daily Growth Layers and Periodical Patterns."
Science 173:1124-1127.

Robertson, D. R. 1983. "On the Spawning Behavior and Spawning Cycles of Eight
Surgeon Fishes (Acanthuridae) From the Indo-Pacific." Environ. Biol. Fish.
9:193-223.

Robinson, A. R., ed. 1983. "Overview and Summary of Eddy Science. 1.1 Eddy
Currents in the Ocean," pp. 3-15. In: Eddies in Marine Science. Springer-
Verlag, Berlin.

Sale, P. F. 1970. "Distribution of Larval Acanthuridae off Hawaii." Copeia
4:765-766.

Schmidt, J. 1922. "The Breeding Places of the Eel." Phil. Trans. Roy. Soc.
Lond., Ser. B, 211:179-208.

Shapiro, D. Y. 1983. "On the Possibility of Kin Groups in Coral Reef Fishes,"
pp. 39-46. In: M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral
Reefs, Symp. Ser. Undersea Res., Vol. (1). NOAXUndersea Res. Progr.,
Rockville, Md.

Shulman, M. J., J. C. Ogden, J. P. Ebersole, W. N. McFarland, S. L. Miller, and
N. G. Wolf. 1983. "Priority Effects in the Recruitment of Juvenile Coral
Reef Fishes." Ecology 64:1508-1513.

Thresher, R. E. 1984. Reproduction in Reef Fishes. Tropical Fish Hobbyist,
Neptune City, NJ. 399 pp.

Victor, B. C. 1983. "Settlement and Larval Metamorphosis Produce Distinct
Marks on the Otoliths of the Slippery Dick, Halichoeres bivattatus," pp. 47-52.
In: M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral Reefs, Symp. Ser.
Undersea Res., Vol. 1(1).NOAA Undersea es.7rogr., Roc-' ' lleF.

GROWTH AND LIFE HISTORY PATTERNS OF CORAL REEF ORGANISMS:
A DISCUSSION GROUP SUMMARY AND OVERVIEW

Ronald H. Karlson
Ecology and Organismic Biology Program
School of Life and Health Sciences
University of Delaware
Newark, Delaware 19716

I would like to take this opportunity to thank all who attended our late
evening discussion group. Although it would have been impossible for us to
deal with all major taxa of coral reef organisms, we discussed a reasonably
diverse array of colonial organisms and a group of solitary echinoderms.
Taxonomic areas of research and a partial list of participants include actinians
(Sebens), bryozoans (Jackson), corals (Brakel, Highsmith, Hughes, Hunter,
Jackson), echinoderms (Highsmith, Keller, Levitan, Mladenov), gorgonians
(Harvell, Lasker, Wahle), sponges (Harvell, Pomponi, Suchanek), and zoanthids
(Karlson, Sebens, Suchanek).

Six short presentations by Brakel, Hughes, Suchanek, Wahle, Karlson, and
Mladenov were given in our discussion session. Each was followed by questions
and/or discussion. Each presentation dealt with variability in one or more of
the following: growth form, growth rate, overgrowth frequency, fecundity,
survivorship, colonial integrity, or reproductive mode. We also discussed size-
specific variation in life histories, fission, fusion, and brooding in coral
reef invertebrates.

Brakel's presentation dealt with depth-related variation in the growth
form of Porites astreoides (Brakel, 1983). He noted that much of the literature
dealing with coral growth form suggests that there may be a simple monotonic
relationship between colony shape and depth. Brakel's analysis indicates that
flattened colonies are typical of high-energy, shallow water habitats and of
low-light, deep water habitats. Between these two extremes, wave energy and
light set upper limits on morphological variation and are not good predictors
of mean morphological parameters.

Three additional citations to those given by Brakel (1983) on the general
relationship between colony morphology and depth are Fricke and Schuhmacher
(1983), Chappell (1980), and Stearn (1982). The first is an interesting paper
on photoadaptations in deep hermatypic corals found down to 145 m. They noted
the prevalence of flat hermatypes in deep water and a great diversity of growth
forms in shallow water. This pattern was not observed for ahermatypic corals.
Chappell (1980) discussed the effects of light and wave stress on coral growth
form and rate and extended these effects to include the growth of coral reefs.
Stearn (1982) critically reviewed the analysis of growth form in corals and
stromatoporoids. He, like Brakel, concluded that the relationship between
growth form and environmental variables is quite complex due to the influences
of genetic and developmental processes.

The presentation by Hughes (1983) dealt with coral growth rates as well
as life history variation over a depth range. The major point of this presentation

was that net coral growth rates do not vary much with depth. Although
calcification rates decrease with increasing depth, rates of injury to colonies _
are much higher in shallow water. This results in several examples of deep
corals at 35 or 55 m growing faster than corals of the same species at 10 or
20 m. These data suggest that we should exercise caution when using calcifica-
tion rates to evaluate coral reef growth.

In our third presentation, Suchanek discussed the relative abundances and
interaction frequencies of corals and demosponges over a 40 m depth range
(Suchanek, et al., 1983). He emphasized the high frequency of demosponge-coral
interactions at-40 m and the increasing prevalence of demosponges as aggressive
spatial competitors with increasing depth. Aggressive interactions in shallow
water involved encrusting gorgonians, hydrocorals, and zoanthids.

The effect of spatial competition on coral or demosponge growth, depth
distribution, and life history variation remains to be determined. Even though
deep hermatypic corals experience relatively low injury rates, a significant
amount of this injury is caused by sponge overgrowth (Hughes, 1983; Suchanek,
et al., 1983; Hughes and Jackson, in review). Lang (1974) has previously noted
FiFe prevalence of demosponges in deep fore-reef (below 55 m) and island slope
zones off Discovery Bay, Jamaica. Hermatypic corals growing under light-limited
conditions near their lower distributional limit may be more susceptible to
overgrowth by demosponges than corals growing at shallower depths.

Wahle presented his data on the relationship between injury, colony size,
and life history variation in the gorgonian Plexaura homomalla (Wahle, 1983). *
Between-colony analysis of variation in fecundity indicated that only colonies
taller than 20 cm contained eggs. Natural and experimentally induced injuries
to large female colonies resulted in physiologically isolated colony branches
which exhibited size-related variation in fecundity. This work demonstrates
that: 1) "colony size and not age controls both onset and continuation of
gametogenesis"; 2) "injury can subtly reduce fecundity without noticeably
affecting (colony) size"; and 3) injuries can alter the level of colony integrity
and "create mosaics of different life history stages coexisting within the same
colony."

My own presentation dealt with an analysis of divergent life history
patterns in Zoanthus sociatus and Z. solanderi. Since these data do not appear
elsewhere, I present them here in more detail than that given to the other five
presentations. Both of these zoanthids are very common at several locations
throughout the Caribbean Sea. These sessile colonial organisms generally
inhabit shallow subtidal and lower intertidal zones where they occasionally are
exposed to extreme wave action (e.g., Hurricane Allen, see Woodley, et al.,
1981) or to desiccation (Sebens, 1982a; Fadallah, et al., 1984). Boi-i species
may successfully escape from such extremes by dispersing gametes and/or fragmented
clusters of adult polyps.

Highsmith (1982) has suggested that fragmentation is an important part of
the life history of many corals, and that one characteristic typical of
fragmenting species is delayed sexual reproduction. The data in table 1 were
collected from 101 zoanthid colonies to determine if either Z. sociatus, Z.
solanderi, or both species delay sexual reproduction. Both species have been _

Table 1.--The frequency of fertile polyps in four colony size classes of
Z. sociatus and Z. solanderi. Colonies were collected in June, 1983,
at Discovery Bay, Jamaica. Numbers of colonies are given in parentheses.

Colony
size Polyp
class condition Z. sociatus Z. solanderi 2
_ _ ~~~~~~x2

A. < 20 polyps fertile 0 8 9.44
nonfertile 227 (N=31) 188 (N=40) p < 0.005

B. 20 < 40 polyps fertile 0 22 22.12
nonfertile 103 (N=4) 92 (N=4) p < 0.001

C. 40 < 60 polyps fertile 0 9 11.63
nonfertile 58 (N=1) 40 (N=1) p < 0.001

D. very large fertile 105 179 26.93
(subsamples nonfertile 395 (N=10) 321 (N=10) p < 0.001
= 50 polyps
per colony)

E. < 60 polyps fertile 0 (N=36) 39 (N=45) 21.36
very large fertile 105 (N=10) 179 (N=10) p < 0.001

observed to fragment either as clusters of loose polyps, analogous to the "polyp
balls" reported by Rosen and Taylor (1969), or as polyps attached to shifting
rubble. The fact that fragmentation occurs, however, does not demonstrate that
it is adaptive or an important life history characteristic.

The frequency of fertile polyps in ten extremely large colonies of Z.
sociatus and Z. solanderi was 21% and 36%, respectively (table 10). These-were
sampled during their peak reproductive season, as indicated by Karlson (1981,
and unpub. data). None of the polyps in 36 small colonies (< 60 polyps) of Z.
sociatus were fertile; 11% of the polyps in 45 small colonies of Z. solanderi
were fertile (table 1E). All three small colony size classes exhi ited
significantly higher frequencies of fertile polyps for Z. solanderi than for Z.
sociatus (table 1A, B, and C). The absence of fertile polyps in small Z.
sociatus colonies suggests that this species delays sexual reproduction while
Z. solanderi does not.

Further examination of the 81 small zoanthid colonies (< 60 polyps)
revealed another significant difference between these two species. Colonial
growth in zoanthids commonly occurs by the budding of new individuals from
stolons. These stolonal connections can degenerate over time (West, 1979),
resulting in very loosely organized colonies with extremely low levels of
integration and asynchronous sexual reproduction (e.g., table 1D). As a measure

Table 2.--Group size as a measure of colonial integrity

Colony
size Group size
class Z. sociatus Z. solanderi t-test

< 20 X = 3.6 polyps per group 2.3 polyps per group
polyps Range = 1 - 13 1 - 8 3.68
n = 63 groups in 31 colonies 84 groups in 40 colonies p < 0.001

20 < 40 X = 6.9 2.3
polyps Range = 1 - 27 1 - 5 4.40
n = 15 in 4 colonies 49 in 4 colonies p < 0.001

40 < 60 X = 5.3 2.7
polyps Range = 1 - 30 1 - 5 1.25
n = 11 in 1 colony 18 in 1 colony N.S.

of colonial integrity, I have counted the number of unconnected polyp groups
and the number of polyps within each group. There were a total of 89 polyp
groups within 36 colonies of Z. sociatus and 151 polyp groups within 45 colonies
of Z. solanderi (table 2). Z3anthus solanderi had significantly fewer polyps
pergroup than did Z. sociatus in the two smallest colony size classes. The
maximum number of polyps per group was also much higher in Z. sociatus than Z.
solanderi.

These significant differences in polyps per group may have important life
history implications. Variation at this low level of colonial integrity is
likely to influence the size of colony fragments and the size frequency
distribution of zoanthid colonies. Among colonial species, survivorship,
fecundity, fragmentation, and fusion are highly dependent on colony size (e.g.,
Hughes and Jackson, 1980, and ms. in review; Buss, 1980, 1981; Sebens, 1982b;
Hughes, 1984). Small fragments of Z. sociatus colonies have higher mortality
rates than do fragments of the same size for L. solanderi (Sebens, 1982a;
Karlson, in preparation). In addition, small colonies of Z. sociatus are
nonreproductive (table 1). The high level of colonial integrity in Z. sociatus
suggests a benefit to large group (and colony) size in this species and,
conversely, a cost (e.g., lower survivorship) associated with small size.
There appears to be little cost to small size in Z. solanderi. This zoanthid
is more resistant to predation than Z. sociatus (�e-bens, 1982a) and does not
delay sexual reproduction. Low colonial integrity in Z. solanderi is likely to
enhance fragmentation of small clusters of polyps. Higher colonial integrity
in Z. sociatus may enhance production of larger clusters by fragmentation or,
alternatively, it may enhance resistance to fragmentation.

In our last presentation, Mladenov briefly reviewed some of the life
history characteristics of seven shallow-water Jamaican ophiuroids. All of the

species studied were small cryptic forms. Three species were fissiparous (see
also Mladenov, et al., 1983) and produced relatively small eggs; two of these
broadcast "exceedingly small numbers of larvae." Two non-fissiparous species
produce and brood a small number of relatively large eggs which undergo direct
devel opment.

Mladenov noted that of approximately 2,000 extant ophiuroid species, only
57 are known brooders (Hendler, 1979) and 37 undergo fission (Emson and Wilkie,
1980; Emson, et al., 1985). While 65% of those exhibiting fissiparity and
12% of the brooders are found in the tropics, brooders like Amphipholis squamata
and fissiparous species like Ophiactis savignyl are among the most widespread
and abundant" tropical ophiuroids. At Discovery Bay, the three fissiparous
ophiuroid species were numerically dominant. Emson, et al. (1985) speculate
that populations of these species have been "maintain~e almost solely by asexual
reproduction."

Among colonial species, growth, fission, fusion, and fragmentation are
important asexual life history processes (e.g., Bak and Engel, 1979; Hughes and
Jackson, 1980; Karlson, 1980, 1983; Bothwell, 1981; Tunnicliffe, 1981; Lasker,
1983; Rylaarsdam, 1983; Hughes, 1984). More critical evaluation of the importance
of sexual reproduction to these colonial species is currently underway. There
has been a relatively recent increase in this particular research area. At the
Third International Coral Reef Symposium held in 1977, there were no presentations
on this topic. Four years later at the Fourth International Coral Reef Symposium,
there were six such presentations (Bothwell, unpub. data; Karlson, 1981; Kojis
and Quinn, 1981; Richmond, 1981; Van Moorsel, 1981; Yamazato, et al., 1981).
Data from the Van Moorsel presentation have now been published--Van Moorsel,
1983). An excellent review of this subject dealing with over 100 scleractinian
coral species has also just been published by Fadallah (1983); also see Harrison,
et al. (1984). Fadalallah (1983) and Szmant-Froelich, et al. (1983) suggest the
neTeior further work on reproductive cycles, reproductive modes, relative
fecundity, and reproductive effort in order to satisfactorily evaluate the
importance of sex to coral reef organisms. At this point we know virtually
nothing about the relationship between reproductive effort and the actual input
of settled larvae into adult populations; this point also is stressed by
McFarland (1985).

The original purpose of this discussion group was to compare growth and
life history patterns of solitary and colonial organisms on shallow and deep
reefs. We did, in part, deal with depth-related variation in coral growth
form, growth rate, and life history patterns. Furthermore, we briefly discussed
some general differences and similarities between colonial and solitary species
(see Jackson, 1977). The remarkable success of fissiparous echinoderms and
colonial corals, sponges, gorgonians, and zoanthids has been noted. This is
suggestive of some level of evolutionary convergence favoring asexual strategies
even in solitary organisms. My own work with zoanthids emphasized low colonial
integrity in two very successful species. These species may be more similar to
aggregating solitary organisms (e.g., barnacles, anemones, polychaetes, and
mussels) than to more highly organized colonial forms. One curious and
potentially significant difference between solitary and colonial organisms is
the ability of colonial species to fuse (Teissier, 1929; Stephenson, 1931;
Schijfsma, 1939; Ivker, 1972; Hughes and Jackson, 1980; Buss, 1982). It is

generally thought that fusion, whether it occurs between clonemates or non-
clonemates, results in the formation of a larger colony with higher survivorship_
and/or fecundity (Hughes and Jackson, in review). It also results in a level
of physiological continuity which even aggregating solitary organisms cannot
mimic. Variation in colonial integrity and the adaptive significance of
fragmentation and fusion among colonial species would appear, then, to be an
interesting new direction for future exploration.

LITERATURE CITED

Bak, R. P. M., and M. S. Engel. 1979. "Distribution, Abundance, and Survival
of Juvenile Hermatypic Corals (Scleractinia) and the Importance of Life
History Strategies in the Parent Coral Community." Mar. Biol. 54:341-352.

Bothwell, A. M. 1981. "Fragmentation, a Means of Asexual Reproduction and
Dispersal in the Coral Genus Acropora (Scleractinia: Astrocoeniida:
Acroporidae) - A Preliminary Report." Proc. 4th Int. Coral Reef Symp.
2:137-144.

Brakel, W. H. 1983. "Depth-Related Changes in the Colony Form of the Reef
Coral Porites astreoides," pp. 21-26. In: M. L. Reaka (ed.), The Ecology of
Deep and Shallow Coral Reefs. Symp. Ser. Undersea Res., Vol. 1i7. NOAA
Unhdersea Res. Progr., Rockville, Md.

Buss, L. W. 1980. "Competitive Intransitivity and Size-Frequency Distributions
of Interacting Populations." Proc. Nat. Acad. Sci. USA 77:5355-5359.

Buss, L. W. 1981. "Group Living, Competition, and the Evolution of Cooperation
in a Sessile Invertebrate." Science 213:1012-1014.

Buss, L. W. 1982. "Somatic Cell Parasitism and the Evolution of Somatic Tissue
Compatibility." Proc. Nat. Acad. Sci. USA 79:5337-5341.

Chappell, J. 1980. "Coral Morphology, Diversity, and Reef Growth." Nature
286:249-252.

Emson, R. H., and I. C. Wilkie. 1980. "Fission and Autotomy in Echinoderms."
Oceanogr. Mar. Biol. Ann. Rev. 18:155-250.

Emson, R. H., P. V. Mladenov, and I. C. Wilkie. 1985. "Patterns of Reproduction
in Small Jamaican Brittle Stars: Fission and Brooding Strategies Predominate,"
pp. 87-100. In: M. L. Reaka (ed.), The Ecology of Coral Reefs, Symp. Ser.
Undersea Res., Vol. 3(1). NOAA Undersea Res. Progr., Rockville, Md.

Fadlallah, Y. H. 1983. "Sexual Reproduction, Development and Larval Biology
in Scleratinian Corals: A Review." Coral Reefs 2:129-150.

Fadlallah, Y. H., R. H. Karlson, and K. P. Sebens. 1984. "A Comparative Study
of Sexual Reproduction in Three Species of Panamanian Zoanthids
(Coelenterata:Anthozoa)." Bull. Mar. Sci. 35:80-89. 0

*_ Fricke, H. W., and H. Schuhmacher. 1983. "The Depth Limits of Red Sea Stony
Corals: An Ecophysiological Problem (a Deep Diving Survey by Submersible)."
Mar. Ecol. 4:163-194.

Harrison, P. L., R. C. Babcock, G. D. Bull, J. K. Oliver, C. C. Wallace, and B.
L. Willis. 1984. "Mass Spawning in Tropical Reef Corals." Science
223:1186-1189.

Hendler, G. 1979. "Reproductive Periodicity of Ophiuroids (Echinodermata:
Ophiuroidea) on the Atlantic and Pacific Coasts of Panama," pp. 145-156.
In: S. E. Stancyk (ed.), Reproductive Ecology of Marine Invertebrates. The
Belle W. Baruch Library in Marine Science, Number 9. Univ. S. Carolina
Press, Columbia.

Highsmith, R. C. 1982. "Reproduction by Fragmentation in Corals." Mar. Ecol.
Prog. Ser. 7:207-226.

Hughes, T. 1983. "Life Histories and Growth of Corals Over a Depth Gradient,"
pp. 17-20. In: M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral
Reefs, Symp. Ser. Undersea Res., Vol. 1(1). NOAA-Undersea Res. Prog.,
Rockville, Md.

Hughes, T. P. 1984. "Population Dynamics Based on Individual Size Rather Than
Age: A General Model With a Reef Coral Example." Amer. Nat. 123:778-795.

Hughes, T. P., and J. B. C. Jackson. 1980. "Do Corals Lie About Their Age?
Some Demographic Consequences of Partial Mortality, Fission, and Fusion."
Science 209:713-715.

Hughes, T. P., and J. B. C. Jackson. In review. "Population Dynamics and Life
Histories of Foliaceous Corals." Submitted to Ecol. Monogr.

Ivker, F. B. 1972. "A Hierarchy of Histo-Incompatibility in Hydractinia
echinata." Biol. Bull. 143:162-174.

Jackson, J. B. C. 1977. "Competition on Marine Hard Substrata: The Adaptive
Significance of Solitary and Colonial Strategies." Amer. Nat. 111:743-767.

Karlson, R. H. 1980. "Alternative Competitive Strategies in a Periodically
Disturbed Habitat." Bull. Mar. Sci. 30:894-900.

Karlson, R. H. 1981. "Reproductive Patterns in Zoanthus spp. from Discovery
Bay, Jamaica." Proc. 4th Int. Coral Reef Symp. 2:699-704.

Karlson, R. H. 1983. "Disturbance and Monopolization of a Spatial Resource by
Zoanthus sociatus (Coelenterata, Anthozoa)." Bull. Mar. Sci. 33:118-131.

Kojis, B. L., and N. J. Quinn. 1981. "Reproductive Strategies in Four Species
of Porites (Scleractinia)." Proc. 4th Int. Coral Reef. Symp. 2:145-152.

Lang, J. C. 1974. "Biological Zonation at the Base of a Reef." Amer. Sci.
*s ~62:272-281.

Lasker, H. R. 1983. "Vegetative Reproduction in the Octocoral Briereum
asbestinum (Pallas)." J. Exp. Mar. Biol. Ecol. 72:157-169.

McFarland, W. N. 1985. "Overview: The Dynamics of Recruitment in Coral Reef
Organisms," pp. 9-15. In: M. L. Reaka (ed.), The Ecology of Coral Reefs,
Symp. Ser. Undersea Res., Vol. 3(1). NOAA Undersea Res. Progr., Rockville,
Md.

Mladenov, P. V., R. H. Emson, L. V. Colpitts, and I. C. Wilkie. 1983. "Asexual
Reproduction in the West Indian Brittle Star Ophiocomella ophiactoides (H.
L. Clark) (Echinodermata: Ophiuroidea)." J. Exp. Mar. Biol. Ecol. 72:1-23.

Richmond, R. 1981. "Energetic Considerations in the Dispersal of Pocillopora
damicornis (Linnaeus) planulae. Proc. 4th Int. Coral Reef S 2mp._21T53-156,

Rosen, B. R., and J. D. Taylor. 1969. "Reef Coral From Aldabra: New Mode of
Reproduction." Science 166:119-121.

Rylaarsdam, K. W. 1983. "Life Histories and Abundance of Colonial Corals on
Jamaican Reefs." Mar. Ecol. Prog. Ser. 13:249-260.

Schijfsma, K. 1939. "Preliminary Notes on Early Stages in the Growth of
Colonies of Hydractinia echinata (Flem.)." Arch. Neerland. Zool. 4:93-102.

Sebens, K. P. 1982a. "Distribution and Spatial Patterns of Zoanthids in the
Coral Reef Flat Community at Galeta, Panama: The Effects of Desiccation and
Predation." Bull. Mar. Sci. 32:316-335.

Sebens, K. P. 1982b. "Competition for Space: Growth Rate, Reproductive
Output, and Escape in Size." Amer. Nat. 120:189-197.

Steam, C. W. 1982. The Shapes of Paleozoic and Modern Reef-Builders: A
Critical Review." Paleobiology 8:228-241.

Stephenson, T. A. 1931. "Development and Formation of Colonies of Pocillopora
and Porites." Great Barrier Reef Expeditions, Sci. Rpts. 3:113-134.

Suchanek, T. H., R. C. Carpenter, J. D. Witman, and C. 0. Harvell. 1983.
"Sponges as Important Space Competitors in Deep Caribbean Coral Reef
Communities," pp. 55-60. In: M. L. Reaka (ed.), The Ecology of Deep and
Shallow Coral Reefs, Symp.Ter. Undersea Res., Vol. 1(1). NOAUndersea
Res. Progr., Rockville, Md.

Szmant-Froelich, A., L. Riggs, and M. Reutter. 1983. "Sexual Reproduction in
Caribbean Reef Corals." Amer. Zool. 23:961, abstract no. 513.

Teissier, G. 1929. "L'Origine Multiple de Certaines Colonies d'Hydractinia
echinata (Fleming) et ses Consequences Possibles." Bull. Soc. Zool. France
54:645-647.

Tunnicliffe, V. J. 1981. "Breakage and Propagation of the Stony Coral Acropora
cervicornis." Proc. Nat. Acad. Sci. USA 78:2427-2431.

Van Moorsel, G. W. N. M. 1981. "Different Reproductive Strategies in Two
Closely Related Stony Corals (Agaricia, Scleractinia)." Proc. 4th Int.
Coral Reef Symp. 2:193 (abstract).

Van Moorsel, G. W. N. M. 1983. "Reproductive Strategies in Two Closely Related
Stony Corals (Agaricia, Scleractinia)." Mar. Ecol. Prog. Ser. 13:273-283.

Wahle, C. M. 1983. "The Roles of Age, Size, and Injury in Sexual Reproduction
Among Jamaican Gorgonians." Amer. Zool. 23:961, abstract no. 514.

West, D. A. 1979. "Symbiotic Zoanthids (Anthozoa: Cnidaria) of Puerto Rico."
Bull. Mar. Sci. 29:253-271.

Woodley, J. D., E. A. Chornesky, P. A. Clifford, J. B. C. Jackson, L. S. Kaufman,
N. Knowlton, J. C. Lang, M. P. Pearson, J. W. Porter, M. C. Rooney, K. W.
Rylaarsdam, V. J. Tunnicliffe, C. M. Wahle, J. L. Wulff, A. S. G. Curtis,
M. D. Dallmeyer, B. P. Jupp, M. A. R. Koehl, J. Neigel, and E. M. Sides.
1981. "Hurricane Allen's Impact on Jamaican Coral Reefs." Science
214:749-755.

Yamazato, K. M. Sato, and H. Yamashiro. 1981. "Reproductive Biology of an
Alcyonacean Coral, Lobophytum crassum Marenzeller." Proc. 4th Int. Coral
Reef Symp. 2:671-678.

OVERVIEW: CORAL REEF COMMUNITY STRUCTURE AND FUNCTION

Mark A. Hixon
Department of Zoology and
College of Oceanography
Oregon State University
Corvallis, Oregon 97331

This paper summarizes an open discussion session which took place during
the "Coral Reef Ecology Workshop" held at the annual meeting of~the American
Society of Zoologists in Philadelphia on December 27, 1983. Due to either
sampling error or a genuine bias in current research emphasis, most of the
participants in this session studied herbivory and grazing patterns on reefs.
Consequently, most of the discussion centered around two related topics: (1)
the recently discovered mass mortality of urchins in the Caribbean and (2) the
relative importance of urchins vs. fishes in structuring benthic reef communities
in areas facing different fishing pressures. A third more general topic was
discussed briefly toward the end of the session: approaches to research on the
community structure of reef systems. I will cover each of these topies in
turn, extending discussion on the final topic with an overview of future needs
in the study of coral reef communities.

MASS MORTALITY OF CARIBBEAN URCHINS
One of the major herbivores on Caribbean reefs is the long-spined urchin
Diadema antillarum. In November 1983, a letter published in Science reported
unprecedented mass mortalities of this species that were sweeping the Caribbean
(Lessios, et al., 1983). Haris Lessios (Smithsonian Tropical Research Institute)
attended the session and provided both a description of the effects of the
apparent pathogen (the origin or nature of which was currently unknown) and
an updated account of the spread of the calamity.

The apparent disease runs a 4- to 5-day course, beginning with the urchins
emerging from coral cover and climbing up any available substrate. The urchins
then discolor, gradually lose their attachment capabilities, and ultimately
lose their spines and die. Up to 98% mortality has been reported in some
populations, with apparently resistant individuals remaining. Curiously, no
other species of urchin has been affected.

The outbreak was first noted near the Caribbean entrance of the Panama
Canal in January 1983. Urchin mortalities were observed subsequently in
Colombia, Costa Rica, and Grand Cayman by June; in Jamaica, Belize, Cancun
(Mexico), and Key Largo (Florida) by July; in the Bahamas by August; in the Dry
Tortugas and Bermuda by September; in Grand Turk and Haiti by October; and
throughout the Greater and Lesser Antilles, including Curacao, Bonnaire, and
part of the coast of Venezuela, between November 1983 and January 1984.

Initially, the spread of the die-off appeared to follow prevailing ocean
currents. However, certain exceptions, such as the outbreak in Barbados in
October (a full month or two before the remainder of the Antilles), suggested a

more complex situation. At present, Lessios reports that the entire Caribbean
except for much of Venezuela, Tobago, and the area between Guadaloupe and the
British Virgin Islands, has been affected, although these areas could be affected
by the time this report is published.

Mark Hay (University of North Carolina at Chapel Hill) noted that, despite
the die-off of Diadema in the Florida Keys, algal abundance on reefs there was
not unusually high 2 months after the die-off. He suggested that either herbivorou
fishes were sufficiently abundant to check algal growth or 2 months may not
have been enough time to detect a noticeable change.

Douglas Morrison (University of Georgia) summarized his quantitative data
on the effects of the die-off on a Jamaican reef. Mean densities of urchins I
year before the die-off were about 15 per m2 at 5 m depth and about 2 per m2 at
16 m. Several months after the die-off, mean densities had declined to about 1
per m2 at 5 m depth, and no urchins were apparent at 16 m. After the urchin die-
off at the shallow site, the coverage of crustose coralline algae (an indicator
of high grazing intensity) declined from 45% to about 18%. Replacing the coral-
lines, coverage by fleshy macroalgae quadrupled (from 7% to about 28%) and coverage
by filamentous algae nearly tripled. These patterns were corroborated by earlier
urchin-removal experiments conducted before the die-off; in the experimental
absence of grazing, crustose corallines were replaced by upright macroalgae.
In contrast, the die-off had little apparent effect upon algal growth at the
deep site, where urchin densities were normally low, and macroalgal abundance
remained relatively high (73% cover). The macroalgae that increased in abundance
at the shallow site are known to be resistant to fish grazing but not to urchi
grazing. These included members of the genera Lobophora, Caulerpa, and DictyoW

Whatever its cause, the mass mortality of Diadema throughout the Caribbean
provides an excellent (although uncontrolled) "naturaFexperiment" for determining
the impact of urchins on entire reef systems. Where adequate baseline data are
available, observations following the die-off will test the widespread applicabilit
of previous experimental analyses of small patch-reef systems (e.g., Ogden, et al.,
1973) and help to resolve the controversy discussed in the next section.

ARE URCHINS IMPORTANT GRAZERS ON UNFISHED REEFS?

Hay summarized a study that was in press at the time of the workshop
(Hay, 1984). He noted that most of the studies of the impact of urchin grazing
on reef benthos have occurred at two sites: Teague Bay, St. Croix (e.g., Ogden,
et al., 1973) and Discovery Bay, Jamaica (e.g., Sammarco, 1980). Such studies
~ve--Ted some to believe that urchins are the most important grazers on "typical"
coral reefs. However, Hay suggested that these particular sites are atypical
in that they have been heavily fished, indirectly resulting in unusually high
densities of urchins which have been freed from predation by (and perhaps
competition with) fishes. Using pieces of the seagrass Thalassia testudinum as
a field bioassay for the intensity of herbivory on macrophytes, he found that
eight lightly fished sites scattered throughout the Caribbean (including Salt
River on St. Croix) showed decreasing grazing intensity with increasing depth
and that almost all herbivory was due to fishes. Moreover, urchin abundances
were low at these sites. In contrast, Haiti and Teague Bay on St. Croix, bothF

heavily fished areas, showed the traditional opposite patterns. Hixon added
that Hay's general observation that urchins are Often rare where large urchin-
eating fishes are abundant is evident on Hawaiian reefs. Where fishing is
prohibited on marine reserves, such as Hanauma Bay and Coconut Island on Oahu,
urchins appear to be relatively rare.

Hay stressed that his major point was that many patterns documented on
human-impacted reefs may be recent, having prevailed only for the past few hundred
years. Because herbivory by fishes may select for different evolutionary responses
in algae than herbivory by urchins, attempts to extract evolutionary implications
from ecological data gathered on heavily fished reefs is not justified.

Les Kaufman (New England Aquarium) objected to a number of Hay's assertions,
particularly the idea that urchins on heavily fished reefs necessarily are
freed from predation by fishes. He took issue with Hay's data which indicated,
on one hand, a decrease in grazing intensity with increased depth at Salt River
on St. Croix but, on the other hand, an increase in grazing intensity with
increased depth at nearby Teague Bay on the same island. Based on his diving
experience at these sites, Kaufman felt that both areas supported fishes capable
of eating urchins. Thus, other factors besides the overall abundance of fishes
may explain the patterns that Hay attributed to fishing pressure. Further,
Kaufman felt that Hay failed to consider the importance of juvenile mortality
in urchin populations. He suggested that predation by wrasses on juvenile
urchins may be considerable on heavily fished reefs, such as Teague Bay (St.
Croix) and Discovery Bay (Jamaica), so that fishing pressure on larger fish
* ~~species may not ultimately affect urchin densities.

Substantial discussion centered on the adequacy of Thalassia as a bioassay
of grazing intensity. Robert Carpenter (University of Georgia) suggested that
different measures of grazing could produce different depth profiles of grazing
intensity (see also Steneck, 1983). Carpenter felt that the Thalassia technique
measured mostly parrotfish grazing and that a more general measure of grazing
is provided by counting the number of herbivore bites per unit reef area. Hixon
agreed, noting that bite marks by fishes from different families (e.g., parrotfishes
vs. surgeonfishes) can be individually identified and counted on flat substrates
(e.g., Hixon and Brostoff, 1983). Morrison also concurred and suggested that,
instead of Thalassia, a food readily consumed by all herbivores should be used
to measure overall grazing intensity. Based on his own and other's work,
Morrison indicated that filamentous algae seem to be preferred by urchins and
nearly all herbivorous fishes (especially parrotfishes and damselfishes).

Sara Lewis (Duke University) noted that some algal species (e.g., Lobophora
variegata) are differentially susceptible to grazing on different reefs. She
suggested that these within-species differences may represent geographical
variation in the development of plant defensive compounds. Hay agreed, noting
that herbivory on a local level possibly could induce the production of chemical
defenses. He went on to stress the problem of interpreting data from herbivore
food-preference observations. Data on food preferences provide information on
the responses of herbivores to different plants, but not necessarily on the
selective pressure that herbivores impose upon the plants. Clearly, more
information will be required to elucidate the reciprocal interactions between
reef herbivores and algae.

STUDYING COMMUNITY STRUCTURE IN REEF SYSTEMS

Eldredge Bermingham (University of Georgia) introduced this general topic
toward the end of the session by suggesting what he believed to be a major
problem with coral reef community ecology. He felt that reef ecologists were
often academic descendants of bird ecologists, but failed to follow their
ornithological forefathers' passion for detailed natural history data. He
stressed that lack of solid baseline data on coral reefs inhibits our ability
to understand these systems.

Both John Ebersole (University of Massachusetts at Boston) and Hixon
agreed in turn that detailed observational data are essential for any convincing
ecological study. In particular, long-term baseline data illustrating the
constancy or variability of observed patterns are needed. However, Hixon
questioned whether most (or even many) reef ecologists were academic descendants
of bird ecologists and, more importantly, noted that reef ecologists have
conducted far more rigorous experimental studies than most bird ecologists.
This disparity is understandable. On one hand, many terrestrial systems
(especially many avian communities) are amenable to long-term observation but
not to experimental manipulation. On the other hand, most coral reefs are both
geographically distant from most research institutions and difficult to observe
for long periods due to diving constraints. However, some reef systems are
small enough or the associated organisms sedentary enough to allow direct
experimentation, as evidenced by the papers in this symposium. Moreover,
artificial reefs constructed from either natural or manmade materials allow
researchers to rigidly control the age and structure of reef habitats, allowing
powerful experimental designs (e.g., Shulman, et al., 1983; Fitz,,et al., 1983;1W
Wolf, et al., 1983).

SYNTHESIS AND PROSPECTUS

It seem~s appropriate to close this paper with an overview of our future
needs in the'study of coral reef communities. It is clear that, despite the
knowledge that has accrued since the first professional ecologists donned SCUBA
gear in the 1950's, we have only begun to scratch the surface of the complexities
of coral reef systems. The number of research possibilities is infinite; the
field is wide open. Thus, rather than suggesting what we need to study in
particular, I would like to review the ideas suggested during this workshop on
how we might improve future studies.

Based an criticisms leveled at current studies and pleas for future changes
aired at various times during this workshop, three basic needs are evident.
First, the complexities of coral-reef community structure cannot be elucidated
without more extensive use of properly controlled field experiments. This
suggestion is not new (e.g., Connell, 1974), and many reef ecologists have
embraced experimentation enthusiastically. However, any experiment, no matter
how elegantly designed and executed, cannot stand alone. As discussed above,
detailed observational knowledge of a system is essential before field experiments
can be properly interpreted.

This brings us to a second need: long-tern field studies. Because
(1) most coral reefs are located far from universities and other research

institutions, (2) most researchers cannot spend large blocks of time away from
home, and (3) travel costs are becoming prohibitive, few current studies have
followed long-term variations in reef systems. Moreover, current policy dictates
that most research grants are limited to durations of several years at most.
The net result is that most ecological studies are relatively static "snapshots"
of intrinsically dynamic systems. Thus, our ability to understand the long -
range consequences of various ecological interactions, and especially major
environmental events (such as the urchin die-off discussed above), becomes
severely limited. Likens (1983) recently has characterized the establishment
and funding of long-term ecological studies in general to be a major priority
for the future. Perhaps reef systems will one day be included in the National
Science Foundation's Long-Term Ecological Research Program. In any event, the
establishment and funding of facilities for detailed studies in prescribed local
areas (such as NOAA's HYDROLAB in Salt River Canyon, St. Croix) or in local
regions (such as NOAA's MAKALI'I submersible program in Hawaii and Johnston
Island, the JOHNSON SEA-LINK submersible program in Florida and the Bahamas,
and the submersible now available at the Discovery Bay Marine Laboratory,
Jamaica) represent important first steps toward obtaining long-term data on
deep reef systems.

The third need is for future studies to include several study areas in
order to determine the ubiquity of observed patterns. Most present studies
occur at single sites, making widespread generalizations about coral reefs
based on a single study tenuous at best (although this fact rarely stops us).
This problem can be rectified by a single project being conducted at a number
of locations, either sequentially by a single research team or simultaneously
by several teams using standardized methods. Present controversies, such as
the importance of urchins as grazers discussed above, could be resolved by such
an approach. NOAAs proposal to establish a saturation facility which can be
moved among different geographical locations for comparative surveys of reef
biota and processes should facilitate standardized studies.

Unfortunately, enacting these last two proposals may require changes in
the current policies of granting agencies. The present system seems geared
toward a fast-turnover "results-now" policy. While programs such as HYDROLAB
provide the potential for long-tern studies of reef systems, few such facilities
exist presently, precluding detailed comparative studies among a number of
study sites. In any case, should the majority of reef ecologists concur that
these needs are important, long-term observational and experimental studies
over wide areas will be realized eventually.

ACKNOWLEDGMENTS

Many thanks to the 'U.C. Irvine Committee on Faculty Travel (especially
Peter Dixon) for funding my participation in the workshop; to the discussion
participants for providing' written summaries of their contributions (my sincere
apologies for any misquotes); to Marjorie Reaka and John Ebersole for constructive
comments on my manuscripts; and, especially, to Marjorie Reaka for bringing us
all together.

LITERATURE CITED

Connell, J. H. 1974. Field Experiments in Marine Ecology," pp. 21-54. In:
R. N. Mariscal (ed.), Experimental Marine Biology. Academic Press, N.Y.

Fitz, H. C., M. L. Reaka, E. B. Bermingham, and N. G. Wolf. 1983. "Coral
Recruitment at Moderate Depths: the Influence of Grazing," pp. 89-96. In:
M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral Reefs, Symp. SeTr.
Undersea Res., Vol. 1T1). NOAA Undersea Res. Progr., Rockville, Md.

Hay, M. E. 1984. "Patterns of Fish and Urchin Grazing on Caribbean Coral
Reefs: Are Previous Results Typical?" Ecology 65:446-454.

Hixon, M. A., and W. N. Brostoff. 1983. "Damselfish as Keystone Species in
Reverse: Intermediate Disturbance and Diversity of Reef Algae." Science
220:511-513.

Lessios, H. A., P. W. Glynn, and D. R. Robertson. 1983. "Mass Mortalities of
Coral Reef Organisms." Science 222:215.

Likens, G. E. 1983. "A Priority for Ecological Research." Bull. Ecol. Soc.
Amer. 64:234-243.

Ogden, J. C., R. A. Brown, and N. Salesky. 1973. "Grazing by the Echinoid
Diadema antillarum Philippi: Formation of Halos Around West Indian Patch
Reefs." Science 182:715-717.

Sammarco, P. W. 1980. "Diadema and Its Relationship to Coral Spat Mortality:
Grazing, Competition, and Biological Disturbance." J. Exp. Mar. Biol. Ecol.
45:245-272.

Shulman, M. J., J. C. Ogden, J. P. Ebersole, W. N. McFarland, S. L. Miller, and
N. G. Wolf. 1983. "Priority Effects in the Recruitment of Juvenile Coral
Reef Fish." Ecology 64:1508-1513.

Steneck, R. S. 1983. "Quantifying Herbivory on Coral Reefs: Just Scratching
the Surface and Biting Off More Than We Can Chew," pp. 103-112. In: M. L.
Reaka (ed.), The Ecology of Deep and Shallow Coral Reefs, Symp. Ser. Undersea
Res., Vol. 1( 1. NOAA Undersea Res. Progr., Rockville, Md.

Wolf, N. G., E. B. Bermingham, and M. L. Reaka. 1983. "Relationships Between
Fishes and Mobile Benthic Invertebrates on Coral Reefs." pp. 69-78. In:
M. L. Reaka (ed.), The Ecologqy of Deep and Shallow Coral Reefs, Symp. Ser.
Undersea Res., Vol.-T1l). NOAAndersea-es. Progr., Tockville, Md.

Note added in proof: A recent update on the mass mortality of urchins can be
found in:

Lessios, H. A., D. R. Robertson, and J. D. Cubit. 1984. "Spread of Diadema
Mass Mortality Through the Caribbean." Science 226:335-337.

SOME THOUGHTS ON THE PAST, PRESENT, AND FUTURE OF STUDIES
ON CORAL REEF COMMUNITY METABOLISM

S. V. Smith
Hawaii Institute of Marine Biology
Kaneohe, Hawaii 96744

Marjorie Reaka, as editor of these two volumes on The Ecology of Deep
and Shallow Coral Reefs, has kindly offered me the opportunit tocment on
where I see studies orreef community metabolism to be coming from--and going
to. The following thoughts arise as I consider, in particular, the seven
papers loosely grouped in "The Organization of Reef Ecosystems" in these
proceedi ngs.

These papers have some aspect of the metabolic performance of coral reef
communities or systems as a broad common theme. Within the context of that
theme, all of the papers can trace their ancestry back to the "foundation
studies" by Sargent and Austin (1949) at Rongelap Atoll and Odum and Odum (1955)
at Enewetak Atoll. Earlier papers in the field of reef community metabolism
can, of course, be identified. They have tended to remain lost in obscurity
compared to the two I have cited. Over the intervening years, the field has
matured and expanded sufficiently that citation of these important papers is
beginning to drop off. Instead, I recognize increasing proportions of second
(or third) generation citations, both in the papers on reef metabolism in these
* ~volumes and in the general literature.

These early papers dealt primarily with oxygen and carbon flux of reef-
flat communities. Several of the papers in the present volumes and a survey
of the literature suggest that we can move beyond a direct concern with average
rates of oxygen and carbon fluxes in studies of reef metabolism. We now can
speak with some confidence about the typical oxygen and carbon metabolic
performance of reef-flat communities. It is, however, worth noting that we
remain without a comparable large data base for other parts of coral-reef
ecosystems.

We are beginning to gain an appreciation for the relationships between
variation in reef metabolism and the controlling roles of external physical,
chemical, and biological variables. Many casual observations about reef
metabolic response to external controls have led to erroneous or trivial
conclusions. Clearly, controlled experiments and carefully designed field
surveys need to be conducted in order to test hypotheses about metabolic
responses to controlling variables.

The study of nutrient fluxes in relation to community metabolism has,
perhaps, been more extensive than other aspects of controls on community
metabolism. This work became prominent during the Symbios Expedition to Enewetak
Atoll (Johannes, et al., 1972). Studies of nutrient flux continue, and most of
the papers on ree~f m tbolism in these volumes at least touch on this subject.
in particular, there is an interest in the effects of both quantity and quality
* ~of nutrient supply on reef metabolism and in the interactive roles of various
communities (especially macrofauna/meiofauna/microfauna in interreef soft-

sediment communities) in nutrient processing of the total system. From my own 1
bias, I perceive the need to "balance the books" of nutrient fluxes in entireW
reef systems as a recurrent--and important--theme.

Solar radiation is clearly fundamental to the community metabolism of
reefs, yet our understanding of metabolic responses to light remains poor.
Traditionally, most studies of coral reef metabolism have not adequately
considered either the quantity or the quality of incident solar radiation,
light attenuation through the water column, or loss of radiant energy through
back-scattering. This situation is beginning to change. We see among the
papers in these volumes several attempts to evaluate the metabolic response of
reefs to light. More work is needed.

Finally, I observe that students of coral reefs tend to emphasize the
unique characteristics which make reefs such pleasurable places to work. I
suggest that we have a great deal more to learn, as ecologists, by considering
coral reefs within the spectrum of other ecosystems than we will ever learn by
treating reefs as unique.

LITERATURE CITED

Johannes, R. E., and the Project Symbios Team. 1972. "The Metabolism of Some
Coral Reef Communities: A Team Study of Nutrient and Energy Flux at Eniwetok."
BioScience 22:541-543.

Odum, H. T., and E. P. Odum. 1955. "Trophic Structure and Productivity of a
Windward Coral Reef Community on Eniwetok Atoll." Ecol. Monogr. 25:291-320.

Sargent, M. C., and T. S. Austin. 1949. "Organic Productivity of an Atoll."
Trans. A~mer. Geophys. Union 30:245-249.

CHAPTER III. CONTRIBUTED PAPERS

Figure 1

Further studies of deep coral reefs and sea mounts have been possible
through the use of submersibles such as NOAA's MAKALI'I, which is shown while
in transit on its launch, retrieval, and transport (LRT) vehicle (fig. 1).
During launch, the LRT is submerged with the submersible; at 50 feet the
submersible is released to continue its descent for research (fig. 2).

Figure 2

Figure 3

Figure 4

NOAA also supports research on the biology of deep water communities in
various localities through the deep diving bell facilities operated off the
R/V SEAHAWK (figs. 3, 4) which is based at the University of North Carolina
at Wilmington. An additional habitat is being constructed for use in research
on marine communities off the southern California coast (fig. 5).

Figure 5

RECRUITMENT OF YOUNG CORAL REEF FISHES FROM THE PLANKTON

William N. McFarland
Section of Ecology and Systematics
Cornell University
Ithaca, New York 14853

John C. Ogden
West Indies Laboratory
Fairleigh Dickinson University
Teague Bay
St. Croix, U.S. Virgin Islands 00820

ABSTRACT

This paper examines the widely held view that the planktonic eggs and/or
larvae of most coral reef fishes represent mechanisms that reduce predation
and/or serve for dispersal of the young. Data on spawning, the pelagic phase,
and recruitment of coral reef fishes are examined. Recruitment represents the
successful end product of a complex suite of adult and larval behaviors and
physical conditions that directly affect survival of offspring. Although
considerable data are available, in no single species has the cycle from spawning
to recruitment been clearly documented. As a result, assessing the various
hypotheses that attempt to explain the almost universal presence of a pelagic
larval phase in coral reef fishes remains tenuous.

Predation on young and ultimate dispersal are both potent selective factors.
The high degree of variation in reproduction, larval characteristics, and
recruitment amongst coral reef species implies, however, that these two selective
factors undoubtedly vary in importance among species.

INTRODUCTION

Virtually all coral reef fishes produce numerous offspring and release
either their eggs or larvae into the offshore plankton (Breder and Rosen, 1966;
Ehrlich, 1975; Goldman and Talbot, 1976; Sale, 1977, 1978a; Johannes, 1978;
Barlow, 1981). There is one known exception, the pomacen-trid Acanthochromis
polyacanthus; members of this species brood their young and disperse them
directly onto the reef (Robertson, 1973). Also, some apogonids (mouth brooders)
and some brotulids (live bearers) have completely eliminated the planktonic
larval phase.

Considering the high diversity of reef fishes (Connell, 1978; Sale, 1980)
and their numerous ecological roles (Hobson, 1974; Sale, 1978b; Smith, 1978),
it is astonishing that so many species use the same reproductive strategy.
Indeed, the central position of a larval pelagic existence in coral reef fishes
led Helfman (1978) to wonder whether answers to the high species diversity of
coral reef fish assemblages "lie in the plankton." One major hypothesis favors
the view that community composition is dependent on the chance recruitment of
each species from the plankton when space becomes available (Russell, et al.,

1974; Sale and Dybdahi , 1975, 1978; Sale, 1977, 1978a). Thus, species compositI
on a reef should change through time, but over years as opposed to centuriesV
or longer. A contrasting hypothesis suggests that species composition is more
deterministic and dependent on competition between adults for resources (Smith,
1973, 1978; Smith and Tyler, 1973). Recruitment from the plankton, therefore,
would not primarily affect diversity but the abundance of each species over
time. No doubt there is truth in each view, but the controversy continues with
additional participants (Gladfelter and Gladfelter, 1978; Holles, 1978; Talbot,
et al., 1978; Gladfelter, et al., 1980; Ogden and Ebersole, 1981; Shulman,
l98TV-Abrams , 1984).

Given the virtually universal use of a pelagic egg and/or larval phase and
its potential significance to explanations of the high diversity amongst
assemblages of coral reef fishes, investigations of the reproductive activities
of adults, the disposition of their eggs and larvae at sea, and the effect(s)
of currents on the dispersal of these propagules are, therefore, not inconsequen-
tial. If, for example, the recruitment of each species is not consistent over
a reasonable time scale, then the high diversity of coral reef fishes could not
be maintained.

In general, data about recruitment processes in coral reef fishes are
sparse and must often be gleaned from literature that primarily deals with
other aspects of coral reef fish biology. Over the last few years, however,
efforts have been directed toward recruitment in particular species. Given the
central importance of recruitment to coral reef fish community structure, we
expect that larval biology at sea will be explored in great detail over the __
coming decade. Here we summarize what is presently known about recruitment andV
closely related processes in coral reef fishes.

REPRODUCT ION

Spawning- activities in coral reef fishes vary considerably. Only over the
last 10 years have sufficient direct observations accumulated to provide a
basis for understanding how the spawning act might be adaptive. Many reef
species spawn over long periods of each year, but they usually show maximum
reproduction in either spring and/or fall (Munro, et al.,, 1973). In Munro's
study, which examined the state of the gonad as annde~x of reproductive
potential, over half of the species were active for longer than 6 months.
Labrids (Warner, et al., 1975) and somne haemulids (McFarland, 1980, 1982) spawn
over the entire y`iar-,-ut many species have highly restricted spawning seasons.
Between these extremes, however, some species produce all of their eggs in a
single or in very few reproductive acts, while others rely on a more consistent
release which involves the release of fewer eggs during each reproductive
bout. It is difficult to generalize with certainty how these highly varied
spawning tactics relate to dispersal and recruitment. This is further complicated
by the fact that virtually every assemblage of coral reef fishes contains
species that utilize each reproductive tactic.

Many species tend to show periodicity of spawning over a season or even
longer. In a survey of reproduction in 52 tropical fishes, ,Johannes (1978) 1

found that 20 tended to spawn near or at new moon, 8 at full moon, 13 at times
of both new and full moon, and the remaining 11 during quarter moons.

Several direct observations have related the timing of spawning to specific
phases of the moon (e.g., quarter moons, Polydactylus sexfilis, May, et al.,
1979; Centropyge potteri, Lobel, 1978). The periodicity in spawning has been
interpreted as an adaptation to enhance the chance that the eggs will be flushed
by tidal currents away from the reef, and therefore reduce mortality by egg
predators, which is commonly observed (Randall and Randall, 1963; Moyer, 1974;
Meyer, 1977; Robertson and Hoffman, 1979; Colin, 1978; Robertson, 1983).

Mortality among the pelagic eggs and larvae is high when compared to the
adults (Sale, 1980). Very likely as a response to this high mortality (but
perhaps also to increase the probability that some propagules will be dispersed)
the fecundity of reef fishes is high, usually in excess of 50,000 eggs per year
per female in pelagic spawners (Randall, 1961; Bryan, et al., 1975; Thompson
and Munro, 1978), and greater than 8,000 eggs in demersal spawners (Warner, et
al., 1975; Bell, 1976).

In addition to varied lunar periodicities, many reef fishes migrate to the
deeper areas of the reefs and/or to pinnacles and reef extensions to spawn, and
there usually invoke discrete behaviors when spawning (e.g., a rush to the
surface to spawn the eggs as far from the reef as possible; Lobel, 1978; Colin,
1982). In addition, spawning often shows a daily cycle (e.g., occurring at
dusk; Lobel, 1978). These behaviors have been interpreted as acts to reduce
egg predation and may be particularly effective because eggs are small,
transparent, and therefore difficult to detect, especially at dusk (Hobson and
Chess, 1976). Many species, however, spawn during midday or the afternoon
(Colin, 1982; Robertson, 1983).

Although the observations on spawning behavior suggest tactics that can
reduce mortality of eggs and improve dispersal away from the reef, few observations
are supported by direct measurements of prevailing tides, currents, and other
conditions to verify that this is what actually happens. A recent paper by
Robertson (1983) combines careful field observations of spawning in seven
species of surgeonfishes at four differen' localities in the Pacific and Indian
Oceans. Readers are referred to the original for the wealth of observations
detailed by Robertson. Several general conclusions, however, are important.
(1) The various species spawned over different intervals of the day (a few in
the morning; most from midday until dusk). (2) Both paired spawnings and group
spawnings occurred in several of the species. (3) Except where territorial
harems were established, most individuals migrated toward outer and deeper
portions of the reefs to spawn. (4) In. almost all instances, spawning fishes
rose toward the surface to spawn. (5) The most common feature among the seven
species was a strong tendency to spawn during ebb tides, as water was flushed
off the reef. All of these-behaviors would minimize the chance that freshly
fertilized eggs would be exposed to planktivorous reef fishes. Spawning as
tides flush water off reefs also was indicated by Randall (1961), Robertson and
Choat (1974), Lobel (1978), and Thresher (1979). In contrast, Robertson (1983)
found no distinct tidal rhythms in surgeonfish spawnings at Lizard Island where
the currents flow parallel to the reef rather than on and off the reef. This

finding suggests that spawning in a species is strongly influenced by local __
conditions. Thus, behaviors that in one location may reduce egg mortality U
might increase egg mortality if repeated precisely at another site.

In summary, reproductive efforts in a species often can be shown to diminish
predation on spawned eggs and enhance the chance of their dispersal. But, many
species show variable spawning behavior. It is therefore uncertain that all
reproductive behaviors are selected to reduce egg predation and/or assure
dispersal.

DISPERSAL -- THE PELAGIC PHASE

Hypotheses to Explain the Pelagic Larval Phase

Why should almost all coral reef fishes cast their eggs and/or larvae
offshore? Three primary hypotheses have been suggested: (1) the antipredator
hypothesis, (2) the food hypothesis, and (3) the dispersal hypothesis (see
Johannes, 1978, for review). Jdhannes, who favors the antipredation hypothesis,
argues that if the eggs and larvae were left on the reef unprotected, they
would be subject to an excessive mortality by reef predators (see also Smith,
1978). Although it is certain that predation on eggs and larvae is very high
by reef planktivorous fishes (Hobson and Chess, 1978) and by sessile filter
feeders (Johannes, 1978), it is not clear how much predation is reduced when
eggs and larvae are offshore (Barlow, 1981). Egg mortality offshore can approach
30% per day (Jones and Hall, 1974; Ware, 1975). Considering the relative merit~
of these factors and others led Barlow (1981) to postulate that the reduction U
in predation achieved from a pelagic phase was of less significance to survival
than the advantages that are derived from dispersal (over both short and long
distances). Barlow sees dispersal as a means to ensure survival of a species
by reducing inbreeding (Bengtsson, 1978), and, particularly, as an adaptation
to the instability of reef communities (Connell, 1978; Sale, 1978a). He argues
that the lor~ger an egg or larvae travels, the less its chance of survival
(Thorson, 1946; Gadgil, 1971). For long-range dispersal to succeed, reef fishes
must produce numerous offspring, which they do, and larvae must be adapted for
a lengthened pelagic life, which in many species is true (e.g., in surgeonfishes,
butterflyfishes, wrasses, and in eels; Barlow, 1981). The possibility exists,
nevertheless, that cohorts fromi a single spawning may remain together as eggs
and larvae in the plankton, and thus maintain genetic relatedness among recruits
(see Shapiro, 1983). This interesting possibility, which would increase inbreeding
success, remains to be documented.

The food hypothesis is dismissed by both Johannes (1978) and Barlow (1981)
because in general planktonic food is less available offshore than inshore;
thus, the advantage of offshore disperal must find another explanation, such as
a reduction in predation or enhanced dispersal.

Although Johannes and Barlow are aware that both recolonization and long-
range dispersal occur, their respective arguments favor different views of the
role of dispersal in larval recruitment from the plankton. Barlow (1981) views
the long-range transport of larvae to other localities than the parent reef
(island, etc.) as the essential selective force that has operated. In contras*

Johannes (1978) emphasizes avoiding predation and invokes data (largely
anecdotal) to suggest that eggs are cast into local current gyres that often
return to the same island or, at least, to the same general locality and not
necessarily to far distant shores. There were some data to suggest that this
was so when Johannes (1978) postulated the role of local currents in recolonization
(Jones, 1968; Sale, 1970; Watson and Leis, 1974; Lobel and Reaka, 1977; and
since then Leis, 1982). More convincing data have now been provided by Lobel
and Robinson (1983) that current gyres do develop off the big island of Hawaii
twice each year and, importantly, coincide with the two major seasonal spawning
peaks for many of the coral reef fishes in Hawaii (Jones, 1968; Sale, 1970).
Lobel and Robinson, using radio-tracked drogues, found circulation times of 7-
14 days (offshore and return inshore) and evidence for at least two cycles.
Thus, there would be time for eggs and larvae to develop to a sufficient size
to settle onto the reef after initially being advected away from the reef.
However, it should be emphasized that in multiple releases of drogues, some
returned close to the inshore release site, but others drifted off and even
headed toward the island of Molokai. Furthermore, as Barlow (1981) emphasizes,
even if major seasonal spawnings at Hawaii do coincide with the weakest offshore
currents and the presence of local gyres (Watson and Leis, 1974) which might
favor recolonization, spawning nevertheless continues in many species at other
times of the year when currents would not favor recolonization, but rather long-
range dispersal.

Duration of Larval Life

Certainly if long-range dispersal is to be invoked, as for example Barlow
(1981) suggests occurs for those coral reef species that have managed to cross
the Indo-Pacific barrier (see Rosenblatt and Walker, 1963), then the length of
larval life must be long. An alternative tactic for long-range dispersal would
invoke rafting of juveniles or adults, a requirement if the larval phase is of
short duration (< 20-30 days). How long, then, are larvae advected at sea?

By comparing spawning dates with recruitment dates, larval duration has
been estimated to vary among coral reef species from a week to several months
(Randall, 1961; Moran and Sale, 1977; Johannes, 1980). Over the last few years,
the otolith daily aging technique has allowed direct estimation of the age of
individual larvae (Brothers and McFarland, 1981). The results to date indicate
that larval duration does vary from weeks to months in different species
(Ralston, 1976; Brothers, et al., 1983; Victor, 1983a; and Brothers and Thresher,
1985). The duration withina species is often variable, as it is in marine
invertebrates (Scheltema, 1968, 1971), but some species have very restricted
lengths of larval life, the cause of which is unknown (Brothers and McFarland,
1981). For the 115 species of Pacific coral reef fishes studied, Brothers and
Thresher (1985) obtained otolith ages for settlement (the so-called settlement
check mark; see Brothers and McFarland, 1981; Brothers, et al., 1983; Victor,
1983b) that range from 15 to 82 days. In some labrids, it can be a week (Victor,
1983b), and in grunts settlement always occurs at about 15 days (McFarland,
1982T. These pelagic periods favor only short-range dispersal. But, some
labrids can have larval durations up to 50 days (Brothers, et al., 1983), which
would allow for long-range dispersal.

Larval Behavior

Another factor that can affect larval mortality and dispersal is the
behavior of fish larvae at sea. It is generally conceded (although with little
direct evidence) that fish larvae disperse passively in offshore currents (Sale,
1970; Leis and Miller, 1976) and, therefore, that their final destinations are
subject to hydrographic conditions. Alternatively, it has been argued that
larval drift is not passive (Sale, 1980). In Hawaii, larvae from species with
pelagic eggs (labrids, chaetodontids, acanthurids) are more common offshore
than species with demersal eggs (pomacentrids and gobies), which are more common
inshore (Miller, 1974; Leis and Miller, 1976; Watson and Leis, 1974). This is
explained by invoking active larval mechanisms. However, it may also be
explained merely by initial passive drift of pelagic eggs. Nevertheless, once
fish larvae begin to swim, vertical movements could place them in currents that
would hold them inshore or advect them away. Unfortunately very little is
known about larval behaviors. Several investigations have established that
many pelagic invertebrates undertake marked diel vertical migrations, ascending
toward the surface at night and to deeper depths during the daytime (see Segal,
1970, for overview). Recent investigations also have revealed that reef
meroplankton emerge high into the water column at night, especially during the
dark phase of the moon (Porter, et al., 1977; Porter and Porter, 1977; Alldredge
and King, 1977, 1980; Hobson and-Chess, 1979; Robichaux, et al., 1981; Ohlhorst,
1982, 1985). These vertical movements take place over a T-i -meters. Similar
vertical distributions seem to occur in coral reef fish larvae, at least nearshore
(Hobson and McFarland, preliminary unpubl. data). It is known, however, that
reverse tidal currents can occur over shallow depths, as they do at St. Croix
(Lee, et al., 1977), and it is possible, even likely, that fish and invertebrate
larvae'-si-hese currents, in combination with vertical migratory behavior, in
their dispersal.

Certainly broad parallels occur between benthic marine invertebrates and
fishes. In tropical regions, 80-85% of the invertebrates produce long lived,
planktotrophic larvae that will hatch, as in fishes, from relatively small eggs
(Thorson, 1950). In general, larval pelagic existence is longer in tropical
invertebrates than in species from higher latitudes (see Thorson, 1950;
Mileikovsky, 1971; and Jablonski and Lutz, 1983, for review). Extended larval
periods appear to enhance dispersion and are associated with broad geographic
distributions in tropical invertebrates (Scheltema, 1968, 1971), and this
pattern seems, at least tentatively, to also hold for coral reef fishes (Brothers
and Thresher, 1985).

In summation, there is no simple explanation in terms of dispersal versus
antipredatory mechanisms that resolves what major selective processes have
caused virtually all coral reef fishes to produce pelagic propagules. In fact
the data cited show such-wide variation that it is likely that in some species
dispersal was the preeminent selective factor, while in other species antipreda-
tion was the most important. In any case, dispersal and antipredation are not
mutually exclusive selective factors, and certainly both act to increase
survival. Nevertheless, it is difficult not to wonder, as Barlow (1981) has,
why more coral reef fishes do not invest in postzygotic activities to reduce
the duration of the larval pelagic phase, or even eliminate it. i

Barlow suggests that postzygotic activity leads to a reduction in lifetime
fecundity in part because of the parental investment required. Most coral reef
fishes that spawn demersal eggs tend to be small, a circumstance that in itself
limits egg production. In addition, most coral reef species are strongly site-
attached and spend their adult lives in a relatively small local area (Sale,
1980). Pelagic dispersal thus can be considered a necessary consequence of
their restricted mobility. If, for example, specific reef habitats tend to be
unstable over time, a circumstance that recent events document (e.g., severe
hurricane damage of reef habitat at Jamaica, Woodley, et al., 1981; Kaufman,
1983; storm effects at Lizard Island, Lassig, 1983; coral-Tight in the Caribbean
and Pacific, Lessios, et al., 1983), then dispersal to colonize adjacent or
distant habitats is an essential element in each species' survival.

RECRUITMENT--HOW DO FISHES SETTLE ON REEFS?

In a review of coral reef fish ecology, -Sale (1980) stated, "The mechanisms
whereby larval fishes make their return to the reef are totally unknown." In
addition, it is only over the last few years that more systematic observations
of recruitment in reef fish larvae have begun to appear in the literature. For
most of the observations, maximum recruitment occurs over a single season and
often is episodic; it may follow a lunar periodicity (Johannes, 1978); it may
be monthly but not coupled to a particular phase of the moon (mixed guilds of
pomacentrids; Williams, 1983), or a rhythm may not be conspicuous (Thalassoma
bifasciatum; Victor, 1982, 1983a).

Larval French grunts (Haemulon flavolineatum) settle from the offshore
plankton following a semi-lunar timetable. This periodicity is most closely
correlated with the quarter moons and/or daily intermediate tidal excursions
that fall between the spring and neap tides (McFarland, et al., 1985; also see
McFarland, 1982). Furthermore, although seasonal, some recruitment continues
throughout the year. Larval life is short (ca. 15 days; Brothers and McFarland,
1981), which implies that spawning also must occur throughout the year; also
ripe eggs are present in adult French grunts at all times of the year (Munro,
et al., 1973).

The extreme variations seen in the spawning periodicities of reef fishes
(see earlier section) are reflected in the variable periodicity in recruitment.
What do larvae select when they settle from the plankton to the reef? Here we
are almost totally in ignorance. Most larval (or postlarval) recruits are
first observed on substrate (coral patches, seagrass beds, sand, etc.). Usually,
it is not possible to state that they have just settled. In a study of recruiting
pomacentrids, Williams and Sale (1981) found that different species tended to
prefer different species of corals. Earlier Sale (1969) found that young
surgeonfish preferred reefs with shaded overhangs. It is generally understood
that each species has certain needs for settlement, but determination of these
needs has barely been initiated. In addition, settlement to a specific site
can be dramatically influenced by resident competitors (Shulman, et al., 1983).

Actual settlement has seldom been observed. During observations of the
acronurus larvae of Acanthurus triostegus in Hawaii, Randall (1961) concluded

that they settle at night. We have observed the same nocturnal timing for
acronurus larvae on settlement reefs at St. Croix. Recruiting young
pomacentrids also seem to appear overnight (Doherty, 1981). But daytime
settlement is not precluded. Given the present status of our knowledge of
larval fish behaviors, it is difficult to envision how a larval fish avoids
settling if by chance it is carried across suitable substrate during the daytime;
but, in contrast, how does a larval fish "know" that it should settle at night?
Are cues visible, olfactory, or, perhaps, sonic? Certainly, planktivorous reef
fishes are less active at night (Hobson, 1974; Hobson, et al., 1981-, Gladfelter
and Johnson, 1983), but filter-feeding sessile invertebrate~s are most active at
night (Sebens and DeReimer, 1977). Obviously, more precise studies are required
to evaluate just how important nocturnal settlement is to survival.

It should be emphasized that recruitment of larvae varies considerably
from year to year (Williams and Sale, 1981; Williams, 1983). As Williams
states, "One large pulse alone is sufficient to make total recruitment in a
year unusually successful." He reminds us as well that a large recruitment of
one species does not mean the same for another species, even in the same guild.

Finally, it has been implicit in the various arguments of recent fish
biologists interested in explaining the diversity in coral reef fish assemblages
that the offshore plankton provides a huge, although variable, reservoir of
potential recruits which is drawn upon when appropriate space opens on the reef
(Dale, 1978; Sale, 1978a; Smith, 1978). Perhaps this is so, but Doherty (1981)
suggests that at least at certain times the larval pool may not be endless and
can be limiting to the maintenance of reef populations. His theoretical model
assumes that larval existence for most species is "precarious" at best (see Is
also Johannes, 1978), and this limiting function introduces a high degree of
chance into the recruitment process.

CONCLUSIONS

From what is currently known about the recruitment of coral reef fishes
from the plankton, it seems likely that strong interactions between local
currents and the times of spawning exist. The functions) of this interaction
is best explained as a mechanism to both remove eggs and larvae from the vicinity
of the reef to reduce predation and to disperse the young to other localities.
The possibility of seasonally developed recirculating eddies could lead to
local recolonization. Even totally isolated islands could sustain their fish
populations by holding larvae in downstream eddies (Emery, 1972), especially if
discrete larval behaviors are used to remain in the eddies. What is missing is
an unambiguous case that this routinely occurs.

ACKNOWLEDGMENTS

Our research as reported in this paper was supported by NSF Research Grant
OCE-79-18569. This paper is Contribution Number 128 of the West Indies
Laboratory, St. Croix, U.S.V.. We thank Drs. John Ebersole and Marjorie Reaka
for their many critical and constructive comments during the preparation of the
manuscri pt.

LITERATURE CITED

Abrams, P. A. 1984. "Recruitment, Lotteries, and Coexistence in Coral
Reef Fish." Amer. Nat. 123:44-55.

Alidredge, A. L., and J. M. King. 1977. "Distribution, Abundance, and Substrate
Preferences of Demersal Reef Zooplankton at Lizard Island Lagoon, Great Barrier
Reef." Mar. Biol. 41:317-333.

Barlow, G. W. 1981. "Patterns of Parental Investment, Dispersal and Size
Among Coral-Reef Fishes." Environ. Biol. Fish. 6:65-85.

Bell, L. J. 1976. "Notes on the Nesting Success and Fecundity of the Anemone
Fish Amphiprion clarkii at Miyake-Jima, Japan." Jap. J. Ichthyol. 22:207-211.

Bengtsson, B. O. 1978. "Avoiding Inbreeding: At What Cost?" J. Theor. Biol.
73:439-444.

Breder, C. M., and D. E. Rosen. 1966. Modes of Reproduction in Fishes.
Natural History Press, Garden City, New York. 941 pp.

Brothers, E. B., and R. E. Thresher. 1985. "Pelagic Duration, Dispersal and
the Distribution of Indo-Pacific Coral-Reef Fishes," pp. 53-69. In: M. L.
Reaka (ed.), The Ecology of Coral Reefs, Symp. Ser. Undersea Res., Vol.
3(1). NOAA Undersea Res. Progr., Rockville, Md.

Brothers, E. B., D. McB. Williams, and P. F. Sale. 1983. "Length of Larval
Life in Twelve Families of Fishes at 'One Tree Lagoon,' Great Barrier Reef,
Australia." Mar. Biol. 76:319-324.

Bryan, P. G., B. B. Madraisau, and J. P. McVey. 1975. "Hormone Induced and
Natural Spawning of Captive Siganus canaliculatus (Pisces: Siganidae) Year
Round." Micronesica 11:199-204.

Colin, P. L. 1978. "Daily and Summer-Winter Variation in Mass Spawning of the
Striped Parrotfish, Scarus croicensis." Fish. Bull. 76:117-124.

Colin, P. L. 1982. "Aspects of the Spawning of Western Atlantic Reef Fishes,"
pp. 69-78. In: G. R. Huntsman, W. R. Nicholson, and W. W. Fox (eds.), The
Biological Bases for Reef Fishery Management. NOAA Tech. Memo NMFS-SEFC--.
Nat. Mar. Fish. Ser., eaufort, N.C.

Connell, J. H. 1978. "Diversity in Tropical Rain Forests and Coral Reefs."
Science 199:1302-1310.

Dale, G. 1978. "Money-In-The-Bank: A Model for Coral Reef Fish Coexistence."m
Environ. Biol. Fish. 3:103-108.

Doherty, P. J. 1981. "Coral Reef Fishes: Recruitment-Limited Assemblages?"
Proc. 4th Int. Coral Reef Symp. 2:465-470.

Ehrlich, P. R. 1975. "The Population Biology of Coral Reef Fishes." Ann.
Rev. Ecol. Syst. 6:211-247.

Emery, A. R. 1972. "Eddy Formation From an Oceanic Island: Ecological
Effects." Carib. J. Sci. 12:121-128.

Gadgil, M. 1971. "Dispersal: Population Consequences and Evolution." Ecology
52:253-261.

Gladfelter, W. B., and E. H. Gladfelter. 1978. "Fish Community Structure as a
Function of Habitat Structure on West Indian Patch Reefs." Rev. Biol. Trop.
26 (Suppl.):65-84.

Gladfelter, W. B., and W. S. Johnson. 1983. "Feeding Niche Separation in a
Guild of Tropical Fishes (Holocentridae)." Ecology 64:552-563.

Gladfelter, W. B., J. C. Ogden, and E. H. Gladfelter. 1980. "Similarity and
Diversity Among Coral Reef Fish Communities: A Comparison Between Tropical
Western Atlantic (Virgin Islands) and Tropical Central Pacific (Marshall
Islands) patch reefs." Ecology 61:1156-1168.

Goldman, B., and F. H. Talbot. 1976. "Aspects of the Ecology of Coral Reef
Fishes," pp. 125-154. In: O. A. Jones and R. Endean (eds.), Biology and
Geology of Coral Reefs, Vol. 3, Biology II. Academic Press, New York.

Helfman, G. S. 1978. "Patterns of Community Structure in Fishes: Summary and
Overview." Environ. Biol. Fish. 3:129-148.

Hobson, E. S. 1974. "Feeding Relationships of Teleostean Fishes on Coral
Reefs in Kona, Hawaii." Fish. Bull. 72:915-1031.

Hobson, E. S., and J. R. Chess. 1976. "Trophic Interactions Among Fishes and
Zooplankters Near Shore at Santa Catalina Island, California." Fish. Bull.
74:567-598.

Hobson, E. S., and J. R. Chess. 1978. "Trophic Relationships Among Fishes and
Plankton in the Lagoon at Enewetak Atoll, Marshall islands." Fish. Bull.
76:133-153.

Hobson, E. S., and J. R. Chess. 1979. "Zooplankters That Emerge From the Lagoon
Floor at Night at Kure and Midway Atoll, Hawaii." Fish. Bull. 77:275-280.

Hobson, E. S., W. N. McFarland, and J. R. Chess. 1981. "Crepuscular and
Nocturnal Activities of Californian Nearshore Fishes, With Consideration of
Their Scotopic Visual Pigments and the Photic Environment." Fish. Bull.
79:1-30.

Jablonski, D., and R. A. Lutz. 1983. "Larval Ecology of Marine Benthic
Invertebrates: Paleobiological Implications." Biol. Rev. 58:21-89.

Johannes, R. E.. 1978. "Reproductive Strategies of Coastal Marine Fishes in
the Tropics." Environ. Biol. Fish. 3:65-84.

Johannes, R. E. 1980. "Using Knowledge of the Reproductive Behavior of Reef
and Lagoon Fishes to Improve Fishing Yields," pp. 247-269. In: J. E.
Bardach, J. J. Magnuson, R. C. May, and J. M. Reinhart (eds.3, Fish Behavior
and Its Use in the Capture and Culture of Fishes. Proc. 5th Int. Center
MMing-Auat. Resource Managemt., ManiT-, Phlippines. 512 pp.

Jones, R. S. 1968. "Ecological Relationships in Hawaiian and Johnston Island
Acanthuridae (Surgeon Fishes)." Micronesica 4:309-361.

Jones, R., and W. B. Hall. 1974. "Some Observations on the Population Dynamics
of the Larval Stage in the Common Gadoids,"- pp. 87-102. In: J. S. Blaxter
(ed.), The Early Life History of Fish. Springer-Verlag, New York.

Kaufman, L. S. 1983. "Effects of Hurricane Allen on Reef Fish Assemblages."
Coral Reefs 2:43-47.

Lassig, B. R. 1983. "The Effects of a Cyclonic Storm on Coral Reef Fish
Assemblages." Environ. Biol. Fish. 9:55-63.

mw Lee, T. N., R. S. C. Monier, and S. Chiu. 1977. Water Mass Structure and
Variability North of St. Croix, U.S. Virgin Islands, AsObserved During the
Summer of 197770r7Tlr-Assessment. Tech. Rept. UM-RSMAS #78004. Univ. -TF
Miam, F-oFida. 81 pp.

Leis, J. M. 1982. "Nearshore Distributional Gradients of Larval Fish (15
Taxa) and Planktonic Crustaceans (6 Taxa) in Hawaii." Mar. Biol. 72:89-97.

Leis, J. M., and J. M. Miller. 1976. "Offshore Distribution Patterns of
Hawaiian Fish Larvae." Mar. Biol. 36:359-367.

Lessios, H. A., P. W. Glynn, and D. R. Robertson. 1983. "Mass Mortalities of
Coral Reef Organisms." Science 222:118.

Lobel, P. S. 1978. "Diel, Lunar, and Seasonal Periodicity in the Reproductive
Behavior of the Pomacanthid Fish, Centropyge potteri, and Some Other Reef
Fishes in Hawaii." Pac. Sci. 32:193-207.

Lobel, P. S., and M. L. Reaka. 1977. "Synchronization of Reproduction With
Seasonal Currents as a Mechanism for Endemism in Hawaiian Marine Fauna."
Unpublished manuscript cited in Lobel, 1978.

Lobel, P. S., and A. R. Robinson. 1983. "Reef Fishes at Sea: Ocean Currents
and the Advection of Larvae," pp. 29-38. In: M. L. Reaka (ed.), The Ecology
of Deep and Shallow Coral Reefs, Symp. Ser'.Undersea Res., Vol. 1(T NOAA
s 5Undersea Res. Progr., ockville, Md.

May, R. C., G. S. Akiyama, and M. T. Santerre. 1979. "Lunar Spawning of the
Threadfin, Polydactylus sexfilis, in Hawaii." Fish. Bull. 76:900-904.

McFarland, W. N. 1980. "Observations on Recruitment in Haemulid Fishes."
Proc. Gulf Carib. Fish. Inst. 32:132-138.

McFarland, W. N. 1982. "Recruitment Patterns in Tropical Reef Fishes," pp. 83-
91. In: G. R. Huntsman, W. R. Nicholson, and W. W. Fox, Jr. (eds.), The
Biological Bases for Reef Fishery Management. NOAA Tech. Memo. NMFS-SEF-i-80,
Nat. Mar. Fish. Serv., eaufort, N.C.

McFarland, W. N., E. B. Brothers, J. C. Ogden, M. J. Shulman, E. Bermingham,
and N. M. Kotchian-Prentiss. 1985. "Recruitment Patterns in Young French
Grunts, Haemulon flavolineatum (Family Haemulidae) at St. Croix, U.S.V.I."
Fish. Bull. (in press).

Meyer, K. A. 1977. "Reproductive Behavior-and Patterns of Sexuality in the
Japanese Labrid Fish Thalassoma cupido." Jap. J. Ichthyol. 24:101-112.

Mileikovsky, S. A. 1971. "Types of Larval Development in Marine Bottom
Invertebrates, Their Distribution and Ecological Significance: A Re-
Evaluation." Mar. Biol. 10:390-404.

Miller, J. M. 1974. "Nearshore Distribution of Hawaiian Marine Fish Larvae:
Effects of Water Quality, Turbidity and Currents," pp. 217-231. In: J. H.
S. Blaxter (ed.), The Early Life History of Fish. Springer-Verlag, Berlin.

Molles, M. C. 1978. "Fish Species Diversity on Model and Natural Reef Patches:
Experimental Insular Biogeography." Ecol. Monogr. 48:289-305.

Moran, M. J., and P. F. Sale. 1977. "Seasonal Variation in Territorial
Response, and Other Aspects of the Ecology of the Australian Pomacentrid
Fish Parma microlepis." Mar. Biol. 39:121-128.

Moyer, J. T. 1974. "Notes on the Reproductive Behavior of the Wrasse Thalassoma
cupido." Jap. J. Ichthyol. 21:34-36.

Munro, J. L., V. C. Gaut, R. Thompson, and P. H. Reeson. 1973. "The Spawning
Season of Caribbean Reef Fishes." J. Fish. Biol. 5:69-84.

Ogden, J. C., and J. P. Ebersole. 1981. "Scale and Community Structure of
Coral Reef Fishes: A Long-Term Study of a Large Artificial Reef." Mar.
Ecol. Prog. Ser. 4:97-103.

Ohlhorst, S. L. 1982. "Diel Migration Patterns of Demersal Reef Zooplankton."
J.. Exp. Mar. Biol. Ecol. 60:1-15.

Ohlhorst, S. L. 1985. "Reef Zooplankton Collected Along a Depth Gradient at
Discovery Bay, Jamaica," pp. 101-116. In: M. L. Reaka (ed.), The Ecology
of Coral Reefs. Symp. Ser. Undersea Res., Vol. 3(1). NOAA Undersea Res.
PFogr., Rockville, Md.

Porter, J. W., and K. G. Porter. 1977. "Quantitative Sampling of Demersal
Plankton Migrating From Different Substrates." Limnol. Oceanogr. 22:553-556.

Porter, J. W., K. G. Porter, and Z. Batac-Catlan. 1977. "Quantitative Sampling
of Indopacific Reef Demersal Plankton." Proc. 3rd Int. Coral Reef Symp.
1:105-112.

Ralston, S. V. D. 1976. "Age Determination of a Tropical Reef Butterflyfish
Utilizing Daily Growth Rings of Otoliths." Fish. Bull. 74:990-994.

Randall, J. E. 1961. "A Contribution to the Biology of the Convict Surgeon
Fish of the Hawaiian Islands, Acanthurus triostegus sandvicensis." Pac.
Sci. 15:215-272.

Randall, J. E., and H. A. Randall. 1963. "The Spawning and Early Development
of the Atlantic Parrotfish, Sparisoma rubripinne, With Notes on Other Scarid
and Labrid Fishes." Zoologica8:49-60.

Robertson, D. R. 1973. "Field Observations on the Reproductive Behavior of a
Pomacentrid Fish, Acanthochromis polyacanthus." Z. Tierpsychol. 32:319-324.

Robertson, D. R. 1983. "On the Spawning Behavior and Spawning Cycles of Eight
Surgeon Fishes (Acanthuridae) From the Indo-Pacific." Environ. Biol. Fish.
9:193-223.

Robertson, D. R., and J. H. Choat. 1974. "Protogynous Hermaphroditism and
Social Systems in Labrid Fish." Proc. 2nd Int. Coral Reef Symp. 1:217-225.

Robertson, D. R., and S. G. Hoffman. 1977. "The Roles of Female Mate Choice
and Predation in the Mating Systems of Some Tropical Labroid Fishes." Z.
Tierpsychol. 45:298-320.

Robichaux, D. M., A. C. Cohen, M. L. Reaka, and D. Allen. 1981. "Experiments
With Zooplankton on Coral Reefs, or, Will the Real Demersal Plankton Please
Come Up?" Mar. Ecol. 2:77-94.

Rosenblatt, R. H., and B. W. Walker. 1963. "The Marine Shore Fishes of the
Galapagos Islands." Occ. Pap. Calif. Acad. Sci. 44:97-106.

Russell, B. C., F. H. Talbot, and S. Domm. 1974. "Patterns of Colonization of
Artificial Reefs by Coral Reef Fishes." Proc. 2nd Int. Coral Reef Symp.
1:207-215.

Sale, P. F. 1969. "Pertinent Stimuli for Habitat Selection by the Juvenile
Manini, Acanthurus triostegus sandvicensis." Ecology 50:616-623.

Sale, P. F. 1970. "Distribution of Larval Acanthuridae off Hawaii." Copeia
1970:765-766.

Sale, P. F. 1977. "Maintenance of High Diversity in Coral Reef Fish Communities."
Amer. Nat. 111:337-359.

Sale, P. F. 1978a. "Coexistence of Coral Reef Fishes--A Lottery for Living 0
Space." Environ. Biol. Fish. 3:85-102. W

Sale, P. F. 1978b. "Reef Fishes and Other Vertebrates: A Comparison of Social
Structures," pp. 313-346. In: E. S. Reese and F. J. Lighter (eds.),
Contrasts in Behavior. Wiley-Interscience, New York.

Sale, P. F. 1980. "The Ecology of Fishes on Coral Reefs." Ann. Rev. Oceanogr.
Mar. Biol. 18:367-421.

Sale, P. F., and R. Dybdahl. 1975. "Determinants of Community Structure for
Coral Reef Fishes in an Experimental Habitat." Ecology 56:1343-1355.

Sale, P. R., and R. Dybdahl. 1978. "Determinants of Community Structure for
Coral Reef Fishes in Isolated Coral Heads at Lagoonal and Reef Slope Sites."
Oecologia 34:57-74.

Scheltema, R. S. 1968. "Dispersal of Larvae by Equatorial Ocean Currents and
Its Importance to the Zoogeography of Shoal-Water Tropical Species." Nature
217:1159-1162.

Sebens, K. P., and K. DeReimer. 1977. "Diel Cycles of Expansion and Contraction
in Coral Reef Anthozoans." Mar. Biol. 43:247-256.

Segal, E. 1970. "Light--Invertebrates," pp. 159-211. In: O. Kinne (ed.),
Marine Biology, Part 1, Environmental Factors, Vol. 1. Wiley-Inter science,
New York.

Shapiro, D. Y. 1983. "On the Possibility of Kin Groups in Coral Reef Fishes,"
pp. 39-45. In: M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral
Reefs, Symp. Ser. Undersea Res., Vol. 1(1). NOAA Undersea Res. Progr.,
Rockville, Md.

Shulman, M. J. 1983. "Species Richness and Community Predictability in Coral
Reef Fish Faunas." Ecology 5:1308-1311.

Shulman, M. J., J. C. Ogden, J. P. Ebersole, W. N. McFarland, S. L. Miller, and
N. G. Wolf. 1983. "Priority Effects in the Recruitment of Juvenile Coral
Reef Fishes." Ecology 64:1508-1513.

Smith, C. L. 1973. "Small Rotenone Stations: A Tool for Studying Coral Reef
Fish Communities." Amer. Mus. Novit. 2512:1-21.

Smith, C. L. 1978. "Coral Reef Fish Communities: A Compromise View." Environ.
Biol. Fish. 3:109-128.

Smith, C. L., and J. C. Tyler. 1973. "Population Ecology of a Bahamian
Suprabenthic Shore Fish Assemblage." Amer. Mus. Novit. 2518:1-38.

Talbot, F. H., B. C. Russell, and G. R. V. Anderson. 1978. "Coral Reef Fish
Communities: Unstable, High Diversity Systems?" Ecol. Monogr. 48:425-440. i

Thompson, R., and J. L. Munro. 1978. "Aspects of the Biology of Caribbean
Reef Fishes: Serranidae (Hinds and Groupers)." J. Fish. Biol. 12:115-146.

Thorson, G. 1946. "Reproduction and Development of Danish Marine Bottom
Invertebrates." Medd. Dan. Fisk. Havunders, Ser. Plankton 4:1-523.

Thorson, G. 1950. "Reproductive and Larval Ecology of Marine Bottom
Invertebrates." Biol. Rev. 25:1-45.

Thresher, R. E. 1979. "Social Behavior and Ecology of Two Sympatric Wrasses
(Labridae: Halichoeres spp.) off the Coast of Florida." Mar. Biol. 53:161-172.

Victor, B. C. 1982. "Daily Otolith Increments and Recruitment in Two Coral-
Reef Wrasses, Thalassoma bifasciatum and Halichoeres bivittatus." Mar.
Biol. 71:203-2081

Victor, B. C. 1983a. "Recruitment and Population Dynamics of a Coral Reef
Fish." Science 219:419-420.

Victor, B. C. 1983b. "Settlement and Larval Metamorphosis Produce Distinct
Marks on the OtoTiths of the Slippery Dick, Halichoeres bivittatus," pp. 47-
51. In: M. L. Reaka (ed.), The Ecology of Deep and Shallow Coral Reefs,
Symp. Ser. Undersea Res., Vol. 1(1).NOA Un'dersea Res. Progr., Rockville,
Md.

Ware, D. M. 1975. "Relation Between Egg Size, Growth and Natural Mortality of
Larval Fish." J. Fish. Res. Bd. Can. 32:2503-2512.

Warner, R. R., D. R. Robertson, and E. G. Leigh. 1975. "Sex Change and Sexual
Selection." Science 190:633-638.

Watson, W., and J. M. Leis. 1974. "Ichthyoplankton of Kaneohe Bay, Hawaii."
UNIHI-Seagrant Rept. pp. 75-101.

Williams, D. McB. 1983. "Daily, Monthly and Yearly Variability in Recruitment
of a Guild of Coral Reef Fishes." Mar. Ecol. 10:231-237.

Williams, D. McB., and P. F. Sale. 1981. "Spatial and Temporal Patterns of
Recruitment of Juvenile Coral Reef Fishes to Coral Habitats Within 'One Tree
Lagoon,' Great Barrier Reef." Mar. Biol. 65:245-253.

Woodley, J. D., E. A. Chornesky, P. A. Clifford, J. B. C. Jackson, L. S. Kaufman,
N. Knowlton, J. C. Lang, M. P. Peterson, J. W. Porter, M. C. Rooney, K. W.
Rylaarsdam, V. J. Tunnicliffe, C. M. Wahle, J. L. Wulff, A. S. G. Curtis, M.
D. Dallmeyer, B. P. Jupp, M. A. R. Koehl, J. Neigel, and E. M. Sides. 1981.
"Hurricane Allen's Impact on Jamaican Coral Reefs." Science 214:749-755.

PELAGIC DURATION, DISPERSAL, AND THE DISTRIBUTION OF
INDO-PACIFIC CORAL-REEF FISHES

Edward B. Brothers
3 Sunset West
Ithaca, New York 14850

Ronald E. Thresher
Division of Fisheries Research
C.S.I.R.O. Marine Laboratories
Cronulla, N.S.W., 2230 Australia

ABSTRACT

The breadth of distribution of most Indo-Pacific coral-reef fishes does
not clearly correlate with the duration of their pelagic developmental stages.
Rather, there is a threshold effect at approximately 45 days, above which
species tend to be broadly distributed, butbelow which pelagic duration and
distribution appear unrelated. Long larval durations also characterize species
that have colonized geographically isolated areas, such as the Hawaiian Islands
and the Eastern Pacific. Comparison of larval durations with estimates of time
required to cross the "East Pacific Barrier" further suggests colonization of
the New World to be an infrequent event, most likely associated with unusually
strong development of the Equatorial Countercurrent.

INTRODUCTION

The pelagic larval stage characteristic of many invertebrates and fishes
is widely assumed to serve as a dispersal mechanism (Scheltema, 1971, 1978;
Crisp, 1976; Barlow, 1981; Smith, 1982). A logical corollary of this assumption,
also widely accepted, is that long larval duration correlates with broad
dispersal and consequent broad distribution (Shuto, 1974; Strathmann, 1974;
Zinsmeister and Emerson, 1979; Reaka, 1980; Ayal and Safriel, 1982; Jablonski
and Lutz, 1983). We tested this hypothesis by comparing the distributions
of 115 species of Indo-Pacific coral-reef fishes, in 22 families, with the
duration of their pelagic developmental stages, determined by examination of
otolith microstructure. The data suggest a threshold effect, with long pelagic
durations correlating with broad distributions only above a species mean value
of approximately 45 days. Only 20% of the species we examined have pelagic
durations this long, a value in surprisingly close agreement with independent
estimates of the frequency of "true long-distance" pelagic larvae of marine
invertebrates (Thorson, 1961). We also examine the question of whether long
pelagic durations are typical of those species that have colonized geographically
isolated areas, such as the Hawaiian Islands, and, particularly, the Eastern
Pacific.

METHODS

Duration of the pelagic larval stages was determined either by aging of
individuals newly recruited to the reef (Brothers, et al., 1983) or by examination

of otolith "settlement marks" in juveniles and adults, the latter involving
techniques similar to those used in other studies (Brothers and McFarland,
1981; Victor, 1982; Thresher and Brothers, in press). Total duration of the
pelagic stages was assumed to be equal to calculated otolith ages for most
demnersal spawning fishes. This assumption is based on limited data (Brothers,
unpublished data) which suggests that daily growth increments in most demersal
spawning fishes begin forming shortly after hatching. The only demersal spawning
families for which this is not likely to be true are Balistidae and Siganidae,
both of which hatch as relatively undeveloped prolarvae (see review in Thresher,
1984). For these two families, 2 days were added to calculated pelagic durations
to account for the likely duration of pelagic pre-feeding stages. For pelagic
spawning fishes, total duration of the pelagic stages was estimated by adding 3
days to otolith ages. This was based on a mean incubation time of 43.1 hours
for pelagic eggs of such fishes (Thresher, 1984), plus limited data suggesting
that another 24 hours may pass before the first growth increment forms (Brothers,
unpublished data).

Species distributions were based on taxonomic reviews and comparison of
faunal lists for different areas. The extent of distribution was quantified by
dividing the tropical Indo-Pacific into 29 "bio-geographical areas" and counting
the number of areas inhabited by each species. Division of the Indo-Pacific
follows Allen (1979). The rationale behind this approach is discussed in
Thresher and Brothers (in press).

Data were obtained for 115 species (see table 1) in 22 families, essentially
all those we had available to us. The taxonomic distribution is not intended
to be qualitatively or quantitatively representative of typical reef
ichthyofaunas, although effort was made to include representatives of as many
families as possible. The only fishes specifically excluded from the analysis
were Hawaiian endemics and a few families, such as kyphosids and carangids,
that are known or strongly suspected to raft as juveniles or adults under
floating objects. The few Hawaiian endemics we had available were excluded
because of difficulty in discerning a single or discrete "settlement mark" (see
Thresher and Brothers, in press, for further discussion of this problem). Some'
material was obtained for us directly in the field (the Philippines, Japan, the
Marshall Islands, and Australia), but the majority were obtained as juveniles
and small adults from the aquarium. As such, most were probably collected in
the Philippines, the Hawaiian Islands, or off the Australian Great Barrier
Reef. Specimens maintained in captivity for a variable period were suitable
for our study since they were initially captured as settled juveniles or adults
(rather than being reared in captivity) and because there are no indications,
nor expectations, that conditions in captivity affect structures already laid
down in the otoliths.

Specific samples sizes were typically small (range I - 15, Y = 2.62, mode
=1), but intraspecific variation in most species examined was also small (see
results; also Thresher and Brothers, in press). Calculated durations, therefore,
are likely to be robust for most species examined. Results of our analysis are
also sufficiently broadly based that minor adjustments of a few species are likely
to have little or no impact on the conclusions drawn.

RESULTS

Table 1 summarizes our data for mean pelagic duration and geographic
distribution for the 115 species examined. Intraspecific variation appears to
be low for most species. Of 57 species with N > 2, this variation ranged from
O to 25 days, with a mode of 3 days, a mean of 5.0 days, and a standard deviation
of 4.97 days. Only seven species (5 broadly distributed) varied intraspecifically
by more than 10 days, and only three, Coris gamaird and Gomphosus varius
(Labridae) and Rhinomuraena amboinensis (Muraenidae), varied by more than 20
days. Based on observations of specimens collected in other areas, la;ge
intraspecific variation appears to be more common in these families and also the
Gobiidae and Scaridae (Brothers, et al., 1983; and Brothers, unpublished data)
than in the remaining families. <i yet, we have no data on geographical or
environmental effects on intraspecific variability in larval duration, if any.

Based on species mean values, there is a conspicuous threshold effect in
the relationship between pelagic duration and species distribution at approximately
45 days total pelagic duration (fig. 1). Below this threshold, duration of the
pelagic stage does not correlate with breadth of distribution (Spearman Rank
Correlation = 0.12, N = 92, p > 0.2), suggesting that for most Indo-Pacific
coral-reef fishes (79.3% of species examined, and all representatives in 14 of
the families), the length of time that pelagic eggs and/or larvae disperse is
not the principal determinant of the extent of species distributions. In
contrast, of the 23 species that have mean pelagic durations longer than 45
days, all but three are broadly distributed in the Indo-Pacific, with more than
half (14 of 23) inhabiting 25 or more biogeographic areas (as compared with 3
out of 92 subthreshold species; this difference is significant at p < 0.001; x2
= 53.7, df = 2). The exceptions to the above are Chaetodon collare (total
duration 46 days), Neocirrhites armatus (51 days), and Rhinomuraena amboinesis
(74 days). Of these, the first two are close to the threshold value. We are
unable, however, to account for the limited distribution of the muraenid, R.
amboinesis. The two specimens examined differed markedly in apparent larval
duration (63 and 85 days), but both are well above the threshold and are con-
sistent with our data for other muraenid species.

The data also suggest that unusually long pelagic durations are critical
for colonization of geographically isolated areas. The average pelagic duration
for Central Pacific species that also are found in the Hawaiian Islands, widely
considered an isolated out-lier of the Indo-Pacific faunal province (Ekman,
1953; Briggs, 1974), is 68.4% longer than those that are not found in Hawaii
(Central Pacific only, X = 49.1 days, S.D. = 9.5, N = 91; species also in
Hawaii, k = 49.1 days, S.D. = 10.3; N = 15; difference significant at p < 0.001,
Kolmolgorov-Smirnov test for unequal sample sizes; fig. 2). The minimum observed
duration of Hawaiian-inclusive species, 35 days, is longer than the mean
durations of 59.5% of the species examined, suggesting that most Central Pacific
species may be incapable of colonizing Hawaiian reefs under normal circumstances.
The pattern for colonizing Eastern Pacific reefs by Indo-West Pacific species
is similar (fig. 2). Trans-Pacific species examined have a mean pelagic duration
25.6% longer than even Hawaiian-inclusive species, and a minimum duration (48
days) that is longer than mean durations of 81.0% of the species we examined.

Table 1.--Duration of pelagic larval stage (in days, as number of pre-transition otolith increments,
age of new recruits, or both) and extent of geographical distribution (number of biogeographic areas
occupied) for 115 species of Indo-Pacific coral reef fish. Location: AT = Aquarium Trade, AUS =
Australia, JA = Japan, MI = Marshall Islands, OT = One Tree Island, PH = Philippines.

Family and species No. bf Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

ACANTHURIDAE

Acanthurus glaucopareius 3 AT 59.7 56-62 18
A. lineatus 2 AT 39.5 38-41 21
A. triostegus 1 AT 51 - 28
Naso lituratus 1 AT 69 - 27
Paracanthurus hepatus 1 AT 37 - 17
Zebrasoma veliferum 1 AT 36 - 26

APOGONIDAE

Apogon cyanosoma 5 OTI 18.2 18-19 20
A. doderleini 5 OTI 22.6 16-27 8
Cheilodipterus quinquelineata 4 OTI 18.8 17-20 24
25 OTI 23.1 17-21
Siphamia orbicularis 1 AT 17 - 13

BALISTIDAE

Balistoides conspiculatum 3 AT 36.0 35-38 18
Rhineacanthus rectangulatus 6 AT 43.5 39-46 24
R. aculeatus 1 AT 48 - 25
Palaluteres prionurus 1 OTI 28 21

gable l.--Continued

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

BLENNIIDAE

Petroscirtes mitratus 2 OTI 24.5 24-25 19

CALLIONYMIDAE

Synchiropus splendens 2 AT 36 33-39 4

CHAETODONTIDAE

Forcipiger flavissimus 3 AT 50.7 41-57 28
Heniochus diphreutes 2 AT 32.0 30-34 8
Chelmon rostratus 2 AT 27.0 26-28 11
2 OTI 25.5 25-26
Chaetodon lineolatus 1 AT 54 - 22
C. kleini 1 AT 56 22
C. speculum 1 AT 26 - 10
C. trifascialis 2 AT 32 30-34 24
C. baronessa 2 AT 38 38 8
C. auriga 2 AT 46.5 40-53 26
C. unimaculatus 2 AT 35.0 35 24
C. collare 1 AT 43 6
C. rainfordi 1 OTI 35 3
C. plebius 1 OTI 39 13

CIRRHITIDAE

Cirrhitichthys oxycephalus 1 AT 51 27

Table I.--Continued

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution
CIRRHITIDAE (continued)

Neocirrhites armatus I AT 48 - 10
Oxyeirrhites Upus 3 AT 69.0 62-73 20

EPHIPPIDAE

Platax pifinatus I AT 25 -21

GOBIIDAE

Amblygobius rainfordi 4 OTI 40.3 40-41 4
Paragobiodon echinocephala I OTI 36 - 19
P. mnelanosoma 3 OTI 42.3 39-47 11

LABRIDAE

Anampses coeruleopunctatus I AT 25 -20
Consl aygula I AT 51 - 26
C.gamnaird I AT 51.3 35-60 26
C.variegata 3 OTI 29.7 28-30 5
Cirrhilabrus temminkii I OTI 24 - 23

1 OTI 30
C. cyanopleura 1 ~ ~~~ AT 30 - 10
GompRhosus Varius 5 AT 52.8 42-75 26
Halichoeres biocellatus I OTI 21 10t

I OTI 27
Hemipteronotus taeniurus 2 AT 45.0 43-48 27

Table l.--Continued

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

LABRIDAE (continued)

Labroides dimidiatus 6 AT, OTI 23.2 22-24 23
L. bicolor 1 AT 24 - 23
Lierdenella fasciata 2 AT 20.0 19-21 16
Macropharygodon meleagris 1 AT 30 - 24
Stethosulis axilaris 1 PH 35 24
Thalassoma hardwicki 6 PH 45.7 44-47 16
T. lunare 4 AT, OTI 47.0 42-50 26

LETHRINIDAE

Lethrinus nebulosus 1 OTI 37 20

LUTJANIDAE

Lutjanus sebae 1 AT 40 17

MALCANTHIDAE

Hoplolatilus purpureus 3 AT 34.0 33-36 4

MURAENIDAE

Echidna nebulosus 1 AT 80 - 27
Rhinomuraena amboinensis 2 AT 71 60-82 9

OSTRACIIDAE

Ostracion meleagris 2 AT 69.0 68-70 25

Table 1.--Continued

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

POMACA NTHIDAE

Apolemichthys trimaculatus 3 AT 24.2 22-26 17
Centropyge acanthops 1 AT 34 - 3
C. bicolor 3 AT 32.0 29-34 11
C. bispinosus 3 AT 32.0 31-33 17
C. eibli 6 AT 25.3 24-27 7
C. ferrugatus 1 AT 38 - 3
C. flavissimus 3 MI, AT 30.3 28-32 7
C. heraldi 4 AT 32.0 32 7
C. interruptus 2 JA 31.5 29-34 2
C. loriculus 1 AT 38 - 6
C. multifasciatus 3 AT 26.2 25-28 7
C. nox 2 AT 31.5 28-33 5
C. tibicen 3 AT 30.0 27-34 10
C. vroliki 1 AT 29 - 10
Chaetodontoplus melanosoma 3 AT 22.2 20-23 6
C. mesoleucas 5 AT 19.9 18-22 7
C. personifer 4 AT, AUS 23.25 23-24 3
C. septemtrionalis 1 AT 23 - 5
Genicanthus bellus 1 AT 25 - 9
G. melanospilos 1 AT 25 - 6
Holacanthus venustus 1 AT 25 - 3
Pomacanthus annularis 2 AT 21.0 20-22 6

Table l.--Continued

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

POMACANTHIDAE (continued)

P. imperator 2 AT 22 22 25
P. navarchus 3 AT 22.8 22-24 3
P. semicirculatus 4 AT 19.5 19-20 21
P. sexstriatus 2 AT, AUS 18.0 17-19 8
P. xanthometopon 4 AT 20.4 19-22 5
Pygoplites diacanthus 2 AT 24.5 23-26 25

POMACENTRIDAE
cn
Chrornis caeruleus 5 PH 25.0 23-26 21
C. atripectoralis (probably) 1 AT 19 - 11
2 OTI 23.0 20-26
Dascyllus aruanus 5 AT 22.4 20-26 25
4 OTI 22.0 20-23
D. melanurus 3 AT 23.3 22-24 5
D. reticulatus 1 AT 23 - 23
D. trimaculatus 5 AT 24.2 21-25 23
Abudefduf saxatilis 1 OTI 23 - 22
Glyphidodontops cyaneus 6 AT 19.5 17-21 7
G. biocellatus 1 AT 19 - 8
G. hemicyaneus 5 AT 19.25 18-21 4
G. rex 3 AT 22.0 19-25 8
3 OTI 19.7 19-20

Table I.--Continued

Family and species No. of Source XOtolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

POMACENTRIDAE (continued)

G. rollandi 3 OTI 23.0 23 2
G. talbot i I OTI 22
Dischistodus pseudochrysopoecilus I OTI 23 - 3
Neopomnacentrus azysron I OTI 24 - 3
Paraglyphidodon behni 4 AT 17.75 15-21 6
Pomacentrus amboinensis 8 OTI 19.6 19-20 6
is OTI 22.5 18-26
N3 ~P. australis 4 OTI 25.0 23-26 1
2 OTI 85.0 85
P. coelestis I AT 20 - 13
P ~~~~~~~10 OTI 19.1 18-21 8
4 OTI 24.0 23-25
P. wardi 7 OTI 20.0 20 1
4 OTI 24.25 23-25
P. vaiuli 6 OTI 24.2 23-26 10
P. Sp. 15 OTI 22.9 20-24 4

SCARIDAE

Bulbometopon muricatum I AT 31 -21

SCOLOPSIDAE

Scolosis dubiosus 3 OTI 19.0 15-23 8

Table l.--Concluded

Family and species No. of Source X Otolith increments Larval Size of
specimens Total for duration geographic
Pre-transition new recruits (Range) distribution

SERRANIDAE

Anthias pascalus 1 AT 26 7
Cromileptis altivelis 1 AT 24 - 9

SIGANIDAE

Siganus (Lo) vulpinus 2 AT 23.5 23-24 8

ZANCLIDAE

Zanclus cornuta 1 AT 57 - 28

A AM
25- o� �
o A� �
c lo
0 O 0 0 0 U
w 20. 0 0 �
0 o
o o o o

1- oo

0 0 08 0
oo o
00 0

000 o o 00 0

~0
*_/ oo o o

'~10 ' o P ooo

�10 io0 30 40 50 60 70 80
X PRE-TRANSITION OTOLITH INCREMENTS

Figure 1.--The relationship between species mean pelagic larval duration, as
indicated by the number of pre-transition daily growth increments, and the
breadth of distribution of 115 species of Indo-Pacific coral-reef fishes.
Total pelagic durations are 0-3 days longer than increment counts, depending t
on the family.

.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

25'-
2 ~ ~5 -Western and Central Pacific Only
X 29.4 S.D.=9.5
N 91

wK
0 . , , J ' .~~~~~~~~.I
Zr 0
L Hawaiian

Ot L
X 50 X= 49.1 S.D. = 160.3
_ a, L N= 15

' 5 X=62.2 S.D.= 11.4
Z N x 9

10 20 30 40 50 60 70 80
DURATION OF PELAGIC STAGE (DAYS)

Figure 2.--Comparison of the duration of the pelagic stages for species showing
different zoogeographic distributional patterns. Values on the abscissa are
adjusted (see text) to convert pre-transition increment counts to pelagic
duration. Open circles = means for species occurring in the Western and
Central Pacific only; solid triangles = species whose range includes the
Hawaiian Islands; solid squares = species distributed trans-Pacifically. All
but one trans-Pacific species are found also in the Hawaiian Islands.

~~~~~~~~I

The mean duration for trans-Pacific species is 62.2 days (S.D. = 11.4, N = 9).
The difference between these and the Hawaiian-inclusive species is not significantW
(p = 0.13; Kolmolgorov-Smirnov test, two-tailed), although the trend in the
data is both evident and consistent with predictions made elsewhere in the
literature (e.g., Rosenblatt and Walker, 1963). All but one trans-Pacific
species also is found in Hawaii.

DISCUSSION

Our results suggest that a general correlation between long pelagic larval
durations and broad geographic distributions is a valid assumption only when
applied to a relatively few species with exceptionally long-lived pelagic stages
(see also Thresher and Brothers, in press). In this respect, our conclusions
are strikingly similar to those of Thorson (1961), who surveyed the pelagf
durations and distributions of marine invertebrates. Thorson concluded
only 20% of all bottom invertebrates with pelagic larvae have "true long-t~tance
larvae." These he defined as those with a planktonic life longer than 6 weeks.
Both of these figures are strikingly close to those we calculate for Indo-
Pacific coral-reef fishes, i.e., a threshold value of about 45 days, with 20.7%
of the species examined exceeding this value. The latter figure obviously
could be manipulated by adding species with high or low pelagic durations. We
strongly suspect, however, that this threshold value will remain robust as more
data are added. It is premature to draw strong conclusions without additional,
detailed work for other families of fishes and other phyla, but the surprisingly
close quantitative agreement between the current study and Thorson (1961)
suggests that some fundamental principle that applies across phyla may be
involved.

Our data are also in substantial agreement with suggestions that geographically
isolated regions of the Indo-Pacific are colonized primarily by species with
longer than average pelagic developmental stages. This point has previously
been raised, particularly with respect to colonization of the Eastern Pacific
Faunal Province by Indo-West Pacific species, both for marine invertebrates
(e.g., Thorson, 1961; Dana, 1975; Zinsmeister and Emerson, 1979; Reaka, 1980)
and coral-reef fishes (Briggs, 1961, 1974; Rosenblatt and Walker, 1963;
Rosenblatt, 1967). Comparison of the minimum observed pelagic durations of
species that occur in Hawaiian waters and in the Eastern Pacific with the large
majority of species found only in the Central Pacific suggests that, although
an element of chance certainly is involved, under normal circumstances the
majority of Central Pacific species may not be capable of colonizing these
outlying areas. Moreover, even for those species that have colonized the
Eastern Pacific, such colonization may well be intermittent rather than
continuous, and dependent on the occurrence of unusual oceanographic conditions.
The minimum time required to cross, the "Eastern Pacific Faunal Barrier" (see
Ekman, 1953) can be estimated from the distance between the likely source of
colonists, the Line Islands, and the westernmost point of the Eastern Pacific
faunal province, Clipperton Island, and peak reported values for the Equatorial
Countercurrent, the presumed route of colonization (e.g., Leis, 1983; Reaka and
Manning, 1980). The distance involved is approximately 5250 kin, and normal
peak current velocities are in the range of 0.46 to 0.7 in sec-1 (Wyrtki, 1965;
Tsuchiya, 1974). Consequently, normal transit time of the Eastern Pacific

Barrier by a passively drifting object will be at least 87 days; given that
currents vary temporally and also decline as the near-shore environment is
reached, a minimum estimate of 100 days is probably more realistic. Yet,
apparent pelagic durations of nine trans-Pacific species examined range from 48
to only 83 days, averaging 24.8 days less than a reasonable minimum. None of
the 302 specimens we examined had a larval duration longer than 87 days,
suggesting that, even if one includes individual variation in larval competance,
few if any species are likely to cross the barrier under normal circumstances.
Our tentative conclusion, therefore, is that such colonization takes place only
during periods when the countercurrent is unusually strong, i.e., during El
Nifo-type events (Wyrtki, 1965). If so, then we would predict that, following
the recent very strong El Nino event in 1982-83 (see Gill and Rasmussen, 1983;
Philander, 1983), unusually large numbers of new recruits of Indo-West Pacific
species should be present on Eastern Pacific reefs.

ACKNOWLEDGMENTS

We thank A. Gronell and N. Wolf for discussions of this work as it
progressed, W. N. McFarland for providing unpublished material for our examination,
L. Bell, 0. Bellwood, P. L. Colon, G. Dingerkus, R. Greenfield, A. Gronell, J.
T. Moyer, P. F. Sale, and D. Williams for providing specimens, and S. Blaber
and J. Stevens for comments on an earlier version of this paper.

LITERATURE CITED

Allen, G. R. 1979. Butterfly and Angelfishes of the World, Vol. 2. Wiley,
New York. 352 pp.

Ayal, Y., and U. N. Safriel. 1982. "R-curves and the Cost of the Planktonic
Stage." Amer. Nat. 119:391-401.

Barlow, G. W. 1981. "Patterns of Parental Investment, Dispersal and Size
Among Coral-Reef Fishes." Environ. Biol. Fish. 6:65-85.

Briggs, J. C. 1961. "The East Pacific Barrier and the Distribution of Shore
Fishes." Evolution 15:545-554.

Briggs, J. C. 1974. Marine Zoogeography. McGraw-Hill, New York. 475 pp.

Brothers, E. B., and W. N. McFarland. 1981. "Correlations Between Otolith
Microstructure Growth, and Life History Transitions in Newly Recruited French
Grunts Haemulon flavolineatum (Desmarest), Haemulidae." Rapp. P.-v. Reun.
Const. Int. Explor. Mer. 178:369-374.

Brothers, E. B., D. M. Williams, and P. F. Sale. 1983. "Length of Larval Life
in Twelve Families of Fishes at 'One Tree Lagoon,' Great Barrier Reef,
Australia." Mar. Biol. 76:319-324.

Crisp, D. J. 1976. "The Role of the Pelagic Larva," pp. 145-155. In: P. S.
Davies (ed.), Perspectives in Experimental Biology, Vol. 1. Pergam-on Press,
New York.

Dana, T. F. 1975. "Development of Contemporary Eastern Pacific Coral Reefs."
Mar. Biol. 33:355-374.

Ekman, S. 1953. Zoogeography of the Sea. Sidgwick and Jackson, London.
417 pp.

Gill, A. E., and E. M. Rasmussen. 1983. "The 1982-83 Climate Anomaly in the
Equatorial Pacific." Nature 306:229-234.

Jablonski, D., and R. A. Lutz. 1983. "Larval Ecology of Marine Benthic
Invertebrates: Paleontological Implications." Biol. Rev. 58:21-89.

Leis, J. M. 1983. "Coral Reef Fish Larvae (Labridae) in the East Pacific
Barrier." Copeia 1983:326-328.

Philander, S. G. H. 1983. "El NiWo Southern Oscillation Phenome-_'." Nature
302:295-301.

Reaka, M. L. 1980. "Geographic Range, Life History Patterns, and Body Size in
a Guild of Coral-Dwelling Mantis Shrimps." Evolution 34:1019-1030.

Reaka, M. L., and R. B. Manning. 1980. "The Distributional Ecology and
Zoogeographical Relationships of Stomatopod Crustacea From Pacific Costa
Rica." Smithsonian Contrib. Mar. Sci. 7:1-29.

Rosenblatt, R. H. 1967. "The Zoogeographic Relationships of the Marine Shore (
Fishes of Tropical America." Stud. Trop. Oceanog. 5:579-592.

Rosenblatt, R. H., and B. W. Walker. 1963. "The Marine Shore Fishes of the
Galapagos Islands." Occ. Pap. Calif. Acad. Sci. 44:97-106.

Scheltema, R. S. 1971. "Larval Dispersal As a Means of Genetic Exchange
Between Geographically Separated Populations of Shallow-Water Benthic Marine
Gastropods." Biol. Bull. 140:284-322.

Scheltema, R. S. 1978. "On the Relationship Between Dispersal of Pelagic
Veliger Larvae and the Evolution of Marine Prosobranch Gastropods," pp. 303-
322. In: B. Battaglia and J. A. Beardmore (eds.), Marine Organisms. Plenum
Press, New York.

Shuto, T. 1974. "Larval Ecology of Prosobranch Gastropods and Its Bearing on
Biogeography and Paleontology." Lethaia 7:239-256.

Smith, C. L. 1982. "Patterns of Reproduction in Coral-Reef Fishes," pp. 49-
66. In: G. R. Huntsman, W. R. Nicholson, and W. W. Fox, Jr. (eds.), The
BiologTcal Bases for Reef Fishery Management, NOAA Tech. Mem. NMFS-SEFT0O.
Nat. Mar. Flsh-. Serv.,Teaufort, N.C.

Strathmann, R. 1974. "The Spread of Sibling Larvae of Sedentary Marine
Invertebrates." Amer. Nat. 108:29-44.

Thorson, G. 1961. "Length of Larval Life in Marine Bottom Invertebrates as
Related to Larval Transport by Ocean Currents," pp. 455-474. In: M. Sears
(ed.), Oceanography. Amer. Assoc. Adv. Sci. Publ. 67, Washington, DC.

Thresher, R. E. 1984. Reproduction in Reef Fishes. Tropical Fish Hobbyist,
Neptune City, N.J. 400 pp. -

Thresher, R. E., and E. B. Brothers. In Press. "Reproductive Ecology and
Biogeography of Indo-West Pacific Angelfishes (Pisces: Pomacanthidae)."
Evolution.

Tsuchiya, M. 1974. "Variation of the Surface Geostrophic Flow in the Eastern
Intertropical Pacific Ocean." Fish. Bull. 72:1075-1086.

Victor, B. C. 1982. "Daily Otolith Increments and Recruitment in Two Coral-
Reef Wrasses, Thalassoma bifasciatum and Halichoeres bivittatus." Mar.
Biol. 71:203-208.

Wyrtki, K. 1965. "Surface Currents of the Eastern Tropical Pacific Ocean."
Bull. Inter-Amer. Trop. Tuna Comm. 9:271-304.

Zinsmeister, W. J., and W. K. Emerson. 1979. "The Role of Passive Dispersal
in the Distribution of Hemipelagic Invertebrates, With Examples from the
Tropical Pacific Ocean." Veliger 22:32-40.

DIURNAL PERIODICITY OF SPAWNING ACTIVITY BY
THE HAMLET FISH, HYPOPLECTRUS GUTTAVARIUS (SERRANIDAE)

Phillip S. Lobell
Center for Earth and Planetary Physics
Harvard University
Cambridge, Massachusetts 02138

and

Steve Neudecker
1301 Union Drive
Davis, California 95616

ABSTRACT

Many coral reef fishes spawn during the afternoon crepuscular period, a
time of increased predatory activity by piscivores. We document the diel
timing and spawning behavior of the hamlet fish, Hypoplectrus guttavarius, at
St. Croix, U.S.V.I. This fish spawns daily during dusk in the water column
above specific reef sites.

The time of spawning relative to sunset was quantified at two sites at
different depths. Spawning commenced earlier at the deep reef site and
continued later at the shallow reef site. A field manipulation evaluated how
this species responded to disburbances during reproduction. When scuba divers
attempted to alter the fish's behavior by swimming vigorously toward mating
pairs, the hamlets did not abort spawning. The pair responded (1) by continuing
to spawn later into the night than those not disturbed and (2) by spawning
progressively nearer the substratum.

INTRODUCTION

Until the 1970s, the reproductive behavior of most tropical coastal
marine fishes eluded observation. This has changed recently with the discovery
that many species spawn during restricted times of ebb tide, off-reef current
flow, or crepuscular periods. A majority of coral reef fishes spawn eggs
that are bouyant and are advected by ocean currents. Having planktonic
propagules is thought to facilitate both rapid advection/dispersal in currents
and reduced predation by other reef fishes (Johannes, 1978; Lobel, 1978;
Barlow, 1981). In general, fishes seek out specific reef localities and/or
pinnacles over which they ascend and spawn. This behavior is particularly
interesting for fishes spawning shortly before and during dusk because
crepuscular periods are generally believed to be times of increased predator
activity and success (e.g., Hobson, 1968, 1972, 1974). Also, to witness

1 Present affiliation: Woods Hole Oceanographic Institution, Woods Hole,
Massachusetts 02543.

natural predation an spawning fishes or their eggs is not rare (Randall and
Randall, 1963; Hobson, 1965; Moyer, 1974, 1975; Meyer, 1977; Colin, 1976,
1978; Robertson and Hoffman, 1977; Jones, 1981; Robertson, 1983). Some
species gather in spawning aggregations and apparently are oblivious to human
or natural piscivore predators even when disturbed or attacked (Johannes,
1978, 1981; Robertson, 1983). In contrast, the mere presence of a diver-
observer will interrupt the mating of some smaller species unless the fish
are approached and observed with stealth (Johannes, 1978; Lobel, 1978).

To date, studies on the reproductive behavior of tropical coastal marine
fishes in the field have been primarily observational. The emphasis has been
on documenting where, when, and how various species spawn. Herein we report
the spawning behavior, diel timing, and a manipulation that attempted to
disturb the behavior of a fish during reproduction. Our goal was to see
whether and how the spawning behavior of the Caribbean hamlet, Hypoplectrus
tuttavarius, might be altered. Because hamlets and many other fishes spawn
uring the evening crepuscular period when predation by piscivores is most
likely, we wanted to determine whether fish that were disturbed by scuba
divers during the spawning act would cease all activity, move sites, depart
and seek other mates, or merely continue despite the disturbance.

The Caribbean hamlet, Hypoplectrus guttavarius (fig. 1) was selected for
study because it maintained stable, identifiable pairs which spawned almost
every evening at specific sites during February 1980, at St. Croix. It was
common, easily observed, and the spawning act was unmistakable and easily
quantified. The diel timing of spawning is compared with that of H. chlorurus,0
H. nigricans, H. puella, and H. unicolor.

The reproductive ecology of hamlets is somewhat enigmatic. It has been
suggested that Hypoplectrus nigricans and other hamlets are simultaneous
hermaphrodites, and that individuals alternate sexual roles between spawning
clasps (Barlow, 1975; Fischer, 1980a, 1980b, 1981). It is also controversial
whether different morphological varTants oT hamlets represent aggressive-
mimetic morphs of the single species, Hypoplectrus unicolor, or are reproductively
isolated incipient species (Thresher, 1978; Graves and Rosenblatt, 1980;
Fischer, 1980b). Hamlets usually mate assortatively by color patterns,
although mixed pairs have been observed, and individuals which exhibit color
markings characteristic of two or more color forms are known (Barlow, 1975;
Thresher, 1978; Graves and Rosenblatt, 1980; Fischer, 1980a). Thus, the
reproductive behavior of hamlets is interesting from several perspectives.
To maintain simplicity and clarity, we refer to the various hamlets by their
established species names (cf. Randall, 1968).

STUDY SITES

Fish were examined at two locations on St. Croix, U.S. Virgin Islands
(64035'W, 17045'N). The first site was the east slope of Salt River canyon
within the excursion limits of the NOAA National Undersea Laboratory System
habitat (HYDROLAB). Since we were saturated, we spent approximately 53 hours
per person observing fish at 20-30 m depths from February 11 to 15, 1980.
The second site was located on several patch reefs at about 10 m depths in

Figure 1.--Adult Hypoplectrus guttavarius (the shy hamlet) approximately
11 cm in length.

the Buck Island Channel outside Teague Bay Reef on the northeast coast of St.
Croix. Observations were made from February 20 to March 3, 1980. Further
site descriptions are given by Neudecker and Lobel, 1982.

METHODS

Specific pairs of H. guttavarius were identified and details of their
spawninq nehavior were recorded before, during, and after the manipulation.
Observers selected one pair each evening and positioned themselves unobtrusively
on site at least 15 min before the fish initiated courtship. The baseline
spawning behavior of each pair was determined during one or two evenings
prior to the disturbance manipulation (3 of 7 pairs were observed for 2
consecutive days before the manipulations; the other 4 pairs were observed
the day before the manipulations). In cases where sites lacked distinguishing
natural features, the specific site at which each pair spawned was marked
with small floats. We spent several days diving at all hours, including
sunrise, and saw hamlets spawn only during dusk (total observation 158.7 man-
hours during the 7 days in HYDROLAB). Data recorded each time a pair spawned
included the time, height in the water relative to the bottom and the
coral/gorgonian structure over which spawning took place, approximate distance
moved between spawning sites, and general notes on the pair's response to
observers and other species.

Figure 2.--A pair of Hypoplectrus guttavarius engaged in courtship at a
typical spawning site (Deep Reef at East Wall of Salt River Canyon,
HYDROLAB site).

The disturbance manipulations were conducted the day after baseline
spawning behavior was determined. The manipulation was to disrupt the spawning
act each time the fish attempted it. As the pair began to enter the spawning
clasp, the observer swam rapidly toward the pair (in an attempt to mimic a
predator's behavior). The pair was harassed repeatedly this way for as long
as spawning was attempted. When mates parted, close watch was maintained
unt~' it was certain that spawning had ceased for the night. Occasionally, a
pair qould separate, swim several meters apart, return several minutes later,
reloin and attempt to spawn again. Follow-up observations were made on the
sarme pair the day after the manipulation. Data analysis of treatments used
tne nonparametric Mann-Whitney U test statistic.

On any given day, at least three diver/observers were engaged in various
phases of the manipulation. To provide control observations, a diver watched
a spawning pair without interfering while other divers were either harassing
a spawning pair or conducting follow-up observations.

Figure 3.--Hypoplectrus guttavarius embraced in the spawning clasp. The
individual facing head-down has bent body in an S-shape.

RESULTS

Typical Spawning Behavior

Our observations provided the following general picture of spawning in
H. guttavarius. Hamlets mate only in pairs. Fertilization is external and
zygotes are planktonic. During the daytime, mates appear loosely associated
in adjacent or nearby home ranges. They begin to interact within 2 hr
of sunset. They first rendezvous, remain close together, and engage in
occasional short chases. The spawning act begins when one fish presents a
late-al display while flicking its pelvic fins (fig. 2). The displaying
ind`iicual slowly rises while continuing its display. The bright blue flank
of this individual fades to a pastel shade. The other fish follows. The two
then rise together about I m above some towering reef structure. At St.
Croix, tnese structures were vertical colonies of the gorgonian, Pseudoplexaura
sp., or the coral, Acropora cervicornis.

Once the proper height (about 1-2 m) is attained over a tall reef
structure, the following fish folds around the pale colored leader, who poses
head down in an S-shaped position (fig. 3). Release of eggs and sperm occurs

Figure 4.--Hypoplectrus guttavarius embraced in the spawning clasp. The
individual folded around the other one facing head-down opens its
mouth while quivering and spawning.

as both fish quiver. Quivering lasts for 2-3 sec. The individual folded
around the other often opens its mouth while quivering (fig. 4). Following
consummation, the fish cease embracing and quickly descend to the bottom.
The embracing behavior of hamlets when spawning has been termed the spawning
clasp (Fischer, 1980a) and occurs many times throughout an evening. Infrequently,
fish would embrace but not quiver, and the release of gametes was not evident.
Cisc-e- 1980a) has described the similar spawning behavior and sexual role of
~. -2ricans.

:ach pair of hamlets initiated spawning at a specific reef location each
-enen5. Pairs showed strong spawning site specificity. They moved between
aew f'from 1 to 4) particular sites each evening, and frequently a pair
-eor-ed to their original site. When some pairs moved between spawning
sies, other individuals occasionally interfered and attempted to steal a
-ate. These 'floater' individuals were seldom successful. Out of all namlet
spawnings observed (N = 314 of 5 spp.), only once was a floater successful in
stealing a mate by breaking up an existing pair (the pair and floater species
was H. unicolor).

25- SHALLOW REEF Full Moon
20- (approx. lO m depth)
a.

25DEEP REEF NAew Moor)
20" (pprox. 25m depth)
'C

._ 10'-CS
z

{

iU 0 3 3 25 20 15 10 5 5 5Moon 20

SUNSET
TIME BEFORE AND AFTER SUNSET(minutes)

Figure s.--oiel reproductive timing of Hypoplectrus guttavarius at two depths.
"The sha llow site (10 m depth) was occupied during the full moon

phase (Feb. 24-27, 1980), while the deep site (22-30 m depth) was
occupied during the new moon (Feb. 11-13, 1980). The shallow
group consisted of 4 pairs observed over 4 days for a total of 67
spawning clasps. The deep group consisted of 7 pairs observed
i0-

over 4 days for a total of 210 spawning clasps. Time is in minutes
45 4 35 $0 25 20 15 10 5 � 5 10 15 20
SUNSET
TIME BEFORE AND AFTER SUNSET(minutes)

beFigure 5.--and after sunset. Seive timing of Hypoplectr discuusi at two depths.

77
The shallow site (10 m depth) was occupied during the full moon
phase (Feb. 24-27, 1980), while the deep site (22-30 m depth) was
occupied during the new moon (Feb. 11-13, 1980). The shallow
group consisted of 4 pairs observed over 4 days for a total of 67
spawning clasps. The deep group consisted of 7 pairs observed
over 4 days for a total of 210 spawning clasps. Time is in minutes
before and after sunset. See text for discussion.

2"el ~eproductive 'imin
Hypoplectrus guttavarius spawned almost every evening during the study
eriod.Fish were studied at two sites which differed in depth. The deep eef
site (20-30 m depth) was occupied by us February 11-15, and the shallow site
tI0 m depth) from February 20-March 3. We did not see hamlets spawn on two
evenings, February 21-22. Observations at the deep site corresponded to the
presence of a new moon (February 15). We were at the shallow site during the
full moon (March 1). The dates on which hamlets were not seen spawning
coincided with the last quarter moon phase (February 22).

The diel timing of reproduction, relative to sunset, differed for H.
guttavarius at the two sites (depths) and lunar phases. The first spawnengs
at the deep reef site began 50 min before sunset (with a mean, X, and stancard
deviation, S.D., of 33 + 1 min) and the last spawnings ended 10 min (X + S.. =
2 + 9 min) after sunset; most spawnings occurred just prior to sunset. Spawning
at the shallow reef site commenced later, 25 min (X + S.D. = 16 + 7 min) oefore
sunset, and ended later, 20 min (X + S.D.-= 3 + 6 min) after sunset; here most
spawnings occurred just past sunset (fig. 5). Reproduction by fish at tne
two sites did not differ in any other aspect.

We did not observe any strong lunar periodicity in the daily spawning
behavior of H. guttavarius. However, our data do not allow distinguishing
between the possible effects of 1) different phases of the moon during the
times that each group was observed vs. 2) the difference in depth and
corresponding changes in light levels during dusk.

Effects of Disturbance on Spawning Behavior

All H. guttavarius persisted in spawning in spite of overt and constant
harassment. Harassed pairs attempted twice as many spawning clasps and
continued for about 25 min later than nonharassed (day before and day after)
pairs (tables 1, 2). However, the frequency of spawning clasps attempted per
5 min remained the same (about 2 + 1 clasps; tables lb, 2b). The number of
times a pair changed spawning sites relative to the number of spawning clasps
also did not vary significantly (tables lc, 2c). One particularly striking
effect of disturbance on these spawning fish was that, as spawning was
continuously disrupted, a pair attempted to spawn progressively nearer the
substratum (based on our qualitative observations). At the end of the
disturbance period, several pairs were trying to spawn next to the bottom
while hidden among coral/gorgonian branches. This result simply may indicate
tnat once ovulation has begun, the fish must spawn.

Overall, the fish persisted in spawning despite the aggressive behavior
of the scuba divers. However, we have no estimate of the fate of zygotes
when released next to the bottom and in between coral/gorgonian branches as
compared to those normally dispersed about a meter above reef structures.
The experience of the disturbance had no apparent lasting effects on the
fish. The fish's behavior did not differ markedly on the day after compared
to the day before the disturbance (tables 1, 2).

Table l.--Effect of disturbance on the spawning behavior of Hypoplectrus guttavarius. Data for shallow anid deep
water groups were pooled except when related to sunset time. In addition to the above changes, whorl Fish were
disturbed, they would attempt to spawn closer to the substrate until they were on the bottom. Notations in last
rows (e, f) indicate before (-) or after (+) sunset.

Baseline before disturbed During disturbance Day after disturbed
(7 pairs; 10 observations) (6 pairs; 6 observations) (5 pairs; 5 observations)

a. X + S.D. (range) no. spawning
clasps per pair 12 + 3 (7 to 16) 25 t 9 (10 to 33) 12 t 2 (9 to 15)

b. X � S.D. (range) no. spawning
clasps per 5 min 2 � 1(0.6 to 3.6) 2 + 1 (1.5 to 3.1) 2 � 1 (1.5 to 3.2)

c. X + S.D. (range) ratio of
no. moves pair made between
spawning sites to total no.
spawning clasps 0.23 � 0.16 (0 to 0.5) 0.48 � 0.30 (0.06 to 0.86) 0.26 t 0.19 (0 to 0.46)

d. X t S.D. min. (range) duration
of spawning period per day 31 + 8 (11 to 91) 57 + 21 (33 to 69) 28 + 8 (21 to 38)

e. Time (X, range) relative to sunset that spawning commenced (min)

shallow group (full moon) -16 (-22 to -6) (N = 4) -24 (-44 to -8) (N = 3) -11 (-19 to -2) (N = 3)
deep group (new moon) -33 (-50 to -19) (N = 6) -50 (-71 to -35) (N = 3) -40 (-44 to -35) (N = 2)

f. Time (X, range) relative to sunset that spawning ceased (min)

shallow group (full moon) +13 (+ 5 to +18) (N = 4) +30 (+20 to +38) (N = 3) t12 (f 2 to +24) (N = 3)
deep group (new moon) - 2 (-19 to + 5) (N = 6) +10 (- 2 to +21) (N = 3) - (- 6 to - 1) (N = 2)

Table 2.--Statistical comparison of data from table 1. Values listed are
calculated values of Z using the Mann-Whitney U test. Values marked by *
are statistically different at p = 0.05 (critical value for Z 1.960.)

Before vs. After vs. Before vs.
During During After
Disturbance Disturbance Disturbance

a. No. spawning clasps per pair *2.490 *2.008 0.245

b. No. spawning clasps per pair
per 5 min. 1.844 0.548 1.347

c. No. moves pair made between
spawning sites per total
no. sPawning clasps 1.627 1.278 0.367

d. Duration of spawning period *2.169 *2.373 0.857

Spawning of Other Hamlets

Four other hamlets also were observed spawning at the shallow reef site:
H. puella, H. unicolor, H. chlorurus, and H. nigricans. We did not see mixed _
mating among hamlets. Data contrasting aspects of reproduction in these W
fishes and of H. guttavarius are presented in table 3. However, the data set
lacks sufficient detail to allow more than a general indication of how the
pattern and timing of reproduction might vary among congeners of hamlets.

DISCUSSION

Many diurnal tropical marine fishes that remain active during the evening
crepuscular period may risk being eated by piscivores, whose feeding activities
generally are believed to increase at this time (Hobson, 1968, 1972, 1974;
Domm and Domm, 1973; Major, 1977). This increased threat of predation is
reflected in the potential prey's behavior, and most species retreat to
shelter, aggregate, or forage closer to the reef (Hobson, 19783, 1978;
Robertson and Sheldon, 1979). However, many species also periodically spawn
during dusk (e.g., Randall, 1961; Lobel, 1978; Moyer and Nakazano, 1978;
Moyer, 1979; Moyer and Zaiser, 1981; Zaiser and Moyer, 1981; Neudecker and
Lobel, 1982; Moyer, et al., 1983). Spawning during the evening crepuscular
period is thought to reduce the likelihood of predation on planktonic eggs,
since most planktivorous fishes that select such small prey are inactive by
this time. It is noteworthy that the eggs of nestbuilding, demersal-spawning
balistid and pomacentrid fishes hatch after dark, whereupon the larvae become
planktonic (Allen, 1972; Moyer and Bell, 1976; Ross, 1978; Fricke, 1980;
Lobel and Johannes, 1980; Doherty, 1983). Diurnal planktivores consume
smaller prey, including eggs, than nocturnal planktivores which infrequently

Table 3.--Reproductive behavior by other hamlets (Hypoplectrus spp.). These fish were observed at 25-30 ft.
depth, February 25-March 2, 1980. For H. nigricans (*), see Fischer, 1980, for comprehensive spawning data.

H. puella H. unicolor H. chlorurus H. nigricans*
(N - 2 pairs) (N = 2 pairs) (N = 2 pairs) TN = 1 pair)

X (range) no. spawning clasps per day 10 (8-12) 12 (6-17) 17 (16-18) 27

X (range) no. spawning clasps per 5 min. 3 (3-4) 2 (2-2) 2 (-2-2) 1

X (range) no. moves pair made between
spawning sites per 10 spawns 5 (4-6) 7 (7-7) 7 (7-7) 7

X (range) min. duration of spawning
per day 15 (14-16) 27 (16-38) 42 (38-45) 101

Minutes before sunset that spawning
commenced 5 + 6 13 + 16 28 + 39 87

Minutes after sunset that spawning
ceased 8 + 11 1 + 24 8 + 11 14

eat eggs (Hobson and Chess, 1978). Most egg-eating fishes cease foraging as *
light levels diminish, since low light levels make visual orientation difficultW
and also predation by piscine predators generally increases at this time
(Hobson, 1972, 1974: Hobson and Chess, 1978). Additionally, any planktivores
still actively feeding through dusk may be quickly satiated by the simultaneous
spawning of many fishes.

Spawning during dusk may reduce mortality of fish eggs by predation, but
it also may increase the potential risk of predation for adult fishes. Hobson
(1968, 1972, 1974, 1978; Hobson, et al., 1981) has documented extensive
evidence suggesting increased predation by fish predators on tropical reef
fishes exposed or otherwise vulnerable during the crepuscular periods. Fishes
also are thought to be more vulnerable to predation when preoccupied with
mating (Johannes, 1978, 1981; Robertson, 1983). Furthermore, it must be
noted that there are many fishes, particularly labrids and scarids, which
spawn throughout the day, appearing more in phase with tidal currents than
crepuscular periods (Barlow, 1981; Kuwamura, 1981; Robertson, 1983).

In this study, we wanted to evaluate how a fish would alter its mating
pattern when disturbed or attacked. The results might then provide some insight
into the selective pressures molding the reproductive tactics of coastal marine
species in the tropics. At the time, the best arrangement we could devise to
test the possible effect of predation was by simulating predator behavior using
scuba divers. Other studies also have considered a fish's response to a human
diver as indicative of reaction to a piscine predator (e.g., Coates, 1980).
Nevertheless, results of this manipulation should be related only tenuously to
how the hamlet might respond to a fish predator. A fish predator probably
would not continue to attack the same prey throughout dusk.

The hamlet, Hypoplectrus guttavarius, altered its normal mating behavior
when harassed by 1) attempting more clasps longer into the night and 2) spawning
progressively nearer to the bottom in and among shelter. In no case was spawning
terminated early; in fact, the opposite occurred, and fish prolonged spawning
attempts. This result simply may indicate that once ovulation has begun, the
fish must spawn. The spawning sites were all reef structures (gorgonian and
coral) taller than the surrounding terrain. Only after a pair had shifted
sites repeatedly did they commence spawning nearer the bottom or in shelter.
Difference in survivorship of the free-floating zygotes when released high
above the reef vs. near the bottom was not determinable. Fish which were
attacked and harassed one day showed no indication of lasting behavioral affects
the next day.

The importance attributed to spawning above a tall reef structure is that
it provides a degree of safety for the adults. It allows the fish to mate
relatively high in the water column while remaining as close as possible to
shelter. Spawning high in the water places the free-floating eggs beyond the
grasp of benthic planktivores and in a position most favorable to advection by
currents.

Hobson (1972, 1974) has suggested that the well defined twilight activities
of tropical reef fishes have been shaped by the threat of crepuscular predators
(Hobson, et al., 1981). If this hypothesis is correct, then it would seem that

fishes spawning during evening crepuscular periods have not been overtly deterred
by the potential threat of predation. Alternatively, the adaptive pressures
favoring zygote survival may have been more influential in molding the diel
timing of reproduction than factors affecting survival of the parents (e.g.,
predation) while spawning during the evening crepuscular period.

ACKNOWLEDGMENTS

We are grateful for the stimulating conversation and help received from
our fellow aquanaut, William Hamilton III, both underwater and topside. The
study was made possible by the tireless support of the HYDROLAB personnel, Bill
and Joan Schane, Barry Walden, Rod Catanach, and Joe Langersdorf, and support
divers, Scott Grace and Don Morris. Valerie Paul, Stacy Tighe, and Nancy Wolf
greatly aided in the field observations. We also received generous support
from the West Indies Laboratory, for which we are grateful. We thank Marjorie
Reaka and William McFarland for their thoughtful reviews of this paper. This
research was funded by a NOAA HYDROLAB (NULS 80-1) grant to S. Neudecker and by
an OPER grant from the Institute of Ecology, University of California at Davis,
to William Hamilton, S. Neudecker, and P. Ward.

LITERATURE CITED

Allen, G. R. 1972. Anemonefishes. T. F. H. Publ., Neptune City, N.J.
288 pp.

Barlow, G. W. 1975. "On the Sociobiology of Some Hermaphroditic Serranid
Fishes, the Hamlets in Puerto Rico." Mar. Biol. 33:295-300.

Barlow, G. W. 1981. "Patterns of Parental Investment, Dispersal and Size
Among Coral-Reef Fishes." Environ. Biol. Fish. 6:65-85.

Colin, P. L. 1976. "Filter-Feeding and Predation on the Eggs of Thalassoma
sp. by the Scombrid Fish Rastrelliger kanagurta." Copeia 1976:596-597.

Colin, P. L. 1978. "Daily and Summer-Winter Variation in Mass Spawning of the
Striped Parrotfish, Scarus croicensis." Fish. Bull. 76:117-124.

Coates, 0. 1980. "Anti-Predator Defense Via Interspecific Communication in
Humbug Damselfish, Dascyllus aruanus (Pisces, Pomacentridae)." Z. Tierpsychol.
52:355-364.

Doherty, P. J. 1983. "Diel, Lunar and Seasonal Rhythms in the Reproduction of
Two Tropical Damselfishes: Pomacentrus flavicauda and P. wardi." Mar. Biol.
75:215-224.

Domm, S. B., and A. J. Domm. 1973. "The Sequence of Appearance at Dawn and
Disappearance at Dusk of Some Coral Reef Fishes." Pac. Sci. 27:128-135.

Fischer, E. A. 1980a. "The Relationship Between Mating System and Simultaneous
Hermaphroditism in the Coral Reef Fish, Hypoplectrus nigricans (Serranidae)."
Anim. Behav. 52:620-633.

Fischer, E. A. 1980b. "Speciation in the Hamlets (Hypoplectrus; Serranidae) _
A Continuing Enigma." Copeia 1980:649-659.

Fischer, E. A. 1981. "Sexual Allocation in a Simultaneously Hermaphroditic
Coral Reef Fish." Amer. Nat. 117:64-82.

Fricke, H. W. 1980. "Mating Systems, Maternal and Biparental Care in Triggerfish
(Balistidae)." Z. Tierpsych. 53:105-222.

Graves, J. E., and R. H. Rosenblatt. 1980. "Genetic Relationships of the
Color Morphs of Serranid Fish, Hypoplectrus unicolor." Evolution 34:240-245.

Hobson, E. S. 1965. "Diurnal-Nocturnal Activity of Some Inshore Fishes in the
Gulf of California." Copeia 1965:291-302.

Hobson, E. S. 1968. Predatory Behavior of Some Shore Fishes in the Gulf of
California. U.S. Fish Wildlife Service Res. Rept. No. 73. 92 pp.

Hobson, E. S. 1972. "Activity of Hawaiian Reef Fishes During the Evening and
Morning Transitions Between Daylight and Darkness." Fish. Bull. 70:715-740.

Hobson, E. S. 1973. "Diel Feeding Migrations in Tropical Reef Fishes."
Helgolander wiss. Meeresunters. 24:361-370.

Hobson, E. S. 1974. "Feeding Relationships of Teleostean Fishes on Coral
Reefs in Kona, Hawaii." Fish. Bull. 72:915-1031.

Hobson, E. S. 1978. "Aggregating as a Defense Against Predators in Aquatic
and Terrestrial Environments," pp. 219-234. In: E. S. Reese and F. J.
Lighter (eds.), Contrasts in Behavior. Wiley and Sons, New York.

Hobson, E. S., and J. R. Chess. 1978. "Trophic Relationships Among Fishes and
Plankton in the Lagoon at Enewetak Atoll, Marshall Islands." Fish. Bull.
76:133-153.

Johannes, R. E. 1978. "Reproductive Strategies of Coastal Marine Fishes in
the Tropics." Environ. Biol. Fishes 3:65-84.

Johannes, R. E. 1981. Words of the Lagoon. University of Calfornia Press,
Berkeley. 245 pp.

Jones, G. P. 1981. "Spawning-Site Choice by Female Pseudolabrus celidotus
(Pisces: Labridae) and Its Influence on the Mating System." Behav. Ecol.
Sociobiol. 8:129-142.

84.

Kuwamura, T. 1981. "Diurnal Periodicity of Spawning Activity in Free-Spawning
Labrid Fishes." Jap. J. Ichthyol. 28:343-348.

Lobel, P. S. 1978. "Diel, Lunar and Seasonal Periodicity in the Reproductive
Behavior of the Pomacanthid Fish, Centropyge potteri, and Some Other Reef
Fishes in Hawaii." Pac. Sci. 32:193-207.

Lobel, P. S., and R. E. Johannes. 1980. "Nesting, Eggs and Larvae of
Triggerfishes (Balistidae)." Environ. Biol. Fish. 5:251-252.

Lobel, P. S., and A. R. Robinson. 1983. "Reef Fishes at Sea: Ocean Currents
and the Advection of Larvae," pp. 29-38. In: M. L. Reaka (ed.), The Ecology
of Deep and Shallow Coral Reefs. NOAA Symp. Ser. Undersea Res., VY3T 1(1).
NOAA Undersea Res. Progr., Rockville, Md.

Major, P. F. 1977. "Predator-Prey Interactions in Schooling Fishes During
Periods of Twilight: A Study of the Silverside Pranesus insulanum in Hawaii."
Fish. Bull. 72:415-426.

McFarland, W. N. 1982. "Recruitment Patterns in Tropical Reef Fishes," pp. 83-
91. In: G. R. Huntsman, W. R. Nicholson, and W. W. Fox (eds.), The Biological
Bases-ior Reef Fishery Management. NOAA Tech. Mem. NMFS-SEFC-80. Nat. Mar.
Fish. Serv., Beaufort, N.C.

Meyer, K. A. 1977. "Reproductive Behavior and Patterns of Sexuality in the
Japanese Labrid Fish Thalassoma cupido." Jap. J. Ichthyol. 24:101-112.

Moyer, J. T. 1974. "Notes on the Reproductive Behavior of the Wrasse Thalassoma
cupido." Jap. J. Ichthyol. 21:34-36.

Moyer, J. T. 1975. "Reproductive Behavior of the Damselfish Pomacentrus
nagasakiensis at Miyake-jima, Japan." Jap. J. Ichthyol. 22:151-163.

Moyer, J. T. 1979. "Mating Strategies and Reproductive Behavior of ostraciid
Fishes at Miyake-jima, Japan." Jap. J. Ichthyol. 26:148-160.

Moyer, J. T., and L. J. Bell. 1976. "Reproductive Behavior of the Anemonefish
Anphiprion clarkii at Miyake-jima, Japan." Jap. J. Ichthyol. 23:23-32.

Moyer, J. T., and A. Nakazono. 1978. "Population Structure, Reproductive
Behavior and Protogynous Hermaphroditism in the Angelfish, Centropyge
interruptus at Miyake-jima, Japan." Jap. J. Ichthyol. 25:25-39.

Moyer, J. T., and M. J. Zaiser. 1981. "Social Organization and Spawning of
the Pteroine Fish, Dendrochirus zebra, at Miyake-jima, Japan." Jap. J.
Ichthyol. 28:52-69.

Moyer, J. T., R. E. Thresher, and P. L. Colin. 1983. "Courtship, Spawning and
Inferred Social Organizations of American Angelfishes (Genera Pomacanthus,
Holocanthus, and Centropyge; Pomacanthidae)." Environ. Biol. Fish. 9:25-39.

Neudecker, S., and P. S. Lobel. 1982. "Mating Systems of Chaetodontid and
Pomacanthid Fishes at St. Croix." Zeit. Tierpsychol. 59:299-318.

Randall, J. E. 1961. "Observations on the Spawning of Surgeon-Fishes
(Acanthuridae) in the Society Islands." Copeia 1961:237-238.

Randall, J. E. 1968. Caribbean Reef Fishes. T. F. H. Publ., Neptune City, N.J.
318 pp.

Randall, J. E., and H. A. Randall. 1963. "The Spawning and Early Development
of the Atlantic Parrotfish, Sparisoma rubripinne, with notes on Other Scarid
and Labrid Fishes." Zoologica 48:49-60.

Robertson, D. R. 1983. "On the Spawning Behavior and Spawning Cycles of Eight
Surgeonfishes (Acanthuridae) from the Indo-Pacific." Environ. Biol. Fish.
9:193-223.

Robertson, D. R., and S. G. Hoffman. 1977. "The Roles of Female Mate Choice
and Predation in the Mating Systems of Some Tropical Labrid Fishes." Zeit.
Tierpsychol. 45:298-320.

Robertson, D. R., and J. M. Sheldon. 1979. "Competitive Interactions and the
Availability of Sleeping Sites for a Diurnal Coral Reef Fish." J. Exp. Mar.
Biol. Ecol. 40:285-298.

Ross, R. M. 1978. "Reproductive Behavior of the Anemonefish Amphiprion
melanopus on Guam." Copeia 1978:103-107.

Thresher, R. E. 1978. "Polymorphism, Mimicry, and the Evolution of the Hamlets
(Hypoplectrus, Serranidae)." Bull. Mar. Sci. 28:345-353.

Thresher, R. E. 1980. Reef Fish: Behavior and Ecology on the Reef and in the
Aquarium. Palmetto Pub-l. C-o.t, St. Petersburg, Florida.- 1- pp.

Zaiser, M. J., and J. T. Moyer. 1981. "Notes on Reproductive Behavior of the
Lizardfish Synodus ulae on Miyake-jima, Japan." Jap. J. Ichthyol. 28:95-98.

PATTERNS OF REPRODUCTION IN SMALL JAMAICAN
BRITTLE STARS: FISSION AND BROODING PREDOMINATE

Roland H. Emson1
Department of Zoology
King's College, University of London
London, WC2R 2LS United Kingdom

Philip V. Mladenov
Biology Department
Mount Allison University
Sackville, New Brunswick, EOA 3C0 Canada

lain C. Wilkie
Department of Biological Sciences
Glasgow College of Technology
Glasgow, G4 OBA United Kingdom

ABSTRACT

Brittle stars were collected from coralline algae and from an adjacent
turtle grass bed in the back reef lagoon at Discovery Bay, Jamaica. Fissiparous
species are numerically predominant in both habitats. These species, although
systematically diverse, share strikingly similar life history characteristics,
including small adult size (< 4.5 mm disc diameter), low fecundity, yet probable
planktotrophic development and asexual propagation. Brooding species are also
well represented. They too are small in adult size (< 4.0 mm disc diameter)
but are hermaphroditic direct developers, brooding small numbers of young in the
bursae. There appears to be selection for small adult size in brittle stars in
the habitats studied here. Small size in turn is correlated with modified
reproductive patterns like fissiparity and brooding. The relative importance
of these patterns in reef habitats is considered and the frequency and geographic
distribution of fissiparity and brooding in brittle stars briefly discussed.

INTRODUCTION

Coral reefs present many microhabitats which may provide refuges from the
pressures of predation generally considered to exist on open surfaces in such
areas. Coralline algae, sponges, and coral boulder scree all offer cryptic
spaces exploitable by certain species. Such crevices potentially provide both
a refuge for juveniles of large species and a habitat in which small species
can complete their entire life history.

During studies on the reproductive biology of the small fissiparous brittle
star Ophiocomella ophiactoides (H. L. Clark) at Discovery Bay, Jamaica, W.I.
(Mladenov, et al., 1983; Mladenov and Emson, 1984), we became aware that in the

1 Contribution number 343 of the Discovery Bay Marine Laboratory of the
University of the West Indies.

coralline algal habitat occupied by 0. ophiactoides and adjacent turtle grass,
a number of other small brittle stars were present, some in abundance. Of
these species, some showed evidence of fissiparity; others, however, were
nonfissiparous brooding species which nevertheless shared certain characteristics
with the fissiparous forms. In this paper, we present data on the relative
abundance of brittle stars in these microhabitats, briefly review some of the
life history characteristics of these brittle stars, discuss the similarities
and differences between them, speculate on the reasons for the abundance of
small brittle stars in these habitats, and comment on the association of small
size with fissiparity and brooding in brittle stars.

STUDY SITES AND METHODS

The study sites consisted of (1) a shallow, sheltered cove (known locally
as Maze Cove or the Blue Maze) located in a part of the back reef just west of
the dive locker of the Discovery Bay.Marine Laboratory, Jamaica, and (2) a
closely adjacent bed of turtle grass.

The rocky walls of the cove are covered with a deep carpet of coralline
and other red algae in which Amphiroa rigida and A. fragillisima are most
prominent. These algae are a l Cdoark incolor and,-except at the bases of the
fronds, relatively free of detritus. During August 1981 and July 1984, clumps
of algae were carefully transferred to polythene bags underwater, removed to
the laboratory, and searched for brittle stars and other fauna. The brittle
stars were identified, examined for evidence of cross-disc division (three long
arms and three shorter regenerating arms on the disc), counted, and measured
(the parameters noted were the greatest dimension of the disc and, in the case
of the fissiparous species, the length of the arms). The sexual condition of
the animals was later examined by dissection. In the case of Ophiocomella
ophiactoides, egg size was obtained by measuring the diameter of spawned,
fertilized eggs. For the other species studied, the diameter of the largest
oocytes present in histological sections of ripe females was measured. For
estimation of total egg number, the ovaries were removed, blotted free of excess
moisture, and weighed. A small fraction was then separated, weighed, and the
number of eggs counted. Total egg number was obtained by multiplication.

The site adjacent to the cove has a sand and silt substratum that supports
an abundance of turtle grass, Thalassia testudinum, as well as a variety of
algae. The bases of the Thalassia are uniformly pale, due to attached detritus
and the presence of dead decaying leaf bases on the outsides of the plants.
Detritus is abundant around and between the plants. Thalassia shoots were
pulled from the substratum and placed in polythene bags underwater along with
samples of the surface of the substratum from immediately around the base of
the plants. This material subsequently was treated similarly to that obtained
from the algal clumps.

RESULTS

The Species and Their General Characteristics

The brittle stars found in the algal turf are listed in table 1A.. This
table also shows their relative abundances in these microhabitats and gives

Table 1.--Characteristics of the brittle stars from two adjacent habitats at Discovery Bay, Jamaica, in 1981.
A. Coralline algal turf (N = 195); B. Thalassia bed (N = 75). Algal volume 1.5 1. Thalassia volume approxi-
mately 6 1.
Relative Maximum Mode of
Species Family Abundance disc size Appearance Reproduction
(% of specimens (mm)
in sample)

A. Ophiocomella Ophiocomidae 69 4.5 Dark green, Fissiparous and broadcasts
o-phi-actoides arms banded planktotrophic larvae
(Mladenov, et al., 1983;
Mladenov an-Em-son, 1984)

Amphipholis Amphiuridae 21 3.0 Brown disc, Brooder (Fell, 1946)
o squamata arms paler
and often
banded

Ophiactis Ophiactidae 9 3.0 Brown or Fissiparous and broadcasts
savignyi mottled disc, larvae; probably plankto-
banded brown trophic (Mortensen, 1931)
and white arms

Ophiostigma Amphiuridae 1 4.0 Sand colored Fissiparous; mode of sexual
isacanthum reproduction unknown

Table 1.--Continued

Relative Maximum Mode of
Species Family Abundance disc size Appearance Reproduction
(% of specimens (mm)
in sample)

B. Ophiostigma Amphiuridae 52 4.0 As in A above As in A above
isacanthum

Ophiolepis Ophiolepidae 21 4.0 Sand colored Brooder (Hendler, 1979)
pauci spina

Amphipholis Amphiuridae 15 3.0 As in A above As in A above
squamata

Ophiocomella Ophiocomidae 5 4.5 As in A above As in A above
ophiactoides

Ophioderma Ophiodermatidae 4 4.5 Sand colored Only juveniles in this habi-
cinereum tat. Mode of reproduction
unknown.

Ophiactis Ophiactidae 1 3.0 As in A above As in A above
savignyi

Ophiothrix Ophiothricidae 1 - Only juveniles in this habi-
oerstedi tat. Adults broadcast
lecithotrophic larvae
(Mladenov, 1979)

Table 2.--Relative abundance (% of specimens) in samples from the coralline
algal turf habitat in 1984. * = all juveniles with disc diameter
less than 5 cm.

A B C D

Total number recovered N = 115 N = 164 N = 50 N = 183

Algal volume (ml) 1,505 1,390 345 1,260

Ophiocomella ophiactoides 76.5 61.0 58 75.4

Amphipholis squamata 10.4 12.2 34 14.2

Amphiura stimpsoni 5.2 16.4 2 9.9

Ophiactis savignyi 3.4 3.0 0 0

Ophiostigma isacanthum 0.9 5.5 4.4 0

Ophiactis algicola - 0.6 2 0.5

Ophiocoma pumila* 0.9

Ophiothrix oerstedi* 0.9

Ophioderma appressum* 1.8 0.6

Ophiolepis paucispina - 0.6

details of their sizes, appearances, and reproductive modes. Of the four
species present in 1981, Ophiocomella ophiactoides was the most abundant and
Ophiostigma isacanthum, the least. Further samples from this habitat in 1984
(table 2) show that the pattern is consistent over time, although in these
later samples rather more occasional species were taken and Amphiura stimpsoni
was moderately abundant. Four of the six most common species (0. ophiactoides,
Ophiactis savignyi, Ophiactis algicola, and 0. isacanthum) in the latter
collections are fissiparous, able to reproduce asexually by cross-disc division,
and two (Amphipholis squamata and A. stimpsoni) are brooders. The fissiparous
species are members of three different families (Ophiocomidae, Ophiactidae,
Amphiuridae). The species represented as juveniles may be found in abundance
as adults elsewhere in Discovery Bay.

A rather similar group of species of brittle stars was found in the samples
taken (in 1981 only) from the detritus surrounding the Thalassia testudinum
shoots and, to a small extent, the axils of Thalassia leaves (table IB). Of
the seven species found, all except Ophioderma cinereum were also in the algal
turf. However, the relative abundances of these species are very different in

the two habitats, since 0. isacanthum, Ophiolepis paucispina, and A. squamata
are the prominent species in this habitat. Only A. squamata achieves a similar
abundance in both habitats, and fissiparous species are numerically predominant
in both habitats.

Two features were common to all the species in these habitats. One is
small body size. No species exceeds 4.5 mm maximum disc di-ameter, and most
species are smaller than this. All species except 0. cinereum and Ophiothrix
oerstedi included reproductively active adult indivTiuals. The second feature
common to these species is the color pattern of the brittle stars. All are
very difficult to see in the natural habitat because of their size and because
the coloration of the animals renders them cryptic; the banded arms of several
species are particularly concealing. Even when the algae were dissected in the
laboratory, many brittle stars were detected only when they moved. Furthermore,
for those residing in coralline algae, the arms were similar in diameter to the
algal strands. It was noticeable that other fauna found within the algal
clumps, notably the small (2-5 cm) holothuroid Synaptula hydriformis and some
unidentified amphipods, also were cryptically patterned. A dark green with
paler blotches, S. hydriformis is very similar to 0. ophiactoides in color.

Asexual Reproduction in the Fissiparous Species

The overall percentage of individuals showing evidence of fission in each
of the three common fissiparous species is very high (table 3) and is clearly
an important means of population increase in all three. This similarity is
only one of several that can be discerned. For instance, the range of disc
size was very similar (table 3), and the size frequency histograms (fig. 1)
were virtually identical in pattern, each population being dominated by
individuals with disc diameters of 1.0-2.5 mm. Few animals of larger disc size
were evident, and very small animals also were missing. Data were available
only for August 1981 and July 1984 for 0. savignyi and 0. isacanthum, but for
0. ophiactoides data were also available-for June, July,-and December of 1981
TMladenov, et al., 1983) and July of 1984; the size frequency patterns were
identical at all times. No evidence of larval recruitment was apparent in that
species at any time. While it remains possible that larval recruitment occurred
in intervening months, we suspect that the populations of all three species
are, at these sites, maintained almost solely by asexual reproduction.

Another similarity between the species is in the relationship between disc
size and evidence of recent fission. Table 4 shows the proportion of animals
with different disc sizes found in each of four regeneration categories. In
all three species, there is a clear trend: most small animals show evidence of
recent splitting (i.e., a low ratio for length of short/long arms), while a
high proportion of large animals do not.

Sexual Reproduction in the Fissiparous Species

Larger individuals in all three fissiparous species may contain mature
functional gonads (table 3 and below). Mladenov and Emson (1984) already have
reported that 0. ophiactoides with disc diameters greater than 2.2 mm can

Table 3.--Some features of fissiparous brittle stars from Discovery Bay, Jamaica. 1 = Diameter of spawned
fertilized eggs; 2 = largest oocytes present in mature females.

% fission Egg number
Range of apparent in Minimum disc size of largest
Species disc size population of mature females Egg diameter specimen sampled
(mm)

Ophiocomella 1.0-4.5 90 2.2 mm 0.08 mmI 7,400
ophiactoides
(N = 359)

Ophiostigma 1.0-4.0 80 3.3 mm 0.12 mm2 5,000
isacanthum
(N = 56)

Ophiactis 1.0-3.0 67 None of the specimens
savignyi- contained gonads
(N = 65)

60.

SO.

~40 ~Ophlocomella Ophiostigma Ophiactis
40.
n=359 n=56 n=65
30

.6 1.6 2-6 36 6 1'6 2.6 3'6 .6 16 2.6 3-6

SIZE(mm)

Figure 1.--Size frequency distributions of discs of the three fissiparous
brittle stars at Discovery Bay (August 1981).

contain ripe gonads, and of 246 animals with disc diameters greater than 2.0
mm, 57 were female (23%). Of the 56 specimens of 0. isacanthum examined during
this study, 11 with disc diameters greater than 2.7-mm had recognizable
gonads. Of the 11, 2 with disc diameters of 3.6 and 3.3 mm were mature
females, 3 with disc diameters of 3.5 and 2.8 mm were mature males, and the
remaining 6 were unsexable. None of the 0. savignyi examined from the
Discovery Bay sites contained gonads (table 3), but Emson and Wilkie (1984)
have shown that at the Bogue Islands in Montego Bay on the Jamaican north
coast, 0. savignyi achieved a larger size (up to 5.5 mm) and specimens with
disc diameters greater than 4.0 mm may contain ripe ova and achieve successful
sexual reproduction. The largest oocytes in these specimens measured 0.10 mm
in diameter, and the largest individual (5.5 mm disc diameter) contained
roughly 10,000 eggs.

The three fissiparous species are similar in producing small eggs within
the size range (0.07-0.20 mm) regarded as resulting in planktotrophic larvae
(Hendler, 1975; Mladenov, 1979).

Since disc size and egg size are similar in the three species, it is not
surprising that reproductive output, as measured by maximum number of eggs
produced, is roughly equivalent (table 3). The maximum egg number for these
species is two to three orders of magnitude lower than that recorded for nonfis-
siparous brittle stars with planktotrophic development (Hendler, 1975; Mladenov,
1979; Mladenov and Emson, 1984). Compared to their nonfissiparous planktotrophic

Table 4.--Comparison of the regeneration patterns of the three fissiparous brittle stars found at Discovery Bay,
Jamaica. Regeneration category refers to the ratio of length of short arms to length of long arms. Data are
given as percent of all individuals with a given disc diameter.

Ophiocomella
Species ophiactoides Ophiactis savignyi Ophiostigma isacanthum
(N = 359) (N = 65) (N= 56)
Regeneration
to category (%) 0-25 26-50 51-75 75+ 0-25 26-50 51-75 75+ 0-25 26-50 51-75 75+

0-0.5 0 0 0 0 0 0 0 0 0 0 0 0
Disc 0.6-1.0 63 37 0 0 42 12 21 25 75 0 0 25
Diameter 1.1-1.5 55 28 14 3 21 35 13 31 53 18 19 10
(mm) 1.6-2.0 51 28 17 4 0 0 33 67 20 20 27 33
2.1-2.5 45 16 26 13 0 0 0 100 40 0 20 40
2.6-3.0 14 42 15 29 0 0 0 0
3.1-3.5 0 28 14 58 0 0 0 100
3.6-4.0 0 0 0 100 0 0 0 0
4.1-4.5 0 0 0 100

counterparts, therefore, these fissiparous brittle stars broadcast exceedingly
small numbers of larvae at each spawning, with important potential consequences
for recruitment. Even if these species spawned 10 times a year, their output
would be low. As only a small proportion of the population is sexually
mature, the output of larvae must be very small.

The three nonfissiparous species present as adults (A. squamata, A.
stimpsoni, and 0. paucispina) share similar reproductive methods. They are
hermaphroditic s-pecies which produce a small number of relatively large yolky
eggs. These are brooded within the genital bursae and released, quite possibly
throughout the year, as competent juveniles (although Hendler, 1975, and
Emson, unpub. obs., have noted that release is reduced at some seasons).

Observations of Potential Predators

Many species of fishes live in the water adjacent to the algal carpets.
During our collections, fish were observed picking at the surface of the
algae on several occasions. Blue Tangs (Acanthurus coeruleus), Surgeon Fish
(Acanthurus chirurgus), and, in particular, Slippery Dicks (Halichoeres
bivittatus) were observed darting at and pushing into the algal clumps. The
latter species is known to eat invertebrates (Randall, 1967).

DISCUSSION

Throughout the Discovery Bay area, brittle stars are an abundant and
obvious component of the benthic macrofauna. In most areas, large species
predominate, although crevices and algae such as Halimeda sp. harbor both
juveniles of large species and individuals of small species. The algal turf
and turtle grass habitats studied here, however, were unusual in having a
large number of very small, inconspicuous species of brittle stars occurring
in abundance. Large species were extremely rare at these sites and, when
present, were represented only by small juvenile forms. It appears that
these microhabitats are particularly favorable for small species and not
suitable for large species. Brittle stars are considered to be subject to
considerable predation pressure in Discovery Bay (Sides, 1982) and our obser-
vations suggest that fish may pose a threat to brittle stars easily visible
in the algal clumps. It seems possible that a similar situation exists
around the bases of Thalassia.

Larger brittle stars survive by concealment beneath rocks, within rubble,
or in crevices, and by burying within the sediment; many also are nocturnal
in activity. They are probably absent from algal turf because it would be
difficult for them to move through, would offer inadequate concealment, and
perhaps would not provide suitable feeding opportunities. Conversely, the
stiff, narrowly spaced strands of the algal turf provide an ideal microhabitat
for the small species which dominate it, a microhabitat that allows movement,
provides concealment, and, if abundance is a guide, offers good feeding
conditions. The juveniles of large species that settle in this microhabitat
presumably emigrate when they "outgrow" it or become victims of predation.

The axils and detrital layer around the bases of Thalassia are also
inaccessible to large brittle stars. There appears, however, to be no reason
why the equally cryptically colored adults of, for example, the species
of Ophioderma should not be successful in the sandy areas between Thalassia
plants.

It seems reasonable to suppose that similar sized crevices elsewhere on
the reef would also be populated by small species, providing that food is
available, and evidence for this has recently been obtained by Hotchkiss
(1982, who found many of the species considered here living in empty conch
shells), Hendler (in press), and the present authors.

The two microhabitats considered in this study, coralline algal turf and
Thalassia beds, are sufficiently different that no single species is predominant
inFbth7. This has several potential causes. Feeding opportunities will
likely be different in the algal turf and around Thalassia beds. Apparent
detritus feeders such as 0. isacanthum are favored in Thalassia, whereas the
scavenging and apparently predatory 0. ophiactoides (Emson and Mladenov, in
prep.) is favored in algal turf. It is interesting that the highly adaptable
onmivorous detritivore A. squamata achieves moderate abundance in both
habitats. Perhaps reflecting predation pressures, body coloration is also
probably important, since the algal turf generally hosts dark colored species
and the Thalassia paler species.

Small size in brittle stars clearly has resulted in reproductive
constraints. Small size inevitably means that small numbers of eggs can be
produced. Thus the production of long lived, planktotrophic larvae is unlikely
to be a successful mode of reproduction for small species, unless perhaps
gametes are shed several times a year. Several alternatives appear to be
open to small species. Two are alternative means of sexual reproduction.
One is abbreviated development (Hendler, 1975) in which rapid development
results in a post larva. The other involves direct development, often with
associated viviparity or brooding. A third alternative is to supplement
sexual reproduction with asexual multiplication.

All the brittle stars in this sample of small adult size either reproduce
asexually or brood the young. The similarity in life history traits among
the three abundant fissiparous species is remarkable. Despite the fact that
they are from three different families, they are small in size, they reproduce
asexually by fission, and they can produce only a very small number of small
eggs which apparently develop into planktotrophic larvae. Rearing experiments
already have demonstrated that fertilized eggs of 0. ophiactoides develop
into typical planktotrophic ophioplutei (Mladenov and Emson, 1984), and the
egg sizes suggest that this is probably the case for the other species as
well.

Asexual reproduction appears to be the dominant mode of propagation in
all three fissiparous species. It is probably responsible for reliable local
recruitment, whereas sexual reproduction may result in dispersal of small
numbers of larvae which, on occasion, found new populations or recruit into
existing populations maintained mainly by asexual reproduction.

Brooding is the other reproductive pattern exhibited by the small adult
brittle stars of this study. This reproductive mode appears to be a clear
adaptation to low fecundity resulting from small size. Hendler (1979) noted
an association between size and brooding in ophiolepid brittle stars and
suggested that there was selective pressure for brooding because of low
fecundity. Also, Menge (1975) and Emson and Crump (1979) found similar
associations between size and brooding in the sea stars and suggested the
same cause.

A general relationship between small adult size and brooding has been
documented for many groups of invertebrates including echinoderms, and
hypotheses explaining the association have been advanced by Strathmann and
Strathmann (1982). However, brooding is rare among brittle stars, although
it is obvious in the present study. Hendler (1979) noted that only 57 of all
known brittle stars brood, and of these only 7 are tropical. It seems that,
although the particular conditions of the habitats sampled here favor this
life history pattern, elsewhere other methods are selected.

The alternative to brooding, abbreviated development (Hendler, 1975), is
also characteristic of species with relatively small disks that are capable
of producing eggs in numbers of the same order of magnitude as fissiparous
species (e.g., Amphiura chiajei; Hendler, 1975). Species with this reproductive
pattern are present on coral reefs but were not found in our study.

Fission is also uncommon in brittle stars. Emson and Wilkie (1980)
reported that only 34 species possess this habitat, although they failed to
record 0. isacanthum as fissiparous (Hotchkiss, 1982). By contrast with
brooding, however, fissiparity is largely a warm water phenomenon. Emson and
Wilkie (1980) recorded that 24 of the 34 fissiparous species were tropical in
distribution. The reasons for these patterns are unknown and would bear
further investigation.

The similarity of these results to the situation described by Boffi
(1972) for Brazil is considerable. Not only are the fissiparous ophiactids
Ophiactis lymani and Amphipholis squamata the predominant species in the
algal turfs examined by Boffi, but the other species in which adults were
present were Amphipholis januarii and Ophiactis savignyi. The remaining four
species were represented by juveniles and uncommon.

It thus seems possible that these reproductive patterns may prevail
among small brittle stars inhabiting algae throughout the tropical and
subtropical Atlantic, since the present authors have established that similar
reproductive patterns are characteristic of small brittle stars from Bermuda
and Carrie Bow Cay, Belize. Hotchkiss (1982) observed many of the same
species in discarded conch shells, demonstrating that these reproductive
patterns probably are adaptive for many cryptic habitats in sheltered areas
of coral reefs.

Evolution to a small size is infrequent in brittle stars, but those
species which achieve it may become very successful. The brooding Amphipholis
squamata and the fissiparous Ophiactis saviqnyi are among the most abundant
brittle stars. Their size apparently allows them access to an underexploited

microhabitat, and their biological characteristics, particularly their
reproductive patterns, enable them to achieve remarkable success in this
environment. Much of contemporary life history theory is concerned with
trade-offs between dispersal ability, fecundity, reproductive effort, and
body size (e.g., Stearns, 1984; Olive, et al., 1984). The situation briefly
examined here illustrates two examples of successful trading and probably
deserves more detailed study.

LITERATURE CITED

Boffi, E. 1972. "Ecological Aspects of Ophiuroids From the Phytal of S.W.
Atlantic Ocean Warm Waters." Mar. Biol. 15:316-328.

Emson, R. H., and R. G. Crump. 1979. "Description of a New Species of
Asterina (Asteroidea) With an Account of Its Ecology." J. Mar. Biol. Assn.
U.K. 54:77-94.

Emson, R. H., and I. C. Wilkie. 1980. "Fission and Autotomy in Echinoderms."
Oceanogr. Mar. Biol. Ann. Rev. 18:155-250.

Emson, R. H., and I. C. Wilkie. 1984. "An Apparent Instance of Recruitment
Following Sexual Reproduction in the Fissiparous Brittlestar Ophiactis
savignyi." J. Exp. Mar. Biol. Ecol. 77:23-28.

Fell, H. B. 1946. "The Embryology of the Viviparous Ophiuroid Amphipholis
squamata Delle Chiaje." Trans. Roy. Soc. New Zeal. 75:419-464.

Hendler, G. 1975. "Adaptational Significance of the Patterns of Ophiuroid
Development." Amer. Zool. 15:691-715.

Hendler, G. 1979. "Sex-Reversal and Viviparity in Ophiolepis kieri, n.sp.,
With Notes on Viviparous Brittlestars from the Caribbean (Echinodermata:
Ophiuroidea)." Proc. Biol. Soc. Wash. 92:783-795.

Hendler, G. 1985. "A Study of Ophiuroid Size, Abundance and Reproductive
Mode in Relation to Coral Reef Zones and Substrata." In: B. F. Keegan
(ed.), Proc. 5th Int. Echinoderm Conference, Galway, Ireland. A. A. Balkema
Press, Rotterdam.

Hotchkiss, F. H. C. 1982. "Ophiuroidea (Echinodermata) from Carrie Bow Cay,
Belize." In: K. Rutzler and I. G. Macintyre (eds.), The Atlantic Barrier
Reef Ecosystem at Carrie Bow Cay, Belize. Smith. Contr. Mar. Sci., Vol. 12:
387-4 12.

Menge, B. L. 1975. "Brood or Broadcast? The Adaptive Significance of
Different Reproductive Strategies in the Two Intertidal Seastars, Leptasterias
hexactis and Pisaster ochraceus." Mar. Biol. 31:87-100.

Mladenov, P. V. 1979. "Unusual Lecithotrophic Development of the Caribbean
Brittle Star Ophiothrix oerstedi." Mar. Biol. 55:55-62.

Mladenov, P. V., R. H. Emson, L. V. Colpitts, and I. C. Wilkie. 1983.
"Asexual Reproduction in the West Indian Brittle Star Ophiocomella ophiactoideW
(H. L. Clark) (Echinodermata: Ophiuroidea)." J. Exp. Mar. Biol. Ecol. 72:1-
23.

Mladenov, P. V., and R. H. Emson. 1984. "Divide and Broadcast: Sexual
Reproduction in the West Indian Brittle Star Ophiocomella ophiactoides and
Its Relationship to Fissiparity." Mar. Biol. 81:273-282.

Mortensen, T. 1931. "Contributions to the Study of the Development and
Larval Forms of Echinoderms. I. The Development and Larval Forms of Some
Tropical Echinoderms; II. Observations on Some Scandinavian Echinoderm
Larvae." K. Danske Vidensk. Selsk. Skr. (Naturv. Math. Afd.), Series 9,
Vol. 4, No. 1, pp. 1-39.

Olive, P. J. W., P. J. Morgan, N. H. Wright, and S. L. Zhang. 1984.
"Variable Reproductive Output in Polychaeta: Options and Design Constraints."
In: W. Engels, W. H. Clark, Jr., A. Fischer, P. J. W. Olive, and D. F. Went
Teds.), Adv. Invert. Reprod. 3:399-408. Elsevier, Amsterdam.

Randall, J. E. 1967. "Food Habits of Reef Fishes of the West Indies."
Stud. Trop. Oceanogr. 5:665-847.

Sides, E. M. 1982. "Estimates of Partial Mortality for Eight Species of
Brittlestars," p. 327 (abstract). In: J. M. Lawrence (ed.), Proc. Int.
Echinoderm Conference, Tampa Bay, Florida. A. A. Balkema, Rotterdam. i

Stearns, S. C. 1984. "The Tension Between Adaptation and Constraint in the
Evolution of Reproductive Patterns." In: W. Engels, W. H. Clark, Jr., A.
Fischer, P. J. W. Olive, and D. F. Went(eds.), Adv. Invert. Reprod. 3:387-398.
Elsevier, Amsterdam.

Strathmann, R. R., and M. F. Strathmann. 1982. "The Relationship'Between
Adult Size and Brooding in Marine Invertebrates." Amer. Nat. 119:91-101.

100

REEF ZOOPLANKTON COLLECTED ALONG A DEPTH GRADIENT
AT DISCOVERY BAY, JAMAICA

Sharon L. Ohlhorst
Fisheries and Wildlife Department
Utah State University
Logan, Utah 84322

ABSTRACT

Reef zooplankton from three sites in the vicinity of Discovery Bay, Jamaica,
were sampled on 29 nights during 1976-77 to determine their abundance and
species composition. DeRth was inversely correlated with zooplankton volume;
10.2 ml of zooplankton/mz/night were collected at 6 m, 5.3 ml at 15 m, and 3.2
ml at 24 m. No seasonality in volume of zooplankton was observed. Variability
in volume both within and between nights at any one site was high, but never
were enough zooplankton collected to meet more than an estimated 12% of the
metabolic requirements of the sessile reef dwellers.

Volume of zooplankton was positively correlated with the number of
individuals in the sample. The unpredictability of zooplankton abundance is at
least partially explained by the swarming behavior of the dominant (usually >
95% of individuals) copepod, the holoplanktonic Oithona colcarva Bowman.
Because of this variability, it is unlikely that local zooplankton abundance
and distribution has any causal effect upon reef species diversity. In addition
to many groups of holoplanktonic copepods, ostracods, isopods, and decapods
were among the most numerous taxa represented. The 6 m site differed from the
sites at 15 m and 24 m in the relative abundance of noncopepod groups.

INTRODUCTION

Zooplankton are an important component of food webs in coral reef
communities. They provide nutrients to corals (Goreau, et al., 1971; Johannes,
et al., 1970; Porter, 1974) and to other benthic suspension _Teders such as
zoanthids (Sebens, 1977), crinoids (Liddell, 1980; 1982) and ophiuroids (Macurda,
1976), and they form a major part of the diet of many reef fish (Hobson, 1974;
Gerber and Marshall, 1974; Hobson and Chess, 1978). Emery (1968) first reported
that many zooplankton reside within coral reefs and, more recently, researchers
using light (Sale, et al., 1976) and emergence traps (e.g., Alldredge and King,
1977, 1980; Hobson Wand c-less, 1979; McWilliam, et al., 1981; Porter and Porter,
1977; Porter, et al., 1977; Walter, et al., 1982T have characterized certain
aspects of the resident zooplankton o 1f hllow reefs and lagoons in the Pacific.
Few studies have examined reef-associated zooplankton in the Caribbean; Robichaux,
et al. (1981) and Youngbluth (1982) placed traps over sand and/or seagrass
beds, while Ohlhorst (1982) reported on diel patterns over reef substrata at
one depth. Only one study (McWilliam, et al., 1981) has addressed the question
of seasonality in zooplankton (from the Pacific), and none have sampled along a
depth gradient on a reef. Further information on the abundance and behavioral
patterns of these zooplankton, especially in the Caribbean, is necessary to the

101

MONTEGO DISCOVERY SAY
BAY

to JAMAICA I ;

KINGSTON
- - - - - - - - - - - - - - -/
A/ ,*24M/
somr EFR

WFR / /E F
Crest . /.

Figure 1.--Map of Discovery Bay, Jamaica, and adjacent fore reef showing the
three sites sampled in this study. EFR = East Fore Reef; WFR = West Fore
Reef.

development of realistic models of reef energetics and an understanding of the
behavioral and feeding strategies of many reef organisms.

This study was undertaken to examine patterns of abundance and distribution
among reef zooplankton at three depths on a Caribbean reef. In this study,
reef zooplankton are those zooplankton [holoplanktonic, meroplanktonic, and
demersal forms (Robichaux, et al., 1981)] which reside for some period in or
near the reef substratum and at some time migrate at least 1 m into the water
column. Most planktivory by the sessile reef community should be upon this
group of zooplankton during their vertical migration. Zooplankton were collected
with emergence traps in a collection program designed to answer the following
questions about the quantity and predictability of zooplankton available as a
potential food source for reef organisms: (1) What quantity of zooplankton is
available to the reef community, especially to sessile reef-dwelling organisms?
(2) How variable is the amount of zooplankton on different nights? (3) Do reef
sites at different depths differ in a predictable way with respect to abundance
of zooplankton? (4) Do reef sites at different depths differ in species
composition of zooplankton? (5) Is there any predictable cycling of zooplankton
abundance?

MATERIALS AND METHODS

A total of 392 zooplankton samples were collected over 29 new and full
moon nights during 1976-77 using a modified version (Ohlhorst, 1980, 1982) of
the demersal trap described by Porter and Porter (1977). Traps were placed
prior to sunset, and collection bottles were removed the following morning.

102

52*
*

28--
E
24-
w
2
20--
0

16-- *
18-. �

632 2L 0 * 2 2 * '
12-- 2 2 2 0

224 00 2 3
23 2 *2 2 22 2 4
�

4-.. 2 2. 2 3 2 . 2
2 4 24* 4 2 * 2 2 2 ~ *
1t **0 3 0t + *s~
I I* * ! , , !,,I
>0 s p to go >a z < � � � a a a a a a a
_ C _ _ _ _ u w 4 - c - X U W 3 N ao - X 0 m M 4 � -

Figure 2.--Volumes of zooplankton (ml/m2) from the 15 m site are plotted for
the 29 nights sampled. Stars indicate full moon nights, and circles indicate
new moon nights.

Samples were preserved in 5-10% formalin immediately upon return to the surface.
Each sample was concentrated over a 90 m mesh sieve, resuspended in formalin
solution and allowed to settle in graduated centrifuge tubes until a constant
volume was reached (usually after 48 hr) (Porter, et al., 1978).

Zooplankton were collected from three reef sites at Discovery Bay, Jamaica
(fig. 1): a shallow site (6 m) immediately behind the reef crest to the east
of the Discovery Bay ship channel; a 15 m site on a broad, shallow sloping fore
reef plain to the east of the boat channel; and a 24 m site on this same plain
(see Ohlhorst, 1980, and Liddell, et al., 1984, for site descriptions). Because
of the variability in abundance of-zoopankton from different locations at any
one reef site on a single night (fig. 2), traps were placed over the same
permanently marked reef locations at each sampling to eliminate substrata
induced variation when comparing one night with the next.

103

80000 -

70000 -
.

60000-

-J .
50000-

- 40000-

a: �
30000- . N
Z N=48
20000- r2: .75
� |1 S.R.= P<.O0
10000- 1

0 I I I I , I
0 X 2 3 4 5 6 7 8
SETTLING VOLUME (ML)

Figure 3.--Correlation of zooplankton settling volume with the number of
individuals counted in the sample. Results are given for regression analysis
(r2) and Spearman Rank Correlation (S.R.).

Caloric determinations were made on six freeze-dried (10.5-15.5 mg) samples
of zooplankton (all dominated by copepods) using a microbomb calorimeter (Tinkle
and Hadley, 1973). Samples of known volume were ashed at 450�C for 5 hr for
ash weight determination. Samples of known settling volume were freeze dried,
weighed, and the weight/individual calculated using the mean value of 1.6 x 10-4-
individuals/ml settling volume determined from a count of individuals from 30
samples of known settling volume. Samples consisting of 20-30 individuals were
removed and weighed on a Cahn Electro-balance for an independent estimate of
individual weight.

In addition to volumetric determinations on all 392 samples, 56 of the
zooplankton samples from 7 nights at the 3 reef sites were enumerated.

RESULTS

Differences in the "packability" of species may result in difficulties in
interpreting settling volume data. However, the dominance of a single species
and the general similarity in morphology, size, and composition of species in
these samples made volume determinations a relatively easy method by which to
compare sites. As illustrated in figure 3, the number of zooplankton is
positively correlated (Spearman Rank Correlation, p < 0.01) with the settling
volume of zooplankton.

104

*^ Variability in Volume of Zooplankton Within Reef Sites

Volumes of zooplankton from the 15 m site ranged from 0.4-50.8 ml/m2 ( =
5.3 ml/m2) (figs. 2, 4) and, on any one night, differed among sites by 0.4-49.2
ml/m2 (X difference in volume/night = 8.0 ml/m2). The variability occurring
between consecutive nights was as great as that between randomly paired nights
(Mann Whitney U Test, p > 0.05). This wide range in volumes was found for all
three sites, with samples from 24 m showing the least variation (fig. 4).

Differences in Abundance of Zooplankton Between Reef Sites

The results of the zooplankton sampling from all three sites are presented
in figure 4. Volume of zooplankton is inversely correlated with depth (Spearman
Rank Correlation, p < 0.01). The shallow (6 m) site yielded the greatest volume
( = 10.2 ml/m2), while the 24 m site yielded the least (X = 3.2 ml/m2).
Pairwise comparison of volumes from all nights shows that volumes from 6 m vs.
24 m and 15 m vs. 24 m are significantly different (Mann-Whitney U Test, p <
0.01), but not 6 m vs. 15 m. However, these relationships do not hold for every
night considered separately. When the sites are paired night by night, the
volume at 6 m is greater than at 24 m only 45% of the nights (N = 11), and the
volume at 15 m is greater than at 24 m only 57% of the nights (N = 14). While
overall the volumes from 6 m and 15 m do not differ significantly, volumes from
6 m are greater than those from 15 m on 45% of the nights.

When the total numbers of individuals captured at each site (table 1) are
compared, similar patterns are observed as for volume. More individuals occurred
at 6 m than at 24 m (Mann Whitney U Test, p < 0.0006) and more at 15 m than at
24 m (Mann Whitney U Test, p < 0.002), but there was no difference between 6 m
and 15 m. However, the pattern changes when the dominant copepod is removed
from calculations. With Oithona colcarva Bowman removed, the site at 6 m does
not have significantly more individuals than 15 m or 24 m (Mann Whitney U Test,
p > 0.05), and, curiously, the 24 m site has significantly more than the 15 m
site (Mann Whitney U Test, p < 0.002).

The Effect of Lunar Phase on Abundance of Zooplankton

The phase of the moon appears to influence volumes of zooplankton, although
the results from different depths are conflicting. The 6 m site yielded a
significantly greater volume on nights of the new moon than full moon, while
the 15 m site yielded a greater volume on nights of the full moon than new moon
(Mann Whitney U Tests, p < 0.01).

The lunar effect on numbers of zooplankton was significant only at the 6 m
site, where significantly greater numbers (Mann Whitney U Test, p < 0.03) were
captured on nights of the new moon than full moon, both for total plankton and
for plankton without 0. colcarva. However, with only 7 nights (2 full moon, 5
new moon) sampled, this deserves further investigation.

When numbers of zooplankton captured in different phases of the moon are
0 compared among sites, collections at 6 m did not differ from those at 15 m or

105

25 -

20 -

MEAN 950/ cl MEDIAN N(SAMPLES) N(NIGHT)
is- ~~~ 10.24 �3.12 1.5 es 11

70-

60 - 15

40 5.28 �0.74 0.9 214 29

:0 z

z20-~

70-

Yo 24 M

50-

40- ~~~3 21 �0.77 0.5 112 14

40-

30-

20-~ ~ ~~~V7

PLANKTON VOLUME (ML) >2

I 0 6

Table 1.--Zooplankton collected over 1 m2 of reef substrata at three depths.
Mean, standard deviation and percent of samples containing each organism
are given for the common categories of organisms. Samples from 6 m included
8 samples from 3 nights; those from 15 m included 27 samples from 7 nights;
and those from 24 m included 21 samples from 7 nights.

6 M 15 M 24 M

ZooDlankton Mean number (St.Oev.) Mean nu Meannumber (St.Dev.) %

Oithona colcarva 124396 (83206) 100 79307 (51392) 100 54381 (118901) 100
Cyclopoid #1 0 32 (64) 27 492 (617) 71
Cyclopoid #2 27 (50) 88 55 (100) 54 87 (116) 52
Cyclopoid #3 26 (74) 38 47 (85) 63 41 (85) 43
Corycaeus sp. 1832 (2215) 100 690 (771) 85 876 (839) 100
Misc. cyclopoids 136 (277) 38 130 (199) 59 243 (346) 81
Harpacticoid #1 26 (74) 63 281 (333) 93 226 (265) 90
Harpacticoid #2 0 15 (40) 52 32 (70) 48
Misc. harpacticoids 68 (148) 38 501 (336) 93 760 (512) 100
Calanoid #1 3046 (5877) 63 177 (319) 52 47 (76) 33
Misc. calanoids 12649 (17242) 88 477 (539) 89 1820 (2280) 100
Copepod nauplii 21 (58) 13 35 (79) 22 110 (140) 52
Barnacle nauplii 74 (85) 75 42 (88) 56 65 (105) 52
Ostracods 0 87 (196) 89 404 (929) 95
Amphipod #1 0 53 (84) 63 21 (37) 43
Amphipod #2 0 8 (29) 30 20 (46) 29
Misc. amphipods 78 (156) 38 62 (107) 89 61 (66) 86
Isopod #1 322 (430) 75 154 (218) 52 324 (376) 71
Isopod #2 84 (162) 75 243 (264) 89 148 (222) 71
Tanaids 53 (107) 38 112 (165) 89 93 (125) 95
Cumaceans 93 (151) 88 23 (60) 70 9 (22) 57
Mysids 53 (148) 88 85 (131) 74 80 (78) 90
Lucifer sp. 1560 (2413) 75 0 0
Misc. decapods 132 (226) 100 88 (381) 70 43 (55) 81
Polychaetes 148 (161) 100 98 (125) 93 126 (142) 86
Appendicularians 0 15 (49) 15 87 (138) 52
Chaetognaths 0 172 (178) 93 105 (113) 95
Gastropods 77 (106) 50 83 (144) 56 94 (90) 38

Total (includes
uncommon species not 146056 (102924) 83244 (51216) 60904 (118592)
included above)

Figure 4 (opposite page).--Volume frequency histogram for the three reef sites.
The number of samples of differing volumes (converted to volume/m2) is plotted.
Volume is plotted in 2 ml increments with the midpoint volume indicated for
each.

107

24 m on new moon nights (Mann Whitney U Test, p > 0.05) but did have greater
numbers than samples from 15 m or 24 m on new moon nights. This holds for
counts both with and without 0. colcarva, except that samples from 6 m have
more individuals than those from 24 m on full moon nights when O. colcarva is
included.

Composition of Zooplankton

The mean, standard deviation, and percent occurrence for 29 categories of
organisms are presented in table 1. No statistically significant differences
between sites in occurrence of taxa were noted across all nights, with the
exception of the absence of several organisms and the exceptional abundance of
a holoplanktonic decapod, Lucifer sp., at 6 m. The relative abundance of
copepod groups was similar (Spearman Rank Correlation, p < 0.01) at all sites,
but the ranking of noncopepod groups differed between 6 m and both 15 m and 24 m
(Spearman Rank Correlations, p > 0.05). The 24 m site possessed a significantly
greater number of organism categories [17.9 + 6.2 (S.D.)] than did the sites at
either 6 m (10 + 3.2) or 15 m (12.5 + 5.9). Evenness and species diversity
were consistentTy higher at 15 m and-24 m than at 6 m (X H' at 6 m = 1.35,
15 m = 2.11, 24 m = 2.46).

Bomb Calorimetry and Biomass Determination

The results of caloric and biomass determinations are presented in table
2. Bomb calorimetry produced caloric values of 5.6 [+ 0.05(S.D.)] calories/mg
dry weight, a figure which agrees well with published values (Cummins and
Wuycheck, 1971). Only 6% of the freeze dried weight was inorganic ash material.
The weighing of a known number of Oithona colcarva individuals (the dominant
organism; Ohlhorst, 1980, 1982) produced weights averaging 2.5 pg/individual,
while the weighing of a known volume with a calculated number of individuals
produced weights averaging 1.6 pg/individual.

DISCUSSION

The highly three-dimensional nature of the reefs in this study made it
impossible to place traps completely flush with the reef floor, and thus some
zooplankton movement undoubtedly occurred both into and out of the traps. A
number of workers have addressed the problems related to demersal trap design
(Robichaux, et al., 1981; Youngbluth, 1982), and their results should be taken
into account when evaluating the results of work based on demersal emergence
traps. Caution is necessary in interpreting the origin of the zooplankton
["demersal traps" catch more than demersal plankton (Robichaux, et al., 1981)]
and their relative abundances due to differing plankton behavior- Robichaux, et
al., 1981; Youngbluth, 1982). Nonetheless, studies such as this, with one
method used throughout, provide data by which different sites can be compared.
The volumes of zooplankton obtained in this study give an indication of the
amount and types of zooplankton available for capture by the sessile inhabitants
of the reef, and serve as a useful basis for comparison of different reef sites..

108

Table 2.--Biomass and caloric value determinations of zooplankton samples (mean + standard deviation).
a: The value of 5.584 calories/mg dry weight agrees well with similar determinations by other workers
(Cummins and Wuycheck, 1971). b: The mean values from weight determinations by two independent
methods (see text). c: Values calculated by Johannes, et al. (1970) for 1 m2 living coral.

Dry Weight Ash Free Calories
(mg) Dry Weight
(mg)

1 ml settling volume 31.5 (�8.1) 29.6 (�+7.5) 175.9 (+45)a
1 Oithona colcarva (0.5 onm)b 0.002 (�+4 x 10-4) 0.0019 (�+4 X 10-4) 0.012 (+3 x 10 3)
Zooplankton over 1 m2 reef
at 15 m 167 (�35) 157 (i33) 928 (�176)
Metabolic requirements of
1 m2 of reefC 6000 33,600

Abundance of Zooplankton

The average volume of 5.3 ml of zooplankton per i m2 reef substratum at 15 m
depth on the East Fore Reef indicates that only 3-4% of the reef energy needs,
as determined by Johannes, et a]. (1970) for an area of high coral cover, could
be met by the capture of alF-othe demersal zooplankton (table 2). While the
6 m site yielded more zooplankton, it was rarely more than 12% of the amount
necessary to sustain the corals. Gut analyses of various reef planktivores,
such as corals (Porter, 1974) and crinoids (Liddell, 1980, 1982), and observations
of feeding by corals (Johannes and Tepley, 1974) confirm that inadequate numbers
of zooplankton are caught to meet the estimated energy needs of these reef
inhabitants.

A comparison of results from this study with those of Alldredge and King
(1977); Porter and Porter (1977); Porter, et al. (1977); McWilliam, et al.
(1981); and Walter, et al. (1982) is presented in table 3. Volume and numbers
of zooplankton collected over Jamaican reefs were larger than those of the
other studies. This difference may reflect either real differences in the
biology of the reefs studied or merely differences in the amount of swarming
copepods captured because of imperfect fit of the traps to the substrata.
Nonetheless, these studies together indicate the range of values for the amount
of demersal zooplankton to be found over coral reefs. Further study using
identical collection methods, including sealing efficiency and plankton mesh
size, is necessary to determine the differences betwen the Caribbean reef
studied in this work and the Pacific reefs studied by others.

Variability in Abundance of Zooplankton

The high degree of spatial and temporal heterogeneity in zooplankton volume
(figs. 2, 4) is probably attributable to the swarming behavior of the dominant
copepod, Oithona colcarva. Although swarms of this species were not observed
by the author at Tamaica, other members of this genus are known to swarm in
large numbers (Emery, 1968; Hamner and Carleton, 1979).

One result of this variability was that patterns of differences in
zooplankton volume between sites varied from night to night, although certain
statistically significant trends were discernible when data from all nights
were pooled. This illustrates how sampling frequency might influence conclusions
about the presence or absence of differences between collection sites; depending
on which nights were sampled, contrary conclusions could be drawn. McWilliam,
et al. (1981) also reported on the variability of zooplankton abundance from
consecutive nights.

Factors Affecting Abundance of Zooplankton

This study indicates that zooplankton volumes for fore reef sites are
negatively correlated with depth within the range of 6-24 m. This is, therefore,
a very important consideration when extrapolating from studies done over shallow
reefs (all of the previous work) to those which may be deeper.

110

Table 3.--Comparative abundances of zooplankton over coral substrata on eight reefs.
a: This study. b: Samples collected as in this study, August 8-9, 1976; N =
24. c: Samples collected as in this study, July 2-3, 1976; N = 12. d: Samples
collected as in this study, July 26-27, 1976; N = 24. e: Alldredge and King
(1977). f: Porter and Porter (1977) and Porter, et al. (1977). g: Walter,
et al. (1982). h: McWilliam, et al. (1981).

Reef location Volume/m2/night (ml) Number/m2/night Ash free dry weight/
m2/night (mg)

Jamaicaa 6m: 10 6m: 1.6x105 6m: 296
15m: 5.3 15m: 805x104 15m: 167
24m: 3.2 24m: 5.3x104 24m: 101

Barbadosb 20m: 3.5 20m: 103x104

Colombiac 15m: 3�4 15m: 1.9xlO4

Curacaod 15m: 4�8 15m: l14x104

Lizard Island 4
Lizard Island 2-5m: l.lx104 2-5m: 95
Lagoone

Philippinesf lOm: 905 lOn: 2O0x104

Philippines9 3m: 5o4x104

Great Barrier
Reefh 4-5m: 3.7x103

The apparent correlation between high zooplankton volume and full moon
periods for collections at 15 m probably reflects the fact that a much higher
proportion of the full moon nights were also "rain" nights (Fisher Exact
Probability Test, p < 0.01). The correlation of volume with "rain" [volumes
correlated only to rainfall (> 0.6 cm) within the 48 hr prior to collection]
is found during both phases of the moon, and in this study it appears that
lunar phase had no effect on the volume of vertically migrating zooplankton on
the East Fore Reef (Ohlhorst, 1980). The implication of rain as a factor
affecting zooplankton volume also has been reported for zooplankton collected
over Puerto Rican reefs (Glynn, 1973) and in Kingston Harbor, Jamaica (Grahame,
1973), where volumes of plankton from surface net samples correlated with
monthly rainfall. It is not clear why rainfall should have the immediate effect
found in this study, unless an increase in floculation of organic matter or
land runoff occurred and produced an immediate increase in the foraging activity

111

of the zooplankton. There was no measurable change in salinity at any site
during rain periods (Ohlhorst, 1980).
Interpretation is further complicated because no correlation between volume
of zooplankton and lunar phase or rainfall was found at the 24 m site, and the
6 m site yielded results opposite to those at 15 m, with more zooplankton
captured on new moon (= "nonrain" in this study) than full moon nights. At the
6 m site, there is an obvious advantage for zooplankton that do not enter the
water column on full moon evenings, as visual predators should be more active
on these nights. It is, therefore, surprising that yields of zooplankton at
15 m were the reverse (highest on full moon evenings), since the reefs at 15 m
were well illuminated on these nights also.

Although Glynn's (1973) collections of plankton from the surface water in
Puerto Rico sampled a different component of the reef zooplankton, his collections
over a shallow reef (< 5 m) through one lunar cycle yielded a pattern consistent
with that found for the shallow (6 m) site in this study; more zooplankton were
collected on the new moon nights than on the full moon nights. In the study by
Alldredge and King (1980) in the Pacific, abundance of zooplankton was not
correlated with lunar phase.

There was no seasonality observed in zooplankton volumes in this study,
although some periods of the year were less intensively sampled (fig. 2).
Periods of high zooplankton yields reflected the occurrence of rain within 48 hr
of sampling; however, collections of very low volumes were interspersed with
these high volumes, casting doubt on any real causative effect (such variation
also is found in other reports of seasonality; Glynn, 1973). It appears, there-
fore, that zooplankton volumes are indicative of previous rainfall, but rainfall
is not necessarily predictive of high zooplankton yields. Moore (1967) and
Grahame (1976) found no patterns of seasonality in their collections of surface
zooplankton from the south coast of Jamaica. In contrast, McWilliam, et al. (1981)
did find seasonality on a Pacific reef, with more zooplankton in the siummer.

With the exception of the trend for greater abundance of zooplankton on
"rain" nights at the 15 m site and for lower yields on full moon ("rain") nights
at the 6 m site, there was no indication that certain of the nights were more
conducive than others for zooplankton migration. Neither the pairing of
collections on the same night from the 15 m and 24 m sites nor the pairing of
collections on the same night from the eight 15 m trap locations produced a
significant correlation (Spearman Rank Correlation, p > 0.05). It appears,
therefore, that abundance of zooplankton cannot be predicted on the basis of
any environmental parameter which would be experienced throughout the reef.

Composition of Zooplankton

As reported earlier for Jamaican reefs (Ohlhorst, 1980, 1982), copepods are
the dominant organism collected in demersal emergence traps placed over the
reef. Because of the great variance, the sites rarely differed significantly
and the ranks of abundances for copepod taxa were correlated between sites.

Due to the absence from the 6 m site of a number of categories of organisms *
and the presence of Lucifer sp. which was rare elsewhere, the rankings of

112

abundances for noncopepod categories differed at 6 m from those at either 15 m
or 24 m. These data indicate differences between shallow and deeper reef sites
which should be investigated further. They also indicate that caution should
be used when extrapolating data from shallow reefs to other reef sites.

Robichaux, et al. (1981) and Youngbluth (1982) discuss the effect of
differing collection.methodology on zooplankton species composition. In light
of the many confounding variables, comparisons between this study and others
will be brief and restricted to shallow reef sites. The data from the 6 m site
in this study compare most closely with those from 3 m in the Philippines
(Walter, et al., 1982). The relative abundance of most plankton groups and the
absolute abundance of certain groups are comparable. Lucifer, for example, is
an important component of the fauna in both shallow reef collections; however,
amphipods are more important in their study than in this study. Although the
abundant taxa from this study are not the same as those from the sealed traps
of Hobson and Chess (1979) and Robichaux, et al. (1981), the values for certain
of the taxa are comparable. Robichaux, et1l. 71981), for example, collecting
over a Thalassia bed (3 m), recovered an average of 56 tanaids and 54 amphipods/m2,
while this study produced an average of 53 tanaids and 78 amphipods/m2 at the 6 m
reef site.

CONCLUSIONS

This study indicates that reef-associated zooplankton can only be of minor
caloric value to sessile reef organisms. While there is a negative correlation
between abundance of zooplankton and depth of the reef site, local abundance is
highly unpredictable. It would appear difficult for planktivores to choose
areas of maximum food potential and unlikely that local zooplankton abundance
has any causal effect upon patterns of species diversity. The polytrophic
habit of many of the reef dwelling invertebrates (Trench, 1974) may reflect the
unpredictability of this food source.

It should be noted that the significance of zooplankton may lie not in
their role as a major caloric source, but in their critical role in the cycling
of nutrients in the reef environment. Zooplankton are able to feed on the very
abundant mucus and organic particles in the water (Johannes, 1967; Gerber and
Marshall, 1979; Richman, et al., 1975; Gerber and Gerber, 1979), much of which
is generated by reef organisms themselves (Lewis, 1973; Coles and Strathmann,
1973). Zooplankton also feed on the bacteria on these organic aggregates
(Sorokin, 1973) and on the phytoplankton. Thus, the zooplankton prevent the
nutrients from all of these sources from being lost from the reef ecosystem.
Not only do these nutrients pass along the food chain through predation upon
zooplankton, but the production of fecal pellets, often composed of only
partially digested material, provides an important food source for certain
benthic invertebrates (Frankenberg and Smith, 1967; Turner and Ferrante, 1979)
and also serves to keep the important nutrients within the reef ecosystem.

ACKNOWLEDGMENTS

I would like to thank J. B. C. Jackson, N. Knowlton, W. D. Liddell., and M.
Youngbluth for their very helpful criticism of this manuscript. K. G. and J.

113

W. Porter provided suggestions concerning the design of the research, and W. D. _
Liddell, V. Tunnicliffe, and B. Westrum provided invaluable diving assistance.
This research was supported through funds provided by the International Women's
Fishing Association, the Society of Sigma Xi, the Organization of American
States, and the Women's Seamen's Friend Society of Connecticut.

LITERATURE CITED

Alidredge, A. L., and J. M. King. 1977. "Distribution, Abundance, and Substrate
Preferences of Demersal Reef Zooplankton at Lizard Island Lagoon, Great Barrier
Reef." Mar. Biol. 41:317-333.

Alidredge, A., and J. M. King. 1980. "Effects of Moonlight on the Vertical
Migration Patterns of Demersal Zooplankton." J. Exp. Mar. Biol. Ecol.
44:133-156.

Coles, S. L., and R. Strathmann. 1973. "Observations on Coral Mucus 'Flocs' and
Their Potential Trophic Significance." Limnol. Oceanogr. 18:673-678.

Cummins, K. W., and J. C. Wychek. 1971. "Calorie Equivalents for Investigations
in Ecological Energetics." Int. Assoc. Theoret. Applied Limnol. 18:1-158.

Emery, A. R. 1968. "Preliminary Observations on Coral Reef Zooplankton."
Limnol. Oceanogr. 13:293-303.

Frankenberg, D., and K. L. Smith, Jr. 1967. "Coprophagy in Marine Animals."
Limnol. Oceanogr. 12:443-450.

Gerber, R. P., and M. B. Gerber. 1979. "Ingestion of Natural Particulate
Organic Matter and Subsequent Assimilation, Respiration and Growth by Tropical
Lagoon Zooplankton." Mar. Biol. 52:33-43.

Gerber, R. P., and N. Marshall. 1974. "Ingestion of Detritus by the Lagoon
Pelagic Community at Eniwetok Atoll." Limnol. Oceanogr. 19:815-824.

Glynn, P. M. 1973. "Ecology of a Caribbean Coral Reef. The Porites Reef Flat
Biotope: Part 2. Plankton Community With Evidence for Depletion." Mar.
Biol. 22:1-21.

Goreau, T. F., N. I. Goreau, and C. M. Yonge. 1971. "Reef Corals: Autotrophs
or Heterotrophs?" Biol. Bull. 141:247-260.

Grahame, J. 1976. "Zooplankton of a Tropical Harbour: The Numbers, Composition
and Response to Physical Factors of Zooplankton in Kingston Harbour, Jamaica."
J. Exp. Mar. Biol. Ecol. 25:219-237.

Hamner, W. M., and J. H. Carleton. 1979. "Copepod Swarms: Attributes and
Role in Coral Reef Ecosystems." Limnol. Oceanogr. 24:1-14.

Hobson, E. S. 1974. "Feeding Relationships of Teleostean Fishes on Coral
Reefs in Kona, Hawaii." Fish. Bull. 72:915-1031.

114

Hobson, E. S., and J. R. Chess. 1978. "Trophic Relationships Among Fishes and
Plankton in the Lagoon at Enewetak Atoll, Marshall Islands." Fish. Bull.
76:133-153.

Hobson, E. S., and J. R. Chess. 1979. "Zooplankters That Emerge From the
Lagoon Floor at Night at Kure and Midway Atolls, Hawaii." Fish. Bull.
77:275-280. -

Johannes, R. E. 1967. "Ecology of Organic Aggregates in the Vicinity of a
Coral Reef." Limnol. Oceanogr. 12:189-195.

Johannes, R. E., and L. Tepley. 1974. "Examination of Feeding of the Reef
Coral Porites lobata In Situ Using Time Lapse Photography." Proc. Second
Int. Coral Ree f Symp. 1:127-131.

Lewis, J. B. 1973. "The Formation of Mucus Envelopes by Hermatypic Corals of
the Genus Porites." Carib. J. Sci. 13:207-209.

Liddell, W. D. 1980. "Population Variability and Resource Utilization Patterns
of Caribbean Comatulid Crinoids (Echinodermata)." Ph.D. Thesis, University
of Michigan, Ann Arbor. 256 pp.

Liddell, W. D. 1982. "Suspension Feeding by Caribbean Comatulid Crinoids,"
pp. 173-182. In: J. M. Lawrence (ed.), Echinoderms: Proceedings of the
International onference, Tampa Bay. Balkema Press, Rotterdam.

Liddell, W. D., S. L. Ohlhorst, and A. G. Coates. 1984. Modern and Ancient
Carbonate Environments of Jamaica. Sedimenta X. Rosenstiel Schol of Marine
and Atmospheric Science, Miami. 101 pp.

Macurda, D. B., Jr. 1976. "Skeletal Modifications Related to Food Capture and
Feeding Behavior of the Basketstar Astrophyton." Paleobiology 2:1-7.

McWilliam, P. S., P. F. Sale, and D. T. Anderson. 1981. "Seasonal Changes in
Resident Zooplankton Sampled by Emergence Traps in One Tree Lagoon, Great
Barrier Reef." J. Exp. Mar. Biol. Ecol. 52:185-203.

Moore, E. A. 1967. "A Study of Surface Zooplankton in the Caribbean Sea Off
Jamaica." Ph. D. Thesis, McGill University. 259 pp.

Ohlhorst, S. L. 1980. "Jamaican Coral Reefs: Important Biological and Physical
Parameters." Ph. D. Thesis, Yale University. 147 pp.

Ohlhorst, S. L. 1982. "Diel Migration Patterns of Demersal Reef Zooplankton."
J. Exp. Mar. Biol. Ecol. 60:1-15.

Porter, J. W. 1974. "Zooplankton Feeding by the Caribbean Reef-Building Coral
Montastrea cavernosa." Proc. Second Int. Reef Symp. 1:111-125.

Porter, J. W., and K. G. Porter. 1977. "Quantitative Sampling of Demersal
Plankton Migrating From Different Coral Reef Substrates." Limnol. Oceanogr.
22:553-556.

115

Porter, J. W., K. G. Porter, and Z. Batac-Catalan. 1977. "Quantitative Sampling
of Indo-Pacific Demersal Reef Plankton." Proc. Third Int. Coral Reef Symp.
1:105-112.

Porter, K. G., J. W. Porter, and S. L. Ohlhorst. 1978. "Resident Reef Plankton,"
pp. 499-514. In: D. R. Stoddart and R. E. Johannes (eds.), Coral Reefs:
Research Method. UNESCO Monogr. Oceanogr. Methodol., No. 5, UNESCT--Paris.

Richman, S., Y. Loya, and L. B. Slobodkin. 1975. "The Rate of Mucus Production
by Corals and Its Assimilation by the Coral Reef Copepod Acartia negligens."
Limnol. Oceanogr. 20:918-923.

Robichaux, D. M., A. C. Cohen, .M. L. Reaka, and D. Allen. 1981. "Experiments
With Zooplankton on Coral Reefs, or, Will the Real Demersal Plankton Please
Come Up?" Mar. Ecol. 2:77-94.

Sale, P. F., P. S. McWilliam, and D. T. Anderson. 1976. "Composition of the
Near-Reef Zooplankton at Heron Reef, Great Barrier Reef." Mar. Biol.
34:56-66.

Sebens, K. P. 1977. "Autotrophic and Heterotrophic Nutrition of Coral Reef
Zoanthids." Proc. Third Int. Coral Reef Symp. 1:397-404.

Sorokin, Y. I. 1973. "On the Feeding of Some Scleractinian Corals With Bacteria
and Dissolved Organic Matter." Limnol. Oceanogr. 18:380-386.

Tinkle, D. W., and N. F. Hadley. 1973. "Reproductive Effort and Winter Activity@
in the Viviparous Montane Lizard Sceloporus jarrovi." Copeia 1973: 272-
277.

Trench, R. K. 1974. "Nutritional Potentials in Zoanthus sociatus (Coelenterata,
Anthozoa)." Helgol. Wiss. Meeresunters. 26:174-216.

Turner, J. T., and J. G. Ferrante. 1979. "Zooplankton Fecal Pellets in Aquatic
Ecosystems." Bioscience 29:670-677.

Walter, C., L. Tulane, and J. N. Pasamonte. 1982. "A Preliminary Quantitative
Study on the Emergence of Reef Associated Zooplankton From a Philippine
Coral Reef." Proc. Fourth Int. Coral Reef Symp. 1:443-451.

Youngbluth, M. J. 1982. "Sampling Demersal Zooplankton: A Comparison of
Field Collections Using Three Different Emergence Traps." J. Exp. Mar.
Biol. Ecol. 61:111-124.

116

TEMPORAL PATTERNS OF ZOOPLANKTON MIGRATION

Sharon L. Ohlhorst
Department of Fisheries and Wildlife
and
W. David Liddell
Department of Geology
Utah State University
Logan, Utah 84322

ABSTRACT

The NOAA/NULS-1 underwater habitat facility at St. Croix was utilized
during July 1982 for a study of the temporal patterns of migration into the
water column by reef zooplankton. Samples were collected for 6 days at 9
daily time intervals using mesh emergence traps, diver pushed plankton nets,
and surface plankton net tows. This is the first study to sample such finely
spaced time intervals. Preliminary analysis from 2 days indicated that there
was migration throughout the night, with increased activity found prior to
sunrise as well as following sunset. This is the first time a predawn rise of
zooplankton has been documented.

The cyclopoid copepod, Oithona colcarva, dominated most time intervals,
although it was not nearly as abundant as in an earlier Caribbean reef zooplankton
study. Most zooplankton taxa followed the same general pattern of activity, and
the relative abundance of the common taxa generally remained the same throughout
the 24 hours. The time intervals around sunset were an exception, however.
For example, during the hour following sunset, calanoids, harpacticoids,
polychaetes, and pagurid larvae were relatively more abundant than during the
hour prior to sunset, while copepod nauplii and amphipods were less abundant
during this postsunset time interval. Significantly more taxa were captured
during the first hour of darkness than during any other time interval.

The effects of different trap designs and types of reef substrata on
zooplankton samples also were examined. There was no significant difference in
the number of individuals collected between treatments using sealed and unsealed
traps or between those using unsealed traps over coral and sand substrata.
This is in contrast to other studies; two possible explanations for these
discrepancies relate to the types of substrata over which sealed and unsealed
traps are compared and to differences in trap sealing efficiency when different
substrata types are being compared.

INTRODUCTION

The zooplankton that reside within or near coral reef ecosystems have only
recently begun to receive the attention of researchers. Zooplankton have been
studied from various wcific reef communities primarily by using various
modifications of emergence traps (Alldredge and King, 1977, 1980; Porter and
Porter, 1977; Porter, et al., 1977; Hobson and Chess, 1979; Birkeland and

117

Smalley, 1981; McWilliam, et al., 1981; and Walter, et al., 1982). Fewer
studies of reef zooplankton from the Caribbean have been conducted (Ohlhorst,
1980, 1982, 1985; Robichaux, et al., 1981; and Youngbluth, 1982). While the
diel migration patterns of these zooplankton have been shown to influence the
behavior of nocturnal planktivorous fish (Hobson, 1974; Hobson and Chess, 1978;
Robertson and Howard, 1978) and might affect the behavior of other reef
planktivores (Porter, 1974; Sebens, 1977; Liddell, 1982), few studies have
investigated the diel migration patterns of reef zooplankton in detail (Alldredge
and King, 1980; Ohlhorst, 1982; Walter, et al., 1982). The only study of
migration by Caribbean reef zooplankton T onducted at Jamaica by Ohlhorst,
1982) suggests that there is no single pulse of migratory activity; rather
zooplankton rise into the water column at variable rates throughout the night
with a peak of activity during the second hour after sunset. Also, different
taxa were shown to exhibit differing migratory patterns. While these observations
are consistent with those from the Pacific, it is important to determine whether
or not such patterns occur elsewhere in the Caribbean. Additionally, more
frequent sampling than previously conducted would be of value in refining the
patterns of zooplankton migration.

One of the reasons for the paucity of detailed studies addressing this
question is the physiological limitation placed upon safely conducting the
repeated sampling dives which are necessary for such studies. Previous studies
which have addressed the question of migratory patterns of reef zooplankton
have been restricted to sampling widely spaced time intervals or sampling very
shallow (< 5 m) sites. Data from Ohlhorst (1985) suggests that caution should
be used when extrapolating data from shallow sites to greater depths. The
present study examines the migratory patterns of reef zooplankton at an
intermediate (15 m) depth at St. Croix through repeated sampling of finely
spaced time intervals. This sampling was made possible by saturation diving
from the NOAA/NULS-1 Underwater Habitat, HYDROLAB.

METHODS

Sampling of reef zooplankton was conducted from the NOAA/NULS-1 HYDROLAB
located at 15 m depth in the Salt River submarine canyon on the north coast of
St. Croix, U.S. Virgin Islands (17045' N, 64�45' W) during July 1982. Samples
were collected from both reef and sand areas located approximately 15 m east
of the Habitat. To eliminate biases caused by the proximity of the study sites
to the Habitat, all of the external lights of the HyDROLAB remained off for the
duration of the study. Saturation diving from the HYDROLAB enabled two teams
of divers to collect plankton samples at closely spaced intervals over a 6-day
period.

Zooplankton were sampled by three methods: (1) Emergence traps, which
covered 0.5 m2, were placed over various types of reef substrata to capture
zooplankton moving from the reef site into the water column. Certain of these
traps were sealed over their substrata by a skirt (Robichaux, et al., 1981),
while others were affixed more loosely over their substrata. 7) Diver pushed

118

Table I. Number of zooplankton captured/m2/hour, beginning with 0030 on July 16 and
continuing through 2300 on July 17, 1982. ST= Students' T test at pgO.05, 0 = no
significant difference between this and the previous time interval, + = a significant
increase, - = a significant decrease; MWU = Mann Whitney U test at psO.05, symbols are
the same as for ST.

Time Interval #Samples Mean (Std. Dev.) Median ST MWU

0030 8 139.2 (78.0) 136.6
0230 8 109.4 (52.0) 100.0 0 0
0500 8 176.0 (97.8) 168.0 0 0
0600 4 26.6 (20.2) 20.0 - -
1200 8 10.6 (9.0) 8.0 0 0
1830 8 40.0 (21.8) 39.0 + +
1930 8 246.8 (103.8) 256.0 + +
2230 8 99.6 (60.0) 92.0 - -
0030 8 348.0 (376.0) 202.0 0 0
0230 8 702.0 (662.0) 54.6 0 0
0500 8 306.0 (364.0) 140.8 0 0
0600 8 55.8 (22.2) 60.0 0 0
1200 8 6.2 (2.0) 6.0 - -
1800 4 92.6 (24.2) 90 0 + +
1930 4 399.2 (199.0) 341.4 + 0
2030 4 189.6 (144.2) 133.0 0 0

2300 4 51.0 (26.0) 40.4 - 0

zooplankton net samples were collected 1 m and 5 m above the reef. (3) A
plankton net was towed at the surface from a boat to sample zooplankton just
below the surface, and vertical net hauls from 15 m to the surface also were
made from the boat. The emergence traps and zooplankton nets were made with 90
micron meshes. Two additional emergence traps were constructed with clear
polyvinyl. The traps were modified (Ohlhorst, 1980, 1982) from those used by
Porter and Porter (1977).

Five hundred and eighty samples were collected over 6 days and 9 time
intervals each day (table 1). Samples were preserved in 5-10% buffered formalin
immediately after collection and counted as in Ohlhorst (1982). Preliminary

119

TEMPORAL PATTERNS OF ZOOPLANKTON MIGRATnON 0

800 -

700 -

'400-

Z300-

200 -

,oo- Legend
A TOTAL
* -.' ". -. * LESS 4Ao" -
0 -

liME

Figure I.--The number of individuals captured per hour during the different
time intervals are plotted, both with and without the copepod Oithona colcarva.
Samples from 2030 on July 16 have not been counted. Refer to table 1 for the
means, medians, and standard deviations. Note that units of time along the x
axis are unequal.

results from 148 samples from mesh emergence traps from 2 days (July 16-17)
will be presented herein. During the collection period sunrise was at 0600,
sunset at 1800, and the lunar phase was new moon.

RESULTS

Effects of Traps and Substrata

No significant differences Cp < 0.05; Mann Whitney U (MWU), Students' T
(ST) tests] in the number of individuals captured per hour were observed between
mesh traps with sealed vs. unsealed bottoms over coral substrata. Additionally,
no significant differences (p < 0.05) in the number of individuals captured per
hour occurred between traps positioned over sand vs. those positioned over coral
or rubble. For all subsequent tests, data from sealed and unsealed traps and
traps located over sand and over coral or rubble were pooled. The polyvinyl
traps tended to capture more zooplankton than mesh traps. Although the difference
between types of trap was not significant, most of the data presented herein
are from mesh traps only.

120

Table 2. Percent occurrence of phyla in demersal traps. Samples were collected on
July 16-17, 1982 and are from both mesh and polyvinyl traps. The taxonomy is
according to Barnes (1980).

time: 0030 0230 0500 0600 1200 1830 1930 2030 2230
# traps : (20) (20) (20) (14) (20) (16) (16) ( 6) (16)
Phyla

ANNELIDA 30 65 35 21 20 6 100 67 56
ARTHROPODA 100 100 100 100 100 100 100 100 100
CHAETOGNATHA 25 30 15 14 5 19 13 0 25
CHORDATA 15 15 25 7 20 19 25 33 31
COELENTERATA 0 25 25 21 25 38 56 67 44
ECHINODERMATA 0 0 0 0 5 0 0 0 0
MOLLUSCA 40 35 35 29 60 94 81 50 38
NEMATODA 0 0 0 0 5 0 13 0 0
PLATYHELMINTHES 0 5 0 0 5 0 19 0 0
SARCODINA 35 80 45 29 80 94 81 67 31
SIPUNCULIDA 25 15 10 21 40 25 63 67 25

Temporal Patterns

The number of zooplankters migrating per hour for the various sampling
intervals is presented in table 1 and figure 1. Data from July 16-17 displayed
significant decreases (p < 0.05; MWU, ST tests) in the total number of individuals
captured per hour between 0500 or 0600 and 1200 hours, and significant (p <
0.05) increases between 1200 and 1800/1830 hours and between 1800/1830 and 1930
hours, followed by a significant (p < 0.05) decrease after 1930 hours. The
pattern differed somewhat when the cyclopoid copepod Oithona colcarva (Bowman)
was removed from analysis (fig. 1).

There was no significant (p < 0.05; MWU, ST tests) difference between the
2 days in the number of individuaTs migrating per hour for most time intervals
(fig. 1, table 1). Significantly more individuals (p < 0.05; MWU, ST tests)
were captured on July 17 than on July 16 in both the intervals from 0030-0230
and 1200-1800/1830. When the samples were considered without 0. colcarva,
numbers of zooplankters differed between days both in the intervals mentioned
above and at 0500-0600.

Crustaceans made up the majority of the zooplankters captured by the traps,
although representatives of 11 phyla, including foraminifera, echinoderm larvae,
sipunculids, Amphioxus, and appendicularians also were collected (table 2). A

121

MEAN NUMBER OF TAXA PER TIME INTERVAL
25-

t
20 -

is-

10-

~t
5-

0o 0 023o 050os 0600 1200 1830 1930 2030 2300
TIME

Figure 2.--The mean number of taxa captured during the various time intervals
(days combined) is plotted. Vertical bars represent 95% confidence intervals.
Units of time along the x axis are unequal.

total of 73 taxa were identified from these samples. Many were rarely encountered,
and 27 taxa was the maximum found in a single sample (July 17, 1982). Figure 2
shows the pattern for the mean number of taxa captured during each time interval
for the 2 days pooled. The only significant differences (p < 0.05, MWU, ST
tests) between the 2 days in the number of taxa captured occurred at 0030 and
0600, when significantly fewer were captured on July 16. When the days were
pooled, there were significant increases (p < 0-.05; MWU, ST tests) in number of
taxa in samples collected at 0030 vs. 0230 and in those collected at 1830 vs.
1930; in addition, significant decreases in number of taxa were observed between
0500 and 0600, 1930 and 2030, and 2030 and 2300. In general, the fewest taxa
were captured during the day, the most during the first hour after sunset, and
intermediate numbers during the rest of the night.

The mean and standard deviation for the abundance of the more common taxa
are presented for the different time intervals in table 3. The cyclopoid
copepod Oithona colcarva (Bowman) was the most abundant of the organisms captured
at all time intervals except those between 1200-2030, although there was some
variation in this pattern when the days were considered separately (fig. 1).
During the morning interval (0600-1200), O. colcarva, foraminifera, and certain
calanoid and harpacticoid species were themost abundant organisms; however,
all occurred in relatively low numbers. In the afternoon (1200-1830), foraminifera_

122

Tab~le 3. Organismus captured at vdrious time intervals (x- = mean, SO standard deviation, * present in 50-75,, of samlples,
present in 75-100'. of samples). Only the 29 most common of the 73 taxa recorded are presented here.

Time: 00 30 0230 0500 0600 1200 1830 1930 2030 2230
#SamplIes: 16 16 16 12 16 12 12 4 12
Taxa x (S) )Sri~ ) i (S) K (SD) 2 (SO) i (SD) i (SO) (so)

Dithona colvarva 200 (272)-* 354 (538)-' 186 (260)"" 20.2 (18.2)1* 2.8 (6.6)** 3.7 (3.1)** 33.0 (3)NP)"" 24.0 (3 )* 46.8 (47.4)**

Corycaeus spp. 0.5 (0.61* 0.9 (1.0)* 1.9 (3.5)* 3.8 (4.3)** 0.1 (0.2) 0.7 (0.7)* 1.8 (2.61* 0 0.1 (0.3)

Calanoid "A" 5.8 (9.8)-* 4.1 (4.3)** 9.8 (21.4)** 2.5 (2.3)* 0.7 (0.61** 0.5 (0.5)- 12.5 (9.9)*- 4.5 (7.7)* 4.1 (5.W)*

Calanoid "B" 5.6 (9.11* 4.5 (6.31* 9.0 (21.8)** 0.5 (0.9) 0.1 (0.2) 0.1 (0.1) 8.8 (9.?)** 7.5 (3J* 0.8 (1.6)

Microsetella spp. 2.3 (2.71* 8.2 (7.2)** 2.5 (3.5) 5.8 (11.21* 0.4 (0.7) 5.0 (13.31* 22.4 (37.2) 43.6 (50.2)*" 0.2 (0.4)

Hlarpacticoid "A" 1.4 (2.41* 2 .7 (2.8)* 5.1 (5.4)* 2.2 (2.31* 0.7 (0.91* 1.3 (3.7) 42.8 (52.4)** 22.6 (25..2)** 5.8 (8.1)**

Ilarpacticoid "8" 0.2 (0.7) 0.4 (1.0) 2.0 (2.8) 0 0.3 (0.41* 1.2 (1.53* 4.5 (6.51" 0 2.8 (5.3)*

liarpacticoid 'C' 0.4 (0.9) 1.2 (2.2) 0.3 (1.0) 0.5 (1.2) 0 0 0.1 (0.4) 3.0 (3.5)-- 0

Copepod "A" 0.6 (0.8) 1.2 (1.3)* 0.8 (1.0) 0.7 (1.3) 0.1 (0.1) 0.1 (0.3) 5.7 (7.0)* 7.5 (4.4)** 0.4 (0.6)

Copepod nauplii 0.2 (0.4) 0.5 (0.9) 3.1 (6.91* 0.7 (1.0) 0.1 (0.2) 0.8 (1.7) 9.1 (16.1) 5.0 (7.61* 0.3 (0.7)

Barnacle nauplii 0 0.6 (1.0) 0.2 (0.6) 0 0.1 (0.1) 0.1 (0.2) 1.3 (1.31* 0 0.1 (0.3)

Ostracods 0.4 (0.8) 0.3 (0.6) 0.5 (1.0) 0.3 (1.2) 0.1 (0.3) 0.1 (0.2) 3.6 (3.6)** 0 0.4 (0.8)

"3 Amphipod "A" 4.2 (4.0)** 4.2 (4.8)** 2.9 (4.9)" 3.0 (7.9) 0.1 (0.2) 0 6.9 (6.7)"" 3.0 h1.1"" 1.9 (2.0)**

Amphipod "B" 0.1 (0.2) 1.3 (1.3)** 0.2 (0.5) 0.7 (1.8) 0.3 (0.4) 0.7 (1.41 8.0 (13.11* 7.0 110.1)-* 0

Isopods 0.3 (0.6) 0.2 (0.4) 0.5 (0.7) 0 0.1 (0.1) 0.2 (0.2 ) 2.4 (1.4)"" 1.0 (1.2)* 1.4 (1.9)*

Cumarean "A" 4.8 (4.31** 4.8 (4.0)** 4.9 (4.7)*- 0 0.1 (0.1) 0.1 (0.5) 10.6 (16.2)** 24.6 (33.2)** 4.1 (4.8)""

Cumacean "8" 0.7 (1.0) 0.2 (0.5) 0.2 (0.4) 0 0 0 1.4 (2.8) 1.0 (2.0) 0.1 (0.3)

Tanaids 0.3 (0.6) 0.4 (0.8) 0.8 (1.11* 0.2 (0.6) 0 0 1.3- (2.?) 0.5 (1.0) 0.3 10.6)

Mysids 0.1 (0.5) 0.6 (0.9) 0.7 (1.2) 0.2 (0.6) 0 0.1 (0.1) 1.6 (2.4) a 2.4 (4.4)**

Shrimp 0.7 (1.0) 1.1 (1.21* 0.3 (0.7) 0 0.1 (0.2) 0.1 (0.1) 0.7 (1.6) 3.0 (2.6)"" 2.0 (2.61*

Shrimp larvae 0.5 (1.2) 0.1 (0.3) 0.4 (0.9) 0 0 0 7.9 (6.9)** 0 0.5 (0.9)

Pagurid larvae 7.0 (15.0)** 2.4 (4.11* 1.8 (2.71* 0.3 (0.8) 0.1 (0.1) 0.1 (0.1) 19.4 (33.2) 5.0 (5.0)** 2.5 (4.21*

Decapod zoea 0.1 (0.3) 1.2 (1.51* 1.8 (3.2) 0.2 (0.6) 0 0.1 (0.1) 0.7 (1.4) 0.5 (1.0) 0.1 (0.3)

Miscellaneous crustaceans 0.7 (1.3) 3.2 (3.5)"" 0.4 (1.2) 0.2 (0.6) 0.1 (0.2) 0.6 (0.81* 15.1 (18.3)** 6.6 (4,l)** 0.7 (0.81*

Chaetognaths 0.5 (1.7) 0.4 (0.7) 0.1 (0.4) 0.3 (0.8) 0 0.1 (0.1) 0.2 (0.6) 00.6 tl.4)

Polyctiaetes 0.2 (0.4) 0.7 (0.81* 0.3 (0.6) 0.3 (0.8) 0.1 (0.1) 0.1 (0.1) 17.7 (1.5)"" 7.0 (6.6)** 0.8 (0.71*

Gastropods 0.4 (0.6) 0.4 (0.7) 0.6 (1.0) 0.3 (1.0) 0.5 (03* 18.8 (23.2)"" 30.2 (27.8)** 3.0 (3.81* 0.6 (0.7)

Sipuncul ids 0.3 (0.6) 0.1 (0.3) 0.2 (0.6) 0.3 (0.8) 0.2 (0.5) 0.1 (0.2) 4.8 (11.21* 2.5 (2.5)"" 0.2 (0.4)

Fordminiferd 1.2 (2.0) 2.2 (2.4)** 1.9 (3.2)" 0.8 (1.6) 1.1 (1.2)*" 22.8 (14.7)** 8.5 (8.1)** 3.0 (2.6)** 1.2 )2.1)

and gastropods dominated the samples. Samples collected in the hour following
sunset were dominated by harpacticoids, with a variety of other organisms
(including 0. colcarva and gastropods) occurring abundantly. Samples taken
during the second hour after sunset again were dominated by harpacticoids,
although the relative abundance of the different harpacticoid species switched
(table 3). Oithona colcarva and a species of cumacean were the other most
abundant taxa during this second hour. Oithona colcarva dominated the other
night time intervals from 2030-0500.

The relative abundance of nine common taxa (0. colcarva, calanoid copepods,
harpacticoid copepods, copepod nauplii, amphipods, cumaceans, pagurid larvae,
polychaetes and gastropods; table 3) were compared using Spearman Rank Corre-
lations (SRC). Although the abundances were greater at night, the relative
abundances (% total sample) of these taxa were correlated (p < 0.05) between
pooled day samples and pooled night samples. When samples from consecutive
time intervals were compared, the relative abundances of taxa in all such pairs
were correlated (p < 0.05) except those from 1800/1830 vs. 1930, and 1930 vs.
2030. For example, harpacticoids, pagurid larvae, calanoids, and polychaetes
were relatively more important, while copepod nauplii, amphipods, and gastropods
were relatively less abundant, at 1930 than at 1800/1830. Cumaceans and
amphipods increased in relative abundance between 1930 and 2030, while gastropods
and pagurid larvae decreased in relative numbers over this interval. Therefore,
the hours around dusk were the only ones where the relative abundance of captured
organisms differed.

Differences in the relative abundances of certain taxa reflect their
different temporal behavior patterns. Sixty-two percent of the 29 common taxa
(table 3) were captured in greatest numbers during the first hour after sunset
(1830-1930), while 17% were captured in greatest abundance during the second
hour after sunset. The only taxon showing peak abundance during the day was
foraminifera. Many of the taxa exhibit sustained vertical migration throughout
the night following the post-sunset pulse (e.g., calanoid "A," amphipod "A,"
cumacean "A," pagurid larvae in table 3). A few taxa (calanoid "B," decapod
zoea) were captured most frequently during the last interval of night (0230-0500),
and many (28%) exhibited a pulse of migration at this time. Only the cyclopoid
Corycaeus spp. peaked in abundance during the hour prior to sunrise (0500-0600);
the harpacticoid Microsetella spp. also exhibited a pulse of vertical migration
at this time. The behavior of 0. colcarva was variable throughout the various
time intervals (fig. 1). The mean number of this species captured for both
nights combined was greatest at 0230; however, that peak reflects an especially
high number captured on July 17 (I = 638) and a similar pattern was not observed
on July 16 (fig. 1). Migration rates for this species were high during all
night hours.

DISCUSSION

The zooplankton captured by emergence traps in this study do not solely
represent, either in numbers or composition, the demersal zooplankton (Hobson
and Chess, 1979; Robichaux, et al., 1981) living within the reef substrata over
which the traps are placed. These data do, however, provide information on the
temporal patterns of migration by the total reef zooplankton which is of value _
to studies of reef energetics and the behavior of reef planktivores.

124

Both Robichaux, et al. (1981) and Youngbluth (1982) addressed the question
of trap design and both found differences in number and composition of zooplankton
between sealed and unsealed traps. Both of these studies, however, were
conducted over sandy bottoms where a complete seal was possible and zooplankton
movement through the underlying substrata unlikely. In this study, both sealed
and unsealed traps were placed over reef substrata; the former were sealed as
well as possible through the use of sand and rubble placed over the trap skirts.
No differences were observed in the numbers of zooplankton captured between the
two treatments, possibly suggesting that there is more movement by zooplankton
through interstices of the reef than previously thought. Analysis of differences
in species composition between the two treatments is currently underway.

While Alldredge and King (1977) and Porter and Porter (1977) found the
most zooplankton over structurally complex coral substrata, this was not found
by Ohlhorst (1980) or Birkeland and Smalley (1981). In the present study,
there also was no difference in the number of individuals captured over different
substrata. One contributing factor may be the differential ability to seal
traps over sand and coral substrata. Robichaux, et al. (1981) and Youngbluth
(1982) both found that more zooplankton were captured by unsealed traps than by
sealed traps when both were placed over sand substrata. Youngbluth (1982)
observed that a gap of 1 cm between trap and substratum resulted in capture of
significantly greater numbers of zooplankton than a total seal, while there was
no difference between samples from traps with gaps of 1 and 10 cm. Therefore,
differences in number of zooplankton collected over different types of substrata
might be expected when certain treatments are well sealed and others not. If
the sand traps of Alldredge and King (1977) and Porter and Porter (1977)
were well sealed in contrast to those over coral substrata, it may be difficult
for these researchers to compare their substratum treatments. Birkeland and
Smalley (1981) compared coral substrata to algal turf pavement and probably
were able to sample both treatments similarly, since in either habitat a total
seal is unlikely. Ohlhorst (1980) and this study used unsealed traps to sample
both coral and sand substrata. These traps captured the zooplankton moving
along the reef bottom over both types of substrata, and the data indicated that
the numbers of zooplankton available to planktivores were comparable in both
habitats. Analysis of the species composition of zooplankton over various reef
substrata is in progress.

Fewer zooplankton were captured per hour in this study than in previous
Caribbean studies by one of the authors (Ohlhorst, 1980, 1982, 1985). The most
probable explanation lies in the sole use of polyvinyl traps in the previous
studies. In this study, two polyvinyl traps were used in addition to mesh
traps (details will be discussed elsewhere); and, while no statistically
significant differences were found in abundance of plankton between these
treatments (due to high variance and small sample size of polyvinyl traps), the
polyvinyl traps usually captured considerably more zooplankton. Both this
study and the earlier work by Ohlhorst (1980, 1982, 1985) found considerable
variability between nights sampled.

The bimodal emergence pattern (post-sunset, pre-sunrise) suggested by
Glynn's (1973) data from plankton tow nets appears to be supported by this
study (fig. 1), although more complete information will be available when all
the samples have been analyzed. The present study provides the first documentation

125

of a presunrise emergence of zooplankton and has considerable implications for
reef bioenergetics and planktivore feeding behavior.

As in the previous work where plankton were collected over Caribbean reef
substrata at different time intervals (Ohlhorst 1982), Oithona colcarva, various
calanoids and harpacticoids, copepod nauplii, amphipods, and polychaetes were
important components of the fauna. The relative abundances of these organisms
from the two studies are correlated (p < 0.05, SRC). There are, however,
differences in the behavior of certain taxa between these studies. In Jamaica,
for example, the harpacticoid Microsetella spp. migrated in greater numbers
during the day than at night, while at St. Croix the peak migration was during
the first and second hour after sunset. Also, isopods were an important
component of the Jamaican fauna but were relatively rare at St. Croix. Oithona
colcarva was less abundant at St. Croix than in Jamaica, and its activity
pattern differed. At Jamaica 0. colcarva was captured in the greatest numbers
during the second hour after sunset, while at St. Croix the capture rate of
this copepod was highest from midnight to just prior to sunrise. The sample
size needs to be increased at both reef locations to determine if these
differences are real.

The differences between this study and those over primarily sand substrata
in the Caribbean (Robichaux, et al., 1981; Youngbluth, 1982) may be related to
habitat differences and/or trap idesign. The relative abundance of nine common
taxa from this study and the unskirted traps of Robichaux, et al. (1981) were
positively correlated (p < 0.05, SRC). There was no correlation, however,
between the relative abundance of taxa in this study and that from any of the
treatments of Youngbluth (1982). While harpacticoids were a very important
component of the reef zooplankton in St. Croix, they did not dominate to
the degree reported by Robichaux, et al. (1981) and Youngbluth (1982) in the
Bahama Islands. The St. Croix sampTes usually were dominated by the cyclopoid
Oithona colcarva. This is a swarming meroplanktonic species unlikely to be
captured in traps sealed over sand. Cumaceans and calanoids were important
over St. Croix reefs, as was found with certain trap designs over sand by
Youngbluth (1982) but not by Robichaux, et al. (1981).

This preliminary analysis of data collected from St. Croix is consistent
with earlier diel studies (Walter, et al., 1981; Ohlhorst 1982) which indicated
that zooplankton move up into the water column throughout the night with a pulse
in activity following sunset. Although zooplankton are therefore available to
planktivores throughout the night, the indication that there are predawn (this
study) and postdusk peaks of migration is consistent with the hypothesis that
fish predation is an important selective factor upon zooplankton behavior since
these dawn and dusk peaks of emergence coincide with periods when there are few
fish predators (Hobson, 1975).

ACKNOWLEDGMENTS

This study was made possible by logistical support and a grant (NA 82-AAA-
01345) provided by NOAA. Additional support and funding were provided by the
West Indies Marine Laboratory of Fairleigh Dickinson University, St. Croix; the t
Ecology Center of Utah State University; and the Lerner-Gray Fund for Marine

126

Research of the American Museum of Natural History, New York. We wish to
express our appreciation to fellow aquanauts Bonnie K. Schwann and Lee H.
Somers, who aided in the sampling program; to Stephen K. Boss, who was responsible
for collecting surface samples and preserving all samples collected; and to
Lloyd Bennett and Mark T. Segars, who assisted in counting the zooplankton
samples.

LITERATURE CITED

Alldredge, A. L., and J. M. King. 1977. "Distribution, Abundance, and Substrat.
Preferences of Demersal Reef Zooplankton at Lizard Island Lagoon, Great Barrier
Reef." Mar. Biol. 41:317-333.

Alldredge, A., and J. M. King. 1980. "Effects of Moonlight on the Vertical
Migration Patterns of Demersal Zooplankton." J. Exp. Mar. Biol. Ecol.
44:135-156.

Barnes, R. D. 1980. Invertebrate Zoology. 4th Edition. Saunders, Philadelphia.
1089 pp.

Birkeland, C., and T. L. Smalley. 1981. "Comparison of Demersal Plankton From
Comparable Substrata From a High Island and an Atoll." Proc. 4th Int. Coral
Reef Symp. 1:437-442.

Glynn, P. M. 1973. "Ecology of a Caribbean Coral Reef. The Porites Reef Flat
Biotope: Part 2. Plankton Community With Evidence for Depletion." Mar.
Biol. 22:1-21.

Hobson, E. S. 1974. "Feeding Relationships of Teleostean Fishes on Coral
Reefs in Kona, Hawaii." Fish. Bull. 72:915-1031.

Hobson, E. S. 1975. "Feeding Patterns Among Tropical Reef Fishes." Amer.
Sci. 63:382-293.

Hobson, E. S., and J. R. Chess. 1978. "Trophic Relationships Among Fishes and
Plankton in the Lagoon at Enewetak Atoll, Marshall Islands." Fish. Bull. 76:
133-153.

Hobson, E. S., and J. R. Chess. 1979. "Zooplankters That Emerge From the
Lagoon Floor at Night at Kure and Midway Atolls, Hawaii." Fish. Bull. 77:275-
280.

Liddell, W. D. 1982. "Suspension Feeding by Caribbean Comatulid Crinoids,"
pp. 173-182. In: J. M. Lawrence (ed.), Echinoderms: Proceedings of the
International Conference, Tampa Bay. A. A. Balkema, Rotterdam.

McWilliam, P. S., P. F. Sale, and D. T. Anderson. 1981. "Seasonal Changes
in Resident Zooplankton Sampled by Emergence Traps in One Tree Lagoon, Great
Barrier Reef." J. Exp. Mar. Biol. Ecol. 52:185-203.

127

Ohlhorst, S. L. 1980. "Jamaican Coral Reefs: Important Biological and Physicala
Parameters." Ph.D. thesis, Yale University. 147 pp.

Ohlhorst, S. L. 1982. "Diel Migration Patterns of Demersal Reef Zooplankton."
J. Exp. Mar. Biol. Ecol. 60:1-15.

Ohlhorst, S. L. 1985. "Reef Zooplankton Collected Along A Depth Gradient at
Discovery Bay, Jamaica," pp. 101-116. In: M. L. Reaka (ed.), The Ecology of
Coral Reefs, Symp. Ser. Undersea Res., Vol. 3(1). NOAA Undersea Res. Progr.,
Rockville, Md.

Porter, J. W. 1974. "Zooplankton Feeding by the Caribbean Reef-Building Coral
Montastrea cavernosa." Proc. 2nd Int. Reef Symp. 1:111-125.

Porter, J. W., and K. G. Porter. 1977. "Quantitative Sampling of Demersal
Plankton Migrating From Different Coral Reef Substrates." Limnol. Oceanogr.
22:553-556.

Porter, J. W., K. G. Porter, and Z. Batac-Catalan. 1977. "Quantitative Sampling
of Indo-Pacific Demersal Reef Plankton." Proc. 3rd Int. Coral Reef Symp.
1:105-112.

Robertson, A. I., and R. K. Howard. 1978. "Diel Trophic Interactions Between
Vertically-Migrating Zooplankton and Their Fish Predators in an Eelgrass
Community." Mar. Biol. 48:207-213.

Robichaux, D. M., A. C. Cohen, M. L. Reaka, and D. Allen. 1981. "Experiments
With Zooplankton on Coral Reefs, or, Will the Real Demersal Plankton Please
Come Up?" Mar. Ecol. 2:77-94.

Sebens, K. P. 1977. "Autotrophic and Heterotrophic Nutrition of Coral Reef
Zoanthids." Proc. 3rd Int. Coral Reef Symp. 1:397-404.

Walter, C., L. Tulane, and J. N. Pasamonte. 1982. "A Preliminary Quantitative
Study on Emergence of Reef Associated Zooplankton From a Philippine Coral
Reef." Proc. 4th Int. Coral Reef Symp., 1:443-451.

Youngbluth, M. J. 1982. "Sampling Demersal Zooplankton: A Comparison of
Field Collections Using Three Different Emergence Traps." J. Exp. Mar. Biol.
Ecol. 61:111-124.

128

FIELD ANALYSIS OF THE DOMINANCE HIERARCHY
OF THE BICOLOR DAMSELFISH
STEGASTES PARTITUS (POEY) (PISCES: POMACENTRIDAE)

Yvonne Sadovy1
Department of Zoology
University of Manchester
Manchester, M13 9PL England

ABSTRACT

The dominance relations of the damselfish Stegastes partitus were examined
in the field. The dominance hierarchy, based on aggressive interactions, was
linear and size-dependent, with most aggression between individuals most similar
in size and least between those most dissimilar in size. These results show
both similarities to and differences from results obtained in a very similar
earlier study on the same species but conducted entirely in the laboratory.
The importance of field studies in assessing the degree of reliability of
results obtained under artificial conditions is emphasized.

INTRODUCTION

Many species of animals display a dominance hierarchy among members of
social units. A group of animals may be said to have a dominance hierarchy if
the interactions between individuals allow some to have priority over others
under certain circumstances. The most simple hierarchy is defined as follows:
individual A dominates all members of a group; another animal B dominates all
members except A, and so on to the last member, omega, which dominates no other
individuals. If this relationship holds, then the hierarchy is said to be
linear (Chase,'1974).

Considerable research has examined the dominance relations of a wide
variety of animal species. However, the majority of studies have been conducted
in the laboratory or under other artificial conditions. In studies carried out
both in the field and in the laboratory on the same species, there often have
been discrepancies in the conclusions. Thompson (1960) found that laboratory
populations of house finches had a hierarchy dominated by females, but that
field populations did not show this pattern. In some primates, a different
form of social structure was evident between captive animals and the same
species in the field (Rowell, 1967; Hinde, 1974). In the baboon, Papio anubis,
interactions were about four times as frequent among captive animals as among
individuals observed in the field (Rowell, 1967). Among the fishes, Myrberg
(1972a) conducted a field and laboratory study of territoriality and social
hierarchy in the bicolor oamselfish (Pomacentridae), Stegastes (= Eupomacentrus)

1 Present address: Department of Marine Sciences, University of Puerto
Rico, Mayaguez, Puerto Rico 00708.

129

partitus. However, limitations of resolution in the underwater camera used for
field observations precluded analyses as detailed as those obtained during the
laboratory section of the work. Therefore, the two parts of the study were not
strictly comparable.

Many damselfish are territorial, and the bicolor, like several other
closely related species, defends a territory which it infrequently leaves. in
the field, several or many such territories may be found in close proximity,
forming colonies on suitable substrate. Reproduction usually occurs within
colonies, which consist of both males and females. Colony members frequently
interact with each other close to territory borders. In small colonies, all
colony members may interact with one another, whereas in very large colonies
many individuals rarely come into contact. Small colonies are, therefore,
particularly useful for studies on intracolony social relationships.

The aims of the present study were: (a) to examine in detail the aggressive
social interactions of small colonies of the bicolor damselfish in the field
and (b) to compare the dominance system of these colonies with hierarchies
found in the laboratory (Myrberg, 1972a).

METHODS

The study was conducted from October 9-16, 1980, from the NOAA underwater
facility, HYDROLAB, situated in 50 feet of water off the north coast of St.
Croix in the U.S. Virgin Islands.

Seven small damselfish colonies, each confined to a patch of coral rubbleW
and separated from each other by clear areas of sand, were selected. The
colonies were small enough to permit interactions between all colony members.
Aggressive social interactions ("chases") were recorded between all individuals
in each colony, where both the giver and receiver of each "chase" were noted.
A chase is defined as a rapid swim towards a fish that is moving away or that
starts to move away from the chaser (Myrberg, 1972b).

Each colony contained six to eight individuals, most of which were
distinguishable on the basis of size. However, very small fish ("tinies") were
difficult to identify individually. Consequently, in the colonies containing
more than one "tiny," i.e., in colonies larger than six fish, the number of
interactions per tiny was determined by dividing the sum of chases given or
received by all tinies of a particular colony by the number of tinies in that
colony. Therefore, in each colony larger than six individuals, the rank (see
below) of six consisted of an average interaction rate per tiny.

Data were taken during both the morning and afternoon over a period of
3-6 days and until at l'east 65 interactions had been noted for each colony
(average observation time per col~ony = 131 minutes; range = 85-240 minutes).
Since colonies were monitored for varying lengths of time, the observations
were standardized to give the number of chases per 100 minutes per colony so
that the colonies could be compared directly.

Matrices were constructed for the chase data from each colony by ass ignin*
dominance ranks to every colony member. For example, rank I chased each other*

130

colony member more than any member chased it; rank 2 chased each colony member
more than any other member chased it, except for rank 1.

For analysis, data on chases were ordered by rank, and all data for each
rank were summed across all 7 groups. The data were then divided into three
categories: (a) the number of chases received by each rank; (b) the number of
chases given by each rank; and (c) the number of chases given by each rank to
each other rank. Each of these categories was analyzed using the Kruskal-Wallis
one-way ANOVA to determine whether or not there was a significant difference
among ranks, i.e., an H-value giving p < 0.05. If so, then the data were
subjected to a nonparametric version of the Newman-Keuls Multiple Range test
(Zar, 1974) to determine how the distribution of chases varied among ranks
(significance level, p < 0.05).

RESULTS

For each colony, a distinct size-dependent social hierarchy was apparent;
it was highly linear with proportionally very few reversals (4.5%). A reversal
is a chase directed at an individual higher in the hierarchy than the chaser.

Figure I shows the distribution of chases received by each rank (category
a). This distribution is not homogeneous (Kruskal-Wallis, p < 0.001). The
nonparametric Multiple Range test showed that ranks 3 and 4 received the most
chases; ranks 2, 5, and 6 received less than 3 and 4 but more than rank I (p <
0.05 in each case).

Figure 2 shows the distribution of chases given by each rank (category b).
This distribution is not homogeneous (Kruskal-Wallis, p < 0.001). The
nonparametric Multiple Range test showed that ranks 1, 2, and 3 gave the most
chases; rank 4 gave fewer chases but more than ranks 5 and 6 (p < 0.05 in each
case).

Figure 3 shows the distribution of chases from each rank to each rank
below it in the hierarchy (category c). The distribution is not homogeneous
for ranks 1 to 4 (Kruskal-Wallis, p < 0.01). In most cases, each rank directed
most chases to its closest ranking subordinate and least to the individual
ranked farthest from itself. See figure 3 for the significant differences
shown by the nonparametric Multiple Range test (p < 0.05 in each case).

CONCLUSIONS

The analyses of the distribution of chases given and received show that
there is a distinct pattern to intracolony aggression:

(a) there is a strong size-dependent social hierarchy which is linear
with few reversals;

(b) individuals chase those closest to and just below themselves in rank
most frequently and those furthest in rank least frequently;

(c) the highest ranking individuals (1, 2, and 3) chase more than other
ranks. Lower ranking individuals chase progressively less the lower their rank;

131

32-
H~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
H~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> I

/ I I I
I ~~~~~~~~~~~~~~I
I ~~~~~~~~~~~~I

'- 11
16 I
II

1 2 3 4 5 6
FISH RANK

Figure 1.--Chases received by each rank plotted as medians and interquartile
ranges. The asterisks denote significant differences between adjacent
points (p < 0.05); e.g., rank 2 received more chases than rank 1.
132

T
40
T
l l

l l
I I

32 4

>~~~~~~~~~~~~~~~
H

C~~~~~~~~~~~
o X 6 II

I~~~~~
$ ~~~~24 6 \ I
4 I l \ I
H F IS IN
I i
0

X ~~~I Ix
I I

(p < 0.05); e~~~~~~~~g., rak4gvsfwrchsstaak3
1.6 II ' ''1

I I J~~~~~~~~~~.

1 2 3 4 5 6
FISH RANK
Figure 2.--Chases given by each rank plotted as medians and interquartile
ranges. The asterisks denote significant differences between adjacent points
(p <~ 0.05); e.g., rank 4 gives fewer chases than rank 3.
133

120
RANK 2
x

100

so 80RANK 3

x
RANK 4

2 3 4 56

FISH RANK

Figure 3.--Chases given by each rank to each rank below it. Data from all seven
groups are summed. Each line represents the chases given by a particular rank
to all ranks below it; e.g., rank I chased rank 2 a total of 96 times. The
asterisks denote significant differences (p < 0.05) between adjacent points;
e.g., rank I chased rank 2 significantly more than it did rank 3.

(d) ranks 2, 3, and 4 receive most chases and ranks 1, 5, and 6, the least.

It is clear that in'the bicolor damselfish, size plays an important role
in intraspecific aggression. As in many other species, dominance relations are
strongly correlated with body size (beetles, Beebe, 1947; lobsters, Fielder,
1965; voles and deerniice, Grant, 1970). Aggression is most intense between
individuals of similar size and least intense between individuals most dissimilar
in size (Noble, 1939; Collias, 1944; Miller, 1964; Coates, 1977).

The importance of size in the dominance relations of damselfish may result i
from size-dependent competition for certain resources. Individuals of different

134

sizes may have different food or shelter requirements (Rasa, 1969; Emery, 1973).
Several studies have reported that in pomacentrids, small individuals often
share the territory of an adult conspecific (Clarke, 1970; Emery, 1973; Sale,
et al., 1980), either because the adult tolerates them or because of "topological
-ce--Tion" (Sale, et al., 1980) whereby they can avoid the adult by using space
in which they cannot t~pursued. The low number of chases given and received
by these small individuals implies that they are not attempting to compete
aggressively with the larger fish for the space they share. Large size in male
bicolors appears to be one of the factors associated with high reproductive
success (Schmale, 1980). In order to establish the significance of size and
how it relates to aggression in this species, detailed studies on the feeding,
reproduction, and use of space and shelter are necessary. Account should be
taken of fish size, stage of maturation (juvenile, adult; reproductive,
nonreproductive), and sex.

The present study demonstrates both similarities and differences between
field and laboratory results on the dominance relations of the bicolor damselfish.
Previous work on this species (Myrberg, 1972a) focused on two distinct time
periods: reproductive and nonreproductive. The distribution of aggression
between colony members was found to differ markedly between these two periods
in the laboratory phase of the study. The field phase of the study covered
only the reproductive period and was much less detailed than the laboratory
phase because of the inability of the underwater camera to record all aggressive
interactions between colony members, especially those involving very small
fish. The present study is considered equivalent to the nonreproductive period
of the Myrberg study, since little courtship activity and no egg-guarding
behavior by males were observed.

The present field analysis and Myrberg's laboratory results for the
nonreproductive period show the following similarities: There is a strong size-
dependent linear dominance hierarchy with a very low percentage of reversals;
the highest ranking individuals did the most chasing and the lowest ranks the
least; the middle ranks received the most chases and rank 1 and the lowest
ranks the least.

However, there were two differences between the two studies. First, the
rate of interactions was markedly higher in the laboratory study, in which a
rate of 10 chases per individual per hour was recorded as opposed to the 1.5
chases per individual per hour in the present study. This is a common difference
between field and laboratory studies and may be due to: (a) the inability of
small individuals to escape those chasing them in an aquarium, (b) the lack of
interspecific interactions in the laboratory, and/or (c) the lack of space
available per individual in an aquarium, although both the present study and
Myrberg's laboratory study had similar fish densities of 0.2 m2 (Myrberg) and
0.1-0.2 m2 (this study).

Second, the relative distribution of chases given by each rank to each
other rank below it in the hierarchy differs between the two studies. In the
laboratory (Myrberg, 1972a), the distribution of chases among individuals showed
no clear pattern, except -or one fish, YB, which directed successively fewer
chases at individuals ranked progressively further away from itself. This 'YB'
pattern of chasing was seen very clearly and consistently in all colony members
of the present study.

135

It is of considerable interest that during the reproductive period of
Myrberg's study, the pattern of intracolony aggression in the laboratory and
possibly in the field differed from that recorded during the nonreproductive
period, with more aggression being directed towards the lowest ranking individuals
than towards those closest in rank. Field work similar to that of the present
study needs to be conducted during the reproductive period to establish if this
difference is significant and related to the phase of the reproductive cycle.

In conclusion, the dominance relations in the bicolor damselfish during
the nonreproductive period show both similarities and differences when laboratory
and field studies on colonies of similar size are compared. The nature of
these differences implies that the laboratory may be used for detailed work
impossible to carry out in the field, but that laboratory results must be
interpreted with caution and should not be used as the sole source of information
on the ecological significance of dominance relations. The importance of
parallel field and laboratory studies is emphasized, since it is evident that
many factors may influence intracolony aggression.

ACKNOWLEDGMENTS

I wish to thank Kimberly Leighton-Boulon for her invaluable help during
the field study and Ileana ClaviJo, who did the organizational work prior to
the HYDROLAB mission. I also wish to thank the HYDROLAB staff in St. Croix for
making the mission so successful. For suggestions and comments on the manuscript,
I am very grateful to D. Y. Shapiro and G. R. Mitcheson.

The study was carried out with the financial assistance of NOAA and of the
SERC of Great Britain Grant No. 78307917.

LITERATURE CITED

Beebe, W. 1947. "Notes on the Hercules Beetle, Dynastes hercules (Linn.) at
Rancho Grande, Venezuela, With Special Reference to Combat Behavior."
Zoologica, New York 32:109-116, .

Chase, I. 0. 1974. "Models of Hierarchy Formation in Animal Societies."
Behav. Sci. 19:374-382.

Clarke, T. A. 1970. "Territorial Behavior and Population Dynamics of a
Pomacentrid Fish, the Garibaldi Hypsypops rubicunda." Ecol. Monogr. 40:189-
212.

Coates, D. 1977. "The Social Behaviour of the Humbug Damselfish Dascyllus
aruanus (Pieces: Polacentridae), With Notes on Two Other Species (D.
reticulatus and D. trimaculatus)." M. Sc. Thesis, Univ. Newcastle Upon
Tyne, England. 750 pp.

Collias, N. E. 1944. "Aggressive Behavior Among Invertebrate Animals."
Physiol. Zool. 17:83-123.

136

Emery, A. R. 1973. "Comparative Ecology and Functional Osteology of 14 Species
of Damselfish (Pisces: Pomacentridae) at Alligator Reef, Florida Keys."
Bull. Mar. Sci. 74:649-770.

Fielder, D. R. 1965. "A Dominance Order for Shelter in the Spiny Lobster
Jasus lalandei (H. Milne-Edwards)." Behaviour 24:236-245.

Grant, P. R. 1970. "Experimental Studies of Competitive Interactions in a Two-
Species System: II. The Behaviour of Microtus, Peromyscus and Clethrionomys
species." Anim. Behav. 18:411-426.

Hinde, R. A. 1974. Biological Bases of Human Social Behaviour. McGraw-Hill
Book Co., Inc., New York. xvi + 462 pp.

Miller, R. S. 1964. "Ecology and Distribution of Pocket Gophers (Geomyiidae)
in Colorado." Ecology 45:256-272.

Myrberg, A. A., Jr. 1972a. "Social Dominance and Territoriality in the Bicolor
Damselfish Eupomacentrus partitus (Poey) (Pisces: Pomacentridae)." Behaviour
41:207-231.

Myrberg, A. A., Jr. 1972b. "Ethology of the Bicolor Damselfish Eupomacentrus
partitus (Pisces: Pomacentridae): a Comparative Analysis of Laboratory and
Field Behavior." Anim. Behav. Monogr. 5:199-283.

Noble, G. K. 1939. "The Role of Dominance in the Social Life of Birds." Auk
56:263-273.

Rasa, O. A. E. 1969. "Territoriality and the Establishment of Dominance by
Means of Visual Cues in Pomacentrus jenkinsi (Pisces: Pomacentridae)." Z.
Tierpsychol. 26:825-845.

Rowell, T. E. 1967. "A Quantitative Comparison of the Behaviour of a Wild and
a Caged Baboon Troup." Anim. Behav. 15(4):499-509.

Sale, P. F., P. J. Doherty, and W. A. Douglas. 1980. "Juvenile Recruitment
Strategies and the Coexistence of Territorial Pomacentrid Fishes." Bull.
Mar. Sci. 30:147-158.

Schmale, M. C. 1980. "A Preliminary Report on Sexual Selection in the Bicolor
Damselfish, Eupomacentrus partitus." Bull. Mar. Sci. 30:328 (abstract).

Thompson, W. L. 1960. "Agonistic Behavior in the House Finch: II. Factors
in Aggressiveness and Sociality." Condor 62:378-402.

Zar, J. H. 1974. Biostatistical Analysis. Prentice-Hall, Inc., Englewood
Cliffs, N.J. xic + 620 pp.

137

I 0

A COMPARISON OF FORE AND BACK REEF POPULATIONS OF DIADEMA ANTILLARUM PHILIPPI
AND EUPOMACENTRUS PLANIFRONS CUVIER AT ST. CROIX, U.S. VIRGIN ISLANDS1

C. Lance Robinson2 and Ann Houston Williams

Department of Zoology-Entomology
Alabama Agricultural Experiment Station
Auburn University
Auburn, Alabama 36849

ABSTRACT

The community structure of populations of threespot damselfish (Eupomacentrus
planifrons) and the black sea urchin (Diadema antillarum) in fore reef and back
reef habitats of Teague Bay reef off the West Indies Laboratory, St. Croix,
U.S. Virgin Islands, was examined during July and August 1981. All populations
studied occupied patches of Acropora palmata at depths of 3-5 m and were observed
with SCUBA.

Field observations indicated that populations of Diadema within back and
fore reef habitats differed in the following ways: 1) mean urchin test diameter
was significantly larger on the back reef (X = 74.48 mm) than on the fore reef
(7 = 47.94 mm), 2) mean urchin density, sampled during daylight hours was lower on
the back reef (X = 2.3 urchins m'3) than on the fore reef (X = 5.04 urchins m'3),
and 3) on the back reef 60.8% of the urchins were found on coral branches more
than 60 cm above the substrate, whereas on the fore reef 74.5% of the urchins
were found below that level. Damselfish densities were significantly lower on
the back reef (X = 0.18 m-3 vs. 0.31 m-3) and algal lawns were significantly smaller
(X = 146.1 cm2 vs. 349.1 cm2) than on the fore reef. Back reef damselfish were
more aggressive (7 bites min'1 = 27.6) than those on the fore reef (X bites
min-1 = 6.6).

These differences in community structure may indicate differing levels of
competitive interaction between threespot damselfish and Diadema. The higher
energy environment of the shallow fore reef may effectively reduce the size of
Diadema by reducing urchin grazing. Thus, competitive interactions between
threespots and Diadema may be lower on the fore reef than on the back reef.

INTRODUCTION

The community structure and competitive interactions between threespot
damselfish (Eupomacentrus planifrons Cuvier) and the herbivorous sea urchin
(Diadema antillarum Philippi) have been documented in the East Back Reef of

1 Contribution No. 130 from the West Indies Laboratory, Fairleigh Dickinson
University.
2 Present address: 41 Wagon Wheel Drive, Semmes, Alabama 36575.

139

Discovery Bay, Jamaica, West Indies (Williams, 1978, 1979, 1980, 1981). The
damselfish maintains algal lawn territories on dead branches of staghorn coral
(Acropora cervicornis Lamarck) and actively excludes herbivores, such as Diadema,
from these territories (Myrberg and Thresher, 1974; Williams, 1979). Through
these competitive interactions, habitat partitioning occurs which affects the
distribution and abundance of Diadema on the reef (Williams, 1979, 1981).
Because of its influence on the distribution of grazing sea urchins, the
damselfish is considered to be a keystone species (Williams, 1980). Sammarco
and Williams (1982) also have suggested that E. planifrons indirectly affects
shallow reef coral community structure by infuencing the differential success
of coral recruitment and algal growth both within and outside damselfish
territories.

This preliminary study examines the structure of two similar communities
and the interactions of these species on the reefs of St. Croix, U.S. Virgin
Islands, to determine whether or not damselfish occupy a similar role in reef
communities other than Jamaica. Both fore and back reef communities were
examined for comparison with each other and with previous observations from
Jamaica.

MATERIALS AND METHODS

Field studies were conducted on the bank barrier reef (Adey, 1978) located
north of West Indies Laboratory on the northeast coast of St. Croix, U.S.V.I.
(17�45.5'N, 64�35.5'W). The reef studied was representative of the area, with
the back reef composed of elkhorn coral (Acropora palmata) in the shallower
areas and various head corals (Montastrea annularis, Diploria spp., etc.) found
in deeper water. Millepora complanata and A. palmata were found in the high
energy areas of the crest, while A. palmata, A. cervicornis, and several species
of head corals prevailed on the fore reef (Ogden, 1974). A study site was
established west of the WIL pier at the site of solar panels on both fore and
back reefs in comparable water depths. This area was chosen because of
accessibility of the fore reef through a swim cut in the reef crest. All
observations and experiments were carried out between 1030 and 1545 hr using
SCUBA.

Study of the interactions between threespot damselfish and Diadema was
restricted to those which took place among branches of A. palmata, an apparent
preferential substrate for threespots in the absence of-. cervicornis (Williams,
pers. obs.; Itzkowitz, 1974). Individual stands of A. paTmata were selected for
observation by the following criteria: at least twothreespot damselfish per
coral stand; coral stand boundaries easily discernible; and coral stands located
in -3 to -5 m of water. Aluminum tags were used to identify the A. palmata
stands. Length, width, and height of each stand were recorded and used in
calculating densities of damselfish and urchins (number of individuals m'3).
Wilcoxon rank sum tests were used to compare densities and coral stand measure-
ments between fore and back reefs. Correlation analysis (Sokal and Rohlf,
1969) was utilized to examine relationships between the species' densities.

Vertical distributions of urchins within coral stands were recorded by
suspending a leaded line, calibrated in centimeters, among the branches of 0

140

coral and counting individuals located within 5 cm intervals above the substrate.
A G-test of independence (Sokal and Rohif, 1969) was used to compare distribu-
tions of urchins above and below 60 cm coral height on both sides of the reef.

A comparison of urchin populations located in the two habitats was made by
examining the size class structure of Diadema. Haphazard samples of individuals
from each population were collected by hand from several areas east and west of
the study sites. Greatest test diameter was measured (to the closest 0.01 cm)
across the ventral surface using vernier calipers for 106 individuals from both
fore and back reefs. Mean test diameters were compared using a Student's t-
test, and size class structures were determined by the method of Harding (1949).

A measure of the relative aggression of threespots toward Diadema in both
habitats was recorded. These data were obtained by placing an urchin representa-
tive of the mean size in the algal lawn of a haphazardly chosen damselfish and
counting the number of bites on urchin spines made by the damselfish in 1 min
or until the urchin was outside the algal lawn. Relative levels of aggression
were compared using a Student's t test. All statistical computations were
completed by using the Statistical Analysis System (SAS).

RESULTS

Stands of elkhorn coral were significantly smaller in total volume on the
fore reef than on the back reef due to significantly smaller width dimensions
(table 1). Densities of Eupomacentrus planifrons associated with the fore reef
stands were significantly higher than those of the back reef. Algal lawns were
significantly larger on the fore reef than on the back reef (table 1). Although
more coral substratum apparently was available in the back reef, fewer fish
maintained smaller territories in this region.

The density of Diadema was higher on the fore reef than on the back reef
(table 1). Additionally, mean test diameter was significantly smaller (t =
18.64; df = 191; p < 0.0001) on the fore reef (: = 47.94 mm) than on the back
reef (: = 74.48 mm). Population size class structure exhibited no apparent
differences between habitats (fig. 1). When all urchins from both habitats are
considered, a probable negative relationship may exist between the size and
density of Diadema. However, because of the limited data available on mean
test diameter and mean density of Diadema in each habitat (N = 2), correlation
analysis could not be completed.

Although significance was not testable due to small sample sizes (N = 3),
an apparent negative correlation was demonstrated between the density of Diadema
and the density of E. planifrons on the back reef, but this was not the case on
the fore reef (fig. 2). Thus, in the back reef habitat where densities of E.
planifrons were low and algal lawns were small, small numbers of large Diadema
were found.

Relative aggression levels of E. planifrons toward Diadema were found to
be much higher (t = 5.25, df = 13, <0.0002) on the back reef (X bites
min-1 = 27.6) than those exhibited by threespots on the fore reef (X bites
min-1 = 6.6) (fig. 3). Thus, intensity of the behavioral interaction between

141

Table 1.--Comparison of communities on fore and back reefs at St. Croix,
U.S.V.I. Data are means (+ SE) and were tested with Wilcoxon rank sum tests
(Z statistic) for all tests except for algal lawn, where a t test was used.

Fore Reef Back Reef Test Statistic P

A. palmata (m3)

Volume 9.5(1.3) 16.79(2.52) Z = 1.75 0.04

N 3 3

A. palmata (m)

Height 1.42(.03) 1.5(.05) Z = 1.03 n.s.

N 3 3

A. palmata (m)

Width 1.87(.19) 2.63(.28) Z = 1.75 0.04

N 3 3

Threespot Density

No. ind.m 0.31(.04) 0.18(.01) Z = 1.75 0.04

N 3 3

Diadema Density

No. ind.m 3 5.04(1.86) 2.30(.47) Z = -1.75 0.04

N 3 3

Algal lawn

(cm2) 349.11(44.57) 146.11(12.89) t = -4.38 < 0.002

N 9 9

142

12 Fore Reef
11
10
9
8
7
6
5
< 4
4 - - -
U 3
<: 2
a , I-
LA.I
0
= 8 Back Reef
.Z 7

5
4
3
2 [i

n mnM n nd l fn
25 30 35 40 45 50 55 60 65 70 75 80 85

TEST DIAMETER [mml

Figure 1.--Frequency distribution of size classes (mm) for Diadema occurring
on the fore reef (top) and back reef (bottom) of St. Croix; N = 106 for
each site.

threespots and Diadema was strongest, and Diadema density varied inversely with
damselfish density on the back reef.
Although the volume of A. palmata per stand differed in the two habitats,
the heights of A. palmata stands were not significantly different (table 1).
However, within T. palmata stands on the fore and back reefs, the vertical
distribution of Dadema was found to differ significantly (table 2). A large

143

o �
8

zoo4 o
a \

a a
31

0-
0.C1 0.20 0. 2 0.30 0.35 0.40

DAMSELFISH DENSITY (no. ind. 3)
Figure 2.--Relationship between density of Diadema (no. inds. m-3) and threespot
damselfish (no. inds. m-3). Closed circles ='back reef populations; open
circles -= fore reef populations. 144

30 S.E.=2.68

28 X 27.6

" 22

20
,I)
I- 18

Z 14
0
12
AdI 10
10 8S.E.=1.57

I X =6.6
6
I
4

0
B F

Figure 3.--Mean relative aggression level (bites min-1) of threespot damselfish
toward urchins on the fore reef (F) and back reef (B) of St. Croix. Fore
reef N = 5, back reef N = 10.

proportion of the Diadema (60.8%) tended to occupy the upper branches of coral
( > 65 cm in height) on the back reef, while 74.5% of the urchins on the fore
reef were located on coral branches less than 60 cm above the substrate.
Differences in sampling times cannot account for these differences, since all
observations but one were made between 1545 and 1830 hr. One back reef

145

Table 2.--Comparison of proportional (%) vertical distributions of Diadema
within stands of A. palmata on fore and back reefs. A G-test compared
individuals found at heights of < 60 cm or > 65 cm (G = 10.76, p < 0.005,
df = 1).

Ht. Fore Reef Back Reef
(cm)

15-30 13.9 19.6

35-60 50.6 19.6

65-90 21.5 12.2

95-135 13.9 48.6

N 79 107

observation was made at 1030 hr; however, it did not differ from samples taken
later in the same area. Thus, differences in vertical distribution must have
been due to differing urchin behavior.

DISCUSSION

Although the damselfish-Diadema assemblages in the fore and back reef
environments at St. Croix appear similar in population composition, the structures
of these two assemblages differ in population abundances, distributions, and
mean size. These differences may be the result of differing physical factors
and/or biological interactions.

Colonies of Acropora palmata were similar in height but significantly
smaller in volume on the fore reef (table 1). The difference in volume was
attributable to differences in width of coral stands. This structural difference
in the corals may be related to the higher wave energy on the fore reef. Waves
can cause considerable damage to fore reef corals, as evidenced by abundant
coral rubble found on the fore reef substratum (Robinson, pers. obs.; Van den
Hoek, et al., 1978). Rogers, et al. (1982) documented that storm damage on the
reefs of St. Croix was due to fractures of the distal ends of A. palmata
branches.

Mean test diameters of Diadema from the fore reef are smaller and those
from the back reef are larger in the present study than those reported by Ogden,

146

et al. (1973) on patch reefs at St. Croix (7 = 59 mm). They also are slightly
larger than mean diameters measured by Bauer (1980) on the back reef at St.
Croix (7 = 52.3 mm). Such differences may indicate differences in the availability
of food, mortality, or adaptations to energy levels in the three different
habitats.

The differences between fore and back reef observed in this study also may
reflect differences in the availability of food, mortality, and/or adaptations
to differing wave energy levels. Preliminary data indicate that smaller urchins,
such as those found on the fore reef, may feed at a faster rate than the larger
urchins of the back reef (Robinson, unpub. data). However, the larger algal
lawns also found on the fore reef indicate a reduced grazing pressure by urchins
(table 1). Van den Hoek, et al. (1975, 1978) attributed a higher biomass of
algal turf to lower grazing pressure from Diadema in similar high energy regimes.
Thus, the smaller body size of urchin populations found on the fore reef may be
due to food limitation (Carpenter, 1981) even though abundant food is present
(i.e., larger damselfish lawns, table 1). High wave energy may force urchins
to remain lower in the coral matrix, preventing them from reaching this potential
food source (table 2). High energy regimes either may select for individuals
with smaller test diameters (i.e., smaller surface area for exposure to wave
energy) or may result in smaller urchins by reducing their effective feeding
periods. Lissner (1980) reported an inverse relationship between turbulence
and activity of the urchin Centrostephanus coronatus.

Williams (1981) demonstrated that Diadema and threespot damselfish in
Jamaica were actively involved in competition for either food or space in the
back reef. This competition affected the amount of algae accumulating outside
of algal lawns. The relative level of competition in differing environmental
conditions was not measured in that study. We believe that the lower levels of
aggression in damselfish (fig. 3) and their larger lawns on the fore reef
provide evidence of reduced competitive interactions in this environment as
compared to the back reef. Additionally, larger lawns and a greater food
supply may be necessary for damselfish to maintain themselves in this high
energy regime. Reduced competition may be the result of environmental mediation,
whereby high wave energy reduces urchin grazing and thus increases the size of
damselfish territories.

Sammarco, et al. (1974) and Sammarco (1980) indicated that, at high
densities of Diadema, successful settlement of corals was inhibited; in contrast,
maximum diversity of successfully settled coral spat was achieved at lower
densities of Diadema. Sammarco and Williams (1982) also indicated that settlement
success varied within and outside of damselfish algal lawns due to alterations
in urchin grazing levels. Similar factors affecting the settlement success of
corals will vary between fore and back reef habitats at St. Croix and may
account for differences in adult coral diversities in these two environments.

ACKNOWLEDGMENTS

We wish to thank the staff and researchers of the West Indies Laboratory
of Fairleigh Dickinson University. Special thanks go to Drs. J. C. Ogden, J.
Ebersole, and J. Morin for their advice and helpful comments. We thank R.

147

Carpenter, J. Sutherland, R. Lishak, W. Mason, and two anonymous reviewers B
for critical reviews of the manuscript. We especially appreciate the assistance
of J. Ebersole and J. Sobel in diving. Financial support was provided for one
of us (AHW) by the Department of Zoology-Entomology, Auburn University, especially
through the generosity of Dr. George Folkerts.

LITERATURE CITED

Adey, W. H. 1978. "Coral Reef Morphogenesis: A Multi-Dimensional Model."
Science 202:831-837.

Bauer, J. C. 1980. "Observations on Geographic Variations in Population
Density of the Echinoid Diadema antillarum Within the Western North Atlantic."
Bull. Mar. Sci. 30:509-515.

Carpenter, R. C. 1981. "Grazing by Diadema antillarum (Philippi) and Its
Effects on the Benthic Algal Community.' J. Mar. Res. 39:749-765.

Harding, J. P. 1949. "The Use of Probability Paper for the Graphical Analysis
of Polymodal Frequency Distributions." J. Mar. Biol. Assn. U.K. 28:141-153.

Itzkowitz, M. 1974. "A Behavioral Reconnaissance of Some Jamaican Reef Fishes."
Zool. J. Linn. Soc. 55:87-118.

Lissner, A. L. 1980. "Some Effects of Turbulence on the Activity of the Sea
Urchin Centrostephanus coronatus Verrill." J. Exp. Mar. Biol. Ecol. 48:185-
191.

Myrberg, A. A., and R. E. Thresher. 1974. "Interspecific Aggression and Its
Relevance to the Concept of Territoriality in Reef Fishes." Amer. Zool.
14:81-96.

Ogden, J. C. 1974. "The Major Marine Environments of St. Croix." In: H. G.
Multer and L. C. Gerhard (eds.), Guidebook to the Geology and EcoT2gy of
Some Marine and Terrestrial Environments. West Indies Laboratory, Fairleigh
Dickinson University, Spec. Pub. 5:5-19.

Ogden, J. C., D. P. Abbott, and I. Abbott. 1973. The Population of Diadema
antillarum on Patch Reef #3. Studies on the Activit and Food of the Echinoid
Diadema antT'larum PRiippi on a West Indian Patch Reef-' West Indies
Laboratory, Fairleigh Dickinson-Un-i versity,Spec. Pu7 -2:24-28.

Rogers, C. S., T. H. Suchanek, and F. A. Pecora. 1982. "Effects of Hurricanes
David and Frederic (1979) on Shallow Acropora palmata Reef Communities:
St. Croix, U.S. Virgin Islands." Bull. Mar. Sci. 32:532-548.

Sammarco, P. W. 1980. "Diadema and Its Relationship to Coral Spat Mortality:
Grazing, Competition, and Biological Disturbance." J. Exp. Mar. Biol. Ecol.
45:245-272.

148

Sammarco, P. W., J. S. Levinton, and J. C. Ogden. 1974. "Grazing and Control
of Coral Reef Community Structure by Diadema antillarum Philippi
(Echinodermata: Echinoidea): A Preliminary Study." J. Mar. Res. 32:47-51.

Sammarco, P. W., and A. H. Williams. 1982. "Damselfish Territoriality:
Influence on Diadema Distribution and Implications for Coral Community
Structure." Mar. Ecol. Pros. Ser. 8:53-59.

SAS Institute, Inc. 1982. SAS User's Guide: Statistics. Cary, North Carolina.
584 pp.

Sokal, R. R., and F. J. Rohif. 1969. Biometry: The Principles and Practice
of Statistics in Biological Research. W. H. Frean, San Francisco. 776 pp.

Van den Hoek, C., and A. M. Breeman. 1978. "The Distribution of Algae, Corals
and Gorgonians in Relation to Depth, Light Attenuation, Water Movement, and
Grazing Pressure in the Fringing Coral Reef of Curacao, Netherlands Antilles."
Aquat. Bot. 5:1-46.

Van den Hoek, C., A. M. Cortel-Breeman, and J. B. W. Wanders. 1975. "Algal
Zonation in the Fringing Coral Reef of Curacao, Netherlands Antilles, in
Relation to Zonation of Corals and Gorgonians." Aquat. Bot. 1:269-308.

Williams, A. H. 1978. "Ecology of Threespot Damselfish: Social Organization,
Age Structure, and Population Stability." J. Exp. Mar. Biol. Ecol. 34:197-
213.

Williams, A. H. 1979. "Interference Behavior and Ecology of Threespot
Damselfish." Oecologia 38:223-230.

Williams, A. H. 1980. "The Threespot Oamselfish: A Non-Carnivorous Keystone
Species." Amer. Nat. 116:138-142.

Williams, A. H. 1981. "An Analysis of Competitive Interactions in a Patchy
Back-Reef Environment." Ecology 62:1107-1120.

149

CARBONATE SEDIMENT PRODUCTION BY THE ROCK-BORING URCHIN
ECHINOMETRA LUCUNTER AND ASSOCIATED ENDOLITHIC INFAUNA
AT BLACK ROCK, LITTLE BAHAMA BANK

Charles M. Hoskin and John K. Reed
Marine Geology Department
Harbor Branch Foundation, Inc.
Fort Pierce, Florida 33450

ABSTRACT

Bioerosion studies were conducted during four cruises in 1982-83 to Black
Rock, a 405 m long island of carbonate eolianite on Little Bahama Bank. The
urchin population (X = 37 adults m'2, 92 x 103 total) bores in a 6 m wide zone
at depths of 0.5-3 m. SCUBA divers using chisels collected rocks with urchins
in their bore holes and similarly sized rocks without urchins and placed them
separately in 20 1 buckets with 0.0625 mm screen-walls for 2 days (13 usable
measurements). Urchins produced spherical to elliptical pellets 1-2 mm in
diameter. Disaggregated pellets contained no particles > 1.00 mm. The pellets
were 46% (wt) unimodal sand, with a modal grain size between 0.125-0.177 mm,
and 54% mud. Urchins produced an average of 242 mg sediment urchin-1 day-1
(dry wt), which is equivalent to 8.9 g m-2 day-1, or 9 tons yr-1 for the entire
population. Volume of individual urchin bore holes averaged 72 cm3. Calculating
from boring rates, the bore holes were excavated in an average of 2.9 years.

Rocks without urchins (controls) produced an average of 0.5 mg organic-
free sediment cm-2 day-1 (dry wt). These particles were produced by bioerosion
of the infauna, which included eunicid and sipunculid worms, sponges, Lithotrya
barnacles, pelecypods, and microborers. Inorganic sediment weight was correlated
(r = 0.968, p < 0.05) with surface area of the control rocks. Infauna from
control rocks produced 5.0 g m-2 day-1, equivalent to 6.5 tons yr-1, for Black
Rock.

INTRODUCTION

Bioerosion has a two-fold role in coral reefs; hard substrate is removed,
creating cavities, and some of the former substrate becomes sedimentary particles
that may be transported to new environments (Hopley, 1982). We have been
studying the area surrounding Black Rock, a small Pleistocene eolianite island
on the southwestern Little Bahama Bank, in order to identify and measure sources
of sediment, to follow dispersal pathways, and to measure the rate at which the
sediments accumulate.

Bioerosion by the urchin Echinometra lucunter (Linnaeus) has been studied
previously in Barbados (McLean, 1967), Bermuda (Hunt, 1969), and the Virgin
Islands (Ogden, 1977). Another species, E. mathaei, has been studied in Persian
Gulf reefs by Shinn (p. 39, in Hughes Clar-e and Keij, 1973) and Enewetak
*s ~(Russo, 1980). The purposes of this study are to determine the rate of sediment
isw production and cavity formation by E. lucunter using a new method, to characterize

151

p M~~~~~~~~~~~pw -~~~~~~~~~

the size distribution of the particles produced, and to compare these data with
the rate of production and size of particles produced by the rock-boring infauna
closely associated with the urchin bore holes.

METHODS

Urchin Population

The distribution and abundance of E. lucunter adults (about 3 cm test
diameter) was determined by SCUBA divers off western Black Rock. Most of the
eastern side is sand, without urchins. Using a 0.25 m2 PVC frame, the divers
counted the adult urchins within each frame (total of 35 frames) on three
separate transects that were continuous from the intertidal zone seaward until
no more urchins were found. The number of adult urchins was calculated for
each linear meter in a seaward direction. These numbers were extrapolated for
the lateral extent of the island to yield the total number of urchins.

Sediment Yield

Using chisels, SCUBA divers collected rocks, each containing one urchin in
its intact bore hole, from 2-3 m depths off western Black Rock. Similarly
sized rocks without urchins were collected for controls. Each rock was brushed
underwater to remove loose particles. The rocks were placed separately, without
exposure to air, in 20 1 plastic buckets. Each rock was placed on a styrofoam
pad and secured to the bottom of the bucket by criss-crossed rubber bands
latched onto plastic hooks attached inside the buckets. The buckets were
ventilated with 1150 cm2 of 0.0625 mm nylon screen on the side and 72 cm2 of
0.5 mm nylon screen at the top. Using an oxygen electrode, measurements made
at the start and end of an experimental run showed only a small decrease (maximum
decrease of 0.4 mg 1-1) in dissolved oxygen within the bucket. Buckets were
secured to the seafloor at the collection site in various ways, all designed to
minimize movement of the rock within the bucket, which otherwise might have
produced sediment particles by abrasion (fig. 1A). Buckets were deployed for
43-50 hr during four separate cruises. Five control and eight experimental
buckets yielded thirteen usable measurements out of the twenty-four buckets
deployed. Some buckets were lost in squalls and others had rips in the nylon
mesh and therefore were not included.

Figure 1 (opposite page).--(A) Buckets (20 1)' in situ at 2.5 m depth used for
urchin study; (B) SEM of limestone (eolianiteT wall in Echinometra lucunter
bore hole. Upper right - encrusted with lithothamnoid algae. Lower left -
algae-free area eroded by urchin and showing spherical grains in the limestone;
(C) Fecal pellets from alimentary tract of E. lucunter. Bar = 0.5 mm;
(D) Broken and partially regrown spine of E.7lucunter; (E) Abraded end of
tooth from Aristotle's lantern of E. lucunter; (F) SEM close-up of fig. 1E
(arrow), showing laminated microstructure of tooth tip visible from abrasion.
Tip pointing to right.

153

The particles produced were recovered by gently washing the rock (and
urchin) while inside the bucket, siphoning off the water, and then washing the
particles into polyethylene bags. Particles, rocks, and urchins were labelled
and then preserved in 70% ETOH. Particles were washed free of salt, oven dried
at 950C, and weighed to 0.01 mg. Twelve free-living urchins were collected and
their alimentary tracts removed by dissection, measured, and the length of
tract containing sediment estimated. The gut contents were removed, washed,
dried, and weighed. The organic content of sediment from the urchin alimentary
tracts, experimental buckets, and control buckets was determined by weight loss
following H202 digestion. Subsequently, inorganic sediment from each sample was
wet sieved (0.0625 mm stainless steel screen) to separate mud. Gravel and sand
size particles (> 0.0625 mm) were oven dried at 95�C and dry sieved at 0.5 Phi
intervals for 5 min using agitation by Vibrapad. Weight of each sieve fraction
was recorded to 0.01 mg. Particle size distributions were generated by the
Cal-Comp plotter from these weights.

Attached organisms (mostly algae) were removed from each rock, washed,
dried, and weighed. Each rock was then oven dried, and the exposed surface
area was measured by the aluminum foil method (Reed, et al., 1982). The volume
of each urchin bore hole was measured by determining t-levolume of water
contained in a latex sheath placed inside the bore hole. Each rock was dissolved
in acid (HCl + HN03) to recover the infauna, which was washed, dried, and
weighed. Urchins were dried and weighed, and test dimensions and volume were
determi ned.

RESULTS

Urchin Population

The total urchin population at Black Rock was 92.4 x 103 adults within an
area of 2521 m2. Average urchin density was 37 adults m-2 (maximum was 100
m-2). Urchins were present between the intertidal zone and 7 m depth. Urchins
never were observed to leave their bore holes; during the day they remained
stationary at the bottom of the hole, but at night they moved within their
individual bore holes. The urchins fed upon both endolithic algae and pieces
of algae that apparently drifted into the bore holes.

Sediment Yield

For control rocks (without urchins), a significant correlation (r = 0.968,
p < 0.05) was found between the total inorganic weight of sediment recovered
from buckets and the total surface area of the rocks. Linear regression gave
the formula for this relationship:

Y = 1.133X - 119.71,

where Y = inorganic sediment weight and X = total rock surface area. The
average organic content of sedimentary particles recovered was 23.74% (range :
14.99 - 41.98%). The mean weight of attached macro-organisms on control rocks

154

was 4.6 g dry weight (range = 1.3 - 13.9). The average yield of inorganic
particles produced by infauna of the control rocks was 0.50 + 0.07 mg cm-2 day-1
(X + 1 SD; table 1). This rate of sediment production was equivalent to 5.0 g
m-2-day-1, and for the area of 2521 m2 in which these rocks occurred, the yield
was 12.6 kg day-1. The average dry weight of infauna recovered by acid dissolu-
tion of each control rock was 8.6 g (range = 4.5 - 13.8).. Dominant infauna
included eunicid polychaetes, sipunculids, boring sponges, Lithotrya barnacles,
pelecypods, and other microborers.

For experimental buckets (rock + urchin), the sediment recovered was a
mixture of particles produced by the rock infauna and by the urchin inside its
bore hole. We assumed that the amount of sediment passing through the mesh
from outside sources was the same for control and experimental buckets. Because
our underwater observations showed that visibility was 12-18 m, we further
assumed that the possibility of mud contaminating the buckets was insignificant.
The average organic content of particles recovered was 15.83% (range = 11.75 -
20.30%). The average weight of attached macro-organisms was 6.7 g dry weight
(range = 1.0 - 13.1). To calculate the weight of sediment produced solely by
the urchin, we used the regression formula (above) and subtracted the weight of
sediment made by the infauna of the experimental rock from the total inorganic
weight of sediment recovered. The calculation for each experimental bucket was
made as follows:

[total organic-free wt] - [1.133 x area of experimental
rock - 119.71]
mg urchin-1 day' 1 = x 24.
duration of experiment in hours

These calculations indicated that the average weight of inorganic sediment
produced by adult urchins was 242.2 + 146.0 mg urchin-1 day- (X + 1 SD, table
1). The total urchin population produced 22.4 kg day-1 or an average of 8.9 g
m12 day-1 within the 2521 m2 area containing urchins. Urchin bioerosion was
64% of the total 13.9 g m'2 day-1 measured by our bucket experiments. In
nature, physical processes of abrasion and other bioeroders (fish) would
contribute additional particles.

Particle size analysis showed that the percentages of sand (2.00 - 0.0625
mm) and mud (< 0.0625 mm) were similar for sediment from alimentary tracts of
uncaged urchins and from control and experimental rocks. The alimentary tracts
of the urchins contained mostly pellets (fig. 1C) which readily disintegrated
into sand and mud. No gravel (> 2.00 mm) was found in alimentary tracts of
urchins. Sediment from control and experimental buckets contained 0 - 28.5%
gravel. Sand fractions from alimentary tracts of uncaged urchins contained
particles most abundantly between 0.125 - 0.177 mm (fig. 2, top) and particles
of this size were also the most abundant in sediment recovered from experimental
buckets which contained eolianite and a living urchin (fig. 2, middle). As
determined with a low-power microscope, the grain size of sand particles in the
eolianite is the same as the modal grain size of sand from urchin alimentary
tracts (0.125 - 0.177 mm). However, sediment from urchin alimentary tracts was
50% mud, which greatly exceeds the mud content of the eolianite. These
observations suggest that urchins cause erosion at Black Rock by breaking
individual eolianite grains away from the substrate, yielding sand, and also by

155

Table 1.--Comparison of sediment production rates of infauna (control rocks)
and urchins (experimental rocks)

Control (rock, no urchin)

Date and Total Organic-free Rock Surface Sediment Production
Duration Sediment Recovered Area (cm2) Rate by Infauna
(dry wt, mg) (mg cm-2 day-1)

25-27 August 1982 442.13 428.1 0.57
43.5 hr
286.63 396.5 0.40

500.19 591.6 0.47

13-15 April 1983 689.78 722.6 0.50
45.5 hr
814.86 802.4 0.54

X = 0.50 + 0.07

Experimental (rock + urchin)

25-27 August 1982 1507.63 640.2 497.65
43.5 hr
1031.92 774.5 169.99

1044.32 621.0 254.03

13-15 April 1983 1146.83 902.4 128.77
45.5 hr

21-23 June 1983 1278.27 1119.5 62.83
49.5 hr
1270.90 795.5 237.24

1588.92 756.0 413.13

1355.85 985.9 173.84

X = 242.2 + 146.0

abrading more tightly cemented eolianite grains, yielding mud (fig. iB).
Infaunal bioeroders produced a much broader spectrum of particle sizes (fig. 2,
bottom), perhaps due to the many species of bioeroders active within these
rocks.

Evacuation of Bore Holes

Inspection of eolianite at the bottom of urchin bore holes by low-power
microsr pe and SEM revealed no scratches on the substrate (fig. 1B). Individual

156

URCHINS

*1~~~~~~~~~~~~~~~~~~~~'
URCHINS \
AND
ROCKS

t~~~~l,.u
N
em &

CONTROL ROCKS

.0 1.0 0.5 0125 0.125 0.0625 mm
4- GRAVEL I SAND I MUD -4
Figure 2.--Size-frequency curves for the sand fraction of sediment from alimentary
tracts of uncaged urchins (upper), from urchins and rocks in bucket enclosures
(middle), and from rocks without urchins in bucket enclosures (bottom). Data
for the gravel and mud fractions are not given; scale of the y-axis is identical
for all sediments. 157

Table 2.--Bioerosion by Echinometra lucunter.

Method Substrate Rate of Erosion Author
(g urchin-1 yr-1)

Gut contents Beachrock 18.6 McLean, 1967

Gut contents Algal ridge on 44.0 Ogden, 1977
coral reef

Gut contents Eolianite 128.0 Hunt, 1969

Bucket enclosure Eolianite 88.0 This paper

teeth from the Aristotle's lantern of the urchin had abraded tips; most spines
were sharply pointed, but a few had broken and re-grown tips (fig. 1D, E, F).
It is possible that E. lucunter may use its spines in addition to teeth in
substrate erosion (MZTean, 1967; Hunt, 1969). The rate of spine re-growth may
be so rapid, averaging 0.3 mm day-1 at 20�C (Davies, et al., 1972, fig. 1, p.
874), that damage to spine tips is not ordinarily obseFvir-Te.

Volume of bore holes for individual urchins averaged 72 cm3 (range = 17 -
126). Using a measured density of 2.13 g cm-3 for the eolianite and the average
rate of substrate erosion by the urchins, we estimated that an average of 2.9 yr
was required to excavate a bore hole (range = 0.7 - 10.3 y).

DISCUSSION

Previous studies (McLean, 1967; Hunt, 1969; Ogden, 1977) measured only the
amount of sediment within urchin alimentary tracts and assumed that frequency
of gut turnover was once each day. Disadvantages of the gut contents method
are that 1) gut turnover frequency is not verified, 2) the material in the gut
is not necessarily produced by urchin activity and may have been ingested after
transport to the urchin bore hole from an allochthonous source, and 3) not all
particles produced by urchin tooth and spine movements may be ingested.

The advantages of the bucket enclosure method are that 1) all sediment
produced by the urchin is captured, 2) rates of infaunal sediment production
and the grain size of their product can be determined, and 3) contamination by
sediment from outside sources is minimized. There is, however, the possibility
of loss of some mud-sized particles through screen windows in the buckets, the
exclusion of drifting algal bits, and reduced water flow inside the buckets.

Measurements of bioerosion by E. lucunter from various places are summarized
in table 2; there are too many variiales to resolve the different rates reported.
To place our measurements in perspective, a comparison with other erosion

158

Table 3.--Comparison of erosion agents and rates for Black Rock
(unpublished data of C. M. Hoskin, D. H. Mook, and J. K. Reed)

Process Agent Average Rate of Substrate Erosion
(mg cm-2 yr-1)

E. lucunter
5bucket enclosures) 667

Bioerosion Rock infauna
(bucket enclosures) 183

Intertidal microborers
(tethered limestone blocks) 26

Physical Waves and currents
abrasion (tethered limestone blocks) 5.5

Chemical Intertidal seawater
solution (filter-protected, suspended
limestone blocks) 0-2

processes at Black Rock (table 3) indicates that biological activities are more
significant than physical and chemical processes in the degradation of reef
substrates.

In terms of paleoecological significance, we suggest that the most
preservable part of the bioerosive activity of E. lucunter at Black Rock will
be the band of bore holes. However, with SEM we searched for and did not find
appropriately sized scratch marks on the bottoms of bore holes which might be
expected from urchin teeth. The sediment produced by E. lucunter is initially
pelleted, but the pellets quickly disintegrate. Sand Tom the pellets is most
abundant at sizes between 0.125 - 0.177 mm, but other processes may produce
sand of the same size. Approximately half (by weight) of the disintegrated
pellets is mud with no distinctive properties. Although the urchin population
density was as high as 100 adults m-2, the total amount of echinoderm skeletal
fragments in the sand fraction of nearby shallow water sediment was < 1%; Hine,
et al. (1981, table 1, p. 288) did not include echinoderm skeletons in their
Tsting of carbonate constituents.

Hine, et al. (1981, fig. 17-B, p. 278) have shown by seismic reflection
profiling tFitte southwest margin of the Little Bahama Bank in the vicinity
of Black Rock is a rocky surface, barren of sediment. Our seismic reflection
profiling confirms the seismic findings of Hine, et al. (1981), but from the
JOHNSON-SEA-LINK submersibles we have seen sediments a few centimeters thick
overlying hardgrounds on the western bank margin. Using the geometry of sand
bodies 3-12 km northeast of the bank margin as an indicator, Hine, et al. (1981,
fig. 14, p. 275, and table 1, p. 268) suggested that sand is transported onto
the Little Bahama Bank. Sediment produced by E. lucunter and the associated

159

ror infauna is predominantly fine sand and mud. The amounts of this sediment
art :onsiderable: 9 tons yr-1 for the urchins plus 6.5 tons yr-1 for the
int una at Black Rock. This sediment probably does not accumulate in the
shallow bank margin environment because currents are strong (Hine, et al.,
1981, fig. 24, p. 287).

From the JOHNSON-SEA-LINK submersibles we observed sediment-containing
grooves passing through the bank-edge reefs and have counted 70 grooves in a
400 m long segment of bank edge west of Black Rock. Hubbard, et al. (1981,
fig. 17, p. 26) have illustrated similar sediment-containing grooves off Lucaya,
Grand Bahama Island. We made sediment trap measurements in the qrooves (32 m
depth) west of Black Rock and found an average of 5.4 kg m-1 yrI of carbonate
sediment moving westward. We suggest that this sediment, in part, represents the
products of urchin and rock infauna bioerosion being transported over the bank
edge and into deeper waters of the Northwest Providence Channel.

ACKNOWLEDGMENTS

We thank the officers and crews of R/V JOHNSON and JOHNSON-SEA-LINK I for
their expert assistance at sea; Craig Caddigan, Chris Chulamanis, and Richard
Morris who assisted with SCUBA diving in deployment and recovery of the bucket
experiments; David Mook for Cal Comp plotting; and Marjorie Reaka and two
anonymous reviewers for manuscript editing. This is Contribution No. 390 from
Harbor Branch Foundation, Inc.

LITERATURE CITED

Davies, T. T., M. A. Crenshaw, and B. M. Heatfield. 1972. "The Effect of
Temperature on the Chemistry and Structure of Echinoid Spine Regeneration."
J. Paleontol. 46:874-883.

Hine, A. C., R. J. Wilber, and A. C. Neumann. 1981. "Carbonate Sand Bodies
Along Contrasting Shallow Bank Margins Facing Open Seaways in Northern
Bahamas." Bull. Amer. Assoc. Petrol. Geol. 65:261-290.

Hopley, D. 1982. The Geomorphology of the Great Barrier Reef: Quaternary
Development of Coral Reefs. Wiley-Interscience, New York. 453 pp.

Hubbard, D. K., L. G. Ward, D. M. Fitzgerald, and A. C. Hine. 1976. "Bank
Margin Morphology and Sedimentation, Lucaya, Grand Bahama Island." Report
No. 7-CRD, Coastal Res. Div., Univ. of South Carolina, Columbia, S.C. 36 pp.

Hughes Clark, M. W., and A. J. Keij. 1973. "Organisms as Producers of Carbonate
Sediment and Indicators of Environment in the Southern Persian Gulf," pp. 33-
56. In: B. Purser (ed.), The Persian Gulf. Springer-Verlag, New York.

Hunt, M. 1969. "A Preliminary Investigation of the Habits and Habitat of the
Rock-Boring Urchin Echinometra lucunter near Devonshire Bay, Bermuda," pp.
35-40. In: R. N. Ginsburg and P. Garrett (eds.), Seminar on Organism-
Sediment-Tnterrelationships. Spec. Publ. No. 2, Bermuda BioT. Sta., St.
George's West, Bermuda.

160

McLean, R. F. 1967. "Erosion of Burrows in Beachrock by the Tropical Sea
Urchin Echinometra lucunter." Can. J. Zool. 45:586-588.

Ogden, J. C. 1977. "Carbonate-Sediment Production by Parrot Fish and Sea
Urchins on Caribbean Reefs," pp. 281-288. In: S. H. Frost, M. P. Weiss,
and J. B. Saunders (eds.), Reefs and Relate4-Carbonates - Ecology and
Sedimentology. Stud. Geol. No. 4, amer. Assoc. Petrol. Geol., Tulsa,
Oklahoma.

Reed, J. K., R. H. Gore, L. E. Scotto, and K. A. Wilson. 1982. "Community
Composition, Structure, Areal and Trophic Relationships of Decapods Associated
With Shallow- and Deep-Water Oculina varicosa Coral Reefs." Bull. Mar. Sci.
32:761-786.

Russo, A. R. 1980. "Bioerosion by Two Rock Boring Echinoids (Echinometra
mathaei and Echinostrephus aciculatus) on Enewetak Atoll, Marshall Islands."
J. Mar. Res. 38:99-110.

161

HIGHLY PRODUCTIVE EASTERN CARIBBEAN REEFS: SYNERGISTIC EFFECTS
OF BIOLOGICAL, CHEMICAL, PHYSICAL, AND GEOLOGICAL FACTORS

Walter H. Adey
Marine Systems Laboratory
Smithsonian Institution
Washington, D.C. 20560

and

Robert S. Steneck
Zoology Department and Oceanography Program
University of Maine
Darling Center Marine Laboratory
Walpole, Maine 05473

ABSTRACT

Studies of reef productivity and community structure were conducted on
three bank barrier reefs along the south shore of St. Croix, U.S. Virgin Islands.
Primary productivity (rate of organic carbon fixation) was measured using
upstream/downstream flow respirometry techniques. Mean annual gross primary
productivity of the St. Croix reef ranged between 17-33 g C/m2/d. These are
. ~among the highest values reported for reefs or for any biological community.

Reef productivity corresponds with the distribution and abundance of turf
algae. These algae are minute filamentous forms which are highly diverse,
productive, and abundant. Algal turf productivity is limited primarily by
available solar energy and is influenced by wave surge and water flow (which
enhance metabolite exchange). Turf cover is limited by the total surface area
of standing dead coral. Nitrogen availability is increased by in situ nitrogen
fixing blue-green algae, as evidenced by increased concentrations -577norganic
nitrogen in water flowing over the most productive reef zones. Phosphorus
availability is enhanced by wave surge and maintained by inflowing water from
the open ocean.

Over geological time, reefs grow toward sea level. The three reefs we
studied represent three stages in reef development. Younger reefs to the east
are deeper, with considerable live branching coral (Acropor Ralmaa) and with
a low abundance of turf algae. The reverse is true for te ole reefs to the
west, where live coral is minimal but the dominant algal turfs occupy a rather
smooth pavement of limited surface area. Intermediate between these tw o types
of reefs, the abundance of live coral is low but biomass of algal turf is very
high due to the large surface area left by the dead branching coral. The
greatest reef productivity was measured under the latter conditions because of
the synergistic effects of large surface area with abundant algal turf cover in
shallow, strongly lighted, and turbulent water. It is clear that, given optimum
conditions, coral reef algal turfs can develop primary productivities that
. ~reach and perhaps exceed the accepted theoretical maximum for primary production
is ~in the sea.

163

INTRODUCTION

Coral reefs have attracted interest over the last 30 yr in part because
they are among the most productive biological communities in the world. Early
studies of reef productivity (Sargent and Austin, 1949; Odum and Odum, 1955)
focused on determining the productivity of individual reefs. There is little
agreement in these and subsequent studies as to how productive reefs can be,
and there is even less agreement concerning why high levels of production are
attained or why reefs differ from each other so much (Lewis, 1977; Sournia,
1977; Mann, 1982). Smith (1981) considered some of these problems in an
examination of the Houtman Abrolhos Islands off western Australia. We address
these questions in some detail in our study of the reefs of St. Croix.

In an examination of the "Potential Productivity of the Sea," Ryther
(1959) concluded that (net) "primary productivity of organic matter of some
10-20 g (dry)/m2/d may be expected," with a maximum of 25 g (dry)/m2/d under
ideal conditions of maximum radiation and no nutrient limitation. Expected
gross primary production (GPP) levels of up to 23-38 g (dry)/m2/d (50-80% higher
than the net values) also were cited. Ryther suggested that, while planktonic
open ocean production (particularly in tropical/subtropical areas) is limited
by nutrients, such limitation in benthic communities is not likely because
nutrients "are continually being replenished as the water moves over the plants
[and] probably prevents their ever being [nutrient] limited."

Recent reviews of coral reef primary productivity by Lewis (1977) and
Sournia (1977), integrating 30 yr of research, reveal a range in GPP of nearly
an order of magnitude (3.4 - 20 g C/m2/d). The mean of these cited reef values
is about 10 g C/m2/d, but only a few points are above 13 g C/m2/d. The mode is
7 g C/m2/d, which, using a typical conversion ratio, would convert to a net biomass
production of about 15 g (dry)/m2/d. Overall, these values are much less than
the theoretical maximum suggested by Ryther (1959). Significantly, all but one
of the 17 coral 'reefs cited by these authors were located in the Indo-Pacific.

Despite Ryther's (1959) suggestion to the contrary, nutrient limitation of
reef productivity has been the focus of numerous studies. Many of these have
focused on explanations of how highly productive reef ecosystems can thrive in
nutrient deserts (Odum and Odum, 1955; Stoddart, 1969; Johannes, et al., 1972;
D'Elia, et al., 1981; Andrews and Muller, 1983). Recently, the roTe-o major
nutrients was questioned due to the finding that benthic algae have lower
nutrient requirements than once thought (Atkinson and Smith, 1983). Also, some
reefs are known to have a surplus of nitrogen (Webb, et al., 1975; Wiebe, et
al., 1975) due to the in situ nitrogen fixation by cyanobacteria (blue-green
algae) (Mague and Holm-Hansen, 1975; Capone, et al., 1977; Hatcher and Hatcher,
1981; Wilkinson and Sammarco, 1983). The benthic algal turf of reef communities
usually is dominated by cyanobacteria (Sammarco, 1983; Van den Hoek, et al.,
1975).

Phosphorus concentrations over reefs generally have been thought to change
little (Pilson and Betzer, 1973), and some authors have suggested that a "tight
recycling" or retention of phosphorus occurs (Pomeroy, et al., 1974; Pomeroy
and Kuenzler, 1969). Atkinson (1982), on the other hand, ha-s shown reactive
phosphate depletion over broad reef flats.

164

Our study examines the productivity of three reef environments in tropical,
oligotrophic seas. Large-scale temporal and spatial variations in their primary
productivity with respect to potentially important biological, chemical,
physical, and geological differences between the reefs are explored.

MATERIALS, METHODS, AND STUDY SITES

Our study was conducted on the southeastern bank barrier reef of St. Croix
from October 1977 to November 1978. It extends 35 km along most of the windward
shore of the island and is the most extensive reef system on the Puerto Rican-
Virgin Island shelf. St. Croix is a small island which is unaffected by
neighboring islands. Temperature, salinity, and wind direction and intensity
are all remarkably constant. The south shore is relatively unpopulated, and it
has no natural rivers affecting salinity or ambient nutrients. The south shore
of St. Croix is ideal for examining the factors that affect reef productivity.
The two major factors that change from east-to west are water motion (decreasing)
and geological reef development (increasing).

The western reefs of St. Croix lie on a shallower limestone basement than
do the eastern reefs; they were the first to grow to sea level (Adey, 1975;
1978a), and they developed broader, more continuous reef flats. Several of
these reefs have become so shallow that the abundance of live coral is reduced
due to the negative effects of intense sunlight (including ultraviolet), tempera-
ture fluctuations, and desiccation during periods of extraordinarily low tides.

We studied three reefs on the south shore of St. Croix intensely and
several others to a lesser extent (fig. 1). The following is a description of
the morphology of the reefs from east to west. Isaac Bay Reef is a "young"
reef with a high coral cover, rapid reef growth rates, and a crest that is just
reaching sea level today. Mean depths are 6.3 m for the fore reef and 0.9 m for
the back reef. The reef structure is open (many channels and breaks in the reef
crest), and there is no continuous reef flat. Robin Bay Reef is more mature,
with a nearly continuous crest and a moderately deep and broad reef flat. Mean
depths are 5.1 m for the fore reef and 0.7 m for the back reef. Halfpenny Bay
Reef is mature, with a continuous, broad, and shallow reef flat. Mean depths
are 5.1 m for the fore reef and 0.3 m for the back reef. We selected these
reefs because they fall along a gradient of reef maturity and because they are
particularly well suited for the upstream/downstream method. All of these reefs
have a continuous unidirection water flow inward across the reef to the lagoon.

A profile of Robin Reef along a permanent transect line established at the
beginning of our study is shown in figure 2. Comparisons with Isaacs and
Halfpenny Reefs are give~n in the accompanying table. The transects were parallel
to the predominant current flow and were established using drogues, to indicate
the exact current flow. We omitted breaker zones due to logistic and scientific
problems created by the turbulence (accelerated diffusion). The outer (seaward)
limit to the fore reef was defined by the first appearance of the dominant reef-
building coral, Acrpora pamaa The inner fore reef and outer back reef zone
were adjacent to te-F-are zoine. The inner (landward) limit of the back reef
was defined by the abrupt transition between the reef and sandy lagoon. Our fore
and back reef zones are consistent with other published studies (Adey, 1975;
Adey and Burke, 1976).

165

nautical miles 1

Airport _ Holocene reef to
~~~~~~~~kilo~ 2meters thick
Island

Tague Bay

ST. CROIX
ISAAC

ROBIN

HALFPENNY
ChAioel LEGEND
O#A irport Holocene reef to
25 meters thick

Figure I.--Locations of transects across bank barrier reefs studied on the south shore of St. Croix,
U.S. Virgin Islands, northeastern Caribbean.

BACK
4 REEF 3 CREST 2 FORE REEF

0-4-A.KAX

4-x
4- REZONES
_ - SACREEF 14-313-2 1 2-1
6- ISAAC |14811251287
I ROBIN 1471701203
fi 8- HALFPENNY 1621951138 I&%

10- oI III

50 METER INCREMENTS

Figure 2.--Typical reef transects (Robin Reef).

Diurnal measurements of changes in oxygen concentration and water flow
were taken in all seasons of the year and were used to derive rates of community
oxygen production (which reflect net or apparent productivity) and consumption
(respiration). Accurate measurements of discharge rates (rate of water volume
flow) are crucial to this upstream/downstream method. For this, we used a
Marsh-McBirney model 527 electromagnetic, x-y current meter. .Water samples
were taken at each transect every 2 hr over a 2- to 3-day period during each
season. Particular care was taken to maximize the level of precision at every
step of analysis (see Adey, et al., 1981). A plot of a typical diurnal cycle
of oxygen concentration at the Robin Bay site is shown in figure 3. Standard
methods of calculating production in upstream-downstream methods were employed
(Marsh and Smith, 1978). Computer plotting and integration of the area under
the curve above the zero oxygen production level provided net daytime primary
productivity values. The area below the zero line provided respiration values.

The biotic composition (percent cover) and surface area of these reefs
were determined with chain transects (methods detailed in Rogers, et al.,
1982). We measured all surfaces along each transect that we could reach, except
for truly cryptic habitats (nearly devoid of light). This method also gave a
measure of surface area per projected square meter of bottom Cm /m , called a
surface area ratio (SAR)].

To examine community structure of algal covered surfaces, random Acropora
p almata substrate samples were collected from both fore and back reef zones.
Algal cover and distribution on the surface were mapped using grids with I cm
squares placed over the top and bottom surfaces. Algal turfs were subsampled
by scraping from 3-12 visually representative, but distinct, 1 cm2 microquadrats.
Each sample included the upper millimeter of calcium carbonate in order to
sample endolithic, crustose, and tightly adherent algae. Solutions of 5% HC1
and 5% formalin were used to decalcify and fix the samples. A representative
subsample of the scraped algal turf samples was spread in a monolayer over a
microscope slide. Species composition of algal turf communities was quantified
using a "point count" technique. This involved counting the number of times
identifiable algae crossed points of intersection on an ocular grid placed
within a compound microscope.

At each of the reef study sites, water samples were collected and analyzed
for nutrients at the beginning and end of each field session, between 1000 and
1400 hr. Standard precision handling techniques were employed at every stage,
and analyses for nitrate, nitrite, phosphate, and total dissolved phosphorus
were accomplished with a Beckman DU spectrophotometer. Extensive descriptions
of methods and data taken during the course of this study appear in Adey, et
al. (1981).

RESULTS AND DISCUSSION

Reef Productivity

The reefs studied were highly productive. Community metabolism, as shown
by mean diel oxygen production at each station, and seasonal metabolism at all
stations is given in figures 4 and 5. Gross primary productivity (GPP), net

168

DISSOLVED OXYGEN

8.6 . .*� backreef
8.2 0 forereef

-o ,
7.8'O~.
.4

6.2.

5.4
o~~~ ., \\ . . ,. . ., ,. . , .

5.0
1200 2400 1200 2400 1200 2400
TIME OF DAY

Figure 3.--Dissolved oxygen concentration across Robin Reef from 1/21/78 to 1/23/78.
Numbers on individual plots refer to standard reef sites.

// AAC ,

2 -

41
s /I

0000 0400 0800 1200 1600 2000 2400

TIME OF DAY

Figure 4.--Mean yearly diurnal oxygen exchange for the reef transects.
/ / %~. / / ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~�4

6- MAY '
A- MAYUGUST

,>ru'~~~ OCTOBER
E
o .....

z . .,_JANUARY

-2- --- ---
* �"/

U'
,./...i/ //S"'JANUARY ' '....

-2-
S/ . .. ,.-..

0000 0400 0800 1200 1600 2000 2400

TIME OF DAY

Figure 5.--Mean seasonal diurnal oxygen exchange: summary of all transects.

Table 1.--Summaries of daily metabolism.

Gross Primary Net Community
Productivity Primary Productivity' Respiration (-)

(9gO2/m2/day) (g02/m2/day) (gO2/m2/hr)

Fore Back Means Fore Means Back Means Fore Back

Jan. 10.02 36.5 6.1 23.0 0.4 1.4

e May 17.4 48.9 9.3 32.7 0.8 1.6
46.7 8.4 27.5
"Aug. 12.4 57.6 6.6 29.1 0.6 2.8

Oct. 13.8 43.7 11.5 25.2 0.2 1.8

Jan. 14.3 56.2 9.2 28.7 0.5 2.8

May 15.2 109.0 10.6 72.4 0.5 3.7
.~ 88.4 10.5 55.7
- Aug. 18.3 115.3 15.0 78.9 0.3 3.6

Oct. 14.6 72.9 7.2 41.0 0.7 3.2

Jan. 17.3 40.2 10.9 22.0 0.6 1.8

' May 20.4 59.8 8.1 32.0 1.2 2.8
tc 50.9 7.9 26.3
r Aug. 17.6 64.7 3.6 32.7 1.4 3.2

Oct. 17.5 38.8 9.2 18.5 0.8 2.02

daytime primary productivity (NPP), and respiration (R) for each reef at each
season appear in table 1. Back reefs are more productive than fore reefs. A
mean annual GPP of about 17.5 g C/m2/d is recorded at Isaac Reef, 19 g C/m2/d at
Halfpenny Reef, and 33 g C/m2/d at Robin Reef (with a mean maximum value of 43.2
g C/ml/d recorded in August and a mean all-reef productivity of 8500 g C/m2/yr).
These values are based on a conversion ratio of 0.375 g C/g902 with a metabolic
quotient for reefs of 1.1 (Marsh and Smith, 1978) applied to the means in table

172

1. They are derived from the molecular weight difference and the ratio of
oxygen to carbon dioxide exchange (the metabolic quotient) for reefs (Marsh and
Smith, 1978). Higher GPP have been reported (Helfrich and Townsley, 1963;
Connor and Adey, 1977), but such high levels of productivity have been subject
to question. Because of this, great care was taken in data collection, and
every potentially error-producing factor has been carefuly considered (see
Adey, et al., 1981). The patterns we observed are not surprising once the
processes affecting reef productivity are examined.

Processes Affecting Reef Productivity

Rates of carbon fixation on reefs reflect a complex interaction of
environmental factors. Slight variations in some of these factors (water depth,
biotic composition, substrate availability, or solar radiation) will affect
these values. In the remaining sections of this paper, we will describe the
results of our studies of biological, chemical, physical, and geological factors
as they affect overall reef productivity.

Biological Components: The Role of Algal Turfs

Benthic algae, as both free-living forms and endosymbionts in coral and
other invertebrates, are the primary producers on reefs. The focus of our
study was on free-living benthic algae because it was the most abundant biological
component of the reefs we studied (see table 2). Large-scale and small-scale
data were integrated to determine community structure. Large-scale cover was
based on chain transects and is presented as a measure of reef complexity (SAR)
for the fore and back reef of each transect in table 2. Small-scale cover is
based on point counts of prepared microscope slides (table 3).

To test the hypothesis that algal turfs are the most important primary
producers of the St. Croix reefs, we compared rates of turf biomass production
with rates of flow respiratory production. To make such a comparison, several
conversions are necessary to obtain common units of measure. All conversions
involve values from either table 1 or 2, or from published literature for a
similar environment. Values for each back reef will be reported in east to
west order (i.e., Isaac, Robin, and Halfpenny Reefs, respectively). Algal NPP
was calculated from community GPP values (table 1) by multiplying by 0.24,
which is the mean NPP/GPP ratio derived from chamber studies of algae by Wanders
and Wanders-Faber (1974) and Rogers and Salesky (1981). ~he resulting algal
NPP values for back reefs are 11.2, 21.2, and 12.2 g 02/m /d. To equate our
projected m2 of a complex reef to a surface rather than a projected area, we
divided by the SaR (table 2). This gave values of 4.1, 7.1, and 9.9
g 02/(surface)/m /d. Assuming that all productivity was from algal turfs, we
divided the surface production values by the proportion of turf in the back
reef (table 2) (values of 15.8, 19.7, and 18.5 g 02/(planar) m /d. Oxygen
production was converted to carbon production by multiplying by the difference
in molecular weight (0.375) and dividing by the photosynthetic quotient for
reefs (1.1) (Smith and Marsh, 1978), giving values of 5.4, 6.7, and 6.3
g C/(surface) m2/d. The conversion of carbon to organic matter involves the
ratio of 2.2 dry wt/g C (Westlake, 1963, and confirmed by our carbon analysis of

173

Table 2.--Mean depth, major benthic components (% cover), and surface area
ratio on chain transects.

Surface
Mean Acropora Other Algal Macro- # Area
Depth palmata Corals Millepora Corallines Turf Algae Sand (m) Ratio

Fore Reefs

Isaacs 6.3 18.9 7.2 5.1 8.4 40.1 10.1 8.2 16 1.87

Robin 5.1 16.2 1.5 0 8.2 65.1 0 0 13 2.36

Halfpenny 4.5 10.3 9.2 11.2 16.2 9.9 8.6 35.0 14 2.51

Back Reefs

Isaacs 0.9 65.9 0.8 0 1.9 25.9 5.7 0 6 2.70
Robin 0.7 19.7 0 0.6 6.5 36.1 10.1 26.7 10 2.98
Halfpenny 0.3 0 8.2 0 0 53.5 8.4 29.9 25 1.23

l
St. Croix turfs). This means that the rates of organic matter production of
10.0, 14.7, and 13.9 g (dry)/m2/d on actual reef surfaces (rather than projected
area) are necessary to produce the rate of oxygen production we observed in the
water flowing over reefs.

Tests of these theoretical rates of biomass production were supported in a
recent study in similar lagoon and back reef environments of Mayaguana in the
Bahamas (Adey and Goertemiller, in ms.). By harvesting algal turfs grown on
screens, actual biomass production was found to be 6-18 g (dry)/m2/d. A mean
rate of 12 g (dry)/m2/d, under a nutrient regime of 0.10-0.13 microgram-at N/1
(NO2 + NO3) (D'Elia, in prep.), was achieved. A similar and more recent study
at Grand Turk, Turk, and Caicos Islands (Peyton, et al., in prep.) gave rates
of algal turf production of up to 31.8 g (dry)/m2d/d tnutrient concentrations
(as measured by NO2 + NO3 nitrogen) of less than 0.2 microgram-at N/1. Steneck
and Porter (in prep.) used a similar harvest technique for turfs growing on
slabs of coral substrata at a depth of 10 m and found a production rate of 6.0
g (dry)/m2/d. These harvest production rates are minimal since they do not
include losses to micro-herbivores, abrasion, leaking of dissolved organics, or
the release of reproductive structures. Thus, the values we obtained using
upstream/downstream flow respirometry correspond well with measured rates of
biomass produced by this "functional group" of algae (sensu Steneck and Watling,
1982) in the Caribbean. This also supports the pattern of high rates of
production (per unit biomass) for turf algae that have a high surface area to
volume ratio (Odum, et al., 1958; Littler, et al., 1983) and, perhaps equally
important, live in a strong wave surge environment.

174

Table 3.--Dominant algal turf genera and abundance as determined by point counts
(% mean abundance top and bottom of samples). Data were generated at the
species level and are in Adey, et al. (1981).

Mean Isaac Robin Halfpenny
Non-Heterocysted Blue Greens 10.7 9.5 12.7 10.1

Heterocysted Blue Greens 7.3 14.6 5.1 2.2

Sphacelaria spp. 7.2 7.3 11.8 2.7

+- Endolithic algae 7.1 7.1 5.1 9.1

Polysiphonia spp. 3.9 2.6 3.1 6.0

�_ Taenioma sp. 3.8 0.7 2.1 1.0

Herposiphonia spp. 4.4 2.1 3.6 7.7

Lobophora sp. 2.7 3.7 4.4 0

Jania spp. 2.5 2.4 4.2 1.0

Non-Heterocysted Blue Greens 18.9 18.0 20.0 18.5

Jania spp. 10.8 8.0 18.7 5.8

Amphiroa spp. 10.3 13.4 14.5 3.2

Endolithic algae 3.7 2.7 0 8.5

Gelidiopsis sp. 2.1 3.8 2.0 0.4

Asparagopsis (Falkenbergia 1.6 0.6 3.9 0.2
stage)
Heterocysted Blue Greens 1.5 3.7 0.3 0.5

Sphacelaria spp. 0.8 2.2 0.4 0

We found over 100 species of algae (mostly in turfs) in the course of this
study. Of these, only 30-50 species were common throughout the year, and fewer
than 10 were common at any site in a given season (table 3). The most
consistent elements on (and in) the upper surfaces of dead coral substrata were
cyanobacteria, boring algae, and filamentous red and brown algae. A high

175

diversity of algae in low abundance was recorded on virtually every specimen
regardless of season, reef, or zone.

Cyanobacteria of 6-8 species formed a large part of the turf biomass on
both fore and back reefs. Heterocysted species (particularly Calothrix spp.)
are far more abundant in fore reefs than back reefs, but nonheterocysted species
(e.g., Oscillatoria) were most abundant overall. Of eukaryotic filamentous
algae, the brown (Sphacelaria spp.) and red (Polysiphonia spp., Herposiphonia
spp., and Taenioma macrourum) species dominate fore reef substrates. Boring
filaments (usually green algae, e.g., Ostreobium) were also ubiquitous. The
predominant macroalgae are calcareous, articulated red algae (e.g., Jania
capillacea and Amphiroa fragilissima) which are found on bottom surfaces in the
fore reef and on both top and bottom surfaces in back reefs. Crustose corallines
(Porolithon pachydermum and Neogoniolithon megacarpum) were most abundant in
the shallow fore reefs, particularly where grazing is intense. Of the fleshy
macroalgae, Lobophora variegata and Acanthophora spicifera were found but they
were not abundant compared to the turf community. Minute macrophytes such as
Gelidiella trinitatensis and Gelidium pusillum commonly were found in the turfs
of the back reef.

As shown in table 3, the differences between fore reef and back reef turf
community structure along the St. Croix reef track are rather consistent.
These numbers and the community structure of the turf species we found agree
with that found on a reef in Curacao by Van den Hoek, et al. (1975). With the
exception of Lobophora variegata (which does not appear in Van den Hoek, et
al., 1975, for Curacao) and Dictyota dichotoma (which was common in Curacao but
not in St. Croix), the floras for Curacao and St. Croix are quite similar. There
are also floristic similarities for the same reef zones between the Caribbean
and Indo-Pacific at the genus level (Hatcher and Larkum, 1983; Sammarco, 1983).

It is beyond the scope of this study to examine the myriad of factors
maintaining turf algal community structure. Most studies indicate that intensive
herbivory in productive environments is important for maintaining a turf
community (Borowitzka, 1981; Carpenter, 1981; Littler, et al., 1983; Sammarco,
1983; Hatcher and Larkum, 1983).

Primary Production: Not Nutrient Limited

High primary productivity on reefs growing in nutrient poor water has
received considerable scientific attention (reviewed by Mann, 1982). The
resulting research has focused on the role of cyanobacteria as in situ nitrogen
fixing organisms (Mague and Holm-Hansen, 1975; Capone, et al., IT77T. Recently,
Sammarco (1983) demonstrated that fish grazing on the Great--arrier Reef can
cause a shift in the community structure of turf algae from filamentous reds
(e.g., Polysiphonia) to minute filamentous cyanobacteria. Other studies have
demonstrated that, as a result of this change in community structure, nitrogen
fixation is significantly enhanced (Wilkinson and Sammarco, 1983). Our results
support their conclusions.

Heterocysted (nostocacean) and nonheterocysted (hormogonalean) cyanobacteria m
are abundant in turfs growing on the upper surfaces of all reefs that we

176

Table 4.--Nutrient levels. Each value is a mean of two replicates taken on
different days over approximately a 10-day period.

Nitrate + Nitrite (ug-at/1 - N as NO3 = NO2)

Yearly
Jan May Aug Oct Mean

Isaac Fore 0.386 X 0.345 X 0.366
Back 0.980 0.302 0.391 .772 0.611

Robin Fore 0.294 0.175 0.549 .075 0.273
Back 1.016 0.328 0.589 .220 0.538

Halfpenny Fore 0.340 X 0.823 .107 0.423
Back 0.650 0.629 0.816 .099 0.549

Channel Fore 0.457 .150
Back 0.686 .130

Airport Fore 0.223 0.228 X .050 0.167
Back 0.466 0.350 X .130 0.315

Grand Mean
Means Fore 0.311 0.,134 .572 0.096 0.283
Back 0.778 0.401 .599 0.270 0.512

Phosphate (ug-at/l - P)

Isaacs Fore 0.024 0.118 0.032 0.058
Back 0.046 0.020 0.038 0.035

Robin Fore 0.010 0.055 0.187 0.084
Back 0.037 X 0.047 0.042

Halfpenny Fore 0.010 0.275 0.049 0.111
Back 0.034 0.392 0.161 0.196

Channel Fore 0.080
Back 0.030

Airport Fore 0.010 0.145 0.052 0.138
Back 0.017 0.165 0.143 0.108

Means Fore 0.013 0.148 0.080
Back 0.033 0.192 0.084

177

studied, but particularly on fore reef and sand substratum habitats of Isaac
Reef. Results of nutrient analyses for nitrate plus nitrite and phosphate are
listed in table 4. The nitrite and nitrate concentration strongly increases in
the water as it flows over the three reefs studied. Ten of the twelve sets of
measurements showed a highly significant yearly increase of 0.35-0.57 microgram-
at/1 across the reef. With a mean yearly flow rate of 6.83 m3/min/m for all
three transects (discussed below), this translates to a net "fertilization
rate" of 0.82 kg N/hectare/d (or 0.08 g N/m2/d). While this is roughly half the
rate estimated by Wieb, et al. (1975) for Eniwetak, the maximum rates we
encountered on the south shore reef of St. Croix were three times the mean.

We assume that this fertilization is due to nitrogen fixation by cyanobacteria
(Wilkinson and Sammarco, 1983), but other possibilities exist. Regeneration of
nutrients from the internal cavity of a reef was demonstrated by Andrews and
Muller (1983). However, this is unlikely to occur on St. Croix reefs, since
the coral of these reefs grows on a pavement, below which the reef structure is
remarkably tight, with cavities usually sand filled (Adey, 1975; Adey and Burke,
1976). Empty cavities for nutrient regeneration are rare in St. Croix, which
is arid, devoid of rivers, and lacking in upwelling aquifers. Thus, nutrient
enrichment from ground water, such as that described in Jamaica (D'Elia, et
al., 1981) is unlikely. Nitrogen fixing algae may achieve a competitive iavan-
tage in more oligotrophic waters, and thus nutrient availability probably has
some significance in determining community structure. Nevertheless, the
fact that this "excess" nitrogen escapes from the reef and flows over the highly
productive back reef suggests that primary production on the reef proper is
not nitrogen limited.

Phosphorus is more problematic, since an atmospheric source is not possible.
Reactive phosphorus as phosphate can be difficult to measure and, as briefly
discussed above, older work did not generally show any consistent pattern of
change in reef environments. A recent paper by Atkinson (1982) does show
significant upstream-downstream depletion of phosphate across 1000 m of reef
flat in Hawaii. However, in our St. Croix situation, with a mean incoming
concentration of 0.084 microgram-at/I and a mean current flow of 6.83 m3/min/m,
phosphorus as phosphate is delivered to a meter width transect of reef at a
rate of about 25 g/d. Even if all net primary productivity were to be exported
from these reefs, using a C/P of 430 (from Atkinson and Smith, 1983, for turfs),
only about 6.4 g P/d/meter width would be removed. Thus, phosphorus is present
in large quantities, though at low concentrations, from the constantly inflowing
waters of the North Equatorial Current as it impinges on St. Croix.

Epibenthic Water Flow and Discharge Rates

Nutrient availability, gas exchange, and metabolite release for aquatic
primary producers is facilitated by water flow past algal thalli. Water flow
over reef substrata (epibenthic water flow) is dependent upon the discharge
rate (movement rate per volume of water), water depth, and on the wave surge or
oscillation rate of the overflowing water. In St. Croix, water movement over
the reefs results primarily from wind-driven wave energy. Thus, measured rates
of epibenthic water flow have geologic (depth) and geographic (wind) components.

Epibenthic flow rates over reefs are variable in space and time. These
variations can affect both the productivity of the reef and the measurements of

178

it. For these reasons, great care was taken in determining water flow and
discharge rates (this information is provided in detail by Adey, et al., 1981).
Our measurements of water motion demonstrate that water flowing over the reefs
oscillates but has a net shoreward flow. Typically, we recorded a landward
surge of 6-8 s duration (reaching speeds of about 0.3-0.6 m/s) and a reverse surge
lasting 2-4 s (reaching 0.05-0.3 m/s). Net epibenthic water flow was determined
from chart recordings of mean maximum currents in both directions (i.e., the
peaks of spikes) integrated over 1 hr intervals from Marsh-McBirney current meter
data. Reversing wave surges and accompanying turbulence produced by waves breaking
over irregular reef surfaces mix the water flowing over the reef, particularly
for back reef habitats where measured rates of production are the greatest. The
turbulent effect of wave surges also significantly increases metabolite exchange
and resulting photosynthesis and growth (Hackney, et al., in ms.).

Epibenthic flow rates increase dramatically from fore to back reef. Means for
all reef transects were 1.2 + 0.066 m/min for fore reefs and 11.0 + 3.5 m/min for
back reefs. This increased Tlow rate results from the progressive-shallowing from
fore reef (overall mean depth of 5.3 m) to back reef (overall mean depth of 0.6 m).
With a constant discharge rate, flow rate varies with depth. Since depth changes
daily with tidal changes, we recorded tidal heights during each sampling period.

Discharge rates are not constant because wind-induced water movements change
geographically along the south shore of St. Croix due to the island's effect of
diminishing the strength of easterly trade winds. Specifically, in the fore
reef environments, Isaacs Reef to the east has the highest discharge rate (mean
8.1 m3/min/m, range 2.5-12.1 m3/min/m, N = 65 1 hr intervals). Robin Reef is
intermediate (mean 6.2 m3/min/m, range 2.8-9.8 m3/min/m, N = 86 1 hr intervals),
and Halfpenny Reef is the lowest (mean 5.3 m3/min/m, range 1.4-9.8 m3/min/m, N = 79
1 hr intervals). These discharge rates created mean back reef epibenthic flow
rates of 9.2, 8.8, and 15 m/min/m, respectively. Although a geological pattern
in exposure from east (most exposed) to west (least exposed) results in a
similar pattern of decreased discharge rates, the magnitude is much less than
the difference observed between back reef and fore reef at any one location.

There is little support for the hypothesis that reef productivity is
controlled by water motion even though differences in flow rates between fore
and back reefs roughly correspond with differences in rates of production. If
reef productivity was primarily dependent upon water flow, diurnal reef produc-
tivity curves (figs. 4 and 5) would level off at the point where light saturation
occurred rather than corresponding closely with variations in light intensity
(discussed below). Also, for any given reef, productivity would correspond
with data on water currents, which we did not observe. Water motion is an
important factor, but the south shore of St. Croix is characterized by consis-
tently high wind-induced wave energy year round (see Adey, 1978b); thus, water
motion does not appear to limit productivity. Note that there Ts no relationship
between reef productivity (table 1) and epibenthic flow rates.

Light

Primary production on the reefs we studied is most likely controlled by
light. Evidence for this comes from the correspondence between diurnal and
seasonal cycles in light intensity (fig. 6) and measured rates of primary

179

MEAN SOLAR RADIATION-ST. CROIX

2 -

500 -
z
0
i- -

x~~~

X
400 -
z
4~~~~~

/
x~~~~~~~~~~~~

300 ' ' ' ' I ' ' X
Dec Jan Feb Mar Apr May June Jul Aug Sept Oct Nov
(1978)

Figure 6.--Monthly means of daily solar radiation.

production (figs. 3, 4, and 5). If any other environmental factor was
limiting, the rate of oxygen production would level off, indicating light
saturation. In contrast to our field measurements, most studies using enclosures,
even with stirring, commonly report light saturation. This is likely an artifact
of insufficient water motion. In more recent work with chambers (e.g., Carpenter,
in ms.), where efforts have been made to simulate surge motion, this lack of
light saturation for algal turfs is repeatedly demonstrated.

In this investigation, daily solar radiation levels were recorded with a
pyroheliometer (see fig. 6 for monthly mean values). During our study we
sampled two periods of high light intensity (May and August) and two periods of
lower light intensity (January and October). Averaging hourly production rates

180

100

w~~~~~~

I-
~' 60.

=c7d
so
0 E
L mean IS

0 20-
0
()annual mean- individual rest
a seasonal mean -all reefs

0 200 400 600 800 1000

TOTAL ALGAL BIOMASS (AREA x SAR. x BIOM.1
[gm dry/M2 projectedl

Figure 7.--Back reef: Gross primary productivity as a function of total algal
biomass and monthly solar radiation

for all reefs reveals a strong seasonal pattern in productivity corresponding
to these seasonal differences in light intensity (fig. 5). Although the seasonal
decrease in primary productivity roughly corresponds to a 40% reduction in
light during that period, the relationship is not exact, since the GPP in October
is reduced to only 30% of that value in May. A better correspondence exists
between the seasonality of primary production and the combination of algal
biomass and solar radiation. Loss of light in autumn is exacerbated on St.
Croix by a higher frequency of rainfall and cloudiness. Thus, incoming light
in October and November is considerably below that of December and January
(fig. 6).

The GPP data, when plotted directly against total algal biomass (per
projected m2, fig. 7), suggests that a weak relationship between productivity
and total biomass exists on these reefs (note that this is not surface biomass
alone, but includes the effects of SAR and percent cover of algae). Also, the

181

plot of biomass as a function of seasonal productivity (line with closed circles,
fig. 7) indicates a cyclic pattern. The sharp reduction in algal biomass
recorded in the fall of 1978 (45% of that recorded in May) cannot be attributed
to wave action, since that is generally a quiet season and no unusual conditions
were experienced during the period of study. Data for nutrients (table 4) and
suspended matter (Adey, et al., 1981) suggest very minimal effects of runoff on
the reefs. We observed, uETid not quantify, changes in the abundance of
micrograzers (amphipods and worms). It is possible that during the spring and
summer algal biomass and the resident micrograzers increase with higher levels
of solar radiation. As light levels fall in autumn, grazing by the resident
micrograzers exceeds rates of algal growth. Such overgrazing could result in a
"crash" in both algal biomass and productivity.

Geological Control of Reef Productivity: Synergy of Multiple Factors

It is to be expected that those reefs which are less complex (lower SAR),
deeper, with lower epibenthic water currents, and have a larger percentage of
living stony coral (relative to algae) will be less productive than the shallow,
complex, algae-dominated reefs of the south shore of St. Croix. All of these
factors, however, are ultimately the result of the reef's geological history,
and thus the relationship between reef morphogenesis and primary production
should be considered.

The south shore bank barrier reef system of St. Croix "matured" (i.e.,
reached sea level) from west to east (see Adey, 1975; Adey and Burke, 1976, for
a summary of the geological history of the reefs of St. Croix). Isaac Reef is
relatively young and is just reaching sea level today. Its live Acropora
palmata cover is high (table 2), but its reef flats are narrow and the crest is
broken (forming numerous channels). Robin Reef is more mature, as evidenced by
its nearly continuous crest and wide, relatively shallow back reef. While
living Acropora palmata is still an important constituent, the percent cover of
A. palmata on Robin back reef is less than a third of that found on Isaac back
reef (table 2). Halfpenny Reef is older and has a continuous crest and a very
shallow reef flat with very few living acroporids. The age difference between
Isaacs and Halfpenny Reefs is about 500-1,000 yr.

Productivity on the young Isaac Reef is low, even though its SAR is quite
high. This is because coral is less productive than algae per unit of surface
area (Rogers and Salesky, 1981). At the other extreme, coral cover is low at
Halfpenny Reef, and this shallow reef flat is dominated by turf algae. However,
this reef is nearly planar (the SAR is half that of the other reefs), and thus
the surface area available for turf photosynthesis is proportionately reduced.

Because carnivorous predators are so abundant on reefs, herbivores do not
stray far from their refuge in the spaces within the reef (e.g., Ogden, et al.,
1973; Talbot and Goldman, 1972). Thus, reefs with a high SAR are more h-eav-T
grazed because they have more habitat space for reef-dwelling herbivores. In
this way, geological events that contribute to a high SAR on reefs also maintain
herbivore populations that intensely graze the dead coral substrate, thereby
maintaining the turf communities.

182

Light levels, depth, current velocity, and abundance of live Acropora
palmata all increase with geological time as a reef grows toward sea level.
As it reaches the surface, several of these trends reverse (i.e., coral dies
and water flow decreases). Aspects of this complex relationship between reef
morphogenesis and primary productivity can be seen along the south shore of St.
Croix. For example, productivity was greatest for the mature Robin Reef and
lower on the younger Isaacs and older Halfpenny Reefs (fig. 7). On Robin Reef,
the SAR of primarily dead Acropora palmata is very high and algal cover is
about 80% greater than that of Halfpenny. Thus, peak productivity probably
occurs soon after the transition from a live, coral-dominated reef to one of
equal surface area but dominated by algae. With time, productivity declines as
destructive forces (e.g., bioeroding sea urchins or severe storms) reduce the
standing dead coral to a pavement-like algal-dominated flat. This pattern of
reef succession is characteristic of many reef systems. Adey, et al.
(1977) described a similar history for the now largely planar aT-gaT dominated
pavement in eastern Martinique. Over the past 500 yr the open Acropora palmata
reef matured and lost its structure, possibly due to a catastrophi event suc
as a hurricane. It has remained an algal pavement, devoid of coral, ever
since. In the case of Martinique, excess sediment and nutrient runoff from
the island following the introduction of intense sugar cane farming in the 17th
century undoubtedly was a major factor in transforming these reefs into macroalgal
pavements. After the completion of this study in 1978, a similar event occurred
at the south shore of St. Croix when a hurricane passed off the coast in 1979.
Rogers, et al. (1982) documented the event and showed a significant reduction
in SAR and live coral following the hurricane.

Differences in reef morphogenesis may be responsible for some of the
reported differences in reef productivity. Holocene sea level characteristics
for the Indo-Pacific differs from that of the Caribbean. Sea level rise stopped
over 5,000 yr ago in the Indo-Pacific (reviewed by Adey, 1978a), resulting in the
characteristic old reefs of the region having emergent, broad reef flats (many
km wide), often with no epibenthic water flow during low tide and with a low SAR.
This extreme "old age" condition of Indo-Pacific reef flats, developed under
stable or dropping sea levels, contrasts with conditions in the Caribbean, where
sea level has been slowly rising until the past several centuries (e.g., Adey,
et al., 1978a). This geological characteristic of reefs could be responsible
for the loweF levels of primary production reported for the Indo-Pacific compared
to results presented here for the St. Croix reefs.

CONCLUSIONS

Fundamentally, reef productivity is controlled by its geology (reef growth
and morphogenesis). Coral reefs grow slowly towards sea level over time,
thereby changing their physical, chemical, and biological environment. Primary
productivity is maximal at the time when the reef surface is becoming too
shallow to support a large cover of live coral but'still deep enough for contin-
uous water flow. Dead branching coral (Acropora palmata) in standing position
provides a complex and convoluted substratum on Which minute algal turfs can
grow and herbivores can take refuge. The structurally complex reefs maintain a
high biomass per projected area of reef, thereby increasing the total turf

183

biomass. The intense grazing maintains the algal community at the turf stage,
wherein the productivity per unit biomass is very high.

Facilitating this high productivity are nitrogen fixing cyanobacteria that
fix more nitrogen than is required by even the most productive reef zones.
Other nutrient requirements, particularly phosphorus, are met by the combination
of continuous and strongly flowing equatorial currents as well as by constant
wave-driven currents. This, with the lower phosphorus requirements of algal
turfs, results in a highly productive community that is not limited by nutrient
availability.

Thus, synergy is attained by the interaction of factors that optimize reef
productivity. If any factor is removed from this system (i.e., water depth,
SAR, wave oscillation, water flow, or algal community structure), the primary
productivity would be less than we found.

The back reefs we studied are among the most productive biological
communities in the world. Using standard conversions, our measurements of a
mean annual gross production of about 23 g C/m2/d and a net production of about
14 g C/m2/d exceed the theoretical maximum gross productivity hypothesized by
Ryther (1959) of 17 g C/m2/d and the net productivity of 12.3 g C/m2/d. On the
other hand, this is not unreasonable, since maximum experimental rates of net
production of 22-31 g (dry)/m2/d (approximately 9-12 g C/m2/d) have now been
found in both laboratory and field environments. Limitations to primary
production due to desiccation, nutrient availability or gas exchange (Lieth and
Whittaker, 1975) are nonexistent or of reduced importance on the reefs of St. 0
Croix. Thus, productivity of these reefs seems to be limited by the ultimate
limiting source: incoming solar radiation.

ACKNOWLEDGMENTS

This study was supported by the Smithsonian Institution, the Virgin Islands
Government and the West Indies Laboratory of Fairleigh Dickinson University.
We were assisted for much of the study by Caroline Rogers and Norman Salesky.
Additional help came from Sara Armstrong, Susan Brawley, Robert Carpenter, and
members of the West Indies Laboratory staff. Caroline Rogers, James Norris,
Susan Brawley, Chris D'Elia, and John Ogden read early drafts of this manuscript
and offered suggestions for its improvement. Steve Smith offered some valuable
suggestions for improvement of the final manuscript.

REFERENCES

Adey, W. 1975. "The Algal :Ridges and Coral Reefs of St. Croix: Their Structure
and Holocene Development." Atoll Res. Bull. 187:1-67.

Adey, W. 1978a. "Coral Reef Morphogenesis: A Multidimensional Model."
Science 202T831-837.

Adey, W. 1978b. "Algal Ridges of the Caribbean Sea and West Indies." Phycologia
17:361-375.

184

Adey, W., and R. Burke. 1976. "Holocene Bioherms (Algal Ridges and Bank
Barrier Reefs) of the Eastern Caribbean." Geol. Soc. Amer. Bull. 87:95-109.

Adey, W., P. Adey, R. Burke, and L. Kaufman. 1977. "The Holocene Reef Systems
of Eastern Martinique." Atoll Res. Bull. 218:1-40.

Adey, W., C. Rogers,.R. Steneck, and N. Salesky. 1981. "The South St. Croix
Reef." Rept. to Dept. Conserv. Cult. Affairs, U.S. Virgin Islands. 64 pp.

Adey, W., and T. Goertemiller. (In ms.). "Coral Reef Algal Turfs - Master
Producers in Nutrient Poor Seas."

Andrews, J., and H. Muller. 1983. "Space-Time Variability of Nutrients in a
Lagoonal Patch Reef." Limnol. Oceanogr. 28:216-227.

Atkinson, M. 1981. "Phosphate Exchange Between Reef Communities and the Water
Column." Proc. 4th Int. Coral Reef Symp. 1:417-418.

Atkinson, M., and S. V. Smith. 1983. "C:N:P Ratios of Benthic Marine Plants."
Limnol. Oceanogr. 28:563-574.

Borowitzka, M. 1981. "Algae and Grazing in Coral Reef Ecosystems." Endeavour,
N.S. 5:99-106.

Capone, D., D. Taylor, and B. Taylor. 1977. "Nitrogen Fixation (Acetylene
Reduction) Associated With Macroalgae in a Coral Reef Community in the
Bahamas." Mar. Biol. 40:29-32.

Carpenter, R. (In ms.). "Relationships Between Primary Production and Irradience
for Shallow Algal Communities on Coral Reefs."

Connor, J., and W. Adey. 1977. "The Benthic Algal Composition, Standing Crop
and Productivity of a Caribbean Algal Ridge." Atoll Res. Bull. 211:1-15.

D'Elia, C., K. Webb, and J. Porter. 1981. "Nitrate Rich Ground Water Inputs
to Discovery Bay, Jamaica: A Significant Source of Nitrogen to Local Reefs?"
Bull. Mar. Sci. 31:903-910.

Hackney, J., W. Adey, D. Lucid, and K. Peyton. (In ms.). "Harvest Production
of Coral Reef Algal Turfs in a Microcosm Environment."

Hatcher, A., and B. Hatcher. 1981. "Seasonal and Spatial Variation in Dissolved
Inorganic Nitrogen in One Tree Reef Lagoon." Proc. 4th Int. Coral Reef
Symp. 1:419-424.

Hatcher, B., and A. Larkum. 1983. "An Experimental Analysis of Factors
Controlling the Standing Crop of the Epilithic Algal Community on a Coral
Reef." J. Exp. Mar. Biol. Ecol. 69:61-84.

Helfrich, F., and S. Townsley. 1963. "The Influence of the Seas," pp. 39-57.
In: F. R. Fosberg (ed.), Man's Place in the Island Ecosystem. 10th Pac.
Sci. Congr. Bishop Museum Press, Honolulu.

185

Johannes, R., and the Project Symbios Team. 1972. "The Metabolism of Some
Coral Reef Communities: A Team Study of Nutrient and Energy Flux at
Eniwetok." Bioscience 22:541-543.

Lewis, J. 1977. "Processes of Organic Production on Coral Reefs." Biol. Rev.
52:305-347.

Lieth, H., and R. Whittaker. 1975. Primary Productivity of the Biosphere.
Springer-Verlag, New York. 315 pp.

Littler, M., D. Littler, and P. Taylor. 1983. "Evolutionary Strategies in a
Tropical Barrier Reef System: Functional Form Groups of Marine Macroalgae."
J. Phycol. 19:229-237.

Mague, J., and 0. Holm-Hanson. 1975. "Nitrogen Fixation on a Coral Reef."
Phycologia 14:579-586.

Mann, K. 1982. Ecology of Coastal Waters: A Systems Approach. Univ. of
California Press, BerkeTey. 322 pp.

Marsh, J., and S. Smith. 1978. "Productivity Measurements of Coral Reefs in
Flowing Water," pp. 361-377. In: D. Stoddart and R. Johannes (eds.), Coral
Reefs: Research Methods. UNESCO, Paris. 581 pp.

Odum, H., and E. Odum. 1955. "Trophic Structure and Productivity of a Windward
Coral Reef Community on Eniwetok." Atoll Ecol. Monogr. 25:291-320.

Odum, H., W. McConnell, and W. Abbott. 1958. "The Chlorophyll 'A' of
Communities." Publ. Inst. Mar. Sci., Univ. Texas 5:65-97.

Ogden, J., R. Brown, and N. Salensky. 1973. "Grazing by the Echiniod Diadema
antillarum Philippi: Formation of Haloes Around West Indian Patch Reefs."
Science 182:715-717.

Pilson, M., and S. Betzer. 1973. "Phosphorus Flux Across a Coral Reef."
Ecology 54:381-388.

Pomeroy, L., and E. Kuenzler. 1969. "Phosphorus Turnover by Coral Reef
Animals," pp. 474-482. In: D. Nelson and F. Evans (eds.), Proc. 2nd Nat.
Symp. Radioecology. Univ. Michigan Press, Ann Arbor.

Pomeroy, L., M. Pilson, and W. Wiebe. 1974. "Tracer Studies of the Exchange
of Phosphorus Between Reef Water and Organisms on the Windward Reef of
Eniwetok." Proc. 2nd Int. Coral Reef Symp. 1:87-96.

Rogers, C., and N. Salesky. 1981. "Productivity of Acropora palmata, Macroscopic
Algae and Algal Turf from Tague Bay Reef, St. Croix, U.S. Virgin Islands."
J. Exp. Mar. Biol. Ecol. 49:179-187.

Rogers, C., T. Suchanek, and F. Pecora. 1982. "Effects of Hurricanes David
and Frederic on Shallow Acropora palmata Reef Communities: St. Croix, U.S.
Virgin Islands." Bull. -r. Sci. -532 -548. 0

186

Ryther, J. 1959. "Potential Productivity of the Sea." Science 130:602-608.

Sammarco, P. 1983. "Effects of Fish Grazing and Damselfish Territoriality on
Coral Reef Algae. I. Algal Community Structure." Mar. Ecol. Prog. Ser.
13:1-14.

Sargent, M., and T. Austin. 1949. "Organic Productivity of an Atoll." Trans.
Amer. Geophys. Union 30:245-249.

Smith, S. 1981. "The Houtman Abrolhos Islands: Carbon Metabolism of Coral
Reefs at High Latitude." Limnol. Oceanogr. 26:612-621.

Smith, S., and J. Marsh. 1978. "Organic Carbon Production on the Windward
Reef Flat of Eniwetok Atoll." Limnol. Oceanogr. 18:953-961.

Sournia, A. 1977. "Analyse et Bilan de la Production Primaire de les Recifs
Coralliens." Ann. Inst. Oceanogr., N.S. 53:47-74.

Steneck, R., and L. Watling. 1982. "Feeding Capabilities and Limitations of
Herbivorous Molluscs: A Functional Group Approach." Mar. Biol. 68:299-319.

Stoddart, D. 1969. "Ecology and Morphology of Recent coral Reefs." Biol.
Rev. 44:443-493.

Talbot, F., and B. Goldman. 1972. "A Preliminary Report on the Diversity and
Feeding Relationships of the Reef Fishes of One Tree Island, Great Barrier
Reef System." Proc. Symp. Corals and Coral Reefs, Mandapam Camp, India:
pp. 425-443. Mar. Biol. Assoc. India, Cochin.

Van den Hoek, C., A. Cortel-Breeman, and J. Wanders. 1975. "Algal Zonation in
the Fringing Coral Reef of Curacao, Netherlands Antilles, in Relationship to
Zonation of Corals and Gorgonians." Aquat. Bot. 1:269-308.

Wanders, J., and A. Wanders-Faber. 1974. "Primary Productivity of the
Southwestern Coast Shallow Reef and the Northeast Coast Brown Algae of
Curacao." Br. Phycol. J. 9:223-224.

Webb, K., W. DuPaul, W. Wiebe, W. Sottile, and R. Johannes. 1975. "Eniwetok
Atoll: Aspects of the Nitrogen Cycle on a Coral Reef." Limnol. Oceanog.
20:198-210.

Westlake, D. 1963. "Comparisons of Plant Productivity." Biol. Rev. 38:385-425.

Wiebe, W., R. Johannes, and K. Webb. 1975. "Nitrogen Fixation on a Coral
Reef." Science 188:257-259.

Wilkinson, C., and P. Sammarco. 1983. "Effects of Fish Grazing and Damselfish
Territoriality on Coral Reef Algae. II. Nitrogen Fixation." Mar. Ecol.
Prog. Ser. 3:15-19.

187

STUDIES ON THE BIO-OPTICS OF CORAL REEFS

Phillip Dustan
Department of Biology
College of Charleston
Charleston, South Carolina 29401

ABSTRACT

This paper examines the possibility of using underwater light measurements
to study the primary productivity of coral reefs. This is a new technique that
may have wide ranging uses as man begins to use remote sensing to study the
processes taking place on the surface of the Earth.

INTRODUCTION

Sunlight is the primary source of energy for the coral reef ecosystem.
Photosynthetic reef organisms utilize light to liberate oxygen from water, fix
carbon, and enhance the rate of calcification. This paper will examine the
concept of using bia-optical techniques (Smith and Baker, 1978a; Gordon, et
al., 1980) to measure the input of light energy to coral reef 'Ecosystems. The
simplicity of the technique, combined with nondestructive sampling, may enable
reef ecologists to study the productivity and flow of energy into coral reefs
. ~~at a higher level of ecosystem organization than is presently possible. In
addition, it is of interest to understand the entrance of energy into these
benthic ecosystems since many reef organisms may be termed polytrophic (Muscatine
and Porter, 1977), and because the structure and dynamics of the primary
production trophic level may have an influence on the structure and/or organization
of subsequent higher trophic levels. Understanding the optics of coral reefs
also holds the promise of one day being able to study these complex and remote
ecosystems from orbiting spacecraft and other remote sensing platforms.

Investigators have long sought a technique that would quantify the primary
productivity of coral reefs. For the most part, estimates have been made using
respirometry and/or radiotracer experiments which have been extrapolated to the
areal extent of the reef (see part III in Stoddart and Johannes, 1978, for
review). While these techniques usually are precise, they are subject to errors
in scale which arise from extrapolating data from an experimental specimen to
the reef community. Another way to estimate primary production is to measure
the input of light energy to the system over a relatively large area of substrate.
Then, using known and theoretical estimates of the quantum efficiency for
photosynthesis, it should be possible to estimate primary production on a scale
commensurate with the scal-e of the reef ecosystem.

The initial step in primary production is the harvesting of light by the
photosynthetic pigments of the primary producers (Rabinowich and Govingee,
1969). This process is a function of the available light, both spectral quality
and intensity, and the absorption characteristics of the algae. The absorption
* ~~and reflectance of light by phytoplankton also can be influenced by the
nutritional state and age of the alga (Kiefer, 1973; Kiefer, et al., 1979;

189

Wilson and Kiefer, 1979). Present models that describe the light-limited growth
rate of cultured phytoplankton have been formulated using the estimated quantum
efficiency of photosynthesis and the cellular cross sectional area for light
absorption (Laws and Bannister, 1980; Kiefer and Mitchell, 1983). The formulation
of similar models for coral reefs might provide estimates of primary production
that would be applicable to different reef systems, habitats, or zones on a
scale that is commensurate with the scale of these ecosystems. Formulation of
such models requires the ability to make reliable estimates of light reflectance and
absorption for the reef community.

Morel (1978) defined a measure of light absorption termed Photosynthetically
Useable Radiation (PUR). PUR defines the fraction of Photosynthetically
Available Radiation (PAR) that a photosynthetic organism can absorb and have
available for chemical work. PUR is the integrated product of normalized cellular
spectral absorption and the spectral irradiance available to the organism
(Morel, 1978). Measures of PUR are relative measures, though, and cannot be
related to an absolute measure of the quantity of light being ~bsorbed. However,
much can be learned by comparing the relative light absorption of photosynthetic
organisms from different light environments.

Calculations of PUR for isolated zooxanthellae from the reef-building coral
Montastrea annularis (Ellis and Solander) have shown that the fraction of light
absorbed varies at different depths, suggesting that the capture of light
energy by these zooxanthellae varies as a function of the available light field
(fig. 1; and Dustan, 1982).

METHODS

Optical oceanographers have developed precise terminology that describes
the behavior and measurement of light in the underwater environment. Spectral
irradiance is defined as the spectral components of the light field that impinge
on a flat-plate, or cosine collector. Across each horizontal plane in the
hydrosol are upward and downward flowing flux defined as upwelling and downwelling
(Tyler and Preisendorfer, 1962; Gordon, et al., 1980). The ratio of upwelling
to downwelling is termed spectral reflectance and is the basis for the measurement
of remotely sensed optical signals (Gordon, et al., 1982; Smith and Baker,
1978b). The difference between upwelling and downwelling at the surface of a
substrate is the amount of light that is absorbed by the substrate. It is
presently impossible to directly measure the upwelling irradiance close to the
substrate because the shadow of the instrument alters the light field. However,
the upwelling and downwelling irradiance at the reef substrate can be estimated
by taking sets of upwelling and downwelling measurements above the substrate.
From these data, the diffuse attenuation coefficient for spectral irradiance is
calculated, and this quantity is used to calculate the light intensities at the
surface of the substrate (Smith and Baker, 1978b). These derived data will
vary with water depth, species absorption properties, and the ambient light
field. Thus, measurements must be taken throughout the daytime period and at
systematically varying depths to use these measurements as an index of light
energy absorption and a community-specific spectral absorption signature. At
this time, we do not fully understand the relationship between this derived measure
and the actual amount of light absorbed by photosynthetic reef organisms, but,
as a first approximation, the measure represents a maximum value which includes

190

1015 i

ALGAL DEPTH / LIGHT DEPTH -

c 1012_ _
6 ' _\ \31.5- 30 m
X, \

Figure .--Photosynthetically Useable Radiation of zooanthell0ae isolated from

spectroradiometer. Note that the algae from different depths vary with
109 l l 191
400 450 500 550 600 650 700

WAVELENGTH rnm)

Figure 1.--Photosynthetically Useable Radiation of zooxanthellae isolated from
Montastrea annularis, Discovery Bay, Jamaica. Relative whole cell absorbance
was measured at low temperature with a modified Cary 14 spectrophotometer.
Measurements of underwater spectral irradience were made with a prototype
spectroradiometer. Note that the algae from different depths vary with
respect to their absorbance of light energy and its spectral quality (see
Dustan, 1982, for experimental details).

the absorption of light by both animals and plants. As such, the data will
overestimate actual primary production until the measure is more fully understood
and the model appropriately refined.

191

Table 1.--Preliminary calculations of light energy absorbed by a coral reef
community. Surface PAR light levels averaged 1.47 x 1017 quanta/cm2/s during
the experiment. Measurements were carried out on Dancing Lady Reef, Discovery
Bay, Jamaica, W.I.

Calculated quanta Percent of Incident
Depth Substrate Absorbed (q/cm2/s) Downwelling PAR
(400-700 nm) (400-700 nm)

10 m sand 2.679 x 1016 87.0

coral 2.429 x 1016 92.7

30 m sand 7.42 x 1015 68.0

coral 1.08 x 1016 957

FIELD STUDIES

In 1978 we developed a small submersible spectroradiometer to study the
photobiology of zooxanthellae (Booth and Dustan, 1979). I used the instrument
to characterize the light field on Dancing Lady Reef, Discovery Bay, Jamaica,
with special emphasis on the photobiology of Montastrea annularis, one of the
principal reef-building corals on both the deep and shallow reefs of the region
(Dustan, 1982). As part of this study, I made systematic measurements of
upwelling and downwelling irradiance and reasoned that it might be possible to
measure the amount of light energy that the reef community "absorbs." In this
experiment measurements were taken over sand, which was reasoned to be an
almost perfect reflective surface, and over areas covered with living coral
(see table 1).

The calculations suggest that corals resemble black bodies in their
absorption properties in both shallow and deep water environments. Reef sand
is not a very good reflective surface, as I had first thought, but does absorb
a smaller fraction of the incident light than corals. The difference in
absorption between coral and sand is greater in deep than shallow water, possibly
suggesting photoadaptive or geometric change in the members of the coral
community. The percentage of incident light absorbed by both substrates seems
high, but since a reference of known standard reflectance was not used to
"calibrate" the methodology, the degree to which light scattering and other
possible nonlinear effects may distort the data is unknown. In addition, the
sandy environment of coral reefs contains a variety of photosynthetic organisms,
each with its own particular light absorption characteristics.

A second set of spectroradlometer measurements was made in August 1982 on
the Molasses Reef in the Key Largo National Marine Sanctuary, Florida. Spectral
measurements were taken with a 15 channel spectroradiometer (Biospherical
Instruments Mer-1015) during a research expedition in support of remote sensing
of coral reefs (P. Dustan, C. R. Booth, and A. R. Hlbbs, unpub. data). Data
from a sandy zone on Molasses Reef gave a net absorbance of 1 x 1014 quanta/cm2/s
which was 51% of the downwelling irradiance at the substrate surface (fig. 2).

192

/ OOWNWELLING

C
E

C3 . UP AND DOWNWELLING SPECTRAL IRRADIANCE \
^ ~~CALCULATED AT THE SUBSTRATE SURFACE \ \
13 ~OVER A SAND BGTTOM ON MOLASSES REEF \ \
L AT A DEPTH OF 7 METERS\ \
L MOLASSES REEF RUN 12 AUGUST 14. 1982\
U

1
cJ

I0
C
0 ~~UP AND DOWNWELLIING SPECTRAL IRRADIANCE
" ~~CALCULATED AT THE SUBSTRATE SURFACE
OVER A SAND BOTTOM ON MOLASSES REEF
L ~~AT A DEPTH OF 7 METERS
L- MOLASSES REEF RUN 12 AUGUST 14. 1982
HER-1015 SPECTRORADIOMETER

400 soo 600 700
Wave length (nm.)

Figure 2.--Calculated upwelling and downwelling spectral irradiance at the
substrate surface of a sandy area 7 m deep on Molasses Reef, Key Largo. The
difference between upwelling and downwelling irradiance equals the light
absorbed by the substrate (1 x 1014 q/cm2/s). Scans were calculated from
sets of spectral irradiance data as described in the text.

This fraction is somewhat higher than the earlier data from Jamaica but is
within a range that could be considered acceptable, since the technique is in
an early stage of development.

Similar measurements made over a nearby grove of living coral on Molasses
Reef revealed somewhat different and unexpected results. The values for

193

upwelling spectral irradiance did not follow the expected pattern of exponential
de ease with depth. As the instrument approached the coral, the spectral
d- ribution of the upwelling irradiance showed a decrease in the blue and
green regions of the spectrum but an increase at wavelengths greater than
589 nm. Particularly striking was an increase in the far red at 694 nm (fig.
3). The total integrated quanta in each scan follows the rule for conservation
of energy and decreases with depth. The increase in the signal between 589 and
671 nm results from reflectance of the yellow brown coral, which begins to
"dominate" the spectral irradiance signature. The far red signal may be
interpreted as an emission of red light due to algal fluorescence, as has been
seen in oceanographic measurements of spectral irradiance (R. C. Smith, C. R.
Booth, and P. Dustan, unpub. obs.). The intensity of the signal increases as
the instrument approaches the coral (depth approx. 4.4 m), suggesting that the
fluorescence may emanate from the zooxanthellae in the tissues of the coral.
These findings complicate the calculation of an energy absorption budget for
the coral, since the organism is simultaneously absorbing and emitting light
and each process is functioning with its own wavelength dependency. However,
the results suggest that we may have uncovered a measurable parameter that may
one day be used to probe the photophysiology of corals and the bio-optics of
coral reef ecosystems.

Researchers are beginning to realize that the fluorescence signature of an
alga contains information about the plant's photosynthetic apparatus, including
information on its potential for primary productivity (Kiefer, 1973; Vincent,
1979; Harris, 1980; Abbott, et al., 1982). The data on corals are particularly
exciting, since they suggest that the spectral signals may contain information
on the physiological state of corals and their zooanthellae in addition to
possible estimates of primary production. This suggests that there is much to
learn about the optics of corals and their associated zooxanthellae using the
passive techniques of bio-optics.

REMOTE SENSING

Coral reefs cover areas of the ocean that are remote and difficult to work
in for extended periods of time. They are also vast in their coverage of the
world's tropical oceans and may play a significant role in the carbonate balance
of the seas. If the optical signal emanating from the reef can be understood
sufficiently to provide interpretation from an orbiting spacecraft, estimates
of coral reef primary productivity and possibly calcification could be made on
a global scale. The importance of this cannot be underestimated when one
considers man's impact on the coral reef ecosystems that are presently in close
proximity to centers of human habitation.

Our research expedition to Florida in 1982 gathered data to test the
hypothesis that coral reefs may be resolvable in spacecraft imagery (Landsat)
(P. Dustan, C. R. Booth, and A. R. Hibbs, unpub. data). We obtained measurements
of upwelling spectral irradiance from a variety of habitats in the shallow
coral reef ecosystem of John Pennekamp Coral Reef State Park, including grass
flats, coral colonies, and sand. We then simulated the spectral sensitivity of
the Multispectral Scanner (MSS) onboard Landsat 3 so that the optical signal we
recorded could be compared to the signal recorded by a Landsat MSS sensor. The
equation for the simulation then "converts" the spectroradiometer signal into

194

4I) 0 \\\ 4
� 0

03 UPWELLINC SPECTRAL IRRAOIANCE OVER Acroporo polmato
0o MOLASSES REEF RUN 12 AUGUST 14. 1982
L MER-lO15 SPECTRORADIOMETER

SCAN FILE DEPTH TOTAL QUANTA
# n (M) (400-700nm)

1 40 1.8 1.9 x 1013
2 41 3.1 1.8 x 1013
3 43 3.6 1.7 x 103
4 42 3.9 1.5 x 1013
* OQ/cm2/suc
-1
10
400 500 600 700
Wovelength (nm.)

Figure 3.--Upwelling spectral irradiance scans taken over a grove of elkhorn
coral, Acropora palmata, at 4.4 m depth on Molasses Reef, Key Largo. Note
that 589 nm is a relatively invariant wavelength. The increase in far red at
694 nm is interpreted as in situ fluorescence by the symbiotic zooxanthellae
of the coral.

the bandwidths of the Landsat MSS sensor:

Band 1 (green) = (.091x488nm + .76x507nm + .913x520nm + .98x540nm
+ .97x570nm + .75x589nm + .05x625nm) / 4.514

195

KEY LARGO. FLORIDA
UPWELLING SPECTRAL IRRADIANCE AT SURFACE
RUN 2 AUGUST 13. 1982 CLEAR SKY
MER-1O0S SPECTRORAOIOMETER

0
0 160

3: 4 4. S ass bed

eMER smulSS d MSS 1/Ha2 b iat
4 4.10 Sao gross ba d
30 4.a97 Swa grace bar d t
109 13.39 Corals/sand

MER iemultatd MSS 1/2 ratio

surface irrodionci(PAR)- 1. 39-l.41xo10 7/cms.

10
400 o500 6o00 700
Wavelength (nm.)

Figure 4.--Upwelling spectral irradiances taken just beneath the sea surface
over a variety of coral reef environments and the western edge of the Gulf
Stream. The Mer simulated MSS ratio is the ratio of Band 1/Band 2 of Landsat,
calculated from measurements taken with the Mer-1015 Biospherical Instruments
spectroradiometer and wavelength weighting functions as described in the text.

Band 2 (red) = (.036x589nm + .927x625nm + .960x656nm + .90x671nm
+ .810x694nm) / 3.633

We then calculated the ratio of the bands (Band 1/Band 2) to allow for
differences in absolute intensity between the ground and spacecraft signals.

196

Figure 4 shows the spectral signals that upwell from different types of reef
habitats in relation to the blue waters of the nearshore Gulf Stream. Different
habitats are clearly distinguishable. The calculated MSS ratios for each habitat
are also different, suggesting that reefs and their associated habitats may be
distinguishable and quantitatively studied using multispectral spacecraft imagery.

CONCLUSIONS
While our work is by no means complete, we are beginning to show that an
understanding of the bio-optics of the system can contribute to our understanding
of the dynamics of coral reef ecosystems. The techniques that I have described
suggest that optical measurements can be used to estimate the input of light energy
into benthic ecosystems at various depths and habitats. Such data are necessary
tounderstand and compare the energetics of coral reef communities and the degree
to which the process of primary production varies and influences the organization
of the ecosystem. The studies described in this communication are parallel to
laboratory and field investigations on marine phytoplankton. The techniques
are nondestructive, which means that the same environment can be repeatedly
sampled. The measurements can be made over a variety of temporal and spatial
scales, providing the potential to investigate large areas of coral reef ecosys-
tems. The results are encouraging and suggest that with more research into the
optics of coral reefs we may be able to use the techniques of remote sensing to
routinely delineate, monitor, and, perhaps, manage these coral reef ecosystems.

ACKNOWLEDGMENTS

This work could not have been completed without the collaboration of
Charles R. Booth, Biospherical Instruments Inc., and Albert R. Hibbs, Jet
Propulsion Laboratory. Edmund Woloszyn and John Halas helped complete the
field work in the Florida Keys. Special thanks go to Captain Mark Glisson,
Florida Department of Natural Resources, for logistical support at John Pennekamp
Coral Reef State Park. This research was funded by the National Science
Foundation (NSF OCE 76-81071), National Oceanic and Atmospheric Administration
(NA 82-AAA-02880), and the College of Charleston. This is Contribution Number
66 from the Grice Marine Biological Laboratory.

LITERATURE CITED

Abbott, M. R., P. J. Richerson, and T. M. Powell. 1982. "In-Situ Response of
Phytoplankton Fluorescence to Rapid Variations in Light." Limnol. Oceanogr.
27:218-225.

Booth, C. R., and P. Dustan. 1979. "Diver-Operable Multiwavelength Radiometer."
Proc. Soc. Photo-Optical Instrumentation Engineers 196:33-39.

Dustan, P. 1982. "Depth-Dependent Photoadaptation by Zooxanthellae of the
Reef Coral Montastrea annularis." Mar. Biol. 68:253-264.

Gordon, H. R., D. K. Clarke, J. L. Mueller, and W. A. Hovis. 1981. "Phytoplankton
Pigments From the Nimbus-7 Coastal Zone Color Scanner: Comparisons With
Surface Measurements." Science 210:63-66.

197

Gordon, H. R. R. Zannifeld, and R. C. Smith. 1980. "Introduction to Ocean
Optics." -oc. Soc. Photo-Optical Instrumentation Engineers 208:14-54.

Harris, G. P. 1980. "The Relationship Between Chl a Flourescence, Diffuse
Attenuation Changes, and Photosynthesis in NaturaT Phytoplankton Populations."
J. Plankton Res. 2:109-127.

Kiefer, D. A. 1973. "Chlorophyll a Fluorescence in Marine Centric Diatoms:
Responses of Chloroplasts to Light and Nutrient Stress." Mar. Biol. 23:39-46.

Kiefer, D. A., and B. G. Mitchell. 1983. "A Simple, Steady State Description
of Phytoplankton Growth Based on Absorption Cross Section and Quantum
Efficiency." Limnol. Oceanog.r. 28:770-775.

Kiefer, D. A., R. J. Olson, and W. Wilson. 1979. "Reflectance Spectroscopy of
Marine Phytoplankton. Part 1. Optical Properties Related to Age and Growth
Rate." Limnol. Oceanogr. 24:664-672.

Laws, E. A., and T. T. Bannister. 1980. "Nutrient- and Light-Limited Growth
of Thalasiosira fluviatilis in Continuous Culture, With Implications for
Phytoplankton Growth in the Sea." Limnol. Oceanogr. 25:457-473.

Morel, A. 1978. "Available, Useable, and Stored Radiant Energy in Relation to
Marine Phytoplankton." Deep Sea Res. 25:673-688.

Muscatine, L., and J. W. Porter. 1977. "Reef Corals: Mutualistic Symbioses
Adapted to Nutrient-Poor Environments." Bioscience 27:454-460.

Rabinowich, E., and Govingee. 1969. Photosynthesis. John Wiley and Sons, New
York. 273 pp.

Smith, R. C., and K. S. Baker. 1978a. "The Bio-Optical State of Ocean Waters
a"d Remote Sensing." Limnol. Oceanogr. 23:247-259.

Smith, R. C., and K. S. Baker. 1978b. "Optical Classification of Natural
Waters." Limnol. Oceanogr. 23:26U-267.

Stoddart, D. R., and R. E. Johannes. 1978. Coral Reefs: Research Methods.
UNESCO, Page Bros. Pub., Norwich, U.K. 581 pp.

Tyler, J. E., and R. W. PreisEndorfer. 1962. "Transmission of Energy Within the
Sea," pp. 397-451. M. N. Hill (ed.), The Sea, Vol. 1. Interscience Pub., N.Y.

Vincent, W. F. 1979. "Mechanisms of Rapid Photosynthetic Adaptation in Natural
Phytoplankton Communities: 1. Redistribution of Excitation Energy Between
Photosystems I and II." J. Phycol. 15:429-434.

Wilson, W. H., and D. A. Kiefer. 1979. "Reflectance Spectroscopy of Marine
Phytoplankton. Part 2. A Simple Model of Ocean Color." Limnol. Oceanogr.
24:673-682.

198

PRELIMINARY STUDIES OF DENITRIFICATION ON A CORAL REEF

Sybil P. Seitzinger
and
Christopher F. D'Elia1
Division of Environmental Research
Academy of Natural Sciences of Philadelphia
Philadelphia, Pennsylvania 19103

ABSTRACT

This preliminary study presents evidence that coral reefs may be the sites
of potentially high rates of denitrification. Qualitative evidence for the
existence of denitrification in dead coral heads and in sediments from a sea
grass community were obtained using the acetylene blockage technique and by
measuring the rate of nitrate reduction occurring in sediment slurries incubated
under anaerobic conditions. The degree to which this denitrification offsets
nitrogen fixation is unknown and awaits quantification with better techniques,
but from rough approximation it appears to be considerable.

INTRODUCTION

* Many studies on coral reefs have shown that while standing stocks for N
species are low, rates of transformations of N species are rapid (Szmant
Froelich, 1983). Coral reefs typically exhibit high rates of nitrogen fixation
(Wiebe, et al., 1975; Mague and Holm-Hansen, 1975; Bunt, et al., 1970; Capone
and TaylorjT77), the process by which atmospheric N2 gas is chemically
transformed into "combined" nitrogen forms (e.g., ammonia) that can be utilized
as nutrients by autotrophs. Indeed, some of the highest rates of N2 fixation
for any ecosystem have been reported for coral reef areas. However, denitrifi-
cation, the process by which combined nitrogen is chemically transformed back
to atmospheric N2 gas, has received much less attention in reef environments.
Denitrification represents a potential removal mechanism for combined N both on
the reef and in lagoons behind reefs. Recent reports of substantial rates of
nitrification (ammonia oxidation to NO2 and NO3) on coral reefs (Webb and Wiebe,
1975; LIMER Expedition, 1977) indicate that a potential exists for denitrifica-
tion to occur.

This paper presents preliminary measurements of denitrification on a coral
reef in Abrahams Bay, Mayaguana, Bahamas, and from sediments behind the reef
in a lagoon with Thalassia beds. We used techniques suitable for qualitative
determination of denitrification under field conditions. This work was carried
out on the R/V MARSYS RESOLUTE with the cooperation and support of Dr. Walter
Adey of the Marine Systems Laboratory of the Smithsonian Institution. Additional
support also was provided by the Ruth Patrick Foundation at the Academy of
Natural Sciences of Philadelphia.

1Present address: Chesapeake Biological Laboratory, Solomons, Maryland
20688.

199

0 1 2 3 4 5 6

Figure 1.--Mayaguana (Abrahams Bay) Coral Reef and Lagoon. Depths in fathoms;
x indicates the study area.

MATERIALS AND METHODS

This study was conducted during May 1983 in Abrahams Bay, Mayaguana, located
in the southern Bahamas (fig. 1).

Dead coral heads, each with a displacement volume of about 250-300 cm3,
were collected by SCUBA divers from several areas behind the reef crest. The
heads which were covered with algae and organic matter were brought back to the
ship's laboratory and placed in gas tight glass incubation chambers (Seitzinger,
et al., 1980) which were then filled to within 1 cm of the top with reef water
'50oml). Denitrification rates were measured using acetylene to inhibit N20
reduction to N2 (Balderston, et al., 1976). The gas phase (60 ml) was flushed
with acetylene, and the water-Tntf-te chambers was stirred with floating magnetic
stir bars to equilibrate the gas and water phases. Samples (30 ml) of seawater
were then withdrawn at time intervals into stoppered 50 ml serum bottles and 1
ml of saturated KOH solution added to each. After 3 hr a concentrated NO3 solution
was added to two of the chambers to make the seawater concentration 40 pM N03.
Time-series samples for N20 continued to be taken at 7, 9, and 17 hr. The N20
concentration in the gas phase of the serum bottles was measured by ECD gas
chromatography within 1 week (Seitzinger, et al., 1980). The total amount of
N20 in the samples was calculated using WeTls and Price's (1980) N20 solubility
equation for the appropriate temperature and salinity. The production of N20
in the presence of acetylene by the coral heads then was calculated based on
the change in concentration between sequential samples and equated to the
denitrification rate.

The potential for denitrification in the sediments from the reef lagoon was
examined using measurements of pore water N02 and N03 concentrations, N03
reduction rate measurements, and measurements of N20 production in the presence
and absence of acetylene. Sed-ment cores (47 cm2 x 25 cm deep) were collected

200

Table 1.--Rate of denitrification by diad coral heads as measured by N20
production (nmol N 0 h- coral head- ) in the presence of acetylene,
but without added N03, except as indicated. Acetylene was added at
time = 0 hr. * after 3 hr a concentrated N03 solution was added to
make seawater 40 pM N03.

Dead Coral Head Incubation Period Rate of N20 Producition
nmol N20 h-

1 2.5 - 5.5 hr 1.3

2 2 - 5 hr 2.1

3 0 - 3 hr 1.6

*3 - 7 hr 57

7 - 9 hr 38

9 - 17 hr 16

4 0 - 3 hr 0.8

*3 - 7 hr 51

7 - 9 hr 16

9 - 17 hr 6

by SCUBA divers from the Thalassia beds in the reef lagoon. One core was
sectioned at 2.5 cm intervals and the pore water N02 and N03 concentrations
determined using a Technicon AutoAnalyzer.

N0O reduction rates were measured in a second core sectioned at intervals
of 0-5 cm, 10-15 cm, and 20-25 cm. Sediment slurries from these depths were
placed in serum bottles with 20 uM NOj seawater solution and the change in N02
+ NO3 concentration with time measured. The incubations were done in the dark
at 29�C.

Denitrification rates in the lagoon sediments were measured, again using
acetylene inhibition of N20 reduction. Sediments from three cores were sectioned
at intervals of 0-5 cm, 10-15 cm, and 20-25 cm. Sediment slurries (14 cm3)
from those depths were incubated aerobically in serum bottles with 15 ml of 20
UM NO3 seawater solution and 0.3 atmospheres of acetylene. N20 production
rates in the presence of acetylene then were measured at each depth after
intervals up to 15.5 hr by injecting 2 ml of a saturated KOH solution into the
serum bottles after approximately 0, 6, and 15.5 hr. The production of N20 in

201

Concentration, .M
0 1 2 3 4 5 6 7 8 9 10
0
a *~ ~ ~ ~ ~ ~ ~~~~~~~ I I I It

E 6-

10-

12-

13-

Figure 2.--Depth profile of pore water nitrite (circles) and NO2 + N03 (triangles)
in sediments of lagoon.

the absence of acetylene was measured in a parallel set ofsamples incubated
with 20 pM NO3 but without acetylene. The reduction of NO3 also was measured
in another parallel set of samples with acetylene and 20 uM NO3. N02 + N03
concentrations were measured immediately after approximately 0, 6, and 15.5 hr
(KOH was not added).

RESULTS

When coral heads were incubated in seawater that was not enriched with
nitrate, the rate of denitrification (N20 production in the presence of acetylene)
was approximately 1-2 nmol N20 per coral head per hour (table 1). When the N03
concentration in the surrounding seawater was increased to 40 pM, the rate of
N20 production increased to approximately 50 nmol per hour per coral head and
then decreased during the subsequent 14 hr.

202

18 -

16 -

14 -

124-

22 -
8~ 8-

0 4 8 12 16 20
24 BB

16-

0 4 8 12 16 20
Elapsed Time, hrs

Figure 3.--(A) N03 depletion in lagoon sediments spiked to 20 uM NO3; (B) N20
production by lagoon sediments incubated with a 20 pM N03 solution and in the
presence of acetylene. Open circles = 0-5 cm depth; triangles = 10-15 cm
depth; and closed circles = 20-25 cm depth.

The concentrations of N02 + NO in the sediments of the Thalassia bed were
higher than the overlying seawater (0.5 4M NOj + NO3), ranging from 3.5 to 9 PM
with a peak concentration at about 11 cm depth (fig. 2).

When sediments were incubated in seawater enriched to a concentration of 20
UM NO3, N03 concentrations decreased rapidly in sediments from all depths to
below concentrations found in the pore waters (fig. 3A). Since this experiment

203

Table 2.--Rate of denitrification as measured by N20 production in lagoon
sediments incubated with 0.3 atm acetylene and 20 OM NO3. Rates are expressed
per cm3 of sediment per hour.

Depth Interval Time Interval N20 Producton
~~(cm) t~(hr) (nmol N20 cm- h1)

0- 5 0- 7 1.9 x 10--

7 - 15.5 -0.3 x 10-1

10 - 15 0 - 6 1.2 x 10-1

6 - 15.5 -0.1 x 10-1

20 - 25 0 - 5.5 2.8 x 10-1

5.5 - 15.5 -0.02 x 10-1

was performed in the presence of acetylene, no new nitrification was occurring,
inasmuch as acetylene also blocks nitrification_(Bremner and Blackmer, 1979).
During that same period of rapid decrease in NO3 concentration (0-6 hr), there
was a rapid increase in N20 concentration in samples incubated with acetylene
(fig. 3B). The highest rates of N20 production during the first 6 hr was seen
in the 20-25 cm depth interval (table 2). The concentration of N20 tended to
decrease during the next 16 hr interval. The rate of NO3 reduction increased
with increasing concentration of substrate (fig. 4).

Concentrations of N20 in sediments incubated with NOT but without acetylene
were negligible and tended to decrease slightly with time (not shown). Changes
were small compared to changes measured in the presence of acetylene.

DISCUSSION

The preliminary measurements of denitrification reported here indicate that
coral reefs and associated lagoon sediments are potentially active sites for
denitrification. Nitrate concentrations inside dead coral heads have been
reported by Webb and Wiebe (1975) and Risk and Muller (1983) to be higher than
the surrounding seawater; this undoubtedly indicates that nitrification is
occurring in dead coral heads. Low oxygen concentrations occur as organic
matter decays in coral heads, thus providing the proper conditions for
denitrification (Risk and Muller, 1983). Although nitrification and denitrification
require different concentrations of oxygen to proceed, both can occur simultaneously
in coral heads due to microdistributional variations in oxygen concentration
that certainly must occur (Risk and Muller, 1983).

Sediments in the lagoon behind the reef also appear to be active areas for
denitrification as well. This was demonstrated by both the N20 production

204

5.0-

c o
a: 4.0
2?
3.0-
ccE

2.0-
Z~~~~~
1.0-

0 2A 10 8l 12

Figure 4.--Rate'of NO reduction (nmo N- 3 h-1) vs concentrateion ofM

~Figure 4.--Rate of NO edcton(no3 h) vs. concentration of
NO3 (>M) in sediment experiment.

rates measured in the presence of acetylene and the NO3 reduction rates. The
source of N03 for denitrification must come from nitrification in the sediments
as opposed to diffusion of N03 from the overlying water, since the N03 concentrations
in the pore waters were much greater than in the overlying water. The N03
concentrations we measured in the pore waters must be maintained by relatively
high rates of nitrification. Two lines of evidence support this: (1) NO3 concen-
trations were shown to decrease rapidly below concentrations measured in the
pore waters when nitrification was inhibited by acetylene, and (2) denitrifi-
cation rates were indicated to be high in the sediments and thus, to maintain
NO3 concentrations in the pore waters with active denitrification occurring,
nitrification must also be rapid.

The rates of decrease of NO3 we observed in lagoon sediments were much
greater than the rates of N20 production. One would expect them to be equal if
all of the NO3 decrease was due to denitrification (see Taylor, 1983, for a
review of the problems associated with these measurements). We can presently
only speculate as to why the discrepancies occurred (figs. 3A and 3B). For
example, acetylene blockage may not be complete or may not yield stoichiometric
quantities of N20. Alternatively, NO3 could be reduced assimilatively, e.g.,
by foraminiferans (Webb and Wiebe, 1978).

We emphasize that the results presented here are preliminary for several
reasons. First, we did only a limited number of experimental measurements and
feel that a more detailed study is needed. Second, the techniques we used to
measure denitrification are adequate for a preliminary survey in the field like
this one but should be supplemented by the preferred, direct measurement of
denitrification (Seitzinger, et al., 1980). We consider the use of sediment

205

slurries that are enriched with nitrate and incubated under anaerobic conditions
(Billen, 1978; Koike and Hattori, 1978, 1979; Oren and Blackburn, 1979; Iizumi,
et al., 1980) to be good qualitative but not necessarily good quantitative
tools--such experimental conditions destroy sediment microenvironmental structure
(e.g., nitrate, oxygen, and organic matter concentrations) that cannot be
simulated exactly in the laboratory. This is particularly problematic in the
Thalassia bed sediments with their complex root structure. The use of the
acetylene blockage technique also incurs problems. For example, acetylene blockage
of N20 reduction may be inhibited by the presence of sulfide (Tam and Knowles,
1979), and acetylene itself may be oxidized anaerobically (Culbertson, et al.,
1981) or may inhibit nitrification (Bremner and Blackmer, 1979; Blackmer, et
al., 1980), methanogenesis (Oremland and Taylor, 1975), methane oxidation
(deBont and Mulder, 1976), and the growth of sulfate-respiring bacteria (Payne
and Grant, 1982). Considering these limitations, it is likely that the
denitrification rates reported here are underestimates of the field rates.

The amount of fixed N being lost through denitrification in the lagoon
behind the coral reef conceivably could be as high as that fixed by N2 fixation
on the reef itself. No measurements of N2 fixation are available for Mayaguana.
However, measurements ofN2 fixation have been made on a number of other ~oral
reefs. Maximum rates found at Enewetak, for example, were 600 4mol N2 m h'
during daylight (Wiebe, et al., 1975; Webb, et al., 1975). For an average rate
over 24 hr, this rate should be decreased by approximately 50% to 300 ,mol m-2
h-1, since such high rates only occur during daylight. The above methodological
limitations aside, we can make a rough approximation of denitrification in the
lagoon. The results from the acetylene inhibition experiments indicated that
denitrification rates were approximately 0.2 nmol cm- h-1 throughout the top
25 cm of lagoon sediment. The integrated value for the top 25 cm is then
50 umol m-2 h-1. The area of the lagoon is estimated to be 10 times the area
of the reef. Therefore, in the reef-lagoon ecosystem, denitrification in the
lagoon would more than balance the N2 fixation on the reef. While these measure-
ments and calculations are only preliminary, they do point to the probable
importance of denitrification in reef ecosystems and to the need for further
measurements.

ACKNOWLEDGMENTS

We thank Walter Adey and the crew of the R/V MARSYS RESOLUTE for giving us
the opportunity to make these measurements. We also thank S. V. Smith and K.
L. Webb for their critical comments and L. Woodbury, D. Engle, and G. Canaday
for technical help.

REFERENCES

Balderston, W. L., 8. Sherr, and W. J. Payne. 1976. "Blockage by Acetylene
of Nitrous Oxide Reduction by Pseudomonas perfectomarinus." Appl. Environ.
Microbiol. 31:504-508.

Billen, G. 1978. "A Budget of Nitrogen Recycling in North Sea Sediments Off
the Belgian Coast." Estuarine Coastal Mar. Sci. 7:127-146.

206

Blackmer, A. M., J. M. Bremner, and E. L. Schmidt. 1980. "Production of
Nitrous Oxide by Ammonia-Oxiding Chemoautotrophic Microorganisms in Soil."
Appl. Environ. Microbiol. 40:1060-1066.

Bremner, J. M., and A. M. Blackmer. 1979. "Effects of Acetylene and Soil
Water Content on Emission of Nitrous Oxide From Soils." Nature 280:380-381.

Bunt, J. S., K. E. Cooksey, M. S. Haeb, C. C. Lee, and B. K. Taylor. 1970.
"Assay of Algal Nitrogen Fixation in the Marine Subtropics by Acetylene
Reduction." Nature 227:1163-1164.

Capone, D. G., and B. F. Taylor. 1977. "Nitrogen Fixation (Acetylene Reduction)
in the Phyllosphere of Thalassia testudinum." Mar. Biol. 40:19-28.

Culbertson, C. W., A. J. B. Zehnder, and R. S. Oremland. 1981. "Anaerobic
Oxidation of Acetylene by Estuarine Sediments and Enrichment Cultures."
Appl. Environ. Microbiol. 41:396-403.

deBont, J. A. M., and E. G. Mulder. 1976. "Invalidity of the Acetylene
Reduction Assay in Alkane-Utilizing, Nitrogen Fixing Bacteria." Appl.
Environ. Microbiol. 31:640-647.

Iizumi, H., A. Hattori, and C. P. McRoy. 1970. "Nitrate and Nitrite in
Interstitial Waters of Eelgrass Beds in Relation to the Rhizosphere." J.
Exp. Mar. Biol. Ecol. 47:191-201.

Koike, I., and A. Hattori. 1978. "Simultaneous Determinations of Nitrification
and Nitrate Reduction in Coastal Sediments by a 15N Dilution Technique."
Appl. Environ. Microbiol. 35:853-857.

Koike, I., and A. Hattori. 1979. "Estimates of Denitrification in Sediments
of the Bering Sea Shelf." Deep Sea Res. 26:409-415.

LIMER 1975 Expedition Team. 1976. "Metabolic Processes of Coral Reef Communities
at Lizard Island, Queensland." Search 7:463-468.

Mague, T., and 0. Holm-Hansen. 1975. "Nitrogen Fixation on a Coral Reef."
Phycologia 14:87-92.

Oremland, R. S., and B. F. Taylor. 1975. "Inhibition of Methanogenesis in
Marine Sediments by Acetyllene and Ethylene: Validity of the Acetylene
Reduction Assay for Anaerobic Microcosms." Appl. Microbiol. 30:707-709,

Oren, A., and T. H. Blackburn. 1979. "Estimation of Sediment Denitrification
Rates at In Situ Nitrate Concentrations." Appl. Environ. Microbiol. 37:174-
176.

Payne, W. J., and M. A. Grant. 1982. "Influence of Acetylene on Growth of
Sulfate-Respiring Bacteria." Appl. Environ. Microbial. 43:727-730.

Risk, M. J., and H. R. Muller. 1983. "Porewater in Coral Heads: Evidence for
Nutrient Regeneration." Limnol. Oceanogr. 28:1004-1008.

207

Seitzinger, S. P., S. W. Nixon, M. E. O. Pilson, and S. Burke. 1980.
"Denitrification and N20 Production in Near-Shore Marine Sediments."
Geochim. Cosmochim. Acta 44:1853-1860.

Szmant Froelich, A. 1983. "Functional Aspects of Nutrient Cycling on Coral
Reefs," pp. 133-139. In: M. L. Reaka (ed.), The Ecology of Deep and Shallow
Coral Reefs, Symp. Ser. Undersea Res., Vol. 1(Ty NOAA Undersea Res. Program,
Rockville, Md.

Tam, T. Y., and R. Knowles. 1979. "Effects of Sulfide and Acetylene on Nitrous
Oxide Reduction by Soil and by Pseudomonas aeruginosa." Can. J. Microbiol.
25:1133-1138.

Taylor, B. F. 1983. "Assays of Microbial Nitrogen Transformations," pp. 809-
837. In: E. J. Carpenter and D. G. Capone (eds.), Nitrogen in the Marine
Environment. Academic Press, New York.

Webb, K. L., W. D. DuPaul, W. Wiebe, W. Sottile, and R. E. Johannes. 1975.
"Enewetak (Eniwetok) Atoll: Aspects of the Nitrogen Cycle on a Coral Reef."
Limnol. Oceanogr. 20:198-210.

Webb, K. L., and W. J. Wiebe. 1975. "Nitrification on a Coral Reef." Can. J.
Microbiol. 21:1427-1431.

Weiss, R. F., and B. A. Price. 1980. "Nitrous Oxide Solubility in Water and
Seawater." Mar. Chem. 8:347-359.

Wiebe, W. J., R. E. Johannes, and K. L. Webb. 1975. "Nitrogen Fixation in a
Coral Reef Community." Science 188:257-259.

208