ABSTRACT 
Title of Thesis: Vertical Resource Partitioning
 and 
Sexuality of Three Sympatric Species of 
Red Sea Sandfishes <~yrisb1Y? 
~~l~~QE~?, Labridae; Trichonotus nikii, 
Trichonotidae; ?Qr9~?l~ sp., Congridae) 
Marianne Martha Krall, Master of Science, 1988 
Thesis directed by: Dr. Eugenie Clark, professor 
Department of Zoology, UMCP 
Three species of marine sandfishes were studied 
from 
1984 to 1986. Their inter- and intraspecific behavior 
was monitored during the daylight hours to exam
ine 
interactions that could result in the vertical 
stratification of the species over the sandy bo
ttom 
within the fringing and patch reefs in the north
 Red Sea. 
orizontal plankton tows were taken at three heig
hts and 
H
three times a day. These samples were compared to 
stomach contents of the fishes to determine the 
trophic 
relationships in the community and their affects
 on 
spatial relations between the species . Prey
 
ecificities of the fishes were determined by usi
ng an 
sp
electivity measure. Using paraffin h istol ogy, ~~CiSQ!~? 
~gl~DQQ~2 wa s determined to be a monan
dric protogynous 
hermaphrodite and Trichonotus nikii, a gonochor
ist. 
ty of 
Previous work on the mating systems and territo
riali
all three sandfish species helped in part to ex
plain the 
vertical spatial arrangement of the sandfish sp
ecies 
ta 
within the community. Effects of pollution 
on the b io
of the Northern Gulf of Aqaba are noted. 
VERTICAL RESOURCE PARTITIONING AND SEXUALITY OF 
THREE SYMPATRIC SPECIES OF RED SEA SANDFISHES (XYRICHTYS 
MELANOPUS, LABRIDAE; TRICHONOTUS NIKII, TRICHONOTIDAE; 
?QB?B?lB SP., CONGRIDAE> 
BY 
MARIANNE MARTHA KRALL 
I I ( 
Thesis submitted to the Faculty of the Graduate School 
of the University of Maryland in partial fulfillment 
of the requirements for the degree of 
Master of Science 
1988 
C,J 
Advisory Committee: 
Professor Eugenie Clark 
Associate Professor Marjorie L. Reaka 
Associate Professor Eugene B. Small M; /1hid 
32 I 
.;VI~ J 
kt , l 
/0.J0. 
Fr 'c 
ACKNOWLEDGEMENT S 
Many people and organizations were e xtremely help f ul 
t o me during the course of this research. I am g r ateful 
t o my advisor, Dr. Eugenie Clark for scientific g uidance 
during my studies and for instilling a great appreciation 
for the Red Sea marine environment in me. I thank my 
committee members: Drs. Marjorie Reaka and Eugene Small 
fo r interesting discussions and for assisting in the 
planning and methods used in the study. Dr. David Allan 
generously allowed me the use of his plankton nets. Dr. 
Harris Linder patiently taught me the art of paraff in 
histotechnique and allowed me unlimited access to his 
laboratory equipment. Dr. Paul Larsen inspired the 
k nowledge of histology in me for which I am truely 
grateful. In my times of dire stress, Dr. David Inouye 
was willing to spend time to help me prepare the numerous 
figures for this paper. I am indebted to the USDA 
statistics group, especially Dr. Larry Douglass, for 
advice, patience, and concern during the frenzy of data 
analysis. I thank David Shen and Dr. David Fridman for 
graciously capturing the much needed specimens of 
X~richt~s melanoeus and Trichonotus nikii. Lucy 
Winchester made it possible for me to spend the summer of 
1985 at the Red Sea by financing my journey. 
Considerable field help in censusing fishes and providing 
logistic support was contributed by Bruce Caylor, Carol 
Falck, Akiva Cohen, Ray Jarvis, Jim Stanton, and Gavin 
l l 
=---
Naylor. I am deeply indebted to Martha J\Jizinski for 
keeping me sane during the preparation 
ii of this paper. The support of my friends, Laurie 
Powers, Kathy Krall, and Paulette Levantine, was much 
appreciated. The Egyptian government donated logistic 
support, and room and board during my two field seasons. 
Funding for the presentation of this study was provided 
by the University of Maryland Graduate School and the 
American Society of Zoologist . 
I reserve the most heartfelt thanks for my parents, 
Albert and Bettijane Krall for all of the intellectual, 
moral, pratical, and generally invaluable support they 
have given me. 
i i i 
Table of Contents
 
Page 
Vi 
List of Figures. 
Vii 
List of Tables 
Introduction. 
l 
 .? 6 
Habitat Descripti
on
 Sites 
sical Characteris
tics of the Study 12 
Phy
of the Study Site
s . 
ics 14 
Biotic Character
ist
cus Species of Fi
shes . 19 
Fo
blage Structure. 29 
Field Observation
s of Assem
Histological Exam
ination of 
Chapter 1 
ermaphroditism in 
~~Cl~b!~? 
H
richonotus 
~~l~~QQ~~ and T
nikii 
33 
Introduction. 
Methods 
hs) 
gonads and stomac 35 
Fish Preparation 
(
36 
Histological Ana
lysis. 
37 
Results . 
48 
Discussion. 
y of the Sandfish
es and 
Chapter 2 Feeding
 Ecol og n the 
Food Resource Av
ailability i
d Environment of 
the Red Sea 
San
52 
Introduction. 
Methods 
ach Contents Ana
lysis. 54 
Stom
5 
Plankton Analysis
. 5
istical Analysis 
? 58 
Sta t
iv 
(Table of Contents , c an ' t) 
El ectivity Analysis . 59 
Results 
Sto mach Co ntents Analysis. . . . . . 6 0 
Plankton Analysis. 67 
Discussi on .. 77 
Literature Cit ed . . . 86 
Appendix 1 94 
Appendi x 2 95 
App endix 3 . . . . 97 
V 
LIST OF FIGURE S 
Figure Page 
1. Map of the Northern Red Sea. 10 
2. Size classes of Trichontus nikii. 25 
3. Diagram of the vertical stratification o f 
X~richt~s melanoeus, Trichonotus ni k ii and 
?QCQ~?i~ sp. in the water column over th e 
sandy bottom. 30 
4. Histological sections of the gonads of 
Tr1chonotus nikii. 
A: T. ~i~ii ovary 
B: T . ~i~ii primary testis 40 
5. Primary testis of Trichonotus Di~ii showing 
the position of the two lobes to the sperm duct 
CSD). 43 
6. Histological sections of the gonads of ~~CiS~~~? 
~~l~~2E~?-
A: X. ~~l~~2E~? ovary 
B : X. ~~l~~Qe~? testis 45 
7 . Chesson ' s (1983) index of electivity measured for 
~ - me lanogus , T. Di~ii, and ?Q[9~2i~ sp.. 74 
vi 
LIST OF TABLES 
Table Page 
 Fishes seen in the san
d environment at the 
1. orth Red Sea. 5 
three study sites in the
 n 1
olume of prey in the stom
achs of 
2. Percent (?) v
species of fishes. 
61 
overlap of habitats of t
hree 
3. Analysis of the 
s of fishes using the fo
rmula for 
specie ition), 
ALPHA(ij) (the coefficie
nt of compet
 Levins (1968) . 66 from
Plankton taxa (? of tota
l sample) found at 
4. e sand bottom. 68 
each site by height over
 th
m the analysis 
5 . Mean abundances of t
he taxa fro
 plankton data by height
 
of variance of the
over the sand bott om . 
70 
vii 
INTRODUCTION 
earch of tropical marin
e fishes focu s ed on 
Early res
hich inhabit coral reefs
 <e.g. Bardach, 
those species w
. 
 Randall, 1961, 1963, 1
965; Wainwr ight, 1965) The 
1958;
in comparison to 
reef community is extrem
ely diverse 
arine communities, thus,
 studies to delineate 
temperate m
pecific interactions of
 
ters
the basis of productivi
ty, in
s were first attempted 
on 
onship
fishes and resource rel
ati
, 
 reef ecosystem (e.g. R
andall, 1961, 1963, 1965
the coral
mming from work by Gau
se 
1967). Competition theory st
e
iamond 
1934), Hutchinson (1959
>, MacArthur (1970>, D
<
d by ecologists, in the
 
(1975) and others, was 
examine
 well 
context of fish commun
ities and was primarily
 food resource limitatio
ns were the 
accepted. Space and
ractions being describe
d by 
inte
basis for the co mpetitiv
e 
 structure 
most researchers invest
igating fish community
Anderson et al., 1981). <e.g. Molles, 1978; 
of 
ical fishes on reefs we
re thought to be part 
Trop
x community", a biologi
cal community that 
a "clima
 and 
ts and changes little fr
om year to year (Odum
persis
oldman and 
Odum, 1955? Hiatt and S
trasburg, 1960; G
' 
In the mid- 1970's, oth
er researchers 
Talbot, 1976). 
tence of space limitatio
ns and 
is
began questioning the e
x
ax assemblages of fishe
s. Their 
the existence of clim
ll 
cale studies (testing h
ypotheses on a few sma
small s
 conspicuous changes 
coral head replicates) 
tended to find
and season to 
in fish assemblages from
 year to year 
1 
season (Sale, 1972, 1975, 1977, 1978a, 1978b, 1980, 1982; 
Sa le and Dybdahl, 1975,; Sale and Williams, 1982; 
Williams, 1980). Opponents to Sale's ideas argued that 
the scale of his studies were too fine and if one 
examined larger ranges, stability of the assemblages 
overwhelms the small scale fluctuations (Gladfelter and 
Gladfelter, 1978; Anderson et al., 1981; Ogden and 
Ebersole, 1981). The movement away from explanations of 
community structure based primarily on competition 
between species had begun. Deemphasis of competition 
structured communities allowed other researchers to 
present their ideas of how communities are structured. 
Among these ideas, Connell (1961) and Paine (1966) 
spearheaded the concept of the keystone predator. The 
keystone predation theory states that, predation on the 
dominant species in the community allows less dominant 
species to exist and thrive where they might not be able 
to if the dominant species numbers were not being cropped 
by the predator. Thus, by cropping the numbers of the 
dominant species, the predator disengages competition 
between the species and allows a higher diversity to 
exist. 
Other factors such as environmental disturbances 
allow increased diversity in coral reef communities 
(Connell, 1975, 1978). Connell 's hypothesis states that 
if enviro mental disturbances occur at an intermediate 
frequency and intensity, competitive exclusion by 
2 
dominant species over less dominant species does not 
The periods of quiesescence (reestablishing of 
occur. 
species ranges after destruction from the disturbance) is 
not long enough to allow for exclusion of subordinant 
species by dominant species. Therefore, according to 
Connell (1978), intermediate disturbance does not allow 
the establishment of a climax community. High diversity 
is maintained by environmental disturbance. 
Other basic environmental variation can cause 
changes in community structure of marine animals. 
Variation and c h ange of current strength can cause shifts 
Also, movement of 
in ranges of planktivorous fishes. 
sand patches and depth variation of sand deposits by 
disturbances can force sand inhabitants to shift their 
home ranges (Clark, pers. comm.). 
Therefore, in the context of marine fish 
communities, many factors may be involved in structuring 
the species ranges and associations. Space and resource 
limitations may cause competition-based community 
Predation may limit a species abundance or 
structure. 
Other 
the plasticity of the ranges of prey species. 
environmental factors such as disturbance intensity and 
frequency, currents, topography, temperature, light 
intensity can influence the assemblage structure of 
marine coral reef fishes . 
The main emph asis of the community ecology of fishes 
has been wit h in coral reef systems. But, within the last 
3 
decade, researchers have begun describing the factor s 
involved in structuring fish populations and communities 
that inhabit the sand environment bordering the reefs 
<e.g. Clark, 1971a, 1971b, 1972, 1975, 1980, 1983a, 
1983b; Clark et al., 1968; Clark et al., in press; Clark 
and Shen, 1986; Clark and von Schmidt, 
1966; Nemtzov ' 
The processes of predation and competition for 
1985). 
food and space resources are also being examined in the 
context of sand fish ecology. 
In the Red Sea, fringing and patch coral reefs are 
numerous. Bordering all of the reef environments, large 
sand patches exist that are inhabited by many 
Field research on the fishes 
invertebrates and fishes. 
of the Red Sea's sand environments have illuminated 
species abundances, intraspecific social systems, and 
sexual life histories (e.g. Clark, 1968, 1971a, 1971b, 
1983; Clark and Shen, 1986; Nemtzov 
1972, 1975, 1980, ' 
Interspecific interactions between species of 
1985). 
sandfishes have been mentioned in passing, but no 
research focusing on the interactions themselves has been 
completed. 
The structure of assemblages of marine sandfishes 
can be influenced by intraspecific and interspecific 
competition, predation, and environmental variations. 
Examination of all of the factors involved in structuring 
Most research on 
a community can be overwhelming, 
communities have isolated one to a few factors which 
4 
influence structure and then manipulate them to see 
changes in community structure subsequent to the 
manipulations (e.g. Paine, 1966; Connell, 1961, Brock ' 
1979) . To begin to investigate a community of sandfishes 
that inhabit the sand patches bordering coral reefs, I 
focused on three species of sandfishes and their spatial ' 
trophic and interspecific relationships. Although many 
more factors are probably intertwined in influencing the 
community structure of these fishes, this study focused 
on the diet and food resource availability in the sand 
environment and its influence on community structure of 
the sandfishes. 
Investigation of these species of fishes has been a 
continuing goal of Clark and others (Clark and van 
Schmidt, 1966; Clark, 1971a, 1971b, 1972, 1980, 1983a, 
1983b; Randall, 1965, 1981; Nemtzov, 1985; Shimada and 
Yoshino, 1984). Within the scope of the investigation of
 
factors that may influence community structure of the 
three species of sandfishes, further characterization of 
sex change within the genus Xyrichtys <Xyrichtys 
mgl~~QQ~?) was also investigated and determined. 
Within the Red Sea sand environment, the genus 
~YilSb!Y? is represented by~- Qentadactylus, ~-
melanoQus, X. Qavo, and~- ~l9~I? Both~- mgl~~QQ~2 and 
~- E~~Q exhibit haremic territoriality <Clark, 1983b; 
Clark and Shen, 1986). This social system is similar to 
X. eentadactylus <Nemtzov, 1985). An inves
tigation of 
5 
hermaphroditism in other species of ~t~i~Q~t? in the Red 
Sea, aside from~- Qentadactylus, has not been attempted. 
X. ~gl~~9e~? and~- e~~2 are rare species in comparison 
to X. eentadactylus. Completion of the analysis of 
hermaphroditism in this genus is hampered by the lack of 
sufficient specimens. However, over two summers, enough 
X. melanoeus specimens were collected to complete a 
histological analysis of the male and female gonads of 
this species. 
The goals of the present study were to (1) observe 
several sympatric species of sandfishes within a 
community, noting interspecific and intraspecific 
interactions; (2) describe the community structure of the 
sandfishes within the sand environment and possible 
structuring forces; (3) examine behavioral limitations to 
the species ranges; (4) relate diet and food resource 
availability from the plankton to the community structure 
of the sandfishes; (5) continue investigation of the 
sexuality of the genus ~Y[lfb!y2 by histologically 
examining the gonads from X. melanoeus for evidence of 
hermaphroditism; and (6) examine I- ~1~11 for evidence of 
hermaphroditism that is theorized by Shimada and Yoshino 
(1984), but has never been experimentally determined. 
HABITAT DESCRIPTION 
The Red Sea is a geologically young, flooded rift 
valley created by the past division of the African and 
Asian continents <Kennett, 1982; Morley, 1975). The 
6 
basin is connected with the Indian Ocean at the straits 
of Bab - el - Mandab in the south and with the Med iterranean 
Sea by the man-made Suez canal in the north. The 
Northern Red Sea is split into two gul fs that surround 
the Sinai peninsula. To the west of the pen insul a is the 
Gulf of Suez, which is fairly shallow (avg. 40m). To the 
east of the Sinai peninsula is the Gulf of Aqaba. This 
gulf is the site of active ocean spreading, thus, the 
trough is very deep ( in excess of 1800m) (Kennett, 1982). 
The Gulf of Aqaba is bordered by fringing reefs in the 
south and patch reefs and c oral heads in the north. The 
salinity of the Northern Red Sea is approximately 41ppt 
The salinity is much higher than the 
<M or 1 ey , 1 975 ) . 
Pacific, Atlantic, or the Indian Oceans which average 
between 33-35ppt (Morley, 1975). The presence of an 
elevated salinity in the Red Sea can in part be 
contributed to comparatively little water runoff into the 
sea from bordering 1and masses and the lack of major 
Interspersed between and 
rivers flowing into the sea. 
along side of the reefs are three types of sand 
formations described by El Baz <Clark, 1983a). 
The sands of the Red Sea are derived from different 
sources and exhibit different ch Ta her  a fc it re sr tis  tics. 
It is derived from flash 
of the sand types is wadi sand. 
flood runoff from desert mountains of the bordering land 
The grains are very angular and loose
m lyas  ses. 
The second type i s "current 
compacted (Clark, 1983a). 
7 
deposit" sand. It is carried great distances by 
nearshore currents and deposited in depressions on the 
fringing and patch reefs and between coral heads. These 
current deposits contain few skeletal or coral fragment s 
and thus, appear darker than the other sand types. Along 
the Gulf of Aqaba, at Ras Mohammed, these "current 
deposit" sands are overlapping and to a degree intermixed 
with other sands, (especially coral sand) (Clark, 1983a). 
The grains of "current deposit" sand are rounded from 
erosion during travel in the currents. The third type of 
sand is coral sand. It is angular and less well - sorted 
than the other types of sand . Coral sand contains many 
foraminifera and other complete and incomplete shells in 
addition to tabular and branching coral skeleta <Clark, 
1983a) . The coral sands are produced by fishes, 
(especially parrot fish, Scaridae; Caris angulata, C. 
1~0~1~, and triggerfishes of the genera Rhinecanthus and 
~~ll?~~?) which ingest stony coral while feeding on the 
symbiotic algae of the corals <Fishelson, 1968). The 
carbonate material is ingested by the fish and is passed 
through its intestinal tract. The ground coral fragments 
are defecated into the water column where they sink and 
collect in depressions on the reef structure. Another 
important source of coral sand is from sponges and other 
boring organisms. The importance of the sand type in an 
organism's habitat selection has not been thoroughly 
investigated . 
8 
s of marine animals are
 residents o r 
Many specie
 to 
Visitors in the sand en
vironment. In comparison
organisms 
reef dwelling organisms
, sand dwelling are 
cies 
dull in coloration. Most of t
he sand dwelling spe
blend into the sand 
are cryptically colored
 to 
e 
environment. Cryptic color
ation is a preventativ
measure to avoid predat
ion (Clark, 1983a). 
in the north Red Sea, th
ree study sites were 
With
h community 
chosen within which to 
study sandfis
d 
eractions. The three sit
es are (1) Ras Mohamme
int ' 
 
a el Mukibela (= Muqab
ila), and (3) Aquasport
(2) Mars
el) 
dive center in the Gulf
 of Aqaba, Eilat, Isra
<a 
ing to the 
<Figure 1). The sites w
ere chosen accord
g criteria: (A) At leas
t two of the three 
followin
n the sand 
sandfish species in the
 study were present i
ss to the 
environment, and (8) we
 could obtain acce
area. 
9 
orthern Red Sea. The
 sites where 
Figure 1. Map of th
e N
are noted on the ma
p 
 
the study was condu
cted
 Northern Red Sea. 
Ras Mohammed is 
1n the
arked by a STAR, Ma
rsa el Mukibela is 
m and Aquasport is 
marked by an ASTERI
SK, 
. 
marked by a solid C
IRCLE
10 
,,.1 
MEDITERRANEAN SEA ( 
() 
\. 
\ . I.  
\ ISRAEL ( 
? I 
\.\  .i  
I?,  / ( 
\ \ 
\ ! 
\_\  .'  
\.  I!  
\ . 
\ I 
SINAI ?, /JORDAN 
SAUDI 
ARABIA 
RED SEA 
* -RAS MOHAMMED ?-MARSA EL MUKIBELA 
e -AQUASPORT (CORAL SEA) 
11 
PHYSICAL CHARACTERISTICS OF THE STUDY SITES 
Ras Mohammed is located on the southern tip of the 
0 0 
Sinai peninsula (27 43'N, 34 16' E T). h e sand slope 
(20 0 ) used during the observations and collections begins 
at the base of a coral cave opening at - Sm and continues 
to a depth of -47m when it begins sloping upward again 
<C Ala prk p roa xn id m aS th ee ln y,   5p 0e  r ?s  . o fc  omm.>. 
this sand environment is surrounded by a fringing coral 
The sand at Ra
r se  e Mf o. h ammed is composed of coral and 
"current deposit" sa Tnd hs e re<  C il sa  rk l, i tt1 l9 e8  3a). 
if any occurence of sea grasses on the slope although an 
abundance of dead sea grass blades are swept on and off 
of the slope by currents. 
Marsa el Mukibela is located approximately 30 km 
south of the Taba border on the east coast of the Sinai 
0 
0 
N, 34 42 ' 50 ' ' E) . The in the Gulf of Aqaba (29 22'20'' 
sand slope used in the study begins at a depth of -2m and 
Near the shore, the 
continues to approximately -48m. 
bo Ftt ao rm th es rl  o fp roe ms  at an angle of five degrees. 
shore the slope drops off at an angle of 22 degrees. The 
sand environment is sprinkled with a few small coral 
heads in the most sh Aa ll sl oo  w regions of the slope. 
immediately to the south of the sand environment in deep 
water (approx. -15m) is a fringing reef. The sand at 
Marsa el Mukibela is composed almost entirely of wadi 
The slope is located at the base of Wadi Magresh. 
sand. 
During winter storms, wadi sand washes down into the bay 
12 
onto the site. The slope is well populated with s ea 
stabilize the sand surface. 
Aquasport is located in the Gulf of Aqaba 
approximately 6 km north of the Taba bord e r in Israel 
0 0 
(34 56' E, 29 30' N >. The site is immediately north of 
the "Coral Sea Reserve" in an area used by the Aquasport 
dive shop. The sand slope, ranging from approximately -
2m to - 30m, begins just beyond a rock and algae covered 
shore. The sandy bottom slopes at an angle of 
approximately eight degrees. Beyond the slope, the 
bottom drops off steeply toward the Gulf's shipping 
channel. The site has numerous coral heads interspersed 
with sand expanses in the shallow and deeper regions. 
The sand at the site is variable. The top layer is silty 
and below are coarse sand grains and rocks mixed in. The 
sand is dominated by coral and wadi sand mixed with small 
rocks from the shore. The grain is the most coarse of 
the three sites in this study. There is a conspicuous 
lack of sea grasses at the Aquasport site. The absence 
of seagrasses, which were previously reported to be 
abundant by Clark (1971a, 1971b, 1980), is thought to be 
due in part from recent oil spillage and record low sea 
levels suffered by the Northern Gulf of Aqaba (Fishelson, 
1973; Loya, 1975). 
13 
BIOTIC CHARACTERISTICS OF THE STUDY SITES 
During the summer of 1984, survey of the fishes that 
were resident and visitors near the sites was begun and 
was finished in the summer of 1985 <Table 1) . The fishes 
r ecorded in the survey are either resident or visitor s in 
the sand environment or on coral heads located in 
intimate contact with the sand environment. Residenc e 
status was given when the fish species was observed on or 
near the sand environment during the majority of the 
observation periods. In no way can this survey be 
considered a complete listing. Rather, the fishes listed 
are the most conspicuous to the researchers . Cryptic 
species and those that hide wi thin the reef are probably 
underrepresented . All of the survey periods were done 
during the daylight hours. 
The following biotic descriptions of the three study 
sites focus on the three species of sandfishes which are 
the basis of the study. The three sandfishes were chosen 
because their territories within the sand environment 
were sympatric, although during daylight they seemed to 
be vertically separated in the water column. The 
sandfishes that were the focus of this study were easy to 
observe because they remained in nearly the same place in 
the sand environment day after day. Also, all of the 
fishes could be observed by SCUBA divers without 
disturbing their natural behavior. If the divers laid 
quietly on the bottom nea r the reef, the animals would 
14 
Tab l e 1 . F i s h es s een i n the san d e n vir onmen t at t h e 
th ree stud y s i t e s in the north Red Sea . The 
fis h es ar e l ist ed b y f a mi l y, gen us, and 
spec ies , where po ss ible. 
V = Vis itor s pec ies ( seen occasi o na ll y) 
R = Resi dent species (seen du ring t h e 
ma j ority of observa ti o n p er i ods ) 
FAMILY GENUS AND SPECIES 
Acan t huri da e Acanthurus spp. (V) 
Naso uni c o r nus (V) 
Zebrasoma xanthurum ( V ) 
Atheri nidae Atherinomorus lacunosus (V) 
Bal is t i dae Balistaeus undulatu s <V> 
Sufflamen alb ic audatus (V ) 
-R-h -i -n-e-c-a-n-th--u-s -a-s-sa--s -i ( V ) 
Belonidae Tylosurus choram (V) 
Bo t h i d ae Bothus eantherinus (R ) 
Caranidae -C-a-r-an-x- -s-e-x-f-a-s-c-i-a-tu- -s (V) 
Caseion i dae -C-a-s-e-io- -lu- -n-a-r-is-  (V) 
Chaetod o ntidae Chaetodon auriga (V) 
-C-h-afb~~-
e
~-
to-d-o-n- austriacus (V) QQQ~ fasciatus ( V) 
Chaetodon lineolatus < V) 
Chaetodon melannotus < V) 
Chaetodon semilarvtus < V) 
Heniochus intermedius ( V) 
Pomacanthus imeerator (V) 
Pygoelites diacanthus (V) 
Cirrhitidae Oxycirrhites tyeus (R) 
Congridae ?Q[9~2i~ sp. (R) 
Fistulariidae Fistularia c ommersoni i (V ) 
Gobidae Amblyeleotris steinit z i (R) 
Haemulidae Plectorhynchus gaterinus (V) 
Hemiramphidae Hyeorhamehus gambarur (V) 
Holocentroidae Holocentrus diadema (R) 
Kyphosidae Kyehosus cinerascens (V) 
Labridae Cheilinus undulatus (V) 
Coris angulata (R) 
Gomehosus coeruleus (V) 
Th allasoma klunzingeri (R) 
~yrichtys melanoeus (R) 
~Y[ifb~Y? ~lg~[ (R) 
~Y[ifb!Y2 E~~Q <R> 
Xyrichtys eentadactylus <R> 
Lutjanidae Lutjanus bohar (V) 
Mullidae Eseudueeneus forsskali (V) 
Muraenidae Siderea grisea (V) 
Ostraciidae Ostracion cubicu s (V) 
15 
<Families of Fishes, cont.) 
FAMILY GENUS AND SPECI
ES 
Eur~eegasus draconi? < V) Pegasidae 
ae Abudefduf saxatil
i? < R) 
Pomacentrid
Amehierion bicinct~? < R) 
Chromis caerulea (R) 
Chromis dimidiatus (R) 
Chromis trialeha <R> 
Dascyllus trimaculat~? (R) 
Dascyllus marginatus <R> 
ae Pseudochromis fri
dma~l (R) 
Pseudochromid
Cetoscarus bicolor <R> Scaridae 
Scaridae ?S:~C~? sp
p. < R > 
rois volitans (V) 
Scorpaenidae E
te
Eterois radiata (V) 
Scoreaenoesis barbatu? < V) 
?~nanceia verrucosa <R) 
Anthias sguamieinnis (R) Serranidae 
~eehaloeholis argus (V) 
Ceehaloeholis miniata (V) 
Plectroeomus truncatus (V) 
~ariola louti <V> 
 ?eh~raena f
lavicauda (V) 
Sphyraenidae thoicthys schultzi (R) 
Sygnathidae 
~ory
ae ?ynodus 
variegatus <R> 
Synodontid Arthron diadematus (V) 
Tetradontidae ecthron hiseidus (V) 
~~nthigaster corinatus < V) 
 Icichono
tus nikii <R> 
Trichonotidae
16 
surface from the sand and feed, mate, etc . Another 
factor used in selecting the focus species was that each 
of the fish species only surfaces from the sand in the 
d Ta hy ul si ,g  h wt h eh n o su tr us d. ying the behavior of 
these fishes, one did not have to worry about missing any 
critical feeding and interspecific behaviors during the 
night. 
On the Ras Mohammed site, the sand slope is covered 
wit Th h ea   e" eg la sr ,d  e ~n Q [o 9f ~ ?e l~e  ls s. p" . , are 
represented on the site by what may be in excess of 9,000 
The eel colony ranges fro
in md  i av  i dd eu pa thls  . o f -7m 
to -47m and then up a slight rise ending before dropping 
The eels thin out as one approache
in s to th ed  eep water. 
coral structures surrounding the sand. The lack of eels 
near the coral is thought to be due in part by the lack 
of current (therefore, a plankton supply) and the 
topography of th Te he  r ae def u ls t tr eu ec lt su  re itself. 
are most abundant toward the center of the colony and the 
immature eels are sometimes seen within the adult 
community but, most often near the outer edges of the 
adult community <Clark, pers. comm.> 
A swarm of Trich
- o- n- o- ~- ~- ?-  - Di~li with approximately 200 
They inhabit a 
individuals is present on the site. 
portion of the sand environment at a d the  p ranging from 
approximately -15m to -28m. 
A pair of ~iCl~QEi~ ~~l~~ee~~ is also present at Ras 
The razorfishes cruise over the sand bottom 
Mohammed . 
17 
and h ol d terri to ries at a dep t h rang e o f -31 m to -40m . A 
close re l ative of ~ i- s also an 
inhabitant of Ra s Mohammed. A ha r em compri sing five 
indi vi d uals (1 male , 4 female s > r a ng es f rom a dep t h of 
-29m t o -37m. 
At t h e Ma rsa e l Mukibe la s ite, a c olony o f ?Qcg~~i~ 
sp . that was previ ousl y n o t ed i n the li te r a t u re by 
Nemtzov (1985) a nd Cl ark (1983a ) ha d disap peared by the 
summer of 1985. 
over 500,000 indivi d uals (Clark, 1983a) do minates the 
water column from a depth of -3m to -15m. 
X~richt~s melanoQ~? is present with approximately 5 0 
i n di v idua l s i n t he depth r ange s of -6m t o -14m. 
Indiv i duals of the ~t~i ~~! t ? genus, ~ - e~~!~Q~~!~l~? , t h e 
focus of Nemt zov's (1985) work on sex c hange and so c ial 
X. e
b ~e ~h Qa ,v  i wo hr i, c h is i s hi ag lh soly   aa  bundant. 
haremic fish species (Clark and Shen, 1986 ) , i s also 
pres ent at Marsa el Mukibela. 
razorfish, has territories which are widely spread out 
over the sand environment at Marsa el Mukibela. 
Aquasport sup ports a colony of approximately 12 0 
?grg~?i~ sp. individuals- The colony was first 
documented by Clark (1971a) as contain i ng over 1,500 
The sharp decrease in numbers of garden 
e eee ll ss  . could 
be tied to the loss of seagrasses and the general poor 
condition of the sea life in this area, but a direct tie 
between these observations has not been clearly 
18 
es t a bli s h ed . Th e c olony of e e : s r ang es 
f r om a dept h of 
- 4m t o - 9 m. 
A small s warm of Ic i c h ono tu s nikii with 
appro x imately 2 0 individuals i s p r e s ent on the northern 
boundary of the eel c olony . Th e T. Dl~ll swarm hovers 
between - 6m and - Sm. There wer e no ~Yrlchtys melanoQ~? 
present at the Aquasport sit e but, ~- E~~Q adults and 
immatures were pre s ent at th e s ite. The razorfish were 
never seen diving within the ee l c olony limits, but they 
were frequently seen cruising through the colony during 
the observation periods . 
All three sites sup port an abundance of marine life 
within and above the sand y bottom. When compared, the 
study sites have major differ enc es. The depth at Ras 
Mohammed is significantly dee p e r than Marsa el Mukibela 
or Aquasport. The sites a r e s iITTilar in the presence of 
the focus species, but the relative abundances of the 
species are not similar betwe n sites. The garden eels 
dominate in abundance at Ras Mohammed, I? Dl~ll dominates 
in abundance at Marsa el Mukibela and at Aquasport, both 
I- Qi~ii and the garden eels are present in low 
abundances in comparison to the other two sites. 
FOCUS SPECIES OF FISHES 
The follo wing is a description of the habits and 
behaviors of the three focu s s pecies in this study, 
X~richt~s melanoeus, Trichonotu s nikii, and ?ecge~ie sp .. 
19 
~yrichtys* melanoeu?, <* Briggs, 1961 points out the 
spelling of ~Yri~b!Y? that must be followed) a razorfish 
of the Family Labridae, inhabits sand environments from 
the Red Sea to Japan (Indopacific). It is a rare species 
of the genus ~Y~l~bit? which has at least 10 Indopacific 
and Red Sea species (Masuda et al., 1985; Dor, 1984) and 
three in the Atlantic and Caribbean (Bohlke and Chaplin, 
1968; Randall, 1981). Razorfishes are known for their 
sharp keel-like foreheads which they use to dive head 
first into the sandy bottom with the approach of danger 
(Randall, 1965). They are diurnal fishes, only surfacing 
from the sand during the daylight hours. 
The razorfishes have prominent canine teeth in the 
front of the mouth, therefore it is not suprising that 
these are carnivorous fishes (Randall, 1967). ~-
~~l~~QE~? has been seen to pick on the sand surface for 
food and also to eat small fishes . 
X. melanoQUS is a territorial haremic labrid with 
dominance hierarchies within the harems (Clark and Shen, 
1986). This haremic social system has been investigated 
by Clark (1983b) in a closely related species, ~-
gentadactylus shown to be a monandr ic protogynous 
hermaphrodite by Nemtzov (1985). Monand ry describes 
hermaphroditism when only one male phase (stage) is 
present within the life history of a sex- changing 
species. Haremi c territories have also been discovered in 
X. Qe~Q in the Red Sea (Clark and Shen, 1986). 
20 
In each ~ - mgl _ ~Q Q~? h r e m, one male controls one to 
five females withi n hi s te rr 1 t o r 1a l boundaries. Al 1 of 
fish on a territory on th sand bottom,
 defend it 
the 
from other~- mgl~~Q~~? ind i v idu a ls present. Th e fish 
cruise at a height of about on e - third of a meter from the 
sand bottom on their territ o r1 e Within a te
rritory the 
razorfish have multiple "di v s l tes" w
here they may 
disappear in ti me s of dange r . dive site i
s a specific 
region in the sand used b y a f i h to bury itself. The 
sites are recognized by the f 1sh and used repeatedly . 
Clark ( 1983b) manipulated t i e 1 rdmarks around many dive 
sites to determine if the fish w re using visual cues to 
unding the 
find the sites. By mo ving the o b jects surro
sites, no noticeable difference in the fish's ability to 
recognize the site was mad e . - u r ther work was attem
pted 
to delineate the sensory cues of dive site recognition, 
but no clear answers ha ve b e en found (Clark, per s comm) . 
X~richt~s melanoeu s is s e xu a lly dimorphic. The 
males are larger than the female s (SL* MALES X = 15.1 
cm, SL FEMALES X = 12 . 7 cm) <* the lengths for all of the 
species were measured after the f ishes had been 
preserved, thus some shrinkage may have occurred) Both 
sexes have a large, c onspicuou s dark red blotch on their 
mid-side below their dorsal fin . Males have a bright 
blue line along the profile of their head (almost absent 
in females) and a more swollen forehead that the females. 
The males lack a mustard yellow blotch on the mid -side ' 
2 1 
The location of t
w heh  ich is characteristic of females. 
mapped, allows for easy identification of the 
males, once 
male during subsequent observations. 
The females have a conspicuous white patch on the 
mid-side under the red blotch and violet triangles on the 
Late in 
s tc ha el  es of the lower part of the white patch. 
day, during spawning season, the females develop a 
swollen belly region due to the large numbers of ripe 
Also, the females have a bright 
eggs that they carry. 
red rim around the anus opening and females show a 
mustard yellow blotch anterior and dorsal surrounding the 
white patch. 
The species is endemic to the sand 
Family Trichonotidae. 
environment of the Red Sea. I? ~l~ll is one of five 
little known species in the genus Trichonotus, whose 
distribution is limited to the Inda-Pacific. 
individuals dive into the sand bottom at the hint of 
The fish do not dive into specific "dive sites" 
danger. 
<as does x. !!)~.!~!: Q!,!?>, but, rather into dive "areas" 
9
This species is also diurnal, 
<Clark, pers. comm.>? 
surfacing from the sand only in the daylight hours. 
Feeding by I? Di~ii is thought to be by selecting 
T.nikii 
plankton from the water column <Randall, 1967). - --- --
forms swarms in the water column of a few to several 
hundred thousand individuals (Clark, pers comm). The 
swarming behavior of the fish tends to confuse pelagic 
22 
 
e over the sand
 environment. Thus, 
h cruis
Predator s whic
ctive to the in
dividual 
nsive behavi o r 
is prote
the defe ii have been 
in the s wa r m. S wa r
ms o f T . ~i~
f i sh wi th
nd s u r fac e , fa
r 
ove the s a
orted to rise 
to over 2m ab
rep
e sites, but a
lso fa r from 
the safety of 
the div
from 
 <Clark and Sh
en, 1986). But, 
ors
l u r k ing benthi
c predat
of 
rs on a height 
ing feeding, t
he swarm cente
dur
ximately 1.7m. appro h has been 
tains a mating 
behavior whic
I? Qi~ii main
n, 1986). Lek
king 
-like (Clark an
d She
described as l
ek
, 1952 ; 
n birds <Patte
rson
avior was firs
t described i
beh
Ballard, 1974; 
and 
am, 1973; Robe
l and 
lli
Lack, 1968; Pu  sand 
f fish lowers 
to the
Lill, 1976). 
The swarm o
-like arena. T
he 
porary lek
surface and se
ts up a tem
h to 
na within whic
claim territor
ies in the are
males ales defend 
ales. The mo
st dominant m
display to the
 fem
f the 
ies in the cen
tral regions o
itor
the prime terr s 
ant males holdi
ng territorie
 domin
arena, with th
e less
ena. The male
s display to 
 to the edges 
of the ar
nearer ose which male
 
 able to cho
 females and t
he females are
the
 period. In a
 true 
uring the lekki
ng
they will mate
 with d
ive, so that on
ce a male is 
 the mating is
 exclus
lek
<Borgia, 1979; 
e mates with th
e female 
chosen, only h
d Bradbury, 19
85). 
, 1983; Gibson 
an
Bradbury and G
ibson
al, it is 
rtilization is 
extern
In -T. nikii 
although fe
-----' 
males' eggs ca
n be 
 that the fe
highly improba
ble
 
an one male be
cause of close
re th
fertilized by 
mo
23 
The external fertilization of eggs in I? -n-i-k-i-i 
pairings. 
lek. 
excludes their mating behavior from being a "true" 
the fact that they dive under the sand in the s a me 
Also, 
site as the leks can exclude this species as a true 
lekking species. 
I? ~i~ii males have been seen to pick at the sand 
This "picking" may 
surface during the lek-like behavior. 
be actual feeding or it may be a "displacement behavior" 
<when males are not chosen for mating), as described by 
True lekking behavior is only for 
Clark and Shen (1986). 
Therefore, if the 
reproduction and not for feeding. 
"picking" behavior of I? DiKii male is actual feeding, it 
could preclude T. ~i~ii from being a true lekking 
species. 
The males are 
T. nikii are sexually dimorphic. 
- -----
larger than the females (SL MALES X = 11.34 cm, SL 
The males are ornately (Figure 2). 
FEMALES X = 6.86 cm> 
The first three dorsal fin rays are elongated 
decorated. 
The anterior base of the dorsal fin is 
and filamentous. 
The male 
dark black in color and rays are striped. 
flares the dorsal fin rays during agonistic and mating 
The pelvic fins of the male are enlarged in 
displays. 
The pelvic fins of the males 
comparison to the female-
are bright white or in sa me cases, bright yellow (on 
Th e body of the male is white with vertical 
dominants). 
saddle marks in brown-b lack, down the entire length of 
The head of the male is decorated with dark 
the body. 
24 
richonotus n
ikij. 
of T 100) 
ure 2. Siz
e classes  
ig of fe
males (N =
F
Standard len
gths (c m) 
. 
and males (N
 = 36)
25 
~a 
(/) S'tl 
w tl 
(/) 
(/) S'tl 
::5 ?l 
u S'?;l 
w ?l 
Nx s? l l (/) 
w 
(/) ~ l l (/) 
(/) 
(/) >- S'Ol 
:::> m 5 
Ol (.) r-- s?s w 0 N 
z (/) 6 
0 
I s?s 
u 8 
et: S'L 
r-
L 
s?g 
9 
s?s 
0 -0 0 N 
AON3n03~j 
26 
s pot s above and behind the eye . 
The female T. 0l~ll is fairly drab in color. The 
body of the female is a plain white-tan. The anterio r 
base of the dorsal fin is dark black, as is the male 's . 
The dimo r phic coloration and size separation o f the 
sizes of the sexes has lead to the speculation of 
possible protogynous hermaphroditism in I? 0l~ll (Shimada 
and Yoshino, 1984) (Figure 2). These characteristics of 
the se xes have been seen in other fishes that are indeed 
sex changing species, but, I? 0l~ll needs further 
testing. 
QQcg~?i~ sp.* (*this species was thought to be Q. 
?ill~~Ci but may be a new species (Fishelson, pers. comm. 
to Clark)) is a garden eel species of the Family 
Congridae, Sub-family Heterocongrinae, which inhabits the 
sand environment. The genus has more than six 
Indopacific species (Randall and Chess, 1979; Abe et al., 
1977; Bohlke, 1957). Garden eels live in colonies on the 
sandy bottom which number from a few to many thousand 
individuals. Each eel burrows tail down in the sand 
vertically. The eel secretes a mucopolysaccaride slime 
which glues the sand grains together to form the walls of 
the burrow (Casimir and Fricke, 1971). The eels do not 
usually leave their burrows . During the daylight hours 
the eels extend over two-thirds of their bodies out of 
the burrow to feed, mate, and defend the territories they 
hold around their burrow opening. When most of the eels 
27 
in a c olony a r e e x tended out of the bur r ows, the c olony 
look s l i ke " blades of grass in a field " r ising f r om t h e 
s a ndy b o ttom. 
If the eels are frightened by a predator, they 
withdraw into the burrows until the danger ha s 
di ss ipated. Also, during the day, portions of the colo ny 
may withdraw at irregular interv?ls. The motivation 
behind this behavior, called a "siesta" by Clark (1980) 
i s still not understood at present (Clark, pers. comm. ) . 
The male eels defend hemispherical territories 
centering on their burrow openings. One to two females 
are allowed within the territorial boundaries defended b y 
a male (Clark, 1980). The territories are fiercely 
defended by the males. Conflicts are characterized by 
two males stretching at least two-thirds of their lengths 
to the disputed boundary while flaring their dorsal fin s 
in a threat display, often followed by strikes at each 
other. 
The eels are thought to be selective plankton 
feeders (Randall, 1967). They extend out of the burrows 
to feed, centering on a height of slightly less than one 
meter* off of the bottom <* within the tables and 
figures the height of ~Q[g~?!~'s sp. feeding is listed as 
one meter, although it is slightly less. Sampling of 
plankton occurred at the height measured in the field 
where most of the eels heads tended to be positioned when 
feeding (slightly less than one meter)). In a strong to 
28 
medium current, their bodies protrude vertically out of 
the burrows, bending so that the head is held 
horizontally facing into the current. The eels are seen 
actively selecting prey from the water column. In weak 
currents, the eels reach to eat prey items passing within 
the bounds of their reach. At dusk, the eels withdraw 
into their burrows until dawn. 
?QC9~?i~ sp. is sexually dimorphic based on size. 
The males are larger than the females <TL MALES X = 88.5 
cm; TL FEMALES X = 63.5 cm>. The males have a swollen 
appearance to the back of their heads in comparison to 
the females. The bodies of both sexes have a mottled 
brown-green color. No dimorphism in coloration is 
apparent. 
FIELD OBSERVATIONS OF ASSEMBLAGE STRUCTURE 
In 1984, field observations of the assemblage 
structure of the sandfishes which was based at Ras 
Mohammed and then subseq uently at the other two study 
sites, illuminated a peculiar stratification of X. 
melanoeus, T. nikii and Gorgasia sp. in the water column 
over the sand bottom. The species ranges <and/or 
territories) were overlapping on the sand bottom 
horizontally, but, when feeding, the species were 
vertically stratified in the water column <Figure 3). 
The spatial arrangement with X. ~~l~~~E~? individuals 
cruising in the water column at a distance of 
approximately 0 . 3m from the sand bottom, ?QC9~?i~ sp. 
29 
Fig ure 3. Diagram of the vertical stratification of 
Xyrichtys melanogus, Trichonotus ni~ii and 
?Q[9~?i~ sp . in the water column over the 
sandy bottom. The stratif ication of the 
fishes is only upheld while the fishes are 
feeding . The height of the center of the 
species distribution when feeding is noted 
on the scale to the left. 
30 
2.0 
1.5 
C 
-E ::, 0 
u 
~
...,,  
 1.0 
"J'  
0.5 
sandy bottom 
31 
r is ing out of their burrows to a height o f appro x imately 
1. 0 m from the sand bottom, and T. ~l~ll swarming at a 
distance of approximately 1.7m (center of swarm) from the 
sand bottom. 
The vertical stratification of the three s pec ies o f 
sandfishes (du r ing feeding) was seen repeatedly du ri ng 
the summers of 1984 and 1985. The vertical layering of 
the fish species broke down during other behaviors such 
as the lek - like mating of I- Di~ii and whenever danger 
( predator lurking or diver disturbance) occurred. 
Possible factors that could be involved in the 
vertical stratification (and possible resource 
partitioning) of the three sympatric fish species 
include foo d reso urc e availability and patchiness, 
interspecific conflict, environmental variation over 
the water c olumn, or physiologic limitations of an 
individual fish species. This study of the assemblage 
structure of X. ~~l~~QQ~?, I? ~i~ii, and ?QCQ~?i~ sp. 
focuses on food resource distribution and 
availability, although the other ideas will be 
discussed. 
In 1985, field observations were combined with 
collections of fish specimens of all three species and 
collections of hori zontal plankton tows. 
32 
Chapter 1 
Histological examination of Hermaphroditism 
in Xtrichtts melanoeus and Trichonotus nikii 
INTRODUCTION 
Hermaphroditism in marine fishes has been discovered 
in at least 10 families (Fishelson, 1970; Fricke,1979; 
Fricke and Fricke, 1977; Hourigan and Kelley, 1985; Leigh 
et al., 1976; Nemtzov, 1985; Policansky, 1982; Robertson, 
1972; Ross, 1978, 1984; Warner, 1978). Sex change can 
occur either by protandry (first male, then female) or by 
protogyny (first female, then male). When sex change 
occurs in a species, one male (monandry) or two males 
(diandry) may occur within the life history. Protandry 
has been found in fish species in which large male size 
is important. When male-male competition is less intense 
and male size has little effect on breeding success, 
large female size outweighs the advantage of large male 
size because female fecundity is always more dependent on 
body size than male fecundity (Warner, 1975). It may 
benefit an individual to start life as a male when it is 
small and change to a fe male when it is older and larger 
Thus, in this case, protandrous 
hermaphroditism acts to increase the inclusive fitness of 
the individual throughout its life span. 
When male-male co mpetition is intense, only the 
largest individuals will be successful at mating. 
Although female fecundity also increases with size 
(larger females produce more eggs), the influence of 
33 
male size on mating success is much stronger. Under 
these conditions it may benefit an individual to be a 
female when sma ll, because all females will breed, and 
a male only when large enough to be a successful 
competitor (Warner, 1975). Protogyny is by far the 
mo st co mmon mode of hermaphroditism in marine fishes 
(Warner, 1978). Within the sand environment, 
p rotogyny was found in ~tCi~b1t? eentadacttlus by 
Nern t z o v ( 1 985) . 
X. eentadacttlus is a haremic ter ri torial species. 
Males defend the territories within which they establish 
" dive sites." Male size is an important factor in mal e-
male competition and therefore, territory size and 
quality. The males with larger territori e s have larger 
harems (Clark, 1983b). Thus, mal e size is directly 
related to mating success. Female size is less important 
to mating success . Clutch volume increases with 
increasing female size, but, mating success is almost 
guaranteed. All of the females within a harem breed with 
the male of that harem. Thus, it is not suprising that 
X. eentadacttlus is a monand ric protogynous 
hermaphrodite. 
Within this s tudy, Xtrichtts melanoeus was 
histologically examined for sex change. 
a haremic social system, thus the idea that X. me lanoeus 
is also a monandr ic protogynous hermaphrodite was assumed 
without previously being tested . An a posteriori 
34 
e xamination of the morphology of the gonads of the males 
and females of X. melanoeus can illuminate the existence 
or absence of hermaphroditism in this species. The 
presence of the "female orientation" o f the testicular 
tissue of the males of the species indicates that se x 
change has occurred (from female to male). The presenc e 
of only one morphotype of the male gonads in the species 
is an indication that monandric hermaproditism occur s . 
Therefore, the gonads of X. melanoeus will be examined 
for morphology and histology to determine the extent (if 
any) that hermaphroditism occurs. 
Also, within this study, Trichonotus ~i~ii will be 
e xamined histologically for hermaphroditism. Without 
testing, I, Dl~ll was considered to be a protogynous 
hermaphrodite on the basis of non-overlapping sizes of 
the sexes (Shimada and Yoshino, 1984) (Figure 2). The 
speculation of sex change can now be replaced by 
morphological and histological evidence in J . -n-i-k-i-i. 
METHODS 
FISH PREPARATION (GONADS AND STOMACHS) 
Each fish was fixed in 10? formalin then transferred 
in stages to 70? ethanol after rinsing. The standard and 
total lengths of the fish were taken after the fish was 
fixed and preserved, thus some shrinkage may have 
occurred, as mentioned previously. After preservation, 
each fish was described morphometricall y and then 
dissected. Trichonotus nikii and Xyrichtys melanoeus 
35 
specimens we r e gutted and their gonads wer e removed. 
Only the alimentary tract was removed from ?9 ~ g~?l ~ s p. 
specimens. Each fish ' s stomach and gonad were placed i n 
a separate vial of 701/. ethanol before being analyzed. 
R ight and left gonads of each fish were stored separatel y 
e xcept in the few cases where the gonads could be r emoved 
together as a unit (esp . Trichonotus nikii). 
HIS TOLOGICAL ANALYSIS 
When histological analysis was begun, the gonad s 
were removed from the alcohol, weighed, and then pla c ed 
into stainless steel screened capsules. These capsules 
were marked to allow easy identification of the gonad' s 
origin. The capsules were placed in a plastic embedding 
basket which was entered into a paraffin embedding cyc le 
on an automatic tissue processor (Autotechnicon 2A). The 
ethanol - clearing solution-paraffin cycle is listed in 
Appendix 1. After 16 hours in the embedding cycle the 
gonads were submerged in a paraffin bath. The basket was 
removed from the tissue processor and each gonad was 
taken out of the capsule and embedded in hot paraffin to 
form a cube. The cube of paraffin surrounding the gonad 
acted to support for the tissue when it is sectioned. 
The cubes were mounted on wooden cutting blocks for 
sectioning with a microtome (AO). The sections were 
sliced at a width of 7 - 10 microns. These sections were 
mounted on slides and stained according to the 
hematoxylin and eosin staining protocol listed in 
36 
Appendix 2. Once the slides were stained, Histoclad wa s 
used to secure a coverslip over the specimens. The 
slides were allowed to dry over a 2-3 day period, then 
the excess mounting glue was cleaned from the slides. 
Once cleaned, the slides were analyzed under a compound 
light microscope (Baush & Lomb). The following items 
were examined: (1) sex; (2) maturity; (3) the presence of 
a lumen and its orientation to mature gametes; and (4) 
the cell stages present. 
RESULTS 
The present system under analysis only involves 
fishes that externally fertilize their gametes, 
therefore, only this mode of reproduction will be 
introduced and discussed. In order to establish any 
morphological difference on the part of either fish 
species being examined histologically, an introduction to 
the morphology of a gonochoristic (non-sex changing) male 
and female gonad will be presented. Groman (1982) 
presented the morphology of the striped bass which will 
be used as a model of gonochorism. Discussion of the 
morphotypes is supplemented with a study by Nagahama 
(1983). 
A gonochoristic (non-sex changing) male has testes 
that are solid masses of germinal tissue. Within a 
testis the seminiferous tubules are packed tightly with 
only a small amount of connective tissue intervening 
between the tubules. A tubule winds around the testis in 
37 
a n d out of the p l a n e T ho uf s ,s  e wc ht ei no  n p. r epared 
hi ts hto e lo teg si tc ia sl  l sy ee, ms to b e fille d wi th 
many 
s em i n 1fe rous t ubules s ome o f whi c h look ci r cular ' 
Each seminif
t eu rb ou ul s ar t, u bo ur l e o v ho asi d  . i t s most 
immat u re g erm s tages near the outer e dges o f the t ubule 
a nd t he more mature stages toward the c ent r al r e g io n . 
Th e matu r e gametes are releas ed i nto the c ent ral region 
These mature gametes 
o <f spt eh re m ) t u ab r eu  le m. o v ed 
toward a du c tu l e and the ductules connect to form a s per m 
d uc t to s t ore the sperm until it is released during 
No evidence of a ce
s npa trw an li  n bg r. a nching lumen 
between the seminiferous tubules of a gono c hor i sti c 
Therefore , the testes of a n
t oe ns -t i s ha s been found. 
se x changing fish is a solid structure devoid o f a 
central lu men, but posessing a duct network to seque s t er 
th e matur e sperm until released. 
The gonochoristic fish ovary has a lamella r 
The ger minal t i ssues are present wi th i
a nr  rangement . 
All of the stages of ova maturatio
l na  m ae rl el  ae. 
The lamellae are 
usually found wi thin each lamella. 
surrounded on three sides by "finger-like " branches of a 
(The fourth si d e of 
central lumen present in the ovary. 
the lamella i s connected to the wal l o f the ovary by 
The 1umen is not always "cent ra l," 
connective tissue). 
but may run do wn one si d e of the o v ary, just below the 
tunica albuginea (t h e dense connective tissue c over i ng o f 
va are stimulated to be 
th 0e ovary) . Wh en ma t u re , 
38 
re lule ma es ne . d during mating and shed into the ovarian 
ma ure ova
T  h ae r e ct a rried down into the "oviduct". The 
receives ova from both ovaries an
" do  v ri ed lu ec at s" e s 
the environment for external fertilization. 
them into 
Gonads of Ici~bQQQi~~ Di~ii a nd x~ci~bi~~ m~l~D~~~2 
were examined morphologically and ultimately compared to 
gonochoristic and hermaphroditic morphologies. 
An examination of female I? Di~ii <N = 23) revealed 
The ovaries of I ? Di~ii 
one mo rphotype of the ovaries. 
The germinal 
have a lamellar arrangement <Figure 4a). 
"Finger-like" 
t i? ssues are lined up within lamellae. 
branches of an ovarian lumen (L) intercede between the 
The lumen lies dorsally within the ovary. I-
lamellae. 
is similar 
-n -i -k l? l?  females have an ovarian morph ? 1 ogy that 
to that which was described previously for the 
gonochoristic ovary. 
The testes of T. Di~ii males <N = 10> have a 
Each testis is packed with 
so 1 l? d arrangement. 
seminiferous tubules which form a solid mass <Figure 
Serial sections of four of the testes examined 
4b>. 
did not show any evidence for the presence of an 
" ovarian-type" lumen within the testicular tissue. 
All stages of spermatogenesis are seen within the 
The mature sperm are shed into 
se m? iniferous tubules. 
The sperm travel 
5
th e central region of the tubule ? 
from the seminiferous tubules to a ductule and then 
into a larger duct located toward the center of the 
39 
--
Figure 4. Histological sections of the gonads of 
Trichonotus niki~. Sections of the gonad s 
were sliced at a thickness of 710 microns and 
stained with hematoxylin and eosin. The 
letter "L" designates th e lumen in the ova r y. 
"TA " represents the tunica albuginea (outer 
covering of the testis) . "S" represents mature 
sperm . "SD" represents the sperm duct. 
A: T. ~i~ii ovary <mag. 13 X) 
8: T. ~i~ .U. primary testis (mag. 13 X) 
t+ O 
a. 
TA-U~~~,&~--~---:::::---,.~~r:Jt..":,,,. 
?, . 
'?-~C
 A?. ; . 
'~ ~ l-?, L 
. . . ~;;. -
--- ------ TA 
b. 
...  
' 
----~s 
41 
whole te s tis. The mature spe r m f r om each testi s is 
s equester e d into a main sperm duct s y s tem (D) whe r e i t 
is stored until spawning by the female (Figure 5). 
Only one testicular morphotype was found upon 
hi s tological examination of Trichonotus Di~ii males. The 
morphology is similar to that which was previous ly 
described for the "model" gonochoristic fish testis. 
Taking into consideration the morphology of the ovary 
and testis of T. ~i~ii examined during the present study, 
T. nikii is a gonochorist and not a protogynous 
hermaphrodite as had been previously theorized by Shimada 
and Yoshino (1984). 
The gonads of the male <N = 3) and female (N = 6) 
Xyrichtys melanOQUS show a striking similarity in 
morphology (Figure 6a and b). Both sexes have their 
germinal tissue in a lamellar array surrounded by a 
"central" lumen (L) which sends "finger-like" branches 
between each lamella <Figure 6a and 6b). 
The morphological similarity of the gonads is not 
carried over physiologically, however. The female sheds 
her mature ova into the lumen for passage into the 
oviduct and out of the body cavity during spawning. The 
male sheds his mature sperm into the central region 
within each seminiferous tubule. All of the seminiferous 
tubules within a lamella eventually connect into one 
large duct, central to each l a mella. The central 
"lamellar" ducts connect to become the sperm duct <SD). 
42 
 
tis o f Trichonotu
s ni k i i s ho wi ng
ry tes er m 
F igure 5 . P rima p
sit i on of the two
 lobes to the s
t he po e s i dua l 
ct <SD) . Note th
e lack o f any r
du nter of the test
is . 
n the ce
ovarian lumen wi
thi
led from t he 
The mature sperm 
(S) is channel  
 tubules into the
 sperm ductul es
seminifer ous
the sperm duct (S
D) which 
and then into es. ( mag. 16 X) 
receives sperm fr
om both lob
43 
s 
44 
Fi gure 6. Histological sections of the gonads of 
Xyrichtys melanoeus. Sections of the gonads 
were sliced at a thickness of 710 microns and 
stained with hematoxylin and eosin. The 
letter " L" designates the ovarian lumen in 
both the ovary and the secondary testes. All 
testes examined showed this feature. "TA" 
represents the tunica albuginea. "S" 
represents mature sperm in the tubules. "SD" 
represents the sperm duct. 
A: X. ~~l~~QE~? ovary (mag. 13 X) 
B: X. ~~l~~QE~? testis (mag . 20 X) 
45 
a. 
b. 
SD 
46 
The sperm duct of each testis is located within the 
tunica albuginea surrounding the testis. The sperm duct 
of each testis connect together to form a ma?i n sperm duct 
which exits the body cavity immediately anterior to the 
Ma
a tun ru es  . s perm is seen in storage within the duct 
No mature sperm or
s  y as nt ye  m o. t her germinal stages 
were seen within the lumen branches surrounding the 
The x. melanoeus testes analyzed during this 
lamellae. 
study revealed only one morphotype. 
The ovary of X. melanoEus (Figure 6a) is 
morphologically similar to the gonochoristic ovary 
described previously. 
6b) is not similar to the morphotype described for the 
The morphology of the testis 
gonochoristic fish testis-
of~- ~~l~ E~? is similar to that which was found in 
09
Nemtzov 
isolated single harems (1 male with 1 to 7 females) of 
~- E~~tadact~lus within aquaria and subsequently removed 
The females of the harem were found to have a 
the male. 
The largest female in 
dominance hierarchy based on size. 
the harem began behaving like the "removed" male had 
Within a few days the female's coloration 
Previously. 
The secondary 
and gonads changed into the male form. 
testes of the "new male" was sectioned histologically 
using the same technique used in the present study. The 
secondary testes were in a 1amellar array with a non -
functional lumen branching between the lamellae. No 
47 
evidence of a male X. eentadact~lus testis without a 
lumen was found. 
The morphology of the secondary testes of~-
E~D~~Q~f~~l~? is similar to that of~- ~~l~DQQ~? testes. 
The lumen found within the testicular tissue of the 
razorfishes is left over from previous ovarian 
morphology. Therefore, due to the close phylogenetic 
relationship of X. eentadact~lus to X. melanoeus, the 
same type of haremic social system, and similar gonad 
morphology, X. melanoeus is being considered a 
"monandric" protogynous hermaphrodite. 
DISCUSSION 
The initial purpose of analyzing X. melanoeus and T. 
Dl~ll for hermaphroditism wa s to end the speculation 
concerning these species. The method of showing the 
presence of sex reversal was an a posteriori one. The 
gonads of both sexes of each species were sectioned and 
stained and the histology and morphology of the organs 
were examined. This method of analyzing sex reversal is 
not complete, however. Using histology, one can only 
examine the end product of a sex change and determine the 
presence or absence of the event, unless all of the 
intermediate cell and morphological stages are sectioned. 
If one is to focus on the sex reversal completely, social 
behavior, the reversal process, and finally, histology of 
the gonads is done. Nemtzov (1985) performed a thorough 
sex reversal study on X. eentadact~lus. The present 
48 
study on X. melanOQUS was intended to add to Nemt z ov ' s 
wo r k by presenting evidence that a close relative off. 
Q~0!~9~S!il~~ is following the same sex change course. A 
complete behavioral study on X. melanoQus' social system 
is in progress by Clark and Shen. 
In T. ~i~ii, a separation of sizes of the sexes is 
not complete evidence of the occurrence of sex reversal, 
as was assumed by Shimada and Yoshino (1984). Obviously, 
the I- Di~ii populations have not been sampled 
sufficiently to find young males. 
Histological examination showed that T . nikii is 
probably a gonochorist. The solid, non-sex reversed 
testes of the males, makes the absence of small males in 
the data (Figure 2) peculiar. Small male T. 0i~ii may be 
pr esent in the marine envi r onment in a place that has not 
been sampled. Because sex reversal is probably not 
occurring, the large males in the samples taken must have 
developed and grown from small males in the population. 
A few possibilities exist that may explain the 
absence of small males in the samples. 
(A) The small, immature (less dominant) male s ma y 
be living separately from the main swarms. The main 
swarms are kno wn to lower to the sand to perform lek-like 
mating behavior. The less do minant males have little 
chance, if any, for potential mating, therefore, it would 
be energetically "smart " for the small males to avoid 
this behavior until their dominance rank was increased. 
49 
During the lekking periods, predators have a good chance 
of catching T. ~i~ii individuals. By avoiding the lek, 
small males are also protecting themselves from 
predators. Small swarms of immature, "bac helor" males 
may be present within the sand environment but, isolated 
from the main swarms that were sampled. Also, the 
"bachelor" swarms could be away from the sand 
environment, thus, they are being missed. 
(8) The small (less dominant) males could be missed 
during collections due to their position in the swarm and 
the lek. The less dominant males on a lek are pushed out 
toward the territories on the outskirts of the arena by 
the dominant males (lack, 1968; Pulliam, 1973). The same 
process may occur in I- -n-i-k-i-i leks. Sampling of the 
population is usually done while the swarm 1s on the sand 
surface or in their dive sites. During sampling, it is 
possible that the outskirts of the lek are not being 
sampled, and therefore, the sma ll males are being missed. 
This idea is presently being examined by Clark and Shen. 
Preliminary observations seem to refute this idea . 
(C) The small males may have drab coloration, as do 
the females. The small males might be discounted for 
females during observations and collections. This idea 
is not completely feasi ble. During the present study, 
all of the fish were dissected and sexed by examining 
their gonads. Thus, within the samples already 
collected, all of the fish with drab female coloration 
50 
are female. 
The present goal with regard to Trichonotus nikii is 
to locate the small males in their natural environment. 
To facilitate searching, basic laboratory studies on the 
development of fertilized T. ~i~ii eggs to adults could 
be completed. Foll owing the fishes' development may lead 
to some clue as to the needs of the individual fish at 
each stage of the process. By taking the information on 
fish development that is learned in the laboratory and 
applying it to the field, it may lead to the possible 
location of the small males in the environment. 
In conclusion, Xyrichtys melanoeus is being 
considered a monandric protogynous hermaphrodite on the 
basis of gonad histology. Further work on the behavior 
of the fishes in the territories and during mating is in 
progress by Clark and Shen. Trichontus nikii is being 
considered a probable gonochorist on the basis of 
preliminary gonad histology. Further work to illuminate 
the location of the small males of the species in the 
environment and to expand the histological sampling of 
the males needs to be completed. 
51 
Chapter 2 
Feeding Ecology of the Sandfishes and Food Resou r ce 
Availablity in the Sand Environment of the Red Sea 
INTRODUCTION 
The vertical stratification of X. melanoeus, T. 
~i~ii, and ?g~g~?i~ sp. in the wate r column above t he 
sand bottom could possibly be influenced by food 
availability and its distribution in the water column; by 
interspecific conflicts or competition for food; by 
physiological requirements of the fish species; or by 
predation. To investigate the structure of this 
assemblage, the factors influencing the fishes 
distributions have to be separated. The most influential 
factor(s) could then be determined without being 
confounded by other variables. To examine all of the 
factors involved in structuring the community all at once 
would be a monu mental task. 
Therefore, the focus of this study of the sandfishes 
community structure is limited to examination of the 
availability of food resources and distribution and its 
relationship to the fishes' spatial arrangement. An 
analysis of food resource availability seems to be the 
logical factor to complete initially, due to the 
concurrence of feeding by the fishes and the vertical 
stratification of the fish species. The other possible 
struct u ring forces mentioned previously will be 
discussed, but no experimentation on them was completed 
within the scope of this study. 
52 
In 1985, field observations were combined with 
collections of specimens of all three species of fishes 
and collections of horizontal plankton tows which were 
taken at the heights which correspond to the heights in 
which the fishes were feeding. The specimens were 
analyzed for stomach contents to determine the diets of 
the three species of fishes. An examination of the diets 
of the fishes can determine food resource-use overlap and 
the possibility of food resource-based competition 
between the fish species. 
Horizontal plankton samples were taken to determine 
the spatial arrangement of various species of zooplankton 
and their abundances within the water column. 
Differences between abundances of species of zooplankton 
at the three heights may be an important factor that 
influences the fishes' distributions in the water column. 
If a certain prey is located only in one area of the 
water column, a fish which preferentially feeds on that 
prey must feed in the area of the water column in which 
the prey is located. Thus, the basis for the fishes ' 
spatial arrangement may be due to the distribution of 
their preferred food organisms . If the abundances of 
species of zooplankton are not vertically patchy in the 
water column, other factors must be involved in the 
vertical stratification of the fish species in the water 
column. 
53 
Also, the plan k to n data wi ll be a used to compa
re 
z ooplankton and abundances at each site the species of 
and within each site, at 
diffe r ent heights. The 
e es y 
horizontal plankton tows 
we r e taken at thre tim of da
any zooplanktor fluctuations could be to determine if 
over time. Fluctuations could
 be due to vertical 
seen 
migration of the zooplankton s pecies over the w
ater 
column, isolated events, or chance alone. 
METHODS 
STOMACH CONTENTS ANALY S I S 
When examined, each stomach was sliced open 
longitudinally and rinsed with 70? ethanol so t
hat all of 
ntents were removed. The sto ma ch con
tents were 
the co
placed in a s mall petri dish marked with a 2mm 
square 
under 
grid on the bottom surface . Each dish wa
s examined 
dissecting microscope <OlympLs) . Every who
le organism 
a 
encountered was identified, measured and enume
rated. 
Only whole items were counted to avoid recounti
ng pieces 
from the same individual item. 
The exception to this rul e were made for stomac
h 
contents from Xyrichtys melanOQUS. The stomach o
f~ -
~~l~~ee~? was filled with bi v a l ve shells, arthro
pod 
appendages and vertebrae. These items may have
 been 
incidentally with other organisms (i . e. bivalve
 
ingested 
shells). Also, it would be un : ikely that the razorfish
 
could ingest a whole brachyuran crab or small f
ish 
without breaking it up int o pieces . Therefore, if on
e 
54 
were to count only whole prey, a significant portion of 
the razorfish ' s diet would be excluded from counting. 
Instead , counts were made of each bivalve s hell fragment, 
arthropod appendage, and vertebrae. These items are 
usually equal or larger in size to the copepods and 
gastropods which were found in the gut as well. 
The counts of the types of prey were co mpiled and 
converted into percentages of the total volume of the 
fish's stomachs (as in Randall, 1967). The percentages 
were used to characterized the fish's diets and then to 
compare to the plankton available at the study sites. 
Electivity indices were calculated from the stomach 
contents and plankton data following Chesson (1983) . 
Feeding niche overlap of the three fish species were 
measured using an overlap index suggested by Levin 
(1968). 
PLANKTON ANALYSIS 
During June 1985, plankton samples were taken to 
characterize the food resources available to the 
planktivorous fishes inhabiting the study sites. At each 
site, a series of three days of collections were made. 
The total collection period was within one lunar cycle 
<no full moon during collections) . Gliwicz (1986) has 
shown that in a freshwater lake, Cahora Bassa, 
fluctuations of abundances of species of plankton due to 
predation seem to be triggered by the full moon. More 
moonlight on nights around the full moon and the vertical 
55 
trajectory of the moon in the tropics permits more 
The ability
f  eeding by nocturnal planktivorous fishes. 
to see their prey items more easily enhances their 
ability to capture them. 
The sandfishes are not night foragers and thus 
moonlight may not affect daytime abundances of 
are in 
zooplankton that they encounter. The study sites 
influx due to current movement ) 
an open system (continual 
in co mparison to Gliwicz's system (closed lake), the 
abundances of zooplankton may not be tied so intimately 
with night predation around the full moon period. The 
zooplankton should be continually renewed by influx due 
to current action at the study sites. 
With the use of SCUBA , three 20 meter long transect 
lines were placed parallel at a distance of five meters 
The lines were 
apart on the sand slopes at the sites. 
Placed in the area of highest abundance of the three 
Twenty meters is more than half 
focus species of fishes-
the distance across the mos t dense region of the species 
The outer limits of the sandfish 
at each site. 
assembl age were avoided during sampling to decrease to 
chance of confounding the collections with the variables 
that cause the decrease in species abundance near the 
edges of the fishes' distribution-
A closed plankton net (#20 mesh, d i ameter 12.7 cm) 
attached to a steel bar extending down its length wa s 
I positioned the 
lowered to a SCUBA diver in the water. 
56 
net above one of the transect lines at a specified 
height, opened the net, and swam, at top speed (0.6m/s), 
the length of the line (method, Porter and Porter, 1977). 
All of the plankton collections were taken horizontal to 
the sand surface to sample layers of zooplankton 
available over the bottom at the sites. The net was kept 
in front so that my movement did not affect the plankton 
collection. The net was closed at the end of each run 
and the sample was brought to the surface to be 
preserved. The seawater-plankton mixture was immediately 
added to 95? ethanol, which diluted to approximately a 
70? ethanol-seawater solution. Each sample contained the 
plankton which was collected from a volume of 0.23m cubed 
from the water column . 
Collections were made at three times of day (morning 
= approximately 0800 hrs . , 1200 hrs., evening= 
approximately 1700 hrs.) at three heights over the sand 
bottom (0 . 3m, 1 . 0m, 1 . 7m) . The heights, originally 
measured in feet, were converted into meters. At each 
time of day, a sample was collected at each height. The 
collections at each height were taken over a different 
transect line. Separation of the collection locations 
was done to avoid disturbance of the zooplankton in the 
area around each line before each collection was made. 
All possible permutations of collection time, column 
height, and transect line were done to decrease the 
chance of sampling bias. 
57 
At the end of each day a set of nine plankton 
samples was obtained (3 times X 3 height s ). After nine 
collection days (3 days at eac h site), 81 plankton 
samples were preserved and ready for analysis. 
At the University of Maryland, each sample was 
sieved through a micro -nylon screen (Nytex), rinsed with 
70? ethanol, and diluted to a volume of 40 cc. All of 
the samples were diluted in the same way, thus, the 
relative concentrations of the zooplank ton in the samples 
were kept in proportion to the original collection 
concentrations. 
A 1 cc Stempel pipette was inserted into the diluted 
sample and used to mix the plankton as evenly as possible 
by collecting and plunging the pipette repeatedly. While 
the sample was agitated, a 1 cc subsample was taken from 
the center of the water mass with the pipette. The 
subsample was transferred into a small petri dish with a 
2 mm squared grid on the bottom of the dish. The pipette 
was rinsed thoroughly into the dish. This process was 
repeated three times so that 3 cc of each sample (3/40ths 
of each, 7.5?) was analyzed for percent composition of 
the major taxonomic groups. The procedure was abstracted 
from Frolander (1968) as being the most accurate use of 
the stempel pipette. 
STATISTICAL ANALYSIS 
To analyze the plankton data for significant 
differences between the sites, sampling heights and time 
58 
of day, the data were entered into SAS (Statistical 
Analysis System) general linear models. An analysis of 
variance <AN0VA) was run on each taxon encountered in the 
plankton. Arith meti c means and standard error were 
measured for each comp a rison. To meet the assumptions of 
the AN0VA the plankton data were transformed . Rare items 
in the plankton were transformed using a square root 
function <SR= SR(n + 0 . 5)). Abundant items in the 
plankton were transformed using a log10 function <L0G10 = 
L0G10(n + 1.0). Repeated day s were treated as replicate 
samples, due to the low variability be tween days of the 
field collections when analyzing daily variations in the 
samples. Significant interactions between site, height 
and time of day of collection were judged by the 
Bonphoroni technique. 
ELECTIVITY ANALYSIS 
Chesson's (1983) index of electivity (+1 to -1) was 
used to quantify the number of prey consumed relative to 
the number available in the plankton, where random 
feeding by a fish on a prey item is denoted by an index 
of zero (0). A positive index indicates that the fis h is 
eating the prey item at a higher proportion than it is 
encountering the prey item in the environment. A 
negative index indicates that the fish is eating the p rey 
item at a lower proportion than it is encountering the 
prey item in the environment. 
59 
The electivity dat a for all of the prey items o f 
all three fishes was used to determine if the fishes 
were feeding selectively from the plankton or if their 
feeding was random. 
RESULTS 
STOMACH CONTENTS ANALYSIS 
All three of the species of fish were examined for 
diet. The stomach contents results of the fishes is 
listed in Table 2 . The counts of prey types are listed 
in percent volume of the total contents found in each 
sto mach. 
~~Ci~b!~? ~~l~~Qe~?' (N = 9) diet consists mainly of 
small fishes (49?) and benthic invertebrates (51?) (Table 
2, column 3). The most common item found in the stomachs 
were vertebrae (49?) from small, elongate, narrow bodied 
fishes. Comparisons with X-rays taken of Trichonotus 
Dl~ll and drawings of J. ?l?9~D? by Shi mada and Yoshino 
(1984) and drawings of I? ??}i9?C by Nelson (1986) 
confirmed that the vertebrae were those of Ici~b9D9}~? 
Rare field observations of Xyrichtys melanoQU S 
preying upon I? Di~ii have been recorded (Clark, 1983a; 
pers. obs.). The extent of X. melanoeus feeding on T. 
~i~ii is much larger than assumed by field observations. 
Out of the nine X. melanOQUS that were examined, seven 
(79?) contained vertebrae of T . ~i~ii in varying 
quantities. 
60 
Table 2. Percent 111 volume of prey in the stomachs of three species of fishes. 
The five most common prey for each species are ranked IA to El. 
STOMACH CONTENTS IY. TOTAL VOL.I 
Gorgasia Trichontus Xyrichtys 
PREY FOUND IN STOMACH sp. nik 11 melanopus 
Planktonic Prey: 
Cyc lopoid copepoda 13.0 B 42.2 A 0.2 
Calanoid copepoda 68.3 A 31. 7 B .o 
Harpacticoid copepoda 2.2 E 11.b C 14) 0.2 
Invertebrate eggs 11.0 C 3.2 E 0.6 E 
Fish eggs 2.4 D 
Ostracoda 0.4 0.2 
Cladocera 1.5 
Gastropoda 1.3 10.b D 1.2 D 
Benthic Prey: 
Ne111atoda 0.2 0.1 
Amphipoda 0.1 0.3 
Polychaeta larvae 0.1 
Bivalve (pieces) Ill 15. 6 C 
Arthropoda (appendages) (2) 33.2 B 
Fish vertebrae (3) 48.b A 
Cumacea .o 
Footnotes: 
1. BIVALVIA PIECES MAY BE INGESTED INCIDEN TALLY OR THEY MAY BE THE 
RESULT OF INGESTING WHOLE BIVALVES WHICH WERE SUBSEQUENTLY CRUSHED 
BY Xyrich tys melanopus' JAWS and PHARYNGEAL TEETH. 
2. LARGE ARTHROPODS WERE INGESTED BY Xyrichtys melanopus, BUT BECAUSE THEY 
WERE BROKEN UP BY THE FISH, COMPLETE IDENTIFICATION WAS IMPOSSIBLE. 
POSITIVE IDENTIFICATION OF PADDLE-LI KE SWIMMING APPENDAGES (5 TH) FROM 
BRACHYURAN CRA BS WAS POSSIBLE. 
3. VERTEBRAE FOUND IN GUT WERE POSI TIVELY IDENTIFIED AS THOSE FROM 
Trichonotus nikii. X-RAYS OFT. niki1 WERE USED FOR COMPARISON. 
4. A FEW HARPACTICOID COPEPODS INT. nikii WERE BENTHI C IN ORIGIN BUT 
BUT WERE NOT CLASSIF IED SEPARATELY FROM THE PLANKTONIC FORMS. 
61 
The remaining prey in the stomach of~- ~gl~~QQUs 
are predominantly benthic in origin. Arthropod 
appendages were the second most common prey of ~ -
The appendages were pieces of 
brachyuran crab. Positive identification was possib
le 
due to the presence of swimming appendages (5th) in 
the 
samples. The exact number of crabs ingested by the 
razorfishes was not possible to determine. As the c
rabs 
were ingested by the razorfishes, they were crushed 
by 
the pharyngeal teeth . It would be impossible for the 
brachyuran crabs to be ingested whole without damage
 by 
either the pharyngeal teeth or those in the jaw. Bi
valve 
pieces were the third most common item found in the 
stomach of X. melanogus (16?). The bivalve pieces may
 
have been ingested incidentally with other prey item
s 
from the bottom surface (Randall, 1967). It is also 
possible that some or all of the bivalve pieces are 
the 
result of the fish ingesting whole bivalves and 
subsequently crushing them to get at the tissues. F
or 
h 
this reason, the bivalve pieces were left in the stom
ac
contents counts . Small gastropods (whole and part
ially 
broken) were the fourth most common prey of X. melan
oeus 
( 1 ?) ? The gastropods were metamorphosed and encased in 
shells, typical of those found on the sandy bottom a
t all 
three sites . Of the remaining prey found in the stomach
, 
1? were also benthic in origin (amphipods, nematodes
, 
harpacticoid copepods, and cumaceans) and 1? were 
62 
plankton1c in origin (cyclopoid and calanoid copepods, 
and invertebrate eggs). 
The stomach contents of Ici~honotus nikii (N = 33) 
is primarily plankton (74?+) and a smaller representation 
of benthic invertebrates <23?) (Table 2, column 2) . The 
most common prey are copepods (cyclopoids, calanoids, 
and harpacticoids, respectively) which comprise 86? of T. 
Di~ii's diet. Planktivory by I , Di~il agrees with 
observations of feeding from the field studies and 
Randall (1967). Harpacticoid copepods are typically 
benthic organisms (Barnes, 1980), thus they were included 
in the benthic component, although a few genera are 
holoplanktonic. Most of the harpacticoids were benthic 
types, but the actual percentage was not separated from 
the total during the laboratory counts. 
Surprisingly, 23? of the stomach contents of T. 
-n-i-k-i-i is benthic invertebrates. The benthic 
invertebrates could have been from the water column due 
to shear flow near the bottom. Organisms that are not 
anchored securely to the bottom can be tossed up into the 
water column as a result of the speed of water flow that 
comes in contact with the sand surface (Palmer, 1986). 
Therefore, part of the benthic co mp onent of I- ~i~ii's 
diet could also, be selected out of the water column. 
Some of the benthic invertebrates may be eaten directly 
off the bottom, such as when I ? ~i~ii is seen to pick at 
the sand surface during the lek - like displays (males ) . 
63 
Thus, the supposed "displacement behavior " may be true 
feeding. The "feeding" by male I- ~i~ii during the lek -
like behavior does not account for the meiofauna found in 
the stomach of female T. ~l~ll- Females also may pick at 
the bottom for food during feeding. Overall, T. nikii is 
a selective plankt on feeder with 85? of their diet 
comprised of copepods . 
?Q~Q~?i~ sp. 's diet is comprised of 971/. plankton and 
3? benthic invertebrates (Table 2, column 1). The four 
most predominant prey are calanoid copepods, cyclopoid 
copepod, invertebrate eggs, and fish eggs, respectively. 
All of these prey (92?) are planktonic in origin 
(invertebrate eggs may have been found on sand surface as 
well as in the plankton). This finding agrees with the 
field observations that the eels actively select prey out 
of the water column. The small amount of benthic 
invertebrates in the eel's diet could either be taken 
from the sand surface, presumably during slack currents 
or from the organisms that are tossed up into the water 
column from the bottom. The exact position from which 
the eels ingested the "benthic" invertebrates is not 
known. Overall, as seen in field observations, ?Qr9~?l~ 
sp. is a selective plankton feeder, maintaining a diet of 
84? copepods (mainly calanoids). 
Comparison of the diets of the three species was 
completed to determine the amount of overlap of food 
resource use by these fishes. If the resources are 
64 
l1m1ting, common resou rce use may indicate co mp etition 
f or prey by the fishes. Where food resources are 
abundant and not limiting, s p ecies will be able to 
coexist without co mp etit i ng from food . In such a 
situation t he ne c essity for fo od resource partitioning 1s 
obviated and a high degree of feeding niche overlap 
between the fish species can be tolerated. Within the 
present study a food overlap inde x is used to give a 
quantitative measure of food usage of each of the three 
species of fishes in relation to the others. The overlap 
values (alphas) of one fish to another are presented in 
Table 3. Each species' diet is compared with the other 
two species. 
The overlap values between X. melanoeus and the 
other two fish species are relatively low (range 0.003 to 
0 .009). Therefore, the measure of alpha introduced by 
Gause (1934) (restated and explained by Levins, 1968) 
indicates that only a small amount or type of food 
resource usage is shared between X. melanoeus and the 
other focus fish species . Conversely, the food overlap 
values (alpha) between I - nikii and Gorgasia sp. is 
relatively high (0.916). Thus, the two fish species are 
preying upon much of the same food resources. Therefore, 
the possibility of competition between I? ~l~ii and 
?9~g~?i~ sp . for prey is likely to be intense if food 
resources are limited. 
65 
Table 3. Analysis of the overlap of habitats of three species of fishes 
using the formula for ALPHA(ij) (The coefficient of competition) , 
from Levins 11968). ALPHA is the overlap of resource usage 
of species "J" on that of species "i" relative to the total 
resource utilization of species "i." Values range from zero 
to slightly greater than one. 
Food Resource Levins (1968) 
Primary Species (il Overlap with (jl Overlap Value 
Gorgasia sp. X. melanopus .0026 
Gorgasia sp. T. nikii .9159 
X. melanopus T. nikii .0075 
X. melanopus Gorgasia sp. .0035 
T. nikii X. 111e l anopus .0091 
T. mkli Gorgasia sp. .9159 
Formulat ion of ALPHA(ijl, the Coefficient of Competition: 
ALPHA(ijl = SUM(Pih*Pihl/SUM(Pih?P1hl 
Pih= proportion of use by species 
of the food resource h 
Pih= proportion of the use by species 
of the food resource h 
66 
The overlap inde x (alpha) does not take into account
 
the quanti ty o f food resour c es ava ilable to the focu
s 
species in t he sand environment. To measure the 
availability of zooplankton, horizontal plankton tow
s 
were taken from each of the three heights over the s
and 
at which the focus species are feeding. 
PLANKTON ANALYSIS 
The composition of the plankton from all three of 
the study sites at the three heights in the water co
lumn 
is listed in Table 4. The plankton samples were analyzed
 
for percentage of the total volume for the major tax
a. 
The taxonomic classification is consistent with that
 used 
in the analysis of stomach contents of the focus spe
cies. 
The three samples that were taken each day, at one s
ite, 
and one height are combined in the table. Also, the taxa
 
found in the plankton samples are ranked (1 -24, 1 = m
ost 
abundant). Of the six most abundant zooplankton taxa 
groups, the first three (copepod nauplii, 
 
dinoflagellates, and radiolarian) ta x a were not foun
d in
the stomach contents of the focus species. 
The plankton data were transformed and then entered 
into an analysis of variance (ANOVA) to answer the 
estions: are significant differences in zooplankton qu
abundances at (1) the different sites used in the stu
dy?, 
(2) the different heights over the sand bottom at wh
ich 
the fishes feed?, and (3) different times of day ? 
67 
Tabl e 4. Plankton taxa II of total lumped sample ) found at each site by height over the sand bottom . 
Sa1p les were col lected in June 1985 with a horizontal diver push net. 
Each sampl e contai ns the plankton fr om 0.231 cubed of seawater. 
PLANKTON SITE= RM Rl1 Rl1 AQ AQ AQ 1111 1111 MM 
RANK Ti\XA COLL. INFO. 0,3?. 1.011 1. 7111 0.311i i .0;; !. 7m 0. 3m !.Om 1. 7lli 
---------------------------------------------------------------------------------------------------------------------------------
Copepod naup l ii I 33. 09 31.43 31.68 17 .23 20.14 16.78 24.72 32 .16 25. 80 
2 flinofiagellates i 6.36 7 .12 5.82 17.12 17 . 12 20.54 15.48 17. 97 14. 09 
'l Inver tebrate eggs I 9.49 9.71:, 10 .83 15.43 13 .38 7.93 10,08 10.21 15.75 " 
4 Calanoid copepoda I 12.93 11. 94 13.93 5.92 8.99 10. 44 8.36 8.65 9. 61 
5 Radiolaria I 12. 41 12 .77 13 .87 4.46 4 .17 3.10 13.27 11.24 11.20 
6 Harpacticoid copepoda I I 6.05 b,51 4.64 9 .67 11. 37 11.7b 11.41 6.56 9.32 
0-- 7 Cyclopoid copepoda I 9.58 7. 81 9.73 8.04 6.26 l 'l 'l~ l..1 1..~ 5.75 5. 17 5.5(1 
{D 8 i3astrcipoda I 3.67 5.74 4.79 6.85 7.55 8.26 3.41 2.87 2. 77 
9 For a11inifera I 2.0b 3.41 2.30 5.54 4.39 3.50 2. 48 1.23 1.24 
10 Polychaete larv ae I 0.58 0.35 0.27 1.36 0.79 1. 19 1.64 1.60 2. 13 
11 Neaatoda l 0,48 (l ,bl 0.46 1.30 0.79 0.79 0.35 0.33 0.73 
12 Ci adt,cer a I 0.64 0.07 0. 11 3.15 0.22 0.79 0.22 0 .12 0.(10 
13 Cr ab 11egalopa I 0.23 0.00 0.13 0.16 0.07 (i ,53 1,72 1.03 1.21 
14 Ostr ac oda I 0.51 0.61 0.55 1.36 0.50 0.59 0.35 0.25 (l , (16 
15 Bivalvia ! 0.23 o. 11 0.15 1.30 1.80 0.66 0.09 0.04 0.06 
16 Barnacle nauplii I 0.39 0.20 0.25 0.49 1. 73 0. 59 0.44 0. 12 0.10 
17 Bryozoa I I 0.51 0.57 (1 .13 (l ,49 0.72 0.20 0.18 0.04 0.06 
18 Larvacea I 0.61 0.44 0 .1 7 0.00 0.00 0.00 0.00 0.29 0.(l(l 
19 Fi sh eggs i I 0.00 0.04 0.04 0.11 0,00 0.00 0. (14 0.12 0.32 
20 Echinoder1ata pleuteus I 0.19 0.13 0.06 0.00 0.00 0.00 (l ,00 0.00 0. (10 
21 Cr ab rnea I 0.00 (1 .07 0.04 0.00 0.00 0.13 0.00 0.00 0.03 
22 Atphipoda I 0.00 0.17 0.02 () , (H) 0,00 0.00 0.00 0.0(1 0.0(1 
23 Tun icate larvae I (l ,(!0 0. 13 0.04 0.00 0.00 0.00 0.00 0. 00 0.00 
24 Nysid larvae I 0.00 0.02 0.00 (l ,(I() 0.00 (1 .00 0.00 0.00 0.00 
---------------------------------------------------------------------------------------------------------------------------------
To answer the first two questions, the three times 
of day were combined for each day as in Table 4 . The 
AN0VA procedures were computed using each taxon as a 
dependent variable. From the AN0VA, using LSD <p > 0.05) 
significance levels, at each of the sites, three 
zooplankton taxa showed significant height differences. 
Gastropods, cyclopoid and calanoid copepods were 
significantly more abundant at 1.7m than at the other 
measured heights <p > 0.05, 0.03, 0.05, respectively) . 
Invertebrate eggs were the only ta xon that was 
significant for the site*height interaction (p > 0 . 02). 
With times and heights combined for each day, only 
invertebrate eggs showed significant variation from site 
to site. Also, when the sites are combined to look at 
variation at different heights, gastropods, cyclopoid 
copepods, and calanoid copepods have marginally 
significantly greater abundances at 1.7m co mpared to 1.0m 
and 0.3m over the sand bottom (Table 5). 
To answer the question if significant variation 
exists in zooplankton abundances at different times of 
day, the times of day were separated and each sampling 
day was treated as a replicate. The assumption of this 
procedure is that a sample taken at Day 1 during the 
sampling regime would be very similar to a sample taken 
from the same site on Day 4 of the nine day collection 
series . The constancy of temperature and weather at all 
of the study sites allows this assumption. Thus in 
69 
Table 5. Mean abundances of the taxa from the anal ysis of variance of the plankton data by height over the sand botto~ 
(SI6N!FICANCE f = .05, If = .01, 1t1 = .001 ) 
HEAN HEIGHT OVER SAND BOTTON HE AN HE IGHT OVER SANDB O TTOM 
TAXA 0.31 1. 011 1.7m TA XA (l.3m I. Om 1.7s 
Copepoda nauplii 56.75 64 .33 65.44 I Crab megalopa 1. 14 0.62 1.51 
I 
Dinofl agellates 26.30 33.67 34.32 I Barnacle naup l ii 0.93 1. 01 0.62 
I 
Invertebrate eggs 28 .24 27 .04 33.01 I Cladocera 1.58 0.25 0.43 
Calanoid copepoda 23 .30 27.06 37 .66 ( 1) I Bivalna 0.87 0.84 0.53 
-..J 
0 Harpacticoid copepoda 22.21 21.87 25.47 I Bryozoa 0.91 0.96 0.35 
I 
Cyc lopoid copepoda 19.88 17.97 29 .18 (2 ) I Larvacea 0.38 0.59 0.23 
Rad iolaria 20.88 21.33 19.86 I Fish eggs 0. 08 0. 14 0.27 
I 
Gastropoda 14.56 14.56 15.80 (3) I Amphipoda .00 0.21 0.03 
I 
Foraminifera 7.55 7.85 7. 10 I Crab zoea .00 0.06 0. 14 
I 
Ech1 noder1ata pleuteus 2.57 2.07 2. 78 I Tun icate larvae 0.00 0.1 0 0. 05 
I 
Ne1a tc,da 1.52 1.26 1. 97 I Mysids 0.01 0.03 0.00 
I 
Ostracoda 1.52 1.08 1. 08 I Nocti luca 0.03 .00 .00 
------------------------------------------------------------------------------------------------------------------------------------
Signi ficant variances: ( ! ) 1-3 t , 2-3 t (2) 1-3 t (3) 2-3 t 
theory, the sampling days are in t erchangeable. Th i s is a 
liberal approach, thus to counteract type I error s , 
Bonphoroni significance levels are used (Bon = 0. 0 5/3 
interactions = 0 . 0167). 
At either the LSD significance level (p > 0. 0 5) or 
the Bonphoroni levels, there are many significant 
differences in zooplankton abundances from one sampling 
time of day to another (Ap pendi x 3). This outcome is 
expected if vertical migrat ion of the zooplankters is 
occurring within the water column . According to vertical 
migration theory (0lhorst, 1982; Robichaux et al., 1981; 
Porter and Porter, 1977; Schmidt, 1973), the zooplankters 
should increase the depth they inhabit during daylight to 
avoid being eaten by planktivores. Increased visibility 
of the zooplankters in dayl ight increases their chance of 
being eaten (Gliwicz, 1986; Porter and Porter, 1977; 
Sameoto, 1974; Sch midt, 1973). At sunset, the 
zooplankters s hould swim, float, etc. toward the surface 
waters to fee d on phytoplankton residing in t h e upper 
photic zone. Darkness provides a refuge from predation 
for zooplankters. Also, fewer planktivores are nocturnal 
feeders, thus, the predation risk is lower at night 
(Co l le tte and Talbot, 1972; Hobson and Chess, 1978). 
Therefore, if vertical migration is being measured 
in samples from different times of day (at the measured 
heights), the early sampling time collection (approx. 
0800 hrs) should have a lower abundance of zooplankton in 
71 
c omp a rison to the later sampling times (1200 , appro x . 
1700 hrs) which are during broad daylight. The a n i ma l s 
should be in transition from the shallows to the depths. 
To accurately measure vertical migration the sampling 
times should ultimately coincide with dawn and dusk. 
Unfortunately, the sampling times of this study were 
restrained by the logistics of transportation to a n d from 
the sites. Thus, within the present sampling regime, I 
should expect to see fairly level values of abundances o f 
zooplankton because the collections do not cover the 
transition periods. 
From the arithmetic means of the abundances of the 
zooplankton taxa in Appendix 3 (tables and representative 
figures ) , there is no evidence of changes in abundance to 
mirror what one might expect if vertical migration is 
occurring, thus reinforcing the idea that the migration 
periods of the zooplankton were not sampled within the 
collection regime (although level abundances of 
zooplankton taxa were also not seen). A few isolated 
means from one taxon at one site show the expected 
shifts, but, are not supported by the same shifts at the 
other sites . The variation in abundances could be due to 
isolated events (spawning, currents sweeping meiofauna 
into the water column, etc.) or to chance alone. 
Appendix 3 lists the AN0VA results, the arithmetic 
means of each of the plankton taxa and a few 
representative graphs of the means. Appendix 3 (figures 
72 
1 - 5 ) shows the variation in the mean abu n danc e s o f the 
planktonic taxa at each site over time. 
The a bundances of each taxon found in the plankton 
were compared to the abundances o f the s ame g r oup s f o und 
in the fishes' stomachs using Chesson ' s (1983) elec ti v ity 
index. The index of electivity incorporates the fi shes ' 
diets with the food available to them (plankton). The 
index is used to answer the question of whether the 
fishes are feeding selectively among the available prey 
or are they feeding randomly? The results of the 
electivity analysis are presented in Fig. 7. The index 
is computed for each taxon found in the species ' stomachs 
that is also found in the plankton <N = 10 taxa) In 
Fig. 7, the five most common planktonic prey of eac h 
species of fish are denoted with a letter (A to E> over 
the index bar. 
The electivity indices for all three species 
indicate that the majority of the prey a r e not being 
selected at random (random = index near or equal to z ero, 
see methods section). Chesson (1983) warns that the 
index should not be analyzed statistically thus, the 
indices are not denoted as either significant or not. 
Most of the prey are being eaten by the fishes in a 
smaller proportion of the diet than the prey is 
represented in the plankton (negative selection or 
avoidance). The remaining prey (gastropods, c yclopoid, 
harpacticoid and calanoid copepods) are being positively 
73 
Figure 7. Chesson's (1983) index of electivity measured 
for~- ~~l~~2~~?, I? ~l~ll, and ?9~g~?l~ sp .. 
The abundances of nine prey taxa which were 
found in the fishes' stomachs were compared to 
the taxa's abundances in the plankton. The 
five most common prey found in each of the 
fishes' diets are marked A-E over the indices. 
Refer to the methods section for more details 
of the index measure. 
74 
CHESSON 1S ELECTIVllY INDEX FOR THE 3 FISHES 
OVER THE MOST COMMON PREY (Top 6 items ranked A-E) 
1 
A ~ XM 
.75 
~ GS 
.5 
~TN 
~ .25 
...J 
tn .>....  0 
(.) 
W...J  -2? 5 
w TT 
-.75 
-1 
-1.25 --.-----,--.----r--.---,--.---.---r----
Q. Q. (/) (/) 
g c., c., ~ ~ ~ ~ c., i ~ c., Q. ~ Iz.L I 0 ~ ~ ~ ~ 0 
PREY TAXA GROUPS 
selected by the fishes. The fishes eat a higher 
proportion of these prey than they occur in the plankton. 
Therefore, the fishes are feeding selectively out of the 
prey choices within the plankton. 
Ninety-seven percent of~- ~~l~~Qe~? ' diet is not 
plankton. Of the remaining 3?, X. melanoeus doe s not 
preferentially choose any one planktonic taxon to 
consume. Thus, t h e planktonic prey consumed must be 
either incidental or supplemental. 
Trichonotus nikii is a planktivore (Table 2) . The 
four most common prey found in the diet of this species 
are selected preferentially from the plankton available 
(gastropods, cyc l opoid, harpacticoid, and calanoid 
copepods). These taxa account for 96? of I ? Dl~il 's 
diet . I- Di~ii consume s invertebrate eggs, nematodes, 
ostracods and amphipo d s that are chosen in a lower 
proportion than the taxa occur in the plankton . The 
nematodes and a mphip o ds that were found in the plankton 
samples probably o riginated from the bottom, but we r e 
tossed into the water column due to turbulence acting on 
the sand surface. T . Di~ii males are seen picking at the 
botto m, but feeding has not been positively determined to 
occur d uring this beh avior. 
?9rg~?i~ sp . is a p lanktivore (Table 2). The two 
most c o mmon prey items of the fishes ' diets are being 
selec t ed preferentially from the plankton (calanoid and 
cycl opoid copepods) . Of the other prey found in ?QC9~?i~ 
76 
. ' s diet, invert
ebrate and fish eggs, gastropods, 
sp
amphipods, cladocerans and harpacticoid 
ostracods 
' in a 
copepods (planktonic or benthic) are being selected 
represented 
lower proportion of the diet than they are 
Overall, Bl? of ?QC9~?i~'s prey are being 
th e plankton. 
The remaining 19? 
selectively chosen from the plankton. 
is not being preferentially sought for food. 
DISCUSSION 
Based on behavioral observations of the 
intraspecific interactions of sandfishes presented 
p reviously by Clark (1971a, 1971b, 1975, 1980, 1983a, 
l983b) and the interspecific interactions presented 
within this paper, I examined the vertical structure of 
an assemblage of sandfishes in the water column over the 
The goals of this study of ~~Ci~Qi~? 
sand bottom. 
sand environment were to relate diet and food resource 
availability fro m the plankton to the structure of the 
assemblage of sandfishes -
Primarily, the investigation of the three fishes' 
nikii 
diets concludes that~- ~~l~~Qe~? consumes small I? -----
and assorted benthiC invertebrates and a relatively 
are 
minute amount of plankton-
Eac h species selectively chooses 
planktivores. 
zooplankton as their primary dietary component (96? and 
The fishes also consume other 
81 ?, respectively). 
A small 
zooplankton prey supplementallY or incidentally. 
77 
proportion of I? ~i~ii ' s diet consists of benth1c 
invertebrates. 
Where food resources are limiting, an assessment of 
the feeding niche overlap of the three species of fishes 
can help determine the level of competition for prey that 
may exist within the assemblage. But, when food 
resources are not limited, as such is probably the case 
in the Gulf of Aquaba, the necessity for food resource 
partitioning is lessened and a high degree of feeding 
niche overlap can be tolerated. Therefore, the food 
overlap index is used in this study to quantify common 
resource usage between the species of fishes, but it 
should not be interpreted as a strict measure of 
competition <Colwell and Futuyma, 1971). However, such 
a situation does not mean that the fishes avoid resource 
partitioning or competition altogether. Competition 
during another phase of life could exercise a regulating 
influence on the numbers of adults in the assemblage. 
The measure of the feeding niche overlap index 
concludes that~- ~gl?DQQ~? has a small dietary overlap 
Therefore, even if food 
resource limited, X. melanoeus 's position, in a lower 
stratum compared to the other fishes' feeding strata, is 
not primarily due to competition for prey. It is 
advantageous for I? -n-i-k-i-i to avoid X. 
the razorfishes eat T. ~i~ii? 
78 
Alternatively, T . nikii and Gorgasia sp. have large 
feeding niche overlaps. Their proportions of various 
prey are vastly different (Table 2) but, the similarity 
of the species of prey is high. The environment in which 
T. nikii and Go rgasia sp. feed does not seem to be food 
resource l imi ted, but alternatively, an overwhelming 
abundance of prey do not seem to be present (based on the 
plankton data) either. Separation of feeding rang es 
based strictly on food resource competition does not seem 
viable, although the foraging efficiency of each 
individual fish should increase as the fishes increase 
their per so nal feeding range s . Therefore, it is probably 
advantageous for I? ~i~ii and ?QI9~~i~ sp. to choose 
feeding strata which are separated either in space or 
time. Both species are diurnal, thus, separation of 
feeding ranges in space is the most viable alternative. 
When analyzing the assemblage structure of these 
sandfishes, an important factor to consider is whether or 
not the fishes are vertically stratified in the water 
column based only on the distribution of prey in the 
plankton. Horizontal plankton tows at the heights at 
which the s pecies feed did not show a significant 
variation in the compositions of plankton at the three 
heights examined (0.3m, 1.0m, 1.7m) or between sites. 
Each of these species was seen feeding at various 
times during the day, thus, the plankton data from each 
day (three samples) were lumped into one for analysis. 
79 
Invertebrate eggs (11? of ~~cg~~i~ s p. ' s d i et, 3? of I -
Qi ~ i i' s diet and 0 .6? of X. melanoQus's diet) was the 
only planktoninc taxon which varied s ignificantly from 
Aquasport, the most polluted site, had 
site to site. 
significantly lower percentages of invertebrate eggs than 
either of the other two sites during the sampling regime. 
Variation in the abundance of invertebrate eggs in the 
plankton should be expected when one considers the 
spatial and temporal variation in spawning by various 
Also, the recent oil spill s in th e 
invertebrate species. 
Northern Gulf of Aqaba may well be affecting the number 
of marine invertebrates that inhabit the area or the 
number of eggs they produce (compared to the less 
polluted areas of Marsa el Mukibela and Ras Mohammed). 
The sandfishes do not rely solely on the nutritional 
supplement of invertebrate or fish eggs for a large 
One would suspect that when 
proportion of their diets. 
the eggs (invertebrate and fish) are available in the 
Plankton, that the fishes would take advantage of the 
high nutritional value and thus increase the proportion 
of the eggs in their diets. 
Based on the results of the plankton data gathered, 
approximately 90? of the taxa (if the samples are lumped 
by day) does .not vary significantly in abundance from 
Therefore, aside from invertebrate egg 
site to site . 
abundances, the sites are considered to be fairly 
comparable in distribution of plankton for purposes of my 
80 
analysis. 
va rious 
When analyzing the spatial arrangement of 
taxa from feeding height to feeding height at each site, 
variation in abundances of gastropods, cyclopoid and 
These 
calanoid copepods were marginally significant. 
taxa we r e more abundant at 1.7m than at the other two 
Invertebrate eggs varied significantly from 
heights. 
height to height also, but, as discussed previously, 
spatial and temporal variation in egg abundances wa s 
Cyclopoid copepods and gastropods are a lar ge 
e xpected. 
proportion of I? 0i~ii ' s diet (42? and 11?, 
Therefore, it is logical that I? ~l~ii is 
respectively). 
found feeding at a height of 1.7m. 
QQcg~?i~ sp. feeds with its body stretched to its 
maximum length (approx- 1m> out of the burrow (except in 
extremely strong currents, the eels "crouch" during 
feeding), thus, it cannot reach higher in the water 
? rg~?i~ sp. preferentially eats calanoid 
column. 9
copepods which are slightly more abundant at 1.7m off the 
bottom than at the other two feeding heights measured. 
If QQcgasia sp. could feed without being in contact with 
its burrow, it might shift its feeding stratum higher in 
search of calanoid copepods- limited 
higher off the sand bottom and if food resource 
' 
it would be in direct competition with I? Oi~ii for 
Planktonic food. 
81 
X. ~~1~09E~2's planktonic prey (less than 3? of 
diet) do not vary significantly in abundance fr om height 
to height in the samples. But, it is not expected that 
X. ~~l~~QE~? would shift its feeding height in response 
to differences in planktonic species abundances. Ninety-
seven percent of~- ~gl~DQQ~2 ?s d i et consists o f small J . 
0i~ii and benthic invertebrates, thus, it would be 
advantageous for~- ~~l~QQE~~ to stay near the sand 
surface to pick at the bottom for food. 
After analyzing the dietary and behavioral 
observations of these three species of sandfishes as some 
of the factors influencing their vertical stratification 
in the water column, the following conclusions become 
apparent. It would be advantageous for J. Dl~ll to stay 
away from X. ~~l~09E~2 when possible to avoid being 
eaten. Also, T. nikii should avoid eating in the same 
stratum of the water column with ?Q~g~?i~ sp . , to avoid 
competition for planktonic prey if in reality the species 
are food resource limited. Feeding higher in the water 
column allows T. Qi~ii to take advantage of slightly 
increased abundances of cyclopoid copepods and 
gastropods. For I ? Q!~ii, being high up in the water 
column away from the safety of the diving areas of the 
sand has other conseq uences. T. Qi~ii is a swarming 
species . The fish's swarm confuses pelagic predators 
which may pass through the sand environment (Table 1). 
Also, T. nikii is safely away from benthic predators 
82 
( e.g. X. ~? l ~~ge~~) which prey on fishes near the sand 
surface. Therefore, based on the present results, T . 
0l~il s h oul d fee d h i gh in the water column to avoid 
pr e d a ti o n, po s sible food res o urce c o mp e t i t io n a n d to gain 
t he b e nefit of the increased abundances o f c y c lop o id 
copepods and gastropods. 
X. ~~l~Q2e~2 should sta y near th e s and surface t o 
feed and to defend its territorial bounda r ies ( and its 
dive si t es) from conspecifics. 
venture up into the water column in pursuit of I ? n ik i i . 
Swarms of I- ~ i ~ii lower to the sand surface to per form 
their lek-like mating displays and to bury in the s and at 
dusk. Thus, X. melanoeus can capture I? ~i~i i fishe s 
near the sand bottom. Therefore, i t is advantageous f or 
X. ~~l~Q2e~2 to stay near the sand surfac e to feed a nd to 
defend its territorial boundaries. 
?g~g~~i~ sp. 's maximum feeding height is cons tr a ined 
by body length. Most of the length of the eel ' s body is 
extended out of the burrow during feeding. Studies on 
the territorial behavior of ?2cg~2 i~ sp. by Clark ( 1980 ) 
show that the longer (and larger) males hold larger 
territories on the sand surface. Dominance behavior may 
explain why ?2cg~2 i~ sp. male e x tend out of the burrow a s 
far as possible to feed. Dominance of males is in part 
due to size (length), thus, the males extend out of the 
burrows as far as possible. 
83 
During slack currents, ?QC9~ 2 i~ sp. males and 
females alike are seen reaching and selecting items from 
the plankton. Maximum extension of the body length out 
of the burrow allows a larger range in which the eels can 
feed. Therefore, at least during slack currents, maximum 
distance of feed ing from the sand bottom (burrow opening) 
is beneficial for prey capture. 
Other factors are probably involved in causing both 
male and female ?Qrg~?l~ sp.to feed at a height of almost 
one meter. There could be a physiological constraint on 
?QC9~2i~ sp. that inhibits the eels from partially 
extending out of the burrow for long periods of time 
(i.e. during feeding) . More research on the ?Qrg~?l~ 
sp. 's stratum limits is needed. 
At the sites where the three species do not coexist, 
(Marsa el Mukibela lacks ?Qrg~?l~ sp. and Aquasport lacks 
similar to X. melanoQU S in that assemblage structure), 
small qualitative differences in the assemblage structure 
of the fishes can be seen. At Marsa el Mukibela, the 
swarms of T . nikii tend to be slightly cl oser to the sand 
surface. I- oi~ii hovers above X. melanoQus in the area 
in which ?Qcg~2 i~ sp . would feed, if present. 
po si tion in the water column at Marsa el Mukibela could 
be due to a lack of competition for food with the eels. 
The decrease in feeding height by I- Qi~ii also gives 
credibility to the minimal difference between feeding on 
84 
plankton at 1.7m and 1.0m. 
be feeding in a stratum that is not noticeably different 
in plankton abundances from any other height that the 
eels are able to span above the sand. 
At Aquasport, X. e~~?'S range does not completely 
overlap the range of T. ~i~ii? T. ~i~ii hovers closer to 
the bottom when they are not actively feeding. This 
difference in height of I ? ~i~ii swarms at Aqua s port may 
indicate that X. melanoeus is forcing I- ~i~ii higher 
into the water column at the other two sites. The exact 
differences seen in the assemblages at the three sites 
when all three species of sandfishes were not present 
were not measured quanititatively, although these data 
would be of interest to add support to the present st udy. 
Within the scope of this study, the vertical 
stratification of X. melanoeus, T. nikii, and ?9~9~?i~ 
sp. in the water column over the sand surface is 
partially explained by int e rspecific interactions and 
predation, and partially by diet limitations and possible 
competition for food resources. Ultimately, more 
research on the sandfishes may reveal other factors 
involved in the structure of the assemblage. 
85 
REFERENCES CITED 
Abe , T., M. Miki, and M. Asai. 1977. Description of a new 
garden eel from Japan. UO. 28:1-8. 
Anderson, G.R.V., A.H. Ehrlich, P.R. Ehrlich, J.D. 
Roughgarden, B.C. Russell and F.H . Talbot. 1981 . Th e 
community structure of coral r eef fishes. Am. Natur. 
11 7 :476-495 . 
Barnes, R.D. 1980. Invertebrate Zo ology. 4ed. Sau nders 
Co l lege. Philadelphia. 1089pp. 
Bardach, J.E. 1958. On the movement of certain Bermuda 
reef fishes. Ecology. 39 : 139. 
Bohl ke, J. 1957. On the occurence of garden eels in the 
western Atlantic, with a synopsis of the 
Heterocongrinae. Proc. Acad. Nat. Sci. Phila. 
109:59-79 . 
Bohlke, J.E . and C . C.G. Chaplin. 1968 . Fishes of the 
Bahamas and adjacent tropical waters. Livingston 
Publ. Co . Wynnewood, PA. 771pp. 
Borgia, G. 1979. Sexual selection and the evolution of 
mating systems. l~ = M & A Blum, eds. Sexual 
Selection and Reproductive Competition in Insects. 
Academy Press, NY. p. 19- 80. 
Bradbury, J.W. and R.M . Gibson . 1983. Leks and mate 
choice . l~ = Bateson, P . P . G. , ed. Mate Choice. 
Cambridge University Press, Cambridge. pp. 109-138 . 
Briggs, J . C. 1961 . The no menclature of ~Yrifb~Y? Cuvier 
1815. Copeia . 1961: 238-239. 
Brock, R.E . 1979 . An experimental study on the effect of 
grazing by parrotfishes and role of refuges in 
benthic community structure. 51 381 - 388. 
Casimir, M.J. and H.W . Fricke. 1971. Zur funktion, 
morphologie und histochemie der schwanzdruse bei 
rohrenaalen (Pisces, Apodes, Heterocongridael. Mar. 
Biol. 9 : 339-346 . 
Chesson, J. 1983. The estimation and analysis of 
preference and its relationship to foraging models. 
Ecology . 64(5) : 1297-1304. 
Clark, E. 1971a . Observations on a garden eel colony of 
Eilat. Sci . News l et . , Hebrew University. 1: 5. 
86 
_____ _ _ _ . 19 7 1b. The Red Sea ga r den e e l. Bull. Ame r. 
Li ttoral Soc., 7 (1): 4 - 10 . 
________ 1972 . The Red Sea ' s garden of eels. Natl. 
Geogr. 142(6): 724- 735. 
____ _ _ _ _ . 197 5. The strangest sea. Natl. Geogr. 148 (3) : 
3 3 8 - 365. 
_ _ _ _ ____ . 1980 . Distribution, mobility, and beh a vior o f 
the Red Sea garden eel. Nat. Geo. Soc. Res . Rep. 12 : 
91 - 10 2. 
1983a. Undersea desert in the Red Sea. Natl. 
Geo. 162: 129-144. 
_____ _ __ 1983b. Sand-diving behavior and territorialit y 
of the Red Sea razorfish, Xyrichtys Eentadact y lu s . 
Bull. Inst. Oceanogr. Fisher., Red Sea. Cairo. 
9:225- 242. 
Clark, E., A. Ben-Tuvia, and H. Steinitz. 1968. 
Observations on a coastal fish community, Dahlak 
Archipelago, Red Sea. Bull. Sea Fish. Res. Sta. 
Haifa. 49: 15-31. 
C lark, E., J.S. Rabin and S. Holderman. in press. 
Reproductive behavior and haremic organization 1n 
the sand tilefish, Malacanthus Qlumieri. Envi r . 
Bio 1 . Fi sh. 
Clark, E. and D. Shen. 1986. Territoriality of Red Sea 
sand-diving fishes of the genera ~Y~ifb~Y? and 
Trichonotus . In: Uyeno, T., R. Arai, T. 
Taniuchi, and K. Matsuura, eds. Inda-Pacifi c Fi s h 
Biology. Proc. Sec. Int. Conf. Inda-Pac. Fish. 
p.937. (abstract) 
Clark, E. and K. van Schmidt. 1966. A new species of 
Trichonotus (Pisces, Trichonotidae) from the Red 
Sea. Bull. Sea . Fish. Res. Sta. Haifa. Israel . S. 
Red Sea Exped., Rep. No. 34:29-36. 
Collette, B. and F.H. Talbot. 1972. Activity patterns of 
coral reef fishes with emphasis on nocturnal - diurnal 
changeover. Bull. Nat. Hist. Mus. Los Angeles 14:98-
124. 
Colwell, R.K. and D.J. Futuyma. 1971. On the measurement 
of niche breadth and overlap. Ecology. 5 2 :567 - 5 76. 
Connell, J.H . 1961. Effects of competition, predation by 
Thais lagillus and other factors on natural 
populations of the barnacle ~~l~Q~? ~~l~DQiQ~?-
87 
Ecol. Monogr. 31:61 - 104. 
_ __ _________ . 197 5 . Some mechanisms producing structure 
in natural communities : a model and evidenc e fr o m 
field experiments . I Q: Ecology and Evolution of 
Communities. M.L. Cody and J.M. Diamond, eds. 
Harvard University Press, Cambridge, Mass. pp. 460-
4 90 . 
____________ 1978. Diversity in tropical rain f or e st s 
and coral reefs. Science . 199:1302- 1310. 
Diamond, J . M. 1975 . Assemblies o f species communi t ies . 
In : Cody, M. L . and J.M. Diamond, eds. Ecology and 
Evolution of Co mmun ities. Belknap Press, Cambridge. 
pp 342- 444 . 
Dor, M. 1984. CL0FRES checklist of fishes of the Red Sea. 
The Israel Academy of Sciences and Humanities. 
Jerusalem, 437pp. 
Fishelson, L. 1968 . Marine biological and oceanographical 
research i n the Red Sea (Final report to 0NR, U.S. 
Navy, Washington, D.C.) 65pp. 
__ _ __ __ _____ . 1970. Protogynous sex reversal in the fish 
Anthias sguamieinnis <Teleostei, Anthiidae) 
regulated by the presence or absence of a male fish. 
Nature. 227 : 90-91 . 
____________ . 1973. Ecological and biological phenomena 
influencing coral species c ompos ition on the reef 
tables at Eilat (Gulf of Aqaba, Red Sea). Mar. Biol. 
19: 183- 196. 
Fricke, H.W. 1979. Mating system, resource defense and 
sex change in the anemonefish, B~EblE~l~~ 
akalloeisos. Zeitschrift Tierpsychol . 50 : 313 - 326 . 
Fricke, H.W. and 5 . Fricke. 1977. Monogamy and se x change 
by aggressive dominance in coral reef fish. Nature . 
266 :830-832. 
Frolander, H.F. 1968. Statistical variation in 
zooplankton numbers from subsampling with a stempel 
pipette. J . Water Poll. Cont. Fed . 40(2): R82-R88. 
Ga use, G.F. 1934. The Struggle for Existence . Williams 
and Wilk ins, Baltimore. Reprinted 1964, Hafner, N.Y. 
163pp . 
Gibson, R .M. and J .W. Bradbury. 1985. Sexual selection in 
lekking sage grouse: phenotypic correlates of male 
mating success . Behav. Ecol. Sociobiol. 18 : 11 7- 123. 
88 
Gladfelter, W.8. and E.H. Gladfelter. 1978. Fish 
c ommunity st r ucture as a fuTiction of habitat 
structure o n West Indian paich reefs . Rev. Tr ap. 
Biol. 26(Suppl. 1):65-84 . 
Gli wicz, Z .M. 1986. A lunar cycle in zo op lankton. 
Ecol ogy. 67(4): 883-897. 
Goldman, B. and F.H. Talbot. 1976 . Aspects of the ecology 
of c oral reef fishes. l ~= Jones, 0.A., and R . 
Endean, eds. Biology and Geology of Coral Reefs. 
Vol. IV. Biology 2. Academic Press , New York, pp. 
125-154 . 
Groman, D.B. 1982. Histology of the str i ped bass. Am. 
Fish. Soc. Monogr. No. 3. 116p. 
Hiatt, R.W. and D.W. Strasburg. 1960. Ecological 
relationships of the fish fauna on coral reef s of 
the Marshall Islands. Ecol. Monogr. 30 :65- 127. 
Hobson, E.S. and J . Chess. 1978. Trophic relationships 
among fishes and plankton in the lagoon at Enewetak 
Atoll, Marshall Islands. Fish. Bull. 76(1 ) :133-153 . 
Hourigan, T.F. and C.D. Kelley. 1985. Histology of the 
gonads and observations of the social behavior of 
the Caribbean angelfish, Holocanthus tricolor. Mar. 
Biol. 88:311-322 . 
Hutchinson, G.E. 1959. Homage to Santa Rosalia, or why 
are there so many kinds of animals ? Amer. Natur. 
93:145-159. 
Kennett, J. 1982 . Marine Geology. Englewood Cliffs: 
Prentice-Hall, Inc . 813pp. 
Lack, D. 1968. Ecological Adaptations in Breeding Birds. 
Methuen, London . 406pp . 
Leigh Jr., E.G., E.L . Charnov, and R.R. Warner. 1976. Sex 
ratio, sex change and natural selection. Proc. Natl. 
Acad. Sci. USA. 73(10):3656-3660. 
Levins, R. 1968. Evolution in Changing Environments. 
Princeton University Press, Princeton. 120pp. 
Lill, A. 1976. Lek behavior in the golden-headed Manakin, 
PiQra er~throceQhala in Trinidad (West Indies). 
Verlag Paul Parey, Berlin. p.1 -84. 
Loya, Y. 1975 . Possible effects of water pollution on the 
co mmunity structure of Red Sea corals. Mar. Biol. 29 
177- 185. 
89 
MacArthu~, _R.H. 1970. Species packing and competitive 
equ1l1br1um for many species. Theoret. Pop. Biol 
1 : 1 - 1 1 . . 
Masuda, H., K. Amaoka, c. Araga, T . Uyeno, and T. 
Yoshino. 1984. The Fishes of the Japanese 
Archipelago . Takai University Press . text: 437pp. 
and 
Molles, M.C . 1978. Fish species diversity on model 
natural reef patches: experimental insular 
biogeography. Ecol, Monogr. 48:289-305. 
Sea. 
Morley, N.J.F. 1975, The coastal waters of the Red 
Bull . Mar. Res. Cent., Saudi Arabia. 5:8-19. 
Nagahama, Y. 1983 , The functional morphology of teleost 
gonads. In : Hoar, w.s., D.J. Randall, and E.M. 
Donalds?~: eds, Fish Physiology. Volume IX. 
Repro ductio n, Part A: Endocrine Tissues and 
Hormones. Academic Press. New York. pp. 223- 277. 
Nelson, J.S. 1986, so me characteristics of Trichonotidae 
with emphasis to those di stinguishing it from ' 
Creedii dae (Perciformes: Trachinoidei) . Jap. J. 
Ichthyo l. 33(1):1-6-
Nemtzov, s.c . 1985, social control of sex change in the 
Red Sea razorfish, ~iric htis Qentadactilus 
(Teleostei, Labridae>- Environ . Biol. of Fish. 
14 ( 2 / 3) : 199-211 . 
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. 
Ogden, J.C. and J.P. Ebersole, 1981, Scale and community 
structure of coral reef fis hes: a long-term study 
of a large artificial reef, Mar. Ecol. Prag. Ser. 
4:97-103. 
Olhorst, s . l982, Diel migration patterns of demersal 
reef z ooplankton. J. Exp, Mar. Biol. Ecol . 60:1 - 15. 
Paine, R.T. 1966. Food web co mplexity and species 
diversity. Amer, Natur. 100:65-75, 
Palmer, M.A. . Hydrodynamics and st~ucture: 
interacti1v9e8 6effects on meiofauna dispersal. J. Exp. 
Mar. Biol. Ecol- 103 :1 -16 , 
Patterson, R.L. 1952. The sage grouse in Wyoming. Wyoming 
Game and Fish commission, Cheyenne. 154pp. 
90 
Policansky, D. 1982. Sex change in plants and animals. 
Ann. Rev. Ecol. Syst. 13:471-495. 
Porter, J. and K .G. Porter. 1977. Quantitative sampling 
of demersal plankton mig r ating from different coral 
reef subsrates. Limnol. 0ceanogr. 22(3):553-556. 
Pulliam, H.R. 1973. The advantage of flocking. J. Theor. 
Biol. 38: 419-422. 
Randall, J.E. 1961. A contribution to the biology of the 
convict surgeonfish of the Hawaiian Islands, 
Acanthurus triostegus sandvicensis. Pac. Sci. 
15:215- 272. 
__________ _ _ 1963. An analysis of the fish populations 
of artificial and natural reefs in the Virgin 
Islands. Carib. J. Sci. 3(1) :31-47. 
____________ . 1965. Grazing effect on sea grasses by 
herbivorous reef fishes in the West Indies. Ecology. 
46:255-260. 
_________ ___ . 1967. Food habits of reef fishes of the 
West Indies. Stud. Trop. 0ceanogr. 5:665-847. 
_____ __ _____ 1981 . Two new species and six new records 
of labrid fishes from the Red Sea. Senckenb. Marit. 
13(1/3) : 79-109 . 
Randall, J . E. and J.R. Chess. 1979 . A new species of 
garden eel (Congridae: Heterocongrinae) of the genus 
?~cg~?!~ from Hawaii. Pac. Sci. 33(1):17-23. 
Robel, R.J. and W.B. Ballard. 1974. Lek social 
organization and reproductive success in the greater 
prarie chicken. Am. Zool. 14 : 121-128. 
Robertson, D.R. 1972. Social control of sex reversal in a 
coral reef fish. Science. 177:1007- 1009. 
Robichaux, D., A. Cohen, M. Reaka, and D. Allen . 1981. 
Experi ments with zooplankton on coral reefs, or will 
the real demersal plankton please come up ? Mar. 
Ecol. 2(1):77-94. 
Ross, R.M. 1978. Reproductive behavior of the anemone 
fish, ~~eQlecl~~ ~~l~~~e~? of Guam. Copeia . 
1978(1):103-107. 
____ _____ . 1984. Anato mical changes associated with sex 
reversal in the fish Thalassoma dueerre~. Copeia . 
1984(1) : 245-248 . 
91 
Sale, P.F. 1972. Influenc e of corals in the dispersion o f 
the pomacentrid fish, Q~?~tll~? ~C~~~~? ? Ecology. 
53:741-7 44. 
______ _ __ 19 75. Patterns of use of space in a guild of 
ter r itorial reef fishes. Mar. Biol. 29:89-97 . 
_________ . 1977 . Maintenace of high divers ity on coral 
reef fish communities. Am. Natur. 111:337- 359 . 
______ _ __ . 1978a. Chance patterns of demographic change 
in populations of territorial fish in coral rubble 
patches at Heron reef. J. Exp. Mar. Biol. Ecol. 
34:233-243 . 
_________ 1978b. Coexistence of coral reef fishes - a 
lottery for living space. Env . Biol. Fish. 3(1):85-
102. 
_________ 1980. The ecology of fishes on coral reefs . 
Oceanogr. Mar. Biol. Ann. Rev. 18 : 367- 421. 
_______ _ _ . 1982. Stock - recruitment relationships and 
regional coexis tence in a lottery competitive 
system : a simulation study. Am. Nat. 120 :139- 159. 
S ale, P.F. and R. Dybdahl. 1975. Determinants of 
community structure for coral reef fishes in an 
e xperimental habitat. Ecology. 56:1343- 1355. 
S a l e, P. F . and D. McB. Williams. 19 8 2 . Community 
struct u re o f coral reef fishes: are t he pat t erns 
mo r e than those e xpec ted by c ha nce ? Am. Na t . 
120 : 121-127 . 
S ameoto, D.D. 1975 . T id a l a n d diu rnal effec ts o n 
zooplankton sample variabil ity in a n earshore marine 
environment. J . Fish. Res. Board. Can. 32 : 347-366 . 
Schmidt, H . E . 1973 . The vertical distribution and diu r na l 
migration of some zooplankton in the Bay of E ilat 
(Red Sea) . Helgolander Wiss. Meeresunters. 2 4: 333-
340 . 
Shimada, K. and T . Yoshino . 1984. A new trichontid fish 
fro m the Yaeyama Islands, Okinawa Prefecture, Japan. 
Japan. J. Ichthyol. 31(1):15-19. 
S mith, C.L. and J.C. Tyler. 1977. Redescription of the 
gobiid fish Cor~QhOQterus liQernes Bohlke and 
Robins, with notes on its habits and relationships. 
Am. Mus. Nat. Hist., Novitates. 2616 :1 - 10. 
92 
Wainwright, S .A. 1965 . Reef communities visited by the 
Israel So uth Red Sea Ex pedition, 1962. Bull. Sea 
Fish. Res. S ta. Haifa. 38:40-53. 
Warner, R .R. 1975. The adaptive sig ni ficance of 
sequential hermaphroditism in animals . Am. Nat . 
109 :61 -82 . 
__ _ ___ _ ___ _ . 1978 . The evolution of hermaphrodi tism and 
unisexuality in aquatic and terrestria l vertebrates. 
In : Contrasts in Behavior. Reese, E .S. and F. 
Lighter, eds. New York: Wiley I nterscience. pp. 77-
101 . 
Williams, D.McB. 1980. Dynamic s of the pomacentrid 
community on small patch reefs in One Tr ee Lagoon 
<Great Barrier Reef). Bull. Mar. Sci . 30 :1 59-170 . 
93 
embedding protocol
 for the 
pendix 1. Paraffi
n 
Ap s are emersed 
Autotechnicon 2A. 
Gonadal tissue
then embedded in a
 
r and 
in the solutions i
n orde
affin block for se
ctioning. 
par
TIME IN ?Q~~IlQ~ 
?QL-U--T-IO-N- <min) 
60 
70 ? Ethanol I 
60 
70 X Ethanol I I 
120 
so X Ethanol I 
60 
95 ? Ethanol I 
60 
100 X Ethanol I 
60 
Clear 100 ? Ethanol 
90 
Clear I 
30 
Clear I I 
60 
Clear III 60 
Paraffin Clear 
60 
Paraffin I 60 
Paraffin II 
--------------------------
---
--- ---------- ------------  
 l tion C650 ( Eth
ylene dichloride,
Clear= Cl . S l 
tate) 
earing o u 
tetrachloride, N-B
uty ace 
Car bon 
plast (melting poi
nt 56-57 
Para
Pa raffin used was 
degrees C) 
94 
Appendix 2 . Hematoxylin and eosin staining protocol for 
light microscopy . 
5 
Xylene in I 
5 
Xylene in II 
5 
Xylene : 100? EtOH 
5 
100? EtOH in I 
5 
100? EtOH in II 
5 
95? EtOH 
5 
70? EtOH 
5 
50? EtOH 
5 
30? EtOH 
0.33 
Distilled water 
3-4 
Hematoxylin 
rinse 
Distilled water 
1- 2 
Acid alchol (HCl:30? Et ?H> 
until 
Alkaline alcohol <NaOH:30? EtOH> blue 
5 
30? EtOH 
5 
50? EtOH 5 
70? EtOH 5 
95? EtOH 2-4 
Eosin Y (5? sol'n) (O.S?M 95? EtOH) 
0.33 
95? EtOH 0.33 
100? EtOH out I 0 .33 
100? EtOH out II 
95 
<Appendix 2, can't) 
5 
Xylene out I 
5 
Xylene out II 
5 
To ulene 
96 
PENDIX 3: ures perfo
rmed on 
AP
s from the AN
OVA proced
s of squares 
and 
The result on data. The
 sum rmed 
th kt sfoe transformed
 plan  tran
values were l
eft in their
th quare represents a 
square-
e mean s SR" on the ta
xa 
 "L" on the 
Values. A p
refix " fix
ion of the da
ta. The pre
mat ) transformat
ion of 
root transfor thmic (base 
10
metic 
ta xa represen
ts a logri
-26, containin
g the arith een 
th tables 2
tw
e data. The or of the int
eractions be
e standard er
r
transformed, 
thus, 
means and th, and time, h
ave been un The arithmeti
c 
Site, height t mean abunda
nces. 
h en lankton
 are 
t e values repr
es  the p
 representati
ve taxa from sence of 
means from s 1-5) to 
show the ab s 
phed (Append
ix 3, Figure ample
Qra tterns within
 the th ree s
ion pa
Vertical mig
rat
y. 
taken during 
each da
97 
Anillysis of pJ
A ilnp kp tt on nd  i1 o n T il lb ll I le   . i;ites 6ene rill Linur "odels Procedure 
Dtptndent Vilriilblt: SRSAST 
Su? of "un 
Squires Squirt F Vi!Ue Pr > F 
Source DF 
0.0001 
26 IOl.2b2B52 3,mm 
45. 75 
"ode! 
54 4.596711 0.085124 Error 
Corrected Toh! BO 
105.859566 
SR6AST "un 
A-Sq Ru oire C. Y. 
ot "SE 
7.73om 4 0.2917bl 
3. 77392 03 1 
o. 956577 
DF P rT  y )p  e F  JI 1 55 "ean Squirt F Vi 1u e 
Source 
0.000
43,385vb 21.msJ 25U3 
: 
SITE 0 .0001 2 3,8198b 1.90993 
22 ,44 
T!IIE 0.0001 4 20.29840 5,07"60 
59.bl 
SJTE?TINE 2,94367 3U8 0.0001 2 5.88i35 
HEIGHT 5,20980 1.30245 15.30 
O.OvOl 
TIIIE?HEJGHT 0.0001 
4 14 ,73036 3.b8259 
43.26 
SITE?HEIGHT 7. 93203 0.99150 ll.b5 
0.0001 
SITE? TINE ?HE J6Hi 8 
-------------------------- - - - -------------------------------
Gene ril Linur Node ls Procedure 
Dtpendent Yirub Je: SRCY CL 
Squir t F Vilue Pr ) F 
DF 
Source 
57.36 0.0001 
Error 
80 227 .231>688 
Corncted Toti! 
SRCYCL IIHn 
R-Squilre C. Y. Root NSE 
U5350853 
o.mos4 8.omm o.383476 
Pr ) F 
Df TY1Jt I JI 55 11,,.n Squ Fi  r Vt 1lue 
Source 
111,86712 55.m5c 
380.36 0.0001 
2 
SITE 25,07758 12,53879 
B5.27 0.0001 
2 6,51252 44,29 0.0001 T1"E 4 21,.05007 6B ,67 0.0001 SITEtTIIIE 2 20,19511 
!O,Of755 
1(!6HT 6,40811 1,60203 
10.B9 0.0001 
0
T .0l1 01 0?t 1H  E16HT 7.51B58 1,B7964 
12. 78 
4 2. 77240 1B.85 0.0001 sm?HEI6HT 8 22.17921 S!TEtT111EtHEI6Hl 
(Appendix Table I con'tl 
6fneril lineu "odels Procedun 
Dtptndent Variable: SRCALA 
Su, of "un 
Source DF Squares Squire F Vilue Pr > F 
"odel 26 413,195007 15.892116 80.60 0.0001 
Error 5" 10.647528 0.197176 
Corrected Tohl BO m.B42535 
R-Squue c.v. Root NSE SRCALA Nun 
o.m879 8.1701449 0.444046 5.43497756 
Source DF Type 111 SS Nun Squue F Vilue Pr ) F 
SITE 207 .17531 103.58765 52~.36 0.000 1 
TINE 33 .44074 16. 72037 8Uo 0.0001 
SITE?TINE 4 123.49332 30 .87333 156.58 C.0001 
HEIGHT 2 24. 16200 12,08100 61.27 0.0001 
Tl"EtHEJ6HT 4 1. 13007 0. 28252 1.43 0 .2357 
SITE?HE16HT 4 7 .00621 I. 75155 8.ee 0,0001 
SITEITINE?HEl6Hl e 16, 7873b 2,09842 10 ,64 0.0001 
?------------------------------------?-???--------------------?--?------
6fntril lintar Nodels Procedure 
Dtptndent Variable: SRHARP 
5111 of "un 
Source DF Squues Squirt F Vi lue Pr > F 
"odel 26 76.1843004 2.9301654 28. I 1 0.0001 
Error 5" 5.6280088 0.1~2224 
Corncted lohl 80 81.8123092 
R-Squire c.v. Root NSE SltttARP Nun 
0.931208 6,6418258 0.322835 4.8606351 4 
5ource Df lYl'I Ill SS llt1n Squar1 F Value Pr > F 
SITE 2 ,.11880 --~940 43.75 0.0001 
TIii? 2 1.36m uem 40.13 0,0001 
&ITE?TI"E 4 24.02640 6,00660 57.63 0.0001 
HEl6Hl 2 2,24594 1.12297 10.77 0.0001 
TIIIE ?HE I &HT 4 7.66520 l.'1630 18.39 0.0001 
SIT?t1El6Hl 4 IU0441 3,22610 30.'5 0.0001 
SI TEt Tl llE ?HE I 6Kl e 11 .115882 1.48235 14.22 0.0001 
99 
(Appendi x hble 1 con'tl 
Seneril Lineu "odeh Procedure 
Dependent Vu i1b It: LCNAUP 
Sui of "tin 
Source DF Squares Squirt F Vilue Pr > F 
"ode! 26 11.3243167 0.4355506 33.31 0.0001 
Erro r 54 0. 7061691 0.0130772 
Cor rec ted Total 80 12 .0304858 
A-Square c.v. Root "SE LCNAUP Nei n 
0.941302 6.3541613 0. I 14356 1.mm31 
Source DF Type III SS Nun Squire F Value Pr > F 
SITE 2 1.mrn 3.507419 268.21 0.0001 
TINE 2 0, 4094 06 0.204703 15 .6~ 0.0001 
SITEtT1"E 4 2.477507 0.619377 47 .36 0.0001 
HE! 6Hl 2 0. 059703 0.029851 2. 28 0.1118 
T!NE?HE!6HT 4 0.046749 0.01166i 0.89 0.4741 
SITE tHE l 6HT 4 0.245423 0,061356 4,69 0. 0025 
SITEtT!NE?HEISHT 8 1. 0706ql o. mm 10.23 0,0001 
-----------------------------------------------------------.. ------------
6tntril Lineu Nodels Proced ur e 
Dependent Vuuble : LIE66 
Sui of "un 
Source DF Squues 5quue F Value Pr > F 
"ode I 26 5.nmm 0.22051785 46 .62 0.0001 
Error 54 0,255446911 o.oomoso 
Corrected Total 80 S.'8891107 
A-Squue c.v. Root NSE LIE6li "un 
0.'157347 umm 0.068779 1.48180179 
Sourt! DF Type III 55 ""n Squne F Vilut Pr > F 
SITE 2 uum 0.B31832 l'r.i.84 0.0001 
T1"E 2 0.37BBIB o.1mo9 40.04 0.0001 
SITEIT1"E 4 1,533905 0.383476 11.06 0.0001 
ll16HT 2 0.104B71 0.052436 II .OIi 0.0001 
TIIIEtHEISHT 4 o.oemo 0.021965 4.64 0.0027 
SITEtllE16HT 4 1,477303 0.369326 78.07 0.0001 
SI TEt TIIIEtHE I &HT 8 0,4B7043 0,0&0880 12.87 0.0001 
100 
(Appendix hblt I con't) 
6tner?l Linear "odels Proctdure 
Dtpendent Vuiable: SRNENA 
Sui of "un 
Source DF Squues Square F Value Pr > F 
"ode! 26 18.9229030 o. 7278040 5.85 0.000! 
Error 54 1i .1221m 0.1244952 
Corrtcted Toh! 80 25.6456465 
R-Square c. v. Root "SE SRNE"~ Nu n 
o. 73781,0 24.510524 0.352839 t.43954019 
Source DF Type 111 SS "ean Square F Value Pr } F 
SITE 2 t.031588 0.515794 4.14 0.0212 
TINE 2 3.483219 I. 741609 13. 99 0.0001 
SITE?Tl"E 4 I. 173793 0.293448 2.36 0.01,SO 
HEi6H1 2 U13980 0.406990 3.27 0.0457 
Tl"E?HEISHT 4 3.ml68 0.962292 7. 73 0.0001 
SlTEtHE16Hi 4 3.mm o. 799006 6.42 0 .0003 
Sl TE t mE ?HE l 6HT 8 5.375131 0.671891 5.40 0.0001 
----------------------?-----------... -------------------------------------
6tneral Linur "odels Proctdure 
Deptndent Vari1ble: LRADIO 
Sui of "'?r, 
Source DF Squirts Square F Value Pr ) F 
ftodel 26 14.2635560 o.~85983 10.85 0.0001 
Error 54 2. 7311288 0.0505765 
Corrtchd Toh! 80 16.,mm 
R-Square c.v. Root "SE LRADIO "un 
0.839295 16.832418 0.224892 I ,331,06537 
Source Df TYJJt Ill SS llt1n Square F Value Pr ) F 
SITE 2 to.mm 5,108340 IOI ,00 0.0001 
TIIIE 2 0.879179 0.43'590 8,1,9 0,0005 
SlTEtTIIIE 4 l,BUOOO 0.4~000 8,98 0.0001 
1?16HT 2 0.012m 0.006177 0,12 0.8853 
TIIIEtllEISHT 4 MUB53 0.065463 l.!9 G.2838 
SITEtlEISHT 4 o.",250 o. 112313 2.22 0.0789 
SITEtTIIIEtHEISHT 8 0.1128239 0.078530 !.~ 0.1612 
101 
<Appendix Tible I co6ne'ntlm 1l Linur "odels Procedure 
Dependent Variable: SRFORA" 
Su? of "ean Pr > F 
DF Squares 
Square F Value 
Source 
2 ,3758887 10.01 0.0001 26 61. 7731060 
"ode! 
54 12,8176807 0.2mm 
Error 
74 ,5907868 
Corrected Total BO 
Root "5E SRFORA" "ean 
R-Square c.v. 
o.m201 2 .82822898 
0.828160 17 .226350 
DF Type Ill SS "un Square F Vilue 
Pr > F 
Source 
25,69661 7 12,848309 
54.13 0.0001 
2 
SITE 2,271663 t.135832 
4. 79 0.0122 
2 5. 16 0,0014 TINE 4 4,897246 
1,224312 
SiTEtT!NE 0.120173 0.51 0.6056 2 0.240346 
HE16HT U79118 1. 144779 
4.8, 0.0021 
4 
Tl"EtHEIGHT 14,188321 3,547080 
14. 94 0,0001 
4 
S! TE tHE I GHT 1.237474 5,21 0.0001 q,899792 
S! TE 8 t Tl"E ?H?l6HT 
---------------------------------------------------- -------------------
General Lintar "odels Procedure 
Dependent Var iable: SRCLAD Su? of "ean F Value Pr > F 
DF Squares 
Square 
Source 
1. 8426263 16, 76 0,0001 
26 47,9082831 
"ode I 
5, 9375705 0, 1099550 54 
Error 
BO 53,8458536 
Corrected Total 
SRCLAD llun 
R-Square C.V. Root "SE 
1.ommo 
0,889730 30,355667 0,331595 
F Value Pr > F 
DF Type 111 SS 11t1n Square 
Source 
1,ffl4191 16,18 0.0001 
2 3,55118383 0,0001 
SITE 3,048m9 1,5244'85 
13,86 
2 33,00 0.0001 
TINE 14,5143'106 3,6285977 4 23,25 0,0001 
sm,mE 2 5, 1118282 
2,5559141 a. 76 0,0001 
HEIGHT Mo274BO 4 3,8509919 U7 t.0003 
TJIIEtHEIGHT 4 2,7596228 
0,  61'1'105 7 
1,8829518 17,12 0.0001 sm?HE16HT 8 15,0636144 
SJTEtT111EtHEJ6HT 
102 
(Appendix hble 1 con'tl 
6enm1l Linear "odels Procedure 
Dependent Vui;ible: SRBNAUP 
Sul of "ean 
Source DF Squares Squ;ire F V;ilue Pr > F 
"ode I 26 16.9312496 0.6512019 5.45 0.0001 
Error 54 6.mern 0.1mm 
Corrected Total 80 23.3B41107 
R-Square c.v. Root "SE SRBNAUP "ean 
0. 724049 2UB364 1 o.mm 1. 16065057 
SourcE DF Type I I I SS "ean Square F Yalu? Pr > F 
SITE 2 1.544665 o. 772332 6.46 O.OQ3C 
TINE 2 2.155491 1.077745 U2 0.00(?4 
SITE?T1"E 3.049343 0. 762336 6, 3B 0.0003 
HEIGHT 0.423270 0.211635 1. 77 0. 1799 
T1"E?HE16HT 2.246589 0.561617 4. 70 0.0025 
SITE?HEISHT 4 I. 723762 0, 430941 3.61 0.0 11 2 
SITE?T 1"E?HE16HT 8 5.788130 0.72351 6 6.05 0.0001 
-------------------------------- ----------------?-----------------------
General Linear "odels Procedure 
Dependent Vari;ible: SROSTR 
Sul of titan 
Source DF Squares Square F Value Pr > F 
"odel 26 21.323391,1 0.8201306 U9 0.0001 
Error 5lt 9,6551122 0. 17879B4 
Corrected Total 80 30 . 97850B2 
R-Square c.v. Root "SE SROSTR llean 
MBB329 32.218383 0.422846 1.31243573 
Source DF Type Ill SS llt1n ~um F Value Pr > F 
SITE 2 6.191611 3.095805 17.31 0.0001 
T1"E 2 2.489716 1.!4485B 6.96 0.0020 
SITE?T1"E 4 3.609051 0,902263 5.05 0.0016 
IIEl6HT 2 0,495305 o.mm 1.39 0,2590 
TIIIE?HEl6HT 4 2.243882 0.560970 3.14 0.021b 
SITEtffE16HT 4 2,937737 0.734434 4.11 0.Offl 
SITEfTIIIEtHEl6HT e 3.356m 0.419512 2.35 0.0304 
103 
(Appendix hble I con'tl 
6entril Linear "odel s Procedure 
Dtpendent Variible: SRPLARV 
Sui of "ean 
Source DF Squares Square F Vi Jue Pr > F 
"ode! 26 37. 7222871 1.4508572 10. 12 0.000) 
Error 54 7.7385188 o. 1rno59 
Corrected Toti! BO 45 .4608059 
R-Square C. V. Root "SE SRPLARV Nm 
o.mm 2U9041 7 0.378558 I. 72146648 
Source DF Type 11 I SS "un Squire F V;.lue Pr ; F 
SITE 2 10. 327138 5.163569 36.03 0.0(101 
T1"E 2 4.656693 2. 32834~ 16.25 0.000! 
SITEtTI NE 13.100594 3.275146 22 .85 0.0001 
HE 16HT 0.620759 0.310380 2. li 0.1245 
T1"E?HEl6HT 2.776274 0. 694068 4.84 0.0021 
SITE?HE16HT 4 2.mm 0.61.9216 4.67 0. 0026 
SITE?mEtHE16HT 8 3.563966 0.445496 3. 11 0 .0058 
------------------------------------------------------------------------
6eneul Linur "ode!, Procedure 
Dtpendent Viriible: SR8I V~l 
Sui of "un 
Source DF Squares Square F Value Pr > F 
"ode! 26 17.4308556 0.6707252 5.48 0.0001 
Erro r 54 o.6130103 0.1221!632 
Corrected Toh! 80 24.o51Bm 
R-Square c.v. Root "SE SRBIVAL """ 
o.7 25052 31.424928 0.mm I. 11359796 
Source DF Type I I I SS """ Squire F V1lue Pr ) F 
SITE 2 7 .556180 3. 778090 30 ,85 0.0001 
T1"E 2 0.mm 0.271807 2.22 0.1185 
SITEtT1"E 4 1.254298 0.313574 2.56 0 ,0488 
!EIGHT 2 Ml5141 0.207m 1.69 0. 1932 
TIIIEtHEl6HT 4 1,534501 0.383625 3.13 0,0217 
SITE?HEI&HT 4 1.5119330 0.3'2333 3.20 0.0197 
SITEtTIIIEtHEl&HT B ~.565789 0.570724 4.66 0.0002 
104 
(Appendix lltlle I con ' tl 
6eneril linen llode ls Procedure 
Dependent Yuiible: SRLARY 
Su? of llun 
Source DF Squues Squire F Yilue Pr ) F 
llodel 26 IB.220B0B7 0. 70 CIB003 7 . 34 0.0001 
Error 54 5. 158768B 0 ,0955328 
Corrected Tohl BO 23. 3795775 
R-Squue c.v. Root IISE SRLARY lle ar. 
o.mm 32 . 755900 0, 3090B4 0,94359715 
Source DF Type I I I SS llean Squa re F Yilue Pr ) F 
SITE 2 5.6620B38 2.B3 10419 29 . 63 0.0001 
TINE 2 2.5B90B63 1.msm 13 .55 0.0(1(11 
SJTEtTIIIE 4 2.6310644 0.6577661 6.B9 0.0001 
HEI GHT 2 0.4B29043 0.2414521 2.53 0.0B93 
TIIIE?HEIGHT 4 3.mB071 O.BB645lB 9.28 O.OOOl 
SJTEtHEIGHT 4 0,47695B3 0. 1192396 1.25 0.30l7 
SITEtTIIIEtHEISHT 8 2.8329045 0.3541131 3. 71 0. 0016 
---------.. --------------------------------------------------------------
General Lineu llode ls Procedure 
Dependent Yuiible : SRAIIPHI 
Su? of llun 
Source DF Squues Square F V?l ue Pr > F 
llode l 26 2 .06B93844 0.07957456 UI 0.0001 
Erro r 5" 0.912666H 0.01690123 
Corrected Tohl 80 2.9816048B 
R-Squue c.v. Root IISE SRAIIPHJ llun 
0.6mOI 17. 154944 0.130005 0.757826 61 
Source Df Type Ill SS IIHn Squm F Vilue Pr ) F 
SITE 2 Ml6?m 0.2083726 12 .33 0.0001 
TIIIE 2 o.~u,20 0.0223310 1.32 0.2753 
SITEtT1"E 4 0.0893239 O.Oc23310 1.32 0.2738 
1?16HT 2 O.c/90662 0.1395331 B,26 0.0007 
TIIIEtHEJ6HT 4 0.2270029 0.0567507 3.36 0.0159 
6ITEt!IEJ6HT 4 0.55111324 0.1mm 1.26 ,.0001 
SITE tT I IIE tllE 16\H a 0.4540059 0.0567507 3.36 8.0034 
105 
IApp,ndix hble I con'tl 
General Linear "odel s Procedure 
Dependent Variable: SRFE66 
Su? of "ean 
DF Squares Square 
F Value Pr > F 
Source 
0,18718741 U9 0.0001 26 4, 8668726 7 Node! 
54 2,15714136 0.03994706 Error 
7 ,02401403 
Corrected Tota l 80 
R-Square c.v. Root "SE 
SRFE66 "ean 
24. 598325 o. 199868 o. 81252530 0,692891 
DF Type Ill Pr > FS  S Nean Square F V;ilue 
Source 
0.3915917 O. 1957959 
4. 90 0.011 1 
2 
SITE o. 1337618 0 .0668809 1.67 
0. 197(? 
2 
TINE u o 0.0029 4 0. 73500B9 o. 1837522 
SITE?TINE o. 1771178 0.0B85589 2.22 0.1188 2 
HEIGHT o. 4.B2 0,0021 7708501 0.1927125 
T1"EtHE16Hl 3.21 o.om 
4 0.5130202 
0.1282551 
SITE?HEIGHT 8 2, 1455220 
0,2b81903 6, 71 0.000 1 
SiTEtTI"EtHEIGHT 
---------------------- -- - ------- -----------------------------
6enu al Li near "ode ls Procedurt 
Dependent Vuiable: SRC"EG Su? of "ear, 
Squart F Value Pr > F 
DF Squares 
Source 
1,3371696 10 , 28 0.000 1 
26 34, 766009 
N~del 
7,0258487 o. 1301083 S4 
Error 
80 41 ,792259b Corncted Total 
C,V, Root NSE SRC"E6 "un 
R-Squ;ire 
1.25071703 
0,831886 211 ,839880 0.360705 
Pr > F 
Df Type Ill SS llun Sq Fu  a Vre 1 Iue 
Source 
1.11mo 66,96 0,0001 
2 17.424480 0,421655 3.24 o.om SITE 2 0.843310 1,93 0, 118B 
TIIIE 1,003946 0,250986 4 
SJTEtTINE !. 7B5697 O,B92849 
6.86 0.0022 
2 12,93 0.0001 
lli6HT 4 6. 727B83 
l,o81971 
0,27 o.am 
TUIEtHE16HT 4 0. 141511 
0,035378 
o.&~948 6,57 0,0001 SITE tHE I6 HT 6.B39583 
SiTE B  t Tl IIE tllE I& HT 
106 
<AppendiK Tible II General Linear "odel5 Procedure 
Dependent Variable: SRTLARV 
Su? of "ean F Value Pr > F 
DF Squnes 
Square 
Source 
I. 29502791 0.04980877 
o. 97 0.5197 
26 
"odel 
0,05133803 
54 2. 77225383 
Error 
Corrected Total BO 
4,06728174 
C, V, Root "SE 
SRTLAR V Ne.in 
R-Square 
o.226579 0, 740643 02 
o.318401 30. 592198 
F Value Pr ) F 
DF Type l l l SS "ean Squire 
Source 
0.1821980 0,091099(
1 I, 77 0.1 m 
!. 77 0.1793 
SITE 2 0,1B21980 
0,091 (?990 
0.1m 
rm 1.77 4 0. 3b439b0 0,091 0990 
sm+mE 0,0314575 0,bl 
0.5456 
2 0.0629151 o.03lm5 0.61 
0, b553 
HEJ6 HT 0, 1258302 o.6m 
TJNEtHEI6Hl 0.1258302 0.0314575 
0.61 
4 0.0314575 0.1,1 0. 7632 SJTE?HEJ6HT 8 0.251bb04 
SI TE? TINE +HEJ6HT 
------------------------------------------------------------------------
General Linear Nodeh Procedure 
Dependent Variable: SREPLEUT Su? of "'an F V;ilue Pr > F 
DF Squares 
Square 
Source 
o.1mm2 2.50 
0.0023 
26 3,53533751 
Nod el 
54 2, 93998452 
0.05444416 
Error 
BO 1,,47532203 
Corrected Total SREPLEUT "ean 
C, Y. Root "SE 
R-Square 
0,233333 o. 777973b! 
0,545971 29. 992369 
F Value Pr > F 
DF Type Ill SS "'an Squm 
Source 7 ,47 0.0014 
2 o.Bl358l4 
M067907 
0, 132b 
0,2284740 0, 1142370 
2.10 
SITE 2 2.10 0.0937 
T1"E 4 0.451,9481 
o.u-2370 o.1rn 
SJTE?Tl"E 0,0312235 0,0!5b118 
0.29 
2 2,97 0.0272 
lll6HT 4 0,1,4~545 
o.WBB81i 
0,29 0.1853 
T1"EtHEl6HT 4 o.0o24m 
0,0151,118 
o.1618886 2,97 
0.0078 
SITE?HE16HT B 1,2951089 
SITE?TIIIE?HEJ6HT 
107 
(Appendix lible I con'tl 
6enerill linur ftodels Procedure 
Dependent Vi1rii1ble: LDINO 
Su? of ftea n 
Source DF Squares Square F Viilue Pr ) F 
ftodel 26 4.01024113 0.1542400~ 15.00 0.0001 
Erro r 54 0.55515714 0,0102B0t9 
Corrected Total BO 4. 56539827 
R-Squilre c.v. Root ftSE LDINO Nur, 
0 .878399 6. 7235856 o. 101394 I. 50B0305t 
Source DF Type II l 55 fteil n Squi re F Viilue Pr > F 
SITE 2 o.mrn 0.231265 22, ~() 0.00(1 1 
T1"E 2 1.mm 0.715376 6UB 0,0001 
SITEtTINE 4 0.482949 0.120737 11. 74 0.0001 
HE16HT 2 0 .209987 0. 104993 10.21 0,0002 
TINE +HEIGHT 4 0.143755 0.035939 3.50 0.0131 
SITE?HEIGHT 4 0.124181 0.031045 3.02 0. 0255 
SITE? TlftE+HEIGHT 8 I. 15608b 0.144511 14 .06 0.0001 
------------------------------------------------------------------------
6enea l Linur ftodels Procedun 
Dtpendent Vi1rii1ble: SRNYSlD 
Su? of ftun 
SourcE DF Squnes Squne F Value Pr ) F 
ftodel 26 0 .08600838 0.00330601 1.00 o.,m 
Error 54 0. I 7863279 0.00330801 
Corrected Total BO 0.2641>4118 
R-Squne c.v. Root ftSE SRftYSID fttan 
0.325000 8.0610448 o.om15 0.71349737 
Source Df Type Ill 55 ftun Squue F V&lue Pr > F 
SITE 2 O.OOUlbO 0.0033080 1.00 0.374!, 
T1"E 2 0.00.6160 0.0033080 1.00 0,3746 
SITEtT1"E 4 0.0132321 0,0033080 1.00 0.4157 
ll16HT 2 0,0066160 0.0033080 1.00 0.3746 
TlftE tHE I 6HT 4 0.0132321 0.0033080 1.00 0.4157 
SITE?HEIGHT 4 0.0132321 0,0033080 1.00 0,4157 
SITEtTIIIEtffE16HT 8 0.0264641 0.0033080 I .00 0.4469 
108 
(Appendix hble I con'tl 
Gener? ! Linen "odels Procedure 
Dependent Virilble: SRCZOE 
Su? of "ean 
Source OF Squires Squire F V,lue Pr ) F 
"ode! 21, t.08147558 0.04159521 1.51, 0 .0838 
Error 54 1.43873084 0.021,64316 
Corrected Toh! BO 2 .52020642 
R-Squire c.v . Root "SE SRCZOE Near, 
0.429122 21.664661 o. 163227 0.753421.b9 
Source OF Type I I I SS Nean Squire F Vilue Pr I F 
SITE 2 o.osmeo 0.0274640 1.03 0.3636 
TINE 2 0. 0549280 0. 0274640 1.03 0, 363b 
sm,mE 4 0.1347770 o.omm 1.26 0 .2952 
HEIG HT 2 o.1m102 0.0622351 2.34 0. 1064 
Tl"EtHE16HT 4 0.0652348 0.0163087 0.61 0.6557 
SITE?HE!6HT 4 0,06523118 0.0163087 0.61 0.6557 
SITE?T1"E?HE16HT B 0.5819027 0. 0727378 2. 73 0.0133 
------------------------------------------------------------------------
6ener?l Linur "odels Proced ur e 
Dependent Vui,ble : SRNO CT 
Sui of "ean 
Source OF Squires Squire F V,ilue Pr ) F 
"ode! 26 o. 08600B3e 0.00330801 1.00 0.4842 
Error 54 0.178o3279 0.00330801 
Corrected Tohl BO 0,26464118 
R-Squ,re C. V. Root "SE SRNO CT "un 
0.325000 B.0610m 0.057515 o. 71349737 
Source DF Type Ill SS lltln ~uue F V,ilue Pr > F 
SITE 2 o.~mo 0.0033080 1.00 0.3746 
mE 2 0.0066160 0.0033080 1.00 0.3746 
SITE?T1"E 4 0.0132321 0.0033080 1.00 o.m7 
1?!6HT 2 0.00&6160 0.0033080 1.00 0.3746 
TJIIEtHEIGHT 4 0.0132321 0.0033080 1.00 o.m7 
SITE?l?16HT 4 0.0132321 0.0033080 1.00 Ml57 
S! TE t TIIIE tltE I 6HT B 0.0264641 0.0033080 1.00 o."119 
109 
(Appendix Tible I con'tl 
6ener?l linen ftodels Procedure 
Dependent Y&ri1ble: SRBRYZ 
Sui of "un 
e DF Squires Squ
ire F Yilue Pr > F 
Sourc
2b 13,5b2lb12 0.521b21b ue 0.0001 ftodel 
54 5 .8975271 0. 1092135 Error 
Corrected Tot1l BO 19,45%883 
SRBRYZ Nein 
A-Squire C.Y. Root ftSE 
o.mm 29 .88b4b9 0,330475 1. 1057bbbb 
DF Type Ill SS Nun Squire F Yilue Pr l F 
Source 
2 3,1mm 1.8938233 
17 .34 0.000! 
SITE b.80 0.0023 
TINE 2 
1,4843097 0.7421549 
4 2.m4039 o.m1 010 
6.85 0 .0002 
sm,mE 
2 I. 362841 0 0.b814205 
6.24 0.0036 
HEIGHT 0.4891595 0.1mm 1.12 0.35b9 
TINEtHE!SHT 2.00 0.1074 
SITE tHE l SHT 4 
0.8743213 0.2185803 
sm,mE?HEISHT 8 2.5714792 
0.3214349 2. 94 0 .0084 
110 
Appendix 
Table 2. The arith metic means and standard 
Error of the interaction between site, 
time, and height for the taxon Gastropoda. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 18 . 65526 
AQ E 1 9.984087 
AQ E 1.67 23 . 62241 
AQ M 0.33 11 . 63393 
AQ M 1 14 . 98923 
AQ M 1.67 7.920358 
AQ N 0.33 11.63393 
AQ N 1 9.984087 
AQ N 1. 67 9.984087 
MM E 0.33 3 . 501969 
MM E 1 6,909334 
MM E 1.67 4 . 969421 
MM M 0.33 13 . 65132 
MM M 1 11.98664 
MM M 1.67 13. 6 5132 
MM N 0.33 8 . 290155 
MM N 1 4 . 254811 
MM N 1.67 10 . 23672 
RM E 0 . 33 5 . 324057 
RM E 1 15 . 98988 
RM E 1. 67 26 . 32494 
RM M 0 . 33 16 . 19953 
RM M 1 30 . 93249 
RM M 1. 67 21.20463 
RM N 0.33 15 . 98988 
RM N 1 40 . 64930 
RM N 1.67 36 . 32724 
Std . error of the mean = 0 . 028374 
111 
Appendi x 
Table 3. The arithmetic mean and s tandard 
error of the interaction be t ween s ite, 
time, and height for the ta xon c yclopoid 
Copepoda. 
S ITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 23.99319 
AQ E 1 20.61602 
AQ E 1.67 22.64968 
AQ M 0.33 14.98923 
AQ M 1 5.292589 
AQ M 1. 67 16.91359 
AQ N 0.33 10.32828 
AQ N 1 2.951281 
AQ N 1.67 21.94645 
MM E 0.33 13.55924 
MM E 1 12.88736 
MM E 1.67 16.87976 
MM M 0.33 6.609860 
MM M 1 15.30908 
MM M 1. 67 25.63949 
MM N 0.33 22.95155 
MM N 1 13.55924 
MM N 1. 67 15.00000 
RM E 0.33 32.30137 
RM E 1 65.98997 
RM E 1 .67 82.32062 
RM M 0.33 25.88124 
RM M 1 17.96384 
RM M 1. 67 23.95097 
RM N 0.33 39.65692 
RM N 1 35.25445 
RM N 1.67 64.29098 
Std. error of the mean= 0.049018 
112 
Appendix 
Table 4. The arithmetic means and standard 
error for the interaction between site, 
time, and height for the taxon calanoid 
Copepoda. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 14.66296 
AQ E 1 15.28619 
AQ E 1. 67 26.31895 
AQ M 0.33 10.98547 
AQ M 1 18.22461 
AQ M 1. 67 15.66319 
AQ N 0.33 10.66161 
AQ N 1 7.870003 
AQ N 1.67 10.63091 
MM E 0.33 14.96658 
MM E 1 9.984087 
MM E 1. 67 12.46249 
MM M 0.33 35.28925 
MM M 1 46.32506 
MM M 1 .67 66.59098 
MM N 0.33 12.58415 
MM N 1 13.98849 
MM N 1.67 21.28536 
RM E 0.33 69.32778 
RM E 1 120.6607 
RM E 1.67 123.9720 
RM M 0 . 33 36.17469 
RM M 1 26.33127 
RM M 1. 67 56.00251 
RM N 0.33 27.99415 
RM N 1 35.21339 
RM N 1. 67 62 . 32716 
Std. error of the mean= 0.065725 
113 
TAapbpleen d5ix.  The arithmeti c means and s tandard 
error of the interaction between si te , 
time, and height for the ta xon harp a ct ic o id 
Copepoda. 
HEIGHT ARITHMETI C S IT E TIME MEANS 
o.33 25.63949 
AQ E 18.62975 
AQ E 1 
E 1. 67 
26.97 572 
AQ 
o.33 16.62553 
AQ M 20.56766 
AQ M 1 
M 1.67 
16.33006 
AQ o.33 16.96176 
AQ N 13.30561 
AQ N 1 1.67 15.98988 
AQ N 19.66388 
E o.33 MM 15.87262 
MM E 1 15.52618 
MM E 
1-67 16.64364 
o.33 
MM M 20.64806 
MM M 1 38.32337 1- 67 
MM M 49.49143 o.33 
MM N 16.62553 
MM N 1 43.66167 1.67 
MM N 21.63467 o.33 
RM E 44.99633 
RM E 1 23.99319 
E 1. 67 RM 12.94704 
M o.33 RM 17.97364 
RM M 1 19.54939 1-67 
RM M 27.65886 o.33 
RM N 36.24157 
RM N 
1 37.31400 
1. 67 
RM N 
Std. error of the mean = o.0347407959 
114 
Appendix 
Table 6. The arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon Copepod 
nauplii. 
ARITHMETIC 
SITE TIME HEIGHT MEANS 
0.33 67.89980 AQ E 
1 41.17388 AQ E 
1.67 54.90459 AQ E 
 0.33 20.3442
8 
AQ M
1 13.75299 AQ M 17.17726 
AQ M 1.67 16.76864 
AQ N 0.33 
1 37.96578 AQ N 
1. 67 11.00231 AQ N 
0.33 37.28797 MM E 9 
MM E 1 
52.9752
1.67 33.60508 MM E 
o.33 82.84827 MM M 
1 183.9444 MM M 
M 1.67 
169.5750 
MM 65.84317 
MM N o.33 24.03249 
1 
MM N 66.92147 
MM N 
1.67 
0.33 117.9123 
RM E 235.6437 
1 
RM E 1.67 213.2952 
RM E o.33 99.7819
3 
RM M 124.7518 
1 
RM M  1.67 136.076
2
RM M 105.2697 o.33 
RM N 119.3212 1 
RM N 174.5648 1-67 
RM N 
:::: 1.164188 
Std. error of the mean 
115 
Appendix 
Table 7. The arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon invertebrat e 
eggs . 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 41. 20029 
AQ E 1 37.62985 
AQ E 1. 67 22.26966 
AQ M 0.33 32.61237 
AQ M 1 13.61228 
AQ M 1. 67 9.670679 
AQ N 0.33 20.60164 
AQ N 1 10.65695 
AQ N 1. 67 7.617738 
MM E 0.33 22.32861 
MM E 1 14.83555 
MM E 1.67 34.58078 
MM M 0.33 16.24669 
MM M 1 43.31590 
MM M 1.67 74.65634 
MM N 0.33 37.29502 
MM N 1 24. 12998 
MM N 1.67 55.61232 
RM E 0.33 36.64583 
RM E 1 74.65634 
RM E 1. 67 101.2932 
RM M 0.33 20.14189 
RM M 1 21.94188 
RM M 1. 67 31. 96844 
RM N 0.33 40.63246 
RM N 1 52.20673 
RM N 1.67 54.33133 
Std. error of the mean= 1.095744 
116 
Appendix 
Table 8. The arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon Bryozoa. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0 . 33 0 .607122 
AQ E 1 1.305107 
AQ E 1. 67 0.607122 
AQ M 0 .33 1.638440 
AQ M 1 0.607122 
AQ M 1. 67 0 .273789 
AQ N 0.33 0.607122 
AQ N 1 1.305107 
AQ N 1. 67 0.000000 
MM E 0.33 0.607122 
MM E 1 0 . 273789 
MM E 1. 67 0 .273789 
MM M 0 .33 0.273789 
MM M 1 0.000000 
MM M 1. 67 0 .273789 
MM N 0.33 0 .273789 
MM N 1 0.000000 
MM N 1.67 0 .000000 
RM E 0.33 0.871233 
RM E 1 0 .273789 
RM E 1. 67 0 .607122 
RM M 0 .33 2.686693 
RM M 1 6.982712 
RM M 1. 67 1.106708 
RM N 0.33 1.305107 
RM N 1 0 .999999 
RM N 1.67 0.273789 
Std. error of the mean= 0 . 036404 
11 7 
Appendix 
Tabl e 9. The arithmetic means and standard 
error of the interaction between si t e , 
time, and height for t he taxon Nematoda. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 2.951281 
AQ E 1 1. 106708 
AQ E 1 .67 0.000000 
AQ M 0.33 3.652 723 
AQ M 1 1.305107 
AQ M 1. 67 2.314684 
AQ N 0.33 1.305107 
AQ N 1 0.871233 
AQ N 1. 67 1.573905 
MM E 0.33 1.999999 
MM E 1 0.607122 
MM E 1.67 1.305107 
MM M 0.33 0.000000 
MM M 1 1.930182 
MM M 1 .67 3.870266 
MM N 0.33 0.607122 
MM N 1 0.000000 
MM N 1.67 2.314684 
RM E 0.33 0.607122 
RM E 1 0.607122 
RM E 1. 67 1.930182 
RM M 0.33 1.660474 
RM M 1 8.012498 
RM M 1. 67 2.686693 
RM N 0.33 2.314684 
RM N 1 0.273789 
RM N 1 .67 2.951281 
Std. error of the mean = 0.041498 
118 
Append ix 
Table 10 . Ar ithmetic means and standard 
e r ror of the interac tion betwee n s i te, 
time, and height f o r the taxon Ra d i olar i a . 
S ITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 6.958114 
AQ E 1 4.768998 
AQ E 1. 67 3.932424 
AQ M 0. 3 3 14.03694 
AQ M 1 8.435387 
AQ M 1.67 7.242570 
AQ N 0.33 5 . 316359 
AQ N 1 4.738793 
AQ N 1.67 2.634241 
MM E 0.33 31. 69320 
MM E 1 39.87768 
MM E 1 . 67 34 . 11418 
MM M 0.33 18.72969 
MM M 1 22.24378 
MM M 1. 67 42 . 65658 
MM N 0.33 48 . 89310 
MM N 1 28 . 48606 
MM N 1.67 40 . 29120 
RM E 0.33 36.99122 
RM E 1 121 . 7246 
RM E 1.67 135 . 2913 
RM M 0 . 33 69.61051 
RM M 1 46 . 85080 
RM M 1.67 42 . 30591 
RM N 0 . 33 16 . 58085 
RM N 1 25 . 84409 
RM N 1. 67 12 . 56484 
Std . error of th e mean = 1.348470 
119 
-----
Appendix 
Table 11. Arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon Foraminifera. 
ARITHMETIC 
SITE TIME HEIGHT MEANS 
AQ E o.33 
8.865732 
5,915322 
AQ E 1 9,98408'7 
AQ E 1. 67 4,585846 
AQ M o.33 6.790546 
AQ M 1 4.322194 
AQ M 1.67 20.23509 
o.33 
AQ N 7,284712 
AQ N 1 3,319389 
N 1.67 AQ 2,713578 
o.33 
MM E 2,713578 
MM E 1 3,098954 
E 1- 67 MM 8,660489 
M o.33 MM 3,501969 
MM M 1 2.698095 1-67 
MM M 6,570147 
N o.33 MM 2.789122 1 
MM N 6,073495 1. 67 
MM N 6,570147 o.33 
RM E 8.290155 
RM E 
1 12.30348 
1. 67 
RM E 7,396571 o.33 
RM M 25,33119 1 
RM M 13,91515 
M 1. 67 RM 6,909334 o.33 
RM N 18,25579 
RM N 
1 13,96658 
1. 67 
RM N 
0.079121 
of the mean 
:::: 
Std. error 
TAappbelen d ix 12 . Ar ithme ti c means and s tandard 
erro r of the interact i on be tween s ite, 
t i me, and height for the ta xo n Cl a doc er a . 
S I TE TIME HEIGHT 
ARITHMETI C 
MEANS 
AQ E o.33 
18. 2,330 6 
0.000000 
AQ E 1 3.962454 
AQ E 1. 6 7 0.607122 
AQ M o.33 o.871233 
AQ M 1 0.000000 
AQ M 1-67 0.000000 
AQ N o.33 0.000000 
AQ N 1 1.67 0.000000 
AQ N 0.000000 o.33 
MM E 0.000000 
MM E 1 0.000000 
E 1.67 MM 0.000000 
MM M 
o.33 0.000000 
MM M 1 0.000000 1-67 
MM M 1.638440 o.33 
MM N o.011233 
MM N 
1 0.000000 
1. 67 
MM N 0.000000 
E o.33 RM 0.000000 1 
RM E 1.638440 
RM E 
1. 67 5.437000 
o.33 
RM M 0.601122 1 
RM M o.273789 
M 1. 67 RM o.273789 
RM N 
o.33 o.273789 
RM N 
1 0.000000 
N 1. 67 RM 
:== o.036651 
Std. error of 
the mean 
App
T ea nb dl ie x  13 . Arithmetic 
e mr er ao nr s  o anf d  th ste a ni dn at re dr  action
ti  m be e, t wa en ed n  h siei tg eh , t for the ta xon bar n a cle 
nauplii. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 0.999999 
AQ E 1 o.273789 
AQ E 1. 67 0.607122 
AQ M 0.33 0.000000 
AQ M 1 2.648017 
AQ M 1.67 0.273789 
AQ N 0.33 1.930182 
AQ N 1 4.874127 
AQ N 1.67 1.930182 
MM E 0.33 0.607122 
MM E 1 0.273789 
MM E 1.67 0.273789 
MM M 0.33 1.163570 
MM M 1 0.607122 
MM M 1.67 0.273789 
MM N 0.33 1. 163570 
MM N 1 0.000000 
MM N 1.67 0.273789 
RM E o.33 0.000000 
RM E 1 0 . 607122 
RM E 1.67 0.000000 
RM M o.33 3.652723 
RM M 1 2.314684 
RM M 1.67 0.000000 
RM N o.33 0.273789 
RM N 1 0.000000 
RM N 1. 67 3.493641 
Std. error of the mean = 0.039832 
122 
Appendi x 
Table 14. Arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon 0str a coda. 
S ITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 1.930182 
AQ E 1 0.607122 
AQ E 1. 67 0.607122 
AQ M 0.33 3.281970 
AQ M 1 0.000000 
AQ M 1.67 0.000000 
AQ N 0.33 2.648017 
AQ N 1 1.638440 
AQ N 1. 67 2.314684 
MM E 0.33 0.000000 
MM E 1 0.273789 
MM E 1. 67 0.000000 
MM M 0.33 2.648017 
MM M 1 1.638440 
MM M 1. 67 0.273789 
MM N 0.33 0.000000 
MM N 1 0.000000 
MM N 1. 67 0.273789 
RM E 0.33 1.999999 
RM E 1 0.000000 
RM E 1. 67 1.421988 
RM M 0.33 1.660474 
RM M 1 5.686629 
RM M 1.67 3.040773 
RM N 0.33 1.163570 
RM N 1 2.951281 
RM N 1. 67 4.054877 
S td. error of the mean= 0.059599 
123 
A pendix Ta
pble 15. The arithmetic means and standard 
error of the interation between site, 
time, and height for the taxon Polychaeta 
larvae. 
ARITHMETIC 
SITE TIME HEIGH
T MEANS 
2.648017 
AQ E o.33 2.648017 
AQ E 1 2.951281 
AQ E 1. 67 3.268489 
AQ M 
o.33 o.871233 
AQ M 1 1.930182 1-67 
AQ M 1.953667 o.33 
AQ N 0.000000 
AQ N 1 o.999999 
N 1-67 AQ 2.240572 
E o.33 MM 3.234138 
MM E 
1 3.234138 
E 1. 67 MM o.871233 o.33 
MM M 2.951281 1 
MM M 2.314684 1. 67 
MM M s.837609 o.33 
MM N 6.508621 
MM N 
1 16.23178 
1.67 
MM N 2 . 999999 o.33 
RM E 2 . 999999 1 
RM E 0 . 000000 1-67 
RM E 1. 106708 o.33 
RM M o.273789 1 
RM M 1.421988 1. 67 
RM M 1.638440 o.33 
RM N 1.999999 
N 1 RM 2.951281 1.67 
RM N 
::: o.047768 
Std. error of 
the mean 
Appendi x The arithmetic means and standard 
Tabl e 16. error of the interaction between site, 
time, and height for the taxon Bivalvia. 
ARITHMETIC 
SI TE TIME HEIGHT MEANS 
3.962454 
AQ E 
o.33 2.648017 
AQ E 1 o.2737B9 
1.67 
AQ E 1.245678 
M o.33 AQ 4.109712 
AQ M 1 0.000000 1,67 
AQ M 1.999999 
N o.33 AQ 1.305107 
AQ N 1 2.999999 1. 67 
AQ N 0.000000 0,33 
MM E 0.000000 
MM E 1
 0.000000 
1- 67 
MM E o.496904 
MM M 
o.33 0.000000 
1 
MM M 0.601122 
M 1. 67 MM 0.000000 o.33 
MM N o.273789 1 
MM N 0.000000 1. 67 
MM N 0.000000 o.33 
RM E 0.000000 1 
RM E 0.000000 1. 67 
RM E 0.000000 3 
RM M 
o.3 1. 163570 
M 1 RM 1.421988 1. 67 
RM M 2.314684 o.33 
RM N o.273789 
RM N 
1 o.699055 
1. 67 
RM N 
0.040021 
the mean == 
Std. error of 
App endi x 
Table 17 . The a rithmetic means and standard 
error of the interaction between site, 
t i me, and height for the taxon Larvacea. 
SI TE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 0.000000 
AQ E 1 0.000000 
AQ E 1.67 0.000000 
AQ M 0.33 0 . 000000 
AQ M 1 0 . 000000 
AQ M 1. 67 0.000000 
AQ N 0.33 0.000000 
AQ N 1 0 . 000000 
AQ N 1. 67 0.000000 
MM E 0.33 0.000000 
MM E 1 2 . 197482 
MM E 1. 67 0 . 000000 
MM M 0.33 0.000000 
MM M 1 0.000000 
MM M 1 . 67 0 . 000000 
MM N 0 . 33 0.000000 
MM N 1 0 . 000000 
MM N 1 . 67 0 . 000000 
RM E 0 . 33 0 . 999999 
RM E 1 6.325386 
RM E 1. 67 1.638440 
RM M 0 . 33 3 . 798147 
RM M 1 0 . 273789 
RM M 1.67 1 . 106708 
RM N 0 . 33 0 . 273789 
RM N 1 0 . 000000 
RM N 1. 67 0 . 000000 
Std. error of t h e mean= 0 . 031844 
126 
Appendix The arithmetic means and standard 
Table 18. error of the interaction between site, 
time, and height for the taxon Amphipoda. 
ARITHMETIC 
SITE TIME HEIGHT MEANS 
0.000000 
AQ E o.33 0.000000 
AQ E 1 0.000000 
AQ E 1. 67 0.000000 
AQ M o.33 0.000000 
AQ M 1 0.000000 
AQ M 1-67 0.000000 
AQ N o.33 0.000000 
AQ N 1 0.000000 
-67 
AQ N 1 0.000000 0,33 
MM E 0.000000 
MM E 1 0.000000 
E 1-67 MM 0.000000 
M o.33 MM 0.000000 
MM M 1 0.000000 
M 1,67 MM 0.000000 o.33 
MM N 0.000000 
MM N 1 0.000000 1-67 
MM N 0.000000 
E o.33 RM o.999999 1 
RM E 0.000000 1. 67 
RM E 0.000000 o.33 
RM M 1,421988 
RM M 
1 0.000000 
1-67 
RM M 0.000000 o.33 
RM N 0.000000 
N 1 RM 0.273789 1. 67 
RM N 
:::: 0.005633 
the mean 
Std. error of 
A
T pa pbel ne d i1x9 . The arithmetic
e  r mr eo ar n so  f a ndth  e s tai nn dte ar ra dc  ti
t oim n e b, eta wn ed e nh  e si ig th et ,  for the taxon fish eggs. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 0.000000 
AQ E 1 0.000000 
AQ E 1.67 0.000000 
AQ M 0.33 0.273789 
AQ M 1 0 . 000000 
AQ M 1. 67 0.000000 
AQ N 0.33 0.273789 
AQ N 1 0.000000 
AQ N 1.67 0.000000 
MM E 0.33 0.000000 
MM E 1 0.871233 
MM E 1. 67 0.000000 
MM M 0.33 0.273789 
MM M 1 0.000000 
MM M 1. 67 0.000000 
MM N 0.33 0.000000 
MM N 1 0.000000 
MM N 1. 67 3.040773 
RM E 0.33 0.000000 
RM E 1 0.000000 
RM E 1. 67 0.000000 
RM M 0.33 0.000000 
RM M 1 0.607122 
RM M 1. 67 0.607122 
RM N 0.33 0.000000 
RM N 1 0.000000 
RM N 1 ? 67 0.000000 
Std. error of the mean= 0.013315 
128 
Appendix 
Table 20. The arithmetic means and standard 
error of the interaction between site, 
time, and height for the taxon crab 
megalopa. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 0.607122 
AQ E 1 0.273789 
AQ E 1. 67 1.930182 
AQ M 0.33 0.000000 
AQ M 1 0.000000 
AQ M 1. 67 0.607122 
AQ N 0.33 0.273789 
AQ N 1 0.000000 
AQ N 1. 67 0.000000 
MM E 0.33 3.563808 
MM E 1 3.040773 
MM E 1. 67 2.448193 
MM M 0.33 0.000000 
MM M 1 4.159635 
MM M 1. 67 5.016430 
MM N 0.33 8.842349 
MM N 1 0.607122 
MM N 1.67 4.435742 
RM E 0.33 0.000000 
RM E 1 0.000000 
RM E 1. 67 1.930182 
RM M 0.33 0.000000 
RM M 1 0.000000 
RM M 1. 67 0.273789 
RM N 0.33 2.314684 
RM N 1 0.000000 
RM N 1 .67 0.000000 
Std. error of the mean= 0.208253 
129 
Appendix 
Tab 1 e 21. The arithmetic means and the standard 
error of the interaction between site, 
time, and height for the taxon crab 
zoea. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 0.000000 
AQ E 1 0.000000 
AQ E 1. 67 0.607122 
AQ M 0.33 0.000000 
AQ M 1 0.000000 
AQ M 1.67 0.000000 
AQ N 0.33 0.000000 
AQ N 1 0.000000 
AQ N 1 .67 0.000000 
MM E 0.33 0.000000 
MM E 1 0.000000 
MM E 1. 67 0.000000 
MM M 0.33 0.000000 
MM M 1 0.000000 
MM M 1. 67 0.000000 
MM N 0.33 0.000000 
MM N 1 0.000000 
MM N 1.67 0.273789 
RM E 0.33 0.000000 
RM E 1 0.699055 
RM E 1. 67 0.000000 
RM M 0.33 0.000000 
RM M 1 0.000000 
RM M 1. 67 0.607122 
RM N 0.33 0.000000 
RM N 1 0.000000 
RM N 1. 67 0.000000 
Std. error of the mean= 0.008881 
130 
Appendix 
Table 22. The arithmetic means and standard 
error of the interaction between site 
time, and height for the taxon Noctilu' ca. 
TIME HEIGHT ARITHMES TI ICT  E MEANS 
AQ E 0.33 0.000000 
E 1 0.0A 0Q 0 000 
AQ E 1.67 
0 . 000000 
AQ M 0.33 
0.000000 
M 1 0.000000 AQ 
M 1.67 0.000000 AQ 
0.33 0.000000 AQ N 
1 0.000000 AQ N 
1.67 0.000000 AQ N 0 .000000 
MM E o.33 0.000000 
MM E 1 
1.67 0.000000 MM E 0.000000 
MM M 0 .33 
1 0.000000 MM M 
MM M 1. 67 
0.000000 
0.33 0.000000 MM N 
1 0.000000 MM N 
1.67 0 .000000 MM N 
0.33 0.000000 RM E 0 . 000000 
RM E 1 
1. 67 0.000000 RM E 
o.33 0.273789 RM M 
1 0.000000 RM M 
1.67 0.000000 RM M 
o.33 0.000000 RM N 
1 0 .000000 RM N 
1.67 0.000000 RM N 
Std. error of the mean = 0.001102 
l 131 
Appendix 
Table 23. The arithmetic means and standard 
error of the interaction bet ween site, 
time, and height for the taxon 
Dinoflagellates. 
SITE TIME HEIGHT ARITHMETIC 
MEANS 
AQ E 0.33 69.41790 
AQ E 1 24.20261 
AQ E 1.67 49.25206 
M 0.33 24.1A 9Q 8 42 
A M 1 17Q .21953 
M 1.67 15.4A 1Q 1 21 1
N 0.33 0.0928A 7Q   37 .2
N 7 1 86A 5Q   
AQ N 1,67 
38.53119 
47,58146 
MM E 0.33 
1 6MM E
5
 .27160 
MM E 1.67 
35.09849 
15.53649 
MM M 0.33 26,69916 
MM M 1 
MM M 1. 67 
23,64896 
52,87993 
MM N o.33 
1 53.31909 MM N 
1,67 87.84043 MM N 
o.33 21.15048 RM E 
RM E 1 
56.70920 
1.67 49,32481 RM E 
o.33 15,08290 RM M 
1 26.30517 RM M 
1,67 23,46990 RM M 
0.33 26 ,75898 RM N 25. 17826 
RM N 1 
1,67 25,95052 RM N 
Std. error of the mean = 1.144299 
132 
Appendix 
The arithmetic means and standard 
Table 24. error of the interaction between site, 
time, and height for the taxon Mysid shrimp. 
ARITHMETIC 
SITE TIME HEIGHT MEANS 
0.33 0.0000 0 0 AQ E 
1 0.000000 AQ E 
AQ E 1.67 
0.000000 
o.33 0.000000 AQ M 
1 0.000000 AQ M 
1. 67 0 . 000000 AQ M 0.000000 
AQ N o.33 
1 0.000000 AQ N 
1. 67 0.000000 AQ N 0.000000 
MM E 0.33 
1 0.000000 MM E 
1.67 0.000000 MM E 
MM M 0.33 
0.000000 
1 0.000000 MM M 
1.67 0.000000 MM M 0.000000 
MM N o.33 0.000000 
MM N 1 
1.67 0.000000 
MM N 
RM E 0.33 
0.000000 
0.000000 
RM E 1 
1. 67 0.000000 
RM E 
0.33 0.000000 
RM M 
1 o.273789 RM M 
1.67 0.000000 
RM M o.33 0.000000 
RM N 0.000000 
RM N 1 t.67 0.000000 
RM N 
the mean :: 0.001102 
Std. error of 
133 
Appendix 
Table 25. The arit
hmetic means and standard 
error of the interaction between site, 
time, and height for the taxon Tunicate 
larvae. 
ARITHMETIC 
SITE TIME HEIGHT MEANS 
0.000000 
AQ E 0.33 
1 0 .000000 AQ E 0.000000 
AQ E 1.67 
AQ M 0.33 
0.000000 
1 0.000000 AQ M 0.000000 
AQ M 1.67 
0.33 0.000000 AQ N 0.000000 
AQ N 1 
1. 67 0.000000 AQ N 0.000000 
MM E o.33 
1 0.000000 MM E 0.000000 
MM E 1.67 
MM M o.33 
0.000000 
0.000000 
MM M 1 
MM M 1.67 
0.000000 
0.000000 
MM N 0.33 
1 0.000000 MM N 1.67 0.000000 
MM N o.33 0.000000 RM E 0000 
RM E 1 
0.00
1.67 0.000000 
RM E o.33 0.000000 RM M 1.245678 
RM M 1 1.67 0.496904 
RM M o.33 0.000000 
RM N 0.000000 
RM N 1 1.67 0.000000 
RM N 
Std. error of the mean= 0.017112 
134 
Appendi x 
The arithmetic means and standard Table 26 . error of the interaction between site, 
time, and height for the taxon Echinodermata 
pleuteus. 
HEIGHT ARITHMETIC SITE TIME MEANS 
0 .33 0.000000 AQ E 
AQ E 1 
0.000000 
1. 67 0.000000 AQ E 0.000000 
AQ M o.33 
1 0.000000 AQ M 
1. 67 0.000000 AQ M 0.000000 
AQ N 0.33 
1 0.000000 AQ N 
AQ N 1. 67 
0.000000 
0.000000 
MM E 0.33 
MM E 1 
0.000000 
MM E 1.67 
0.000000 
0.33 0.000000 MM M 
MM M 1 
0.000000 
0.000000 
MM M 1. 67 
0.33 0.000000 MM N 
1 0.000000 MM N 
1. 67 0.000000 MM N 
RM E 0
.33 0.000000 
1 1.660474 RM E 
t.67 0.000000 
RM E 
0.33 1,660474 
RM M 
1 0.000000 RM M 
1.67 0.699055 
RM M 3 0.000000 
RM N 0
.3
1 0.000000 RM N 0 .000000 
RM N 
1. 67 
Std. error of the mean= 0.018148 
135 
endix 3 f
or 
App an abundances o
f gastropoda 
e were 
Figure 1 . The daily
 m
mean abundance
s 
s . The 
each study site ontained in 
taken from the 
ANOVA data c
he daily variat
ion in mean 
Appendix 3. T y different at significant l
abu ndances was 
p > 0.05. 
136 
!50_-------------------
45 Site = Ros Moh oam -m oed  0 .33m 
?-? 1.00m 
40 A-A 1.87m 
35 
30 
25 
20 
Q) 15 
(.) 10 
C 5 
-0  !O5+0-,-------,-------------r----------------_-_C o-o-_--J ::::, 45 Sit =  e 0. 33mM  arso el Mukibelo 
?-? 1.00m 
.0 40? A-A
0  1.87m 
35 
C 30? 
0 
Q) 25 
E 20 
15 
:(;.;)  10? 
Q) o5 J----------,-
E !
-
5 --0 --- -,- r- --- -- -- --- -- -- 1-  -----o---o- 0.-.r:. 33m- 45. Site ? Aq
+ uoJ eport ? ?c - ? 1.00m A-A 1.87m<  ( 40? 
35 
30? 
25 
20 
15? 
10 
o!5 ? J-----------,--.-----,-.r-------t 
morning noon 
evening 
Time of Day 
137 
Appendix 3 
 2. The daily mean abund
ances of cyclopoid 
Figure udy sites. The mean
 
copepoda for each st
ances were taken from
 the ANDVA data 
abund ly 
contained in Appendi
x 3. The dai
 was 
variation in mean ab
undances
0.05. 
significantly differ
ent at p> 
138 
90 Site = o-o 0.33m Marso el Mukibela 
eo ?-? 1.00m 
A-A 1.S7m 
70 
eo 
so 
40 
Q) 30 
() 20 
C 
0 10 
-0 0 
C 90 o-o 0.33m 
::::, Site = Marso el Mukibela 
..0 eo ?-? 1.00m 
0 70 A-A 1.S7m 
C eo 
0 so 
Q) 
E 40 
30 
?(-) :::::-----,, 20 ~ +-' 
Q) 10 ~ 
E 0 
.c 90 o-o 0.33m 
?+--' Site = Aquasport 80 ?-? 1.00m L. A-A 1.S7m 
<{ 70 
eo 
so 
40 
30 
20 
10 
0 noon evening morning 
Time of Day 
139 
Appendix 3 ances of calanoid an abund
Figure 3. The daily m
e The mean 
copepoda for each 
study sites, 
ces were taken from
 the AN0VA data 
abundan ily 
contained in Appen
dix 3. The da
nces was 
variation in mean 
abunda
0.05. 
significantly diff
erent at p> 
140 
O-O 0.JJm 
120 Site= Ras Mohammed ?-? 1.00m 
A-A 1.157m 
100 
80 
80 
40 
~ 20 
oC oJ----------_,..-----,----___. 
o-o O.JJm 
"cO  120 Site = Marso el Mukibela ?-? 1.00m 
:::, A-A 1.157m 
.0 100 
0 
C 
0 
Q) eo 
E 40 
(.) 
20 
0 J------.-----_,.-----,--------f 
o-o O.JJm 
120 Site = Aquasport ?-? 1.00m 
A-A 1,157m 
100 
80 
eo 
40 
2o0 J...----I -------,-----r-------t moming noon evening 
Time of Day 
141 
 harpacticoi
d 
Appendix 3  mean abunda
nces of
 The mean 
Figure 4. The 
daily ch study sit
es.
data 
copepoda for
 ea A 
ere taken fro
m the AN0V
abundances w  dail
y 
ed in Appen
dix 3. The
contain s abundances w
a
variation in
 mean p> 0.05. 
ificantly di
fferent at 
sign
142 
eo 
Site - Ras Mohammed o-o 0.33m 
so ?-? 1.00m 
A-A 1.87m 
40 
30 
20 
Q) 10 
(.) 
C: 
0 0 
-0 eo 
C: Site ? Marso el Mukibela o-o 0.33m 
::) so 
..0 ? -? 1.00m A
0 -A 1.87m 40 
C: 
0 30 
Q) 
E 20 
(.) 
10 
-+-' 
Q) 
E 0 eo 
L: 
?-+--' Site ? Aquasport 0-0 O.JJm L. so ?-? 1.00m 
<( A-A 1.87m 
40 
30 
20 
10 
0 
morning noon evening 
Time of Day 
143 
3 of invertebr
ate 
Appendix an abundances
 
 
Figure 5. The d
aily me s. The meantudy site ta 
eggs for each
 s da
ere taken from
 the AN0VA 
abundances w  daily 
ined in Appe
ndix 3. The
conta bundances was
 
variation in 
mean a p > 0 . 05. 
nificantly d
ifferent at 
sig
144 
120 
Site ? Ras Mohammed o-o 0,33m 
100 ?-? 1.00m 
4-4 1.87m 
80 
80 
Q.) 40 
(.) 
C 20 
0 
-0 0 
C 120 
::, Site = Marso el Mukibela o-o 0.33m 
..0 100 ?-? 1,00m 
0 4-A 1,87m 
C 80 
0 
Q.) 80 
E 40 
?...u.-..,  20 
Q.) 
E 0 
....
120 
..c ite = Aquasport o-o 0.33m ?L-
.., S
100 ?-
? 1.00m 
. 7m 
<( 4-4 1.8
80 
80 
40 
20 
0 
morning noon evening 
Time of Day 
145