Diets and metabolic rates of four Caribbean tube blennies, genus Acanthemblemaria (Teleostei: Chaenopsidae)

Abstract

In order to determine the extent to which tube blennies depend on food derived from within or outside of the reef system, the diets of Acanthemblemaria spinosa, A. aspera, A. greenfieldi and A. paula were compared with food availability using plankton, benthic and gut sampling. All species fed primarily on copepods, but A. spinosa consumed calanoids and cyclopoids of planktonic origin whereas the other species consumed harpacticoids of benthic origin. This pattern correlates with the occurrence of A. spinosa in tall corals whereas the other species are found in corals close to the reef surface. Calanoids were more abundant 1 m above the reef surface than <0.5 m above the reef surface. Oxygen consumption rates were determined for the above species and Emblemariopsis pricei and E. ruetzleri. On the oxygen consumption versus body weight graph, A. spinosa shares a high slope with the Emblemariopsis species, which was three times greater than the slopes of other species of Acanthemblemaria. It is not known if the high metabolic rate of A. spinosa is a consequence of the greater food availability in its microhabitat, an adaptation for acquiring the higher quality microhabitat, or both. In any case, A. spinosa is the only Acanthemblemaria species in this study that imports nutrients to the reef by consuming plankton.

Full text

BULLETIN OF MARINE SCIENCE, 65(1): 185–199, 1999
185
DIETS AND METABOLIC RATES OF FOUR CARIBBEAN TUBE
BLENNIES, GENUS ACANTHEMBLEMARIA
(TELEOSTEI: CHAENOPSIDAE)
Raymond D. Clarke
ABSTRACT
In order to determine the extent to which tube blennies depend on food derived from
within or outside of the reef system, the diets of Acanthemblemaria spinosa, A. aspera, A.
greenfieldi and A. paula were compared with food availability using plankton, benthic
and gut sampling. All species fed primarily on copepods, but A. spinosa consumed
calanoids and cyclopoids of planktonic origin whereas the other species consumed
harpacticoids of benthic origin. This pattern correlates with the occurrence of A. spinosa
in tall corals whereas the other species are found in corals close to the reef surface.
Calanoids were more abundant 1 m above the reef surface than <0.5 m above the reef
surface. Oxygen consumption rates were determined for the above species and
Emblemariopsis pricei and E. ruetzleri. On the oxygen consumption versus body weight
graph, A. spinosa shares a high slope with the Emblemariopsis species, which was three
times greater than the slopes of other species of Acanthemblemaria. It is not known if the
high metabolic rate of A. spinosa is a consequence of the greater food availability in its
microhabitat, an adaptation for acquiring the higher quality microhabitat, or both. In any
case, A. spinosa is the only Acanthemblemaria species in this study that imports nutrients
to the reef by consuming plankton.
Chaenopsid blennies of the genus Acanthemblemaria constitute 18 species (Almany
and Baldwin, 1996) and are among the most inconspicuous diurnal fishes on coral reefs.
They can be very abundant, reaching densities as high as 60 m−2 (Lindquist, 1985) and
200 m−2 (J. Buchheim, pers. comm.) and may play a significant role in the nutrient dy-
namics of reef communities. They feed primarily on copepods and other small crustacea
(Lindquist and Kotschral, 1987). One of the features that determines their role in the
community is the extent to which their food is derived from within the reef ecosystem as
compared to the water mass over the reef, but there is little information on this topic for
any fishes (Sedberry and Cuellar, 1993). Studies of currents upstream and downstream of
reefs have shown that in general there is a net removal of plankton and therefore an im-
port of nutrients to reefs (Glynn, 1973). Many organisms, including a subset of the fish
assemblages, participate in this process. Whereas the mobile planktivorous fishes have
received much attention (Hamner et al., 1988; Hobson, 1991), the role of chaenopsid
blennies in this process may have been overlooked.
The inconspicuousness of chaenopsid blennies derives from their small size (generally
20–30 mm standard length) and their almost infaunal nature. Many species live in small
cavities in hard corals created by boring organisms. Some species seem to saturate the
environment, avoiding competition through precise spatial relations within and between
habitats (Clarke, 1989, 1994; Buchheim and Hixon, 1992). One would expect to find a
relationship between reef zone, microhabitat, and the extent to which their food is derived
from the pelagic or benthic environments. This study is designed to evaluate that expecta-
tion, specifically to test three hypotheses: (1) that species living furthest from the shore
will consume more plankton than those living on the backreef or the shallower portions of

186 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
the forereef; (2) that the species living above the reef surface in tall corals will eat more
plankton than those living on the reef surface; and (3) that metabolic rates will be higher
for those species eating plankton because plankton are delivered constantly whereas benthic
prey may become depleted in the vicinity of the cavities where the fish live.
In addition to testing the above hypotheses, this paper clarifies the relationship between
two species by showing which one has diverged from the general adaptive pattern for the
genus. In St. Croix, U.S. Virgin Islands, I have previously shown that Acanthemblemaria
spinosa lives higher in the corals than A. aspera, that it eats more food of planktonic
origin, and that it has a higher metabolic rate (Clarke, 1989, 1992, 1996). This study
builds on those observations by broadening them with more extensive sampling in a dif-
ferent locality, by including sampling of food availability, and by including the diets of
two additional species and the metabolic rates of four additional species. With these data,
this paper shows that A. aspera is typical of other members of its genus and that A. spinosa
possesses unique adaptations to occupy a different microhabitat and to feed on different
food.
MATERIALS AND METHODS
All field work took place at Carrie Bow Cay, on the Belizean barrier reef (Rützler and Macintyre,
1982).
DIET.—The design was devised to sample the potential food organisms in the same locations
where the fishes feed, thus avoiding a weakness of many food habit studies (Gerking, 1994). In
March 1994, I investigated Acanthemblemaria paula, A. greenfieldi, A. spinosa and A. aspera in the
pavement zone, surge zone, high spur and groove zone and low spur and groove zone (Rüetzler and
Macintyre, 1982; Clarke, 1994). Two species were studied in each zone, except for the low spur and
groove which supported only A. aspera. Each zone was sampled twice at early (08:05–08:40), mid
(13:20–13:30) and late (17:05–18:00) day for a total of 24 sample sets. As described below, a
sample set consisted of: (1) a fish collection, (2) a benthic collection and (3) one or two plankton
collections.
(1) Five fish of each species were collected using quinaldine sulfate (see below). Immediately on
surfacing, I fixed them in 10% formalin. After 2 d I placed them in 70% ethanol with a change to
fresh 70% ethanol after 1 d. The complete guts were subsequently dissected out and the contents
were identified to various taxonomic levels. The data for the five individuals were pooled and
treated as a single sample.
(2) Three pieces of coral rubble were collected by dislodging with a hammer after covering them
with plastic bags. Immediately on surfacing, I fixed them in 10% formalin. Soon afterward I vigor-
ously agitated the sample to dislodge the organisms and decanted the supernatant, saving the settled
material in 5% formalin buffered with borax for later identification. Subsequently I randomly se-
lected one of the samples, stained it with rose bengal, and identified the organisms to various
taxonomic levels. The pieces of rubble were selected to resemble those locations occupied by
Acanthemblemaria and although no fish were seen in the pieces before collecting, several yielded
fish. The presence of fish in the samples confirms that I was collecting from locations that repre-
sent the benthic food available to the fish.
(3) Plankton were sampled with a diver-towed 243 μm 1:3 plankton net. The 50 cm diameter ring
was formed into a 38 cm square to allow sampling from a more precise distance above the reef
surface. Tows were either (a) low, ranging from the reef surface to 0.5 m above the reef surface,
which is below the tops of the abundant gorgonians, or (b) high, ca 1 m (always <1.5 m) above the
reef surface, which is above the tops of the gorgonians but at the level of the Acropora palmata
occupied by A. spinosa. All tows were for 5 min. A current meter was not used, but as part of a
different study I used the same net towed in the same way with a current meter and determined that

187
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
the volume sampled in 5 min was 12.01 m3 (SE = 0.25, n = 12). Immediately on surfacing, I fixed
the samples in 10% formalin. They were later transferred to 5% formalin buffered with borax.
Subsequently, the samples were divided with an Alan plankton splitter (range: 1/2 to 1/32), placed
in a plankton counting wheel, and identified to various taxonomic levels. The same taxonomic
categories were used for gut, plankton and benthic samples.
The pavement zone, which averaged ca 0.5 m in depth, was sampled differently. The plankton net
was held stationary, supported with a stake, with the lower edge of the ring resting on the reef
surface. The current usually kept the net filled and trailing. The net was left out for 20–30 min and
sampled very different volumes from the other habitats. Current speed was not measured; it was
generally weak, but it varied considerably, so the sampling effectiveness in this habitat was vari-
able.
METABOLIC RATES.—In March 1997, I captured the fish each morning (between 08:48 and 10:38)
by squirting a 0.1% solution of quinaldine sulfate at the shelter hole and covering the opening with
a test tube. The fish emerged into the tube in a few seconds. To minimize exposure to the anesthetic,
I immediately transferred the fish to ventilated vials which I kept in a mesh bag, thus allowing a
free exchange of seawater.
Within 1 to 1.5 h of capture, I placed each fish, still in its vial, in a 300 ml BOD bottle and left it
for an average of 5.2 h (range 3.5 to 8.8 h). The smallest fishes were “incubated” for the longest
times. O2 concentration was measured to the nearest 0.01 mg L−1 before and after “incubation” with
a digital electronic O2 meter. The O2 concentrations fell to an average of 86% (5.3 mg L−1) of the
initial values, which is not considered to produce stress for the fish. With each run, I also incubated
two bottles without fish to use as controls; the average O2 consumption in the controls was sub-
tracted from the consumption in the bottles containing fish. The temperatures during incubation
averaged 27.5°C ± 0.16 SE (range 25.3 to 29.6°). The fishes generally remained quiescent while in
the BOD bottles, so the O2 consumption can be considered an index to standard metabolic rate
(SMR). SMR is positively correlated with maximum metabolic rate (Brett and Groves, 1979) thus
SMR is a useful index of capacity to engage in a variety of energetically expensive behaviors.
After the fish were “incubated”, I measured their standard lengths (SL) and weighed them.
ANALYSIS.—Overlap between samples was measured using Schoener’s (1968) overlap index:
Pjk = Σ min(pij,pik),
where pij is the proportion of the ith item in sample j and pik is the proportion if the ith item in sample
k. This index varies between 0 (no overlap) and 1.0 (complete overlap).
Electivity was measured by Ivlev’s (1961) index:
Ejk = (pij−pik)/(pij+pik),
where pij is the proportion of the ith item in the gut sample and pik is the proportion of the ith item in
the environmental sample. This index varies from −1 (none in gut) to +1 (only ith item in gut and
none in environmental sample).
RESULTS
BENTHIC AND PLANKTONIC SAMPLES.—The number of taxa recognized in the benthic
samples was 30. The five most numerous taxa constituted 83.3% of the 11,642 individu-
als counted; in decreasing order (with percent of individuals) were harpacticoid copepods
(58.6), nematodes (7.6), ostracods (5.9), foraminifers (5.9) and polychaetes (5.3). Calanoid
copepod fragments represented 0.05% of the total and cyclopoid copepods represented

188 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
0.59%. With the exception of the high samples in the high spur and groove zone,
harpacticoids dominated all samples (Fig. 1). The high samples were more variable than
the others with two having large numbers of nematodes and three having large numbers
of ostracods. Even in these samples, harpacticoids were the single most numerous group
with one exception: in row 1, column 4 (R1C4) of Figure 1 where nematodes constituted
40% of the sample compared to 37% for harpacticoids.
Figure 1. Relative abundances of the five most common taxa in benthic samples from four habitats
(one with two microhabitats) at three times during the day. Each habitat and time was sampled on
two days and these replicates are placed one above the other.

189
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
The number of taxa recognized in the plankton samples was 48. The five most numer-
ous taxa constituted 95.8% of the 32,488 individuals counted; in decreasing order (with
percent of individuals) were calanoid copepods (82.6), cyclopoid copepods (9.5), eggs
(2.1), mysids (1.1) and foraminifers (0.5). Harpacticoid copepods represented 0.28% of
the total. The plankton samples were more variable than the benthic samples both be-
tween habitats and over time. With the exception of the pavement zone, most samples
were dominated by calanoid and cyclopoid copepods (Fig. 2). The variation over time
Figure 2. Relative abundances of the five most common taxa in plankton tows from four habitats
(one sampled at two levels) at three times during the day. Each habitat and time was sampled on two
days and these replicates are placed one above the other.

190 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
was considerable, occasionally with complete dominance by one group such as cyclopoid
copepods (82.4% of R4C2 in Fig. 2) or calanoid copepods (98.5% of R2C5 in Fig. 2).
This temporal variability was most pronounced during 7 and 8 March when there was a
striking aggregation of the oceanic calanoid copepod, Acartia spinata. The percent of
calanoid copepods in the samples those days (Fig. 2) was 97.6 (R1C4), 95.1 (R5C4) and
98.5 (R2C5). The aggregation dispersed on 8 March and the percents dropped as the day
progressed with the early, mid and late day samples from the low spur and groove zone
being 98.5 (R2C5), 87.3 (R4C5) and 67.0 (R6C5). The density dropped from 103,744
individuals per tow in the morning to 7309 at midday, whereas cyclopoids remained steady
with 592 in the morning and 619 at midday. Calanoid densities during the aggregation
averaged 60,226 individuals per tow as compared to 1468 for the remaining tows in the
high and low spur and groove zones. It is also notable that the only days in which calanoids
were abundant in the pavement zone were 6 March and 9 March (R5C1 and R4C1 in Fig.
2, respectively), the day before and the day after the aggregation was apparent in the two
spur and groove zones. Sampling did not occur in the pavement zone on 7 and 8 March.
During the aggregation, calanoid copepods were denser in high than in low samples in
the high spur and groove zone. The two high tows on 7 March averaged 38,467 calanoids
whereas the two low tows averaged 788 calanoids. This pattern held at other times but
was not a result of sampling bias because other organisms, such as mysids and isopods,
were denser in low tows (Fig. 3). The two points in the upper right corner of Figure 3
represent calanoid densities during the aggregation when the high-low difference was
greatest (high-low ratios 103 and 33.3 as compared with a range of 0.99 to 7.71 for the
other times). Unlike calanoids, cyclopoid and harpacticoid copepods were not more abun-
dant in high tows but larvaceans and eggs were (Fig. 3). The density of all plankton
Figure 3. Abundances of five taxa in high plankton tows versus low plankton tows in the high spur
and groove zone. Curve is for high = low.

191
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
combined was greater in the high tows (high-low ratios = 1.43 ± 0.20 SE for tows taken
when there was no calanoid aggregation).
The average overlap of the planktonic and benthic samples was 0.11. This f igure masks
the unusual nature of the pavement zone, however. Because the plankton sample in this
zone was taken with a stationary net in contact with the substrate (see Methods), the
plankton sample is probably intermixed with some benthic organisms. This was evident
in the average similarities of the benthic and plankton samples in the different zones: for
the pavement zone the value was 0.23 whereas it was 0.08 for the other zones combined
(P < 0.0001, m = 6, n = 24, Mann-Whitney Test). The real overlaps were lower than
reported here because many groups were recognized only to order, thus failing to distin-
guish between planktonic and benthic species; the three groups for which this was most
evident (with their average benthic and planktonic percents were: Isopoda (2.4, 3.9), Fora-
minifera (6.1, 13.8), and Ostracoda (4.9, 1.5).
Figure 4. Relative abundances of the five most common taxa in gut samples from four species of
Acanthemblemaria in four habitats at three times during the day. Each habitat and time was sampled
on two days and these replicates are placed one above the other.

192 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
GUT SAMPLES.—The number of taxa recognized in the gut samples was 33 and the
number of items identified in the samples ranged from 59 to 586 per five fish. The five
most numerous taxa constituted 84.3% of the 8000 individuals counted; in decreasing
order (with percent of individuals) were harpacticoid copepods (79.1), calanoid copep-
ods (5.5), ostracods (4.7), tanaids (2.1) and cyclopoid copepods (1.9). When combined,
copepods constituted 61 to 96% of the food items in the 42 five-fish gut samples. A.
greenfieldi, A. paula and A. aspera guts averaged >80% harpacticoid copepods in all
habitats whereas A. spinosa contained an average of 18.6% harpacticoids (Fig. 4). Con-
versely, A. spinosa guts averaged 46.3% calanoid copepods whereas the other species
averaged <3% in all habitats. All copepods in all f ish species ranged from 84.5 to 90% on
average.
For benthic and gut samples, Ivlev’s index of electivity indicates that harpacticoids are
actively selected by all species except A. spinosa (P = 0.02, Sign test) whereas isopods are
actively rejected by all species except A. spinosa (Table 1). The apparent preference of A.
spinosa for isopods is misleading because the isopods consumed by that species are of
planktonic origin and therefore not represented in the benthic samples used in this analy-
sis. Ostracods are selected in all cases except the high spur and groove zone, and polycha-
etes and nematodes are selected against in all cases.
For plankton and gut samples, Ivlev’s index of electivity indicates that A. spinosa does
indeed select isopods whereas none of the other species do (P = 0.02, Sign test, Table 2).
f
oseicepsruofrofaxatyerpcihtnebevifrofytivitcelefoxedni)1691(s’velevI.1elbaT airamelbmehtnac
A
seulavevitisoP.enozevoorgdnarupsehtotsrefer”G&S“.senozfeerruofni rellamsaetacidniseulavevitagendnaelpmasehtninahttugehtninoitroporpretaergaetacidni .noitroporp
enoZfeeR,seicepSsdiocitcapraHsdoposIsdocartsOsete
ahcyloPsedotameN
aluap.
A
tnemevap,02.0 −06.0 40.0 −56.0 −49.0
aluap.
A
egrus,31.0 −37.0 45.0 −86.0 −68.0
idleifneerg.
A
tnemevap,12.0 −44.0 21.0 −56.0 −00.1
idleifneerg.
A
egrus,11.0 −33.0 45.0 −65.0 −49.0
arepsa.
A
G&Shgih,20.0 −12.0 −61.0 −05.0 −95.0
arepsa.
A
G&Swol,51.0 −36.0 20.0 −38.0 −79.0
tpecxe(naem asonips.A )41.0 −94.0 81.0 −56.0 −88.0
asonips.
A
G&Shgih, −04.0 75.0 −66.0 −96.0 −99.0
f
oseicepsruofrofaxatyerpcinotknalpowtrofytivitcelefoxedni)1691(s’velvI.2elbaT airamelbmehtnac
A
seulavevitisoP.enozevoorgdnarupsotsrefer”G&S“.senozfeerruofni rellamsaetacidniseulavevitagendnaelpmasehtninahttugehtninoitroporpretaergaetacidni .noitroporp
enoZfeeR,seicepSsdionalaCsdoposI
aluap.
A
tnemevap, −00.1 −14.0
aluap.
A
egrus, −00.1 −72.0
idleifneerg.
A
tnemevap, −89.0 −78.0
idleifneerg.
A
egrus, −97.0 −15.0
arepsa.
A
G&Shgih, −78.0 −57.0
arepsa.
A
G&Swol, −59.0 −10.0
tpecxe(naem asonips.A )−39.0 −74.0
asonips.
A
G&Shgih, −90.0 97.0

193
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
In addition, all the species have very few calanoids in their guts relative to their abun-
dance in the water column but A. spinosa, while having fewer calanoids than the water
column, has a considerably lesser reduction than the other species. A. spinosa fed prima-
rily on plankton whereas the other three species fed primarily on benthos (Fig. 5). For all
taxa, the gut-plankton overlap was 0.56 for A. spinosa and 0.09 for the other species
combined (t = 14.2, df = 40, P < 0.0001). In contrast, the gut-benthos overlap was 0.29 for
A. spinosa and 0.56 for the other species combined (t = 8.29, df = 40, P < 0.0001). There
is a positive correlation of percent calanoids in the plankton tows and in the A. spinosa
guts (Spearman rank correlation = 0.9, one-tailed P = 0.042, n = 5).
Based on copepods, gut contents reflected an increasing planktonic proportion as samples
were taken from greater distances from shore (Fig. 6). Fish from the pavement zone,
which is on the backreef, had many fewer calanoids and cyclopoids relative to harpacticoids
than fish taken from the forereef. A. spinosa guts were very different from the others with
a planktonic to benthic copepod ratio of 3.5 as compared with 0.02 (range 0.006 to 0.03)
for the other species on the forereef.
The mean number of items in the guts at early, mid and late day were 156, 250 and 166
per five fish. The mean number of items varied with fish species, so the numbers were
expressed as percents (with the highest value on a given day = 100%) to give equal weight
to each species. Mean percents for early, mid and late day were 62, 95 and 66 (F = 10.3, df
Figure 5. Relative abundance of planktonic copepods (calanoids and cyclopoids) vs abundance of
benthic copepods (harpacticoids) in 42 gut samples from four species of Acanthemblemaria. A.
spinosa is indicated by closed circles and the other species are indicated by open circles. Descending
solid line indicates that combined copepods represent 100% of gut contents; descending dashed
line indicates 60% of gut contents. Ascending line indicates equivalency of planktonic and benthic
copepods.

194 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
= 2,39, P = 0.0003, ANOVA). The reason for a significantly greater number of food items
at midday is unclear.
METABOLIC RATES.—The data presented include four species of Acanthemblemaria and
two species of Emblemariopsis. A seventh species, Acanthemblemaria maria, was repre-
sented by only two specimens and is omitted from the analysis. For each species, the O2
consumption rate vs live weight showed a straight line relationship (Fig. 7, Table 3). The
weight of the fishes varied by more than an order of magnitude (11 to 220 mg), yet the O2
consumption rates fell into three groups based on comparisons of the slopes of O2 con-
sumption vs weight: (1) A. spinosa, E. pricei and E. ruetzleri had steep slopes (0.34 com-
bined) whereas (2) A. aspera and A. greenfieldi had lower slopes (0.10 combined) and (3)
the diminutive A. paula, with a narrow range of weights, fell into its own group (Fig. 7).
The five significant differences in slopes are all between species of groups 1 and species
Figure 6. Planktonic-benthic copepod ratio [(calanoids+cyclopoids)/harpacticoids] vs reef zone for
four species of Acanthemblemaria. “S&G” refers to spur and groove. Curve is fitted by eye for all
points except A. spinosa. Note logarithmic scale on ordinate.
OrofstneiciffeocnoissergerdnasepolS.3elbaT
2
f
oseicepsrofthgiewevilsvnoitpmusnoc
airamelbmehtnac
A
dna sispoiramelbmE H(epolsfotset-troferaseulavP.
0
epols: =.)0
seicepSNepolsESrP
asonips.
A
44942.0430.057.0 <1000.0
arepsa.
A
44631.0220.096.0 <1000.0
aluap.
A
61962.0731.074.0960.00
idleifneerg.
A
81490.0310.078.0 <1000.0
iecirp.
E
9914.0090.078.04200.0
irelzteur.
E
8623.0390.028.0310.00

195
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
of group 2 (Table 4); there were no significant differences between species within these
groups. Although the slope of group 3 (A. paula) did not differ significantly from any of
the other species (Table 4), all 16 points fell below the regression curves for both groups
1 and 2 (Fig. 7), which placed the cluster significantly outside the range of either group (P
= 0.516 = 0.0002).
DISCUSSION
DIET.—Acanthemblemaria spp. are basically copepod eaters, with numerical percents
of gut contents being 63 to 85 (Greenfield and Greenfield, 1982), 65 to 90 (Lindquist and
Kotrschal, 1987), 70 to 83 (Clarke, 1992), and 85 to 90 (this study). A. spinosa is unusual
Figure 7. O2 consumption vs live weight for six species of chaenopsid blennies in the genera
Acanthemblemaria and Emblemariopsis. Fitted curves are for two groups: (1) A. spinosa, E. pricei,
E. ruetzleri; and (2) A. aspera, A. greenfieldi.
,lauqetonerewsnoitaiveddradnatsesuaceb(tset-tetanretlas’hcleWnodesabseulavP.4elbaT Orofsepolsneewtebsecnereffidehtrof)3991,.lateykslutoM
2
rofthgiewevilsvnoitpmusnoc
foseicepsruof airamelbmehtnacA foseicepsowtdna ispoiramelbmE eraseulavtnacifingiS.s
.sksiretsaybdetacidni
seicepS aluap.Aarepsa.Aidleifneerg.Aasonips.Aiecirp.E
arepsa.
A
353.0
idleifneerg.
A
222.0301.0
asonips.
A
988.0*700.0 <*1000.0
iecirp.
E
173.0*610.0700.00* 011.0
irelzteur.
E
437.0880.0340.00* 164.0884.0

196 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
in that it feeds primarily on calanoid and cyclopoid copepods whereas the other species
feed on harpacticoid copepods. This correlates with its occurrence in tall corals that place
it higher in the water column (Clarke, 1989; 1996) where planktonic organisms are 43%
more abundant on average (see Results). Calanoid copepods may be even more stratified,
with higher samples being 4900% more abundant during the Acartia spinata aggregation
(Fig. 3). This difference in density occurs over only 1 m, a remarkable gradient that places
a potentially high premium on occupying the high locations. Indeed, field experiments
have shown that fish 1 m above the reef surface have higher feeding rates (3.6 times
higher), growth rates (6.3 times higher) and fecundities (6.3 times higher) than fish living
on the reef surface (Clarke, 1992).
The advantage of high locations might be greater than prey counts or feeding rates
indicate because calanoid copepods are larger than harpacticoids and they also have
thinner exoskeletons. Harpacticoids are found intact without distortion in the guts of
Acanthemblemaria, whereas calanoids are disarticulated and the cephalothoraxes are
misshapen. Consequently, calanoids probably provide more energy per unit catch, but
more variables than size and chitin content must be considered to definitively establish
this (e.g., Kaiser et al., 1992). Calanoids may also take more energy to capture (see
below).
The advantage of high locations may be variable as plankton abundances and composi-
tion change with time. The high-low ratios for calanoid copepods was much greater dur-
ing the Acartia spinata aggregation than at other times (Fig. 3). Unlike the situation for
Chromis chrysurus in Japan however (Noda et al., 1992), no guts in this study were found
empty as a result of low plankton abundance. In addition, many calanoid copepods swarm,
forming dense, stationary aggregations on the one to several meter size scale (Emery,
1968; Hamner and Carleton, 1979; Davis at al. 1992), and may thus add considerable
spatial variability to food availability. A sudden increase in one prey organism could cause
a lag in gut contents as compared to tows either because the guts represent an accumula-
tion over some time in the past or because it takes some time to learn a new search image
(Marcotte and Browman, 1986). Indeed, the lowest plankton-gut overlap for A. spinosa
occurred on the first sampling during the Acartia spinata aggregation.
In spite of the potential for great variability, the patterns between the species are re-
markably consistent in this study, with calanoid and cyclopoid copepods predominating
in A. spinosa guts and harpacticoid copepods predominating in the guts of the other spe-
cies (Fig. 4.). As expected, the fishes living furthest from shore consumed the greatest
proportion of plankton (Fig. 6), but this only represented 2 to 3% of the copepods in their
guts. Only A. spinosa, whose gut contents consist of 33 to 80% calanoid and cyclopoid
copepods, has the potential to import significant amounts of nutrients to the reef from
pelagic plankton. Its higher metabolic rate (see below) correlates with its ingestion of
food at a higher rate than the other species (Clarke, 1992), also contributing to its greater
role in importing nutrients.
METABOLIC RATES.—The slopes of O2 consumption vs weight fall into two clear groups
and a remaining cluster consisting of A. paula (Fig. 7, Table 4). A. paula is the smallest
member of the genus (Johnson and Brothers, 1989); consequently it spans a relatively
narrow range of sizes and is at the limits of resolution for the method of O2 consumption
used herein. As a result, there is a high SE for the regression coefficient and the correla-
tion with size is not statistically significant (Fig. 7, Table 3), but the points do fall signifi-
cantly below the regression curves for the other species.

197
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
The group of species with the highest slopes consists of the two species of
Emblemariopsis and A. spinosa. E. pricei is known to live on surfaces of live coral and sea
fans (Tyler and Tyler, in press) and E. ruetzleri lives on algal turfs (Tyler and Tyler, 1997);
only reproductively active males of these species are found in cavities. Living exposed on
surfaces probably requires a higher metabolic rate than the more sedentary “hemisessile”
manner of Acanthemblemaria, which is found in cavities during all post-larval stages.
The metabolic similarity of A. spinosa with the Emblemariopsis species, therefore, places
it in an uncharacteristic group for its genus. It is also uncharacteristic in occupying corals
that extend high above the reef surface (Clarke, 1989; 1996), a position it shares with the
Emblemariopsis species whose coral and sea fan substrates also extend well above the
reef surface (Tyler and Tyler, 1997; in press). Another feature that A. spinosa shares with
the Emblemariopsis spp. is that they probably both have a higher rate of food encounter
than do the other species of Acanthemblemaria. Feeding above the reef surface on plank-
ton results in a flow of prey past the cavities occupied by A. spinosa and the mobility of
Emblemariopsis allows them to forage over a wider area than the species of
Acanthemblemaria which remain in their cavities, extending for only one, or at most two,
body lengths from the entrance.
The O2 consumption of A. spinosa relative to A. aspera in Belize is the same as in St.
Croix (1.54 and 1.57 times greater respectively, for a median size fish of 80 mg). But the
rates for both species were 60% greater in Belize than in St. Croix, and this is associated
with both higher slopes (0.25 vs 0.15 and 0.14 vs 0.10 for A. spinosa and A. aspera
respectively) and higher intercepts (19.2 vs 11.9 and 14.2 vs 7.2 for A. spinosa and A.
aspera respectively; this study and Clarke, 1992). The causes of these differences are
unknown; the temperatures were not higher in Belize. The only differences in technique
were that (1) the fish in Belize were incubated in BOD bottles while still in their venti-
lated vials whereas the fish in St. Croix were placed directly in the BOD bottles, and (2)
the average time that incubation started was 1030 in Belize and 1410 in St. Croix.
ADAPTIVE STRATEGIES.—Except for A. spinosa, members of the genus Acanthemblemaria
are found on or very near to the reef pavement where plankton densities are low. They
feed primarily on benthic organisms (this study; Lindquist and Kotrschal, 1987) and be-
cause they are not mobile, they are dependent on the movement of their prey to provide a
continuous supply of food. This may result in a low rate of food delivery, so they may
have low metabolic rates as an adaptation to this resource, as has been suggested for
terrestrial ectotherms (Pough, 1980).
A. spinosa stands out from other members of the genus that have been studied in three
ways: (1) higher metabolic rate, (2) higher locations in dead corals, and (3) greater con-
sumption of planktonic prey. Diurnal planktivorous fish are generally vulnerable to preda-
tors and possess anti-predator adaptations such as streamlining for fast swimming or a
body form that makes them difficult for predators to handle (Hobson, 1991). A. spinosa
seems to have found a different strategy: it protects itself from predators by living in
cavities in the coral yet is in a position to eat plankton where they are most abundant by
living almost exclusively in branching corals that elevate them well above the reef sur-
face. To occupy this guild, A. spinosa may need a higher metabolic rate than other mem-
bers of the genus because calanoid copepods, with their long antennae, are more difficult
to catch than the harpacticoids eaten by other members of the genus (J. Cordell, pers.
comm.). Its higher metabolic rate may also bestow a competitive advantage in agonistic
encounters for preferred high locations (Clarke, 1989; 1992). In addition, planktonic food

198 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 1, 1999
is probably less subject to local depletion than benthic food, which may enable the higher
food consumption necessitated by a higher metabolic rate. In this way, A. spinosa may be
like a homeotherm in that, at a metabolic cost, it can occupy a different habitat and catch
different food.
It is possible to test the requirements and consequences of high metabolic rates; for
example (1) the capture success for harpacticoid and calanoid copepods can be deter-
mined for different blenny species in laboratory studies, (2) the caloric values of ingested
food and feces can be determined for harpacticoid and calanoid copepods in different
blenny species, (3) the outcomes of agonistic encounters can be compared between blenny
species, and (4) harpacticoid copepod density can be compared through precise sampling
in the agal turf near and far from blenny shelters.
With its conspicuous nodding behavior and prominant microhabitat, A. spinosa is more
likely than other members of its genus to be observed by divers. It is thus important to
note that the most widespread and well-known species of Acanthemblemaria in the Car-
ibbean, indeed the type species for the genus, is atypical. Therefore we should exert cau-
tion in choosing representative species in comparative studies of reef community dynam-
ics.
ACKNOWLEDGMENTS
I thank K. Clarke, J. Tyler, D. Tyler, M. Phillips, A. Sundberg, and C. Clark for assistance in the
field and C. Muelders for help with the gut analyses. J. Cordell, T. Sutton and J. Grieve aided with
identification. J. Tyler and D. Tyler provided helpful comments on an earlier draft of this paper. This
is Contribution No. 557 from the Caribbean Coral Reef Ecosystem Program of the National Mu-
seum of Natural History, Smithsonian Institution.
LITERATURE CITED
Almany, G. R. and C. C. Baldwin. 1996. A new Atlantic species of Acanthemblemaria (Teleostei:
Blennioidei: Chaenopsidae): Morphology and relationships. Proc. Biol. Soc. Wash. 109: 419–
429.
Brett, J. R. and T. D. D. Groves. 1979. Physiological energetics. Pages 279–352 in W. S. Hoar, D. J.
Randall and J. R. Brett, eds. Fish physiology, vol. VIII. Academic Press, New York.
Buchheim, J. R. and M. A. Hixon. 1992. Competition for shelter holes in the coral-reef fish,
Acanthemblemaria spinosa Metzelaar. J. Exp. Mar. Biol. Ecol. 164: 45–54.
Clarke, R. D. 1989. Population fluctuation, competition and microhabitat distribution of two spe-
cies of tube blennies, Acanthemblemaria (Teleostei: Chaenopsidae). Bull. Mar. Sci. 44: 1174–
1185.
__________. 1992. Effects of microhabitat and metabolic rate on food intake, growth and fecun-
dity of two competing coral reef fishes. Coral Reefs 11: 199–205.
__________. 1994. Habitat partitioning by chaenopsid blennies in Belize and the Virgin Islands.
Copeia 1994: 398–405.
__________. 1996. Population shifts in two competing fish species on a degrading coral reef. Mar.
Ecol. Prog. Ser. 137: 51–58.
Emery, A. R. 1968. Preliminary observations on coral reef plankton. Limnol. Oceanogr. 13: 293–
303.
Davis, C. S., S. M. Scott and A. R. Solow. 1992. Microaggregations of oceanic plankton observed
by towed video microscopy. Science 257: 230–232.
Gerking, S. D. 1994. Feeding ecology of fish. Academic Press, San Diego, California.

199
CLARKE: DIETS AND METABOLIC RATES OF BLENNIES
Glynn, P. W. 1973. Ecology of a Caribbean coral reef. The Porites reef-flat biotope: Part II. Plank-
ton community with evidence for depletion. Mar. Biol. 22: 1–21.
Greenfield, D. W. and T. A. Greenfield. 1982. Habitat and resource partitioning between two spe-
cies of Acanthemblemaria (Pisces: Chaenopsidae) with comments on the chaos hypothesis.
Smithson. Contrib. Mar. Sci. 12: 499–507.
Hamner, W. M. and J. H. Carleton. 1979. Copepod swarms: Attributes and role in coral reef ecosys-
tems. Limnol. Oceanogr. 24: 1–14.
____________, M. S. Jones, J. H. Carleton, I. R. Hauri and D. McB. Williams. 1988. Zooplankton,
planktivorous fish, and water currents on a windward reef face: Great Barrier Reef, Australia.
Bull. Mar. Sci. 42: 459–479.
Hobson, E. S. 1991. Trophic relationships of fishes specialized to feed on zooplankters above coral
reefs. Pages 69–95 in P. F. Sale, ed. The ecology of fishes on coral reefs. Academic Press, San
Diego, California.
Ivlev, V. S. 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven,
Connecticut.
Johnson, G. D. and E. B. Brothers. 1989. Acanthemblemaria paula, a new diminutive chaenopsid
(Pisces: Blennioidei) from Belize, with comments on life history. Proc. Biol. Soc. Wash. 102:
1018–1030.
Kaiser, M.J., A. P. Westhead, R. N. Hughes and R. N. Gibson. 1992. Are digestive characteristics
important contributors to the profitability of prey? Oecologia 90: 61–69.
Lindquist, D. G. 1985. Depth zonation, microhabitat, and morphology of three species of
Acanthemblemaria (Pisces: Blennioidea) in the Gulf of California, Mexico. Mar. Ecol. 6: 329–
344.
_____________ and K. M. Kotschral. 1987. The diets in four Pacific tube blennies
(Acanthemblemaria: Chaenopsidae): Lack of ecological divergence in syntopic species. Mar.
Ecol. 8: 327–335.
Marcotte, B. M. and H. I. Browman. 1986. Foraging behavior in fishes: perspectives on variance.
Environ. Biol. Fish. 16: 25–33.
Motulsky, H. J., J. R. Pilkington and P. Stannard. 1993. InStat for MacIntosh, ver. 2.0. GraphPad
Software.
Noda, M., K. Kawabata, K. Gushima and S. Kakuda. 1992. Importance of zooplankton patches in
foraging ecology of the planktivorous reef fish Chromis chrysurus (Pomacentridae) at
Kuchinoerabu Island, Japan. Mar. Ecol. Prog. Ser. 87: 251–263.
Pough, F. H. 1980. The advantages of ectothermy for tetrapods. Am. Nat. 115: 92–112.
Rützler, K. and I. G. Macintyre. 1982. The habitat distribution and community structure of the
barrier reef complex at Carrie Bow Cay, Belize. Smithson. Contrib. Mar. Sci. 12: 9–45.
Schoener, T. W. 1968. Sizes of feeding territories among birds. Ecology 49:123-141.
Sedberry, G. R. and N. Cuellar. 1993. Planktonic and benthic feeding by the reef-associated vermil-
ion snapper, Rhomboplites autorubens (Teleostei, Lutjanidae). Fish. Bull. 91: 699–709.
Tyler, D. M. and J. C. Tyler. 1997. A new species of chaenopsid fish, Emblemariopsis ruetzleri,
from the western Caribbean off Belize (Blennioidei), with notes on its life history. Proc. Biol.
Soc. Wash. 110: 24–38.
Tyler, J. C. and D. M. Tyler. in press. Natural history of the sea fan blenny, Emblemariopsis pricei
(Pisces, Chaenopsidae), in the western Caribbean. Smithson. Contrib. Zool.
DATE SUBMITED: July 23, 1998. DATE ACCEPTED: November 23, 1998.
ADDRESS: Department of Biology, Sarah Lawrence College, 1 Mead Way, Bronxville, New York 10708.