312 
Fishery Bulletin 109(3) 
heavily rely on crypsis or camouflage provided by mac- 
roalgae and on rugosity for either refuge or ambush. 
Large pools near the infralittoral are expected to 
have a larger proportion of transient fishes (Gibson 
and Yoshiyama, 1999). However, on low gradient reefs, 
pool size is probably more important than connectiv- 
ity to the sea or other pools because many such spe- 
cies are schooling or active swimmers (e.g., families 
Atherinopsidae, Carangidae, Gerreidae, Haemulidae, 
Mugilidae, and Sparidae). No distributional pattern 
clearly linked community structure to distance from 
the sea and the most suggestive patterns (C, A, and 
D) probably include the influence of parameters, as yet 
unaccounted for, that act upon the fauna. Although 
distance from the forereef may be one of these deter- 
minant parameters, we suspect that “aloneness” of a 
tidepool may play an important role in concentrating, 
at low tide, many fish (herbivores and nonherbivores 
alike) of large size that were roving over the reef flat 
at high tide. Thus, if juvenile rovers avoid the forereef, 
a relatively small pool alone in a large area suitable for 
roving (i.e., the reef flat at high tide) would at low tide 
“drain” as many roving transient fish as several large 
pools dispersed over this same (and now unsuitable) 
area. Experimental substrate manipulations, such as 
those proposed by Griffiths (2003) and performed by 
Griffiths et al. (2006), Arakaki and Tokeshi (2010) and 
Rojas and Ojeda (2010), would permit evaluation of the 
importance of each factor in determining the composi- 
tion and structure of pool communities. 
Another factor that could influence the spatial dis- 
tribution of species is intra- or interspecific resource 
partitioning, and either of these may cause segregation 
(Gibson, 1986; Faria and Almada, 1999; Zander et al., 
1999; Davis, 2000; Faria and Almada, 2001; Arakaki 
and Tokeshi, 2010). The sympatric gobiid species B. 
soporator and C. boleosoma display frequent intra- and 
interspecific agonistic behavior. In particular, B. sopora- 
tor feeds on C. boleosoma and, occasionally, cannibalizes 
smaller individuals (senior author, personal observ.). For 
both species, territoriality explains agonistic behavior 
between conspecific individuals. The two species share 
the same preference for eurythermal-euryhaline pools 
and interspecific interactions result from niche overlap. 
These interactions could theoretically result in spatial 
segregation, but there is no clear evidence that such a 
process occurs at Praia do Castelhanos. Further study 
is necessary to comprehend the importance of interspe- 
cific interactions in the distribution of these species. 
Conclusions 
The main factors structuring intertidal fish communi- 
ties on rocky shores (e.g., isolation, height, gradient, 
and exposure of tide pools to waves) were held constant 
at Praia dos Castelhanos. Spatial distribution of fishes 
only partially depended upon the physicochemical set- 
ting and was apparently independent of distance of the 
tide pools from the sea. The structure of the fish corn- 
el Permanent residents □ Opportunists 
Transients 
100 %. 
In 80% - 
3 
*o 
CD 
Q. 
40%- 
20 % - 
0 % 
B 
Figure 5 
(A) Percentage of the number of individuals, and (B) 
percentage of total weight of each residency category in 
pools at Praia dos Castelhanos, Espirito Santo, Brazil. 
munity within each physicochemical setting principally 
resulted from the synergistic interaction of niche avail- 
ability (that directly depends upon pool morphometry) 
and ecological relationships among species (e.g., com- 
petition, territoriality, predation). Mason et al. (2008) 
suggested that niche complementarity (i.e. niche differ- 
ences between species) prevents competitive exclusion 
(when one species outcompetes and displaces another 
by making a better use of the same critical resource) 
and can increase ecosystem function. Further studies 
should focus on niche overlap within and among species 
for an understanding of the influence of competition in 
the structuring of intertidal fish communities. Moreover, 
we hypothesize that pool “aloneness” (the concept here 
includes the degree of connectivity with other pools) 
influences the interactions within pools at low tide by 
incorporating into the pool-effect the space and niches 
available at high tide. The use of out-of-pool space would 
widely vary between direct pool-to-pool passage (e.g., 
B. soporator, Aronson, 1951) and extensive roving over 
nonpool space (e.g., Faria and Almada, 2006). The para- 
digm therefore needs to shift from isolated tidepools 
to pools integrated into their surroundings (pools and 
