Madeira and Joyeux: Distribution patterns of tidepool fishes on a tropical flat reef 
309 
weight among pools ( B . mystacium, L. nuchipinnis, C. 
boleosoma, A. saxatilis, S. fuscus, and A. bahianus ). 
The nMDS ordination analysis on both abundance 
and mean weight showed that the fish assemblage, 
when all taxa are considered, is structured differently 
among pools (Fig. 7). In both cases, pools 1 and 3 were 
segregated from all others (and each other), 3 because 
of a low total abundance and a low number of taxa and 
1 because of the absence of some common species (H. 
poeyi, S. axillare, and M. delalandei ) and the dominance 
of B. soporator and C. boleosoma. The other pools, loose- 
ly grouped, shared similar values in number of taxa, 
Shannon-Weiner diversity, total length (Fig. 4), and 
higher percentages of opportunist and transitory spe- 
cies (Fig. 5). Although differences in the physicochemi- 
cal setting of pools were low (except for pool 1), the 10 
most abundant taxa presented distinctive distribution 
patterns for abundance related to the morphometric 
characteristics of each pool, such as position on the 
reef, depth, surface area, volume, substrate composition, 
algal cover and rugosity. 
Discussion 
Spatial distribution of fishes 
Species living in intertidal ecosystems are distributed 
along a vertical gradient according to their tolerance of 
physical factors and their response to ecological inter- 
actions (Raffaelli and Hawkins, 1996). Although the 
vertical distribution of sessile organisms has been stud- 
ied for many years, the distribution of fishes and other 
mobile organisms that take refuge in pools has not been 
investigated as extensively because it is much more 
dynamic and thus more difficult to study (Zander et al., 
1999; Thompson et al., 2002). The shape and volume 
of a pool, its degree of isolation from the sea, and its 
connectivity with other pools determine the amplitude 
of the fluctuation in physicochemical characteristics 
of the water (Mahon and Mahon, 1994; Davis, 2000; 
Castellanos-Galindo et al., 2005). Consequently, the 
occurrence of each species is dependent more upon pool 
characteristics than upon vertical position of the pool on 
the rocky shore, i.e., species occurrence is nearly azonal 
(Zander et ah, 1999). Other factors, such as exposure to 
waves (Gibson, 1972; Grossman, 1982) and algal cover 
(Bennett and Griffiths, 1984), have been investigated, 
but their influence on the distribution of fishes may be 
secondary. 
Habitat heterogeneity is intimately associated with 
variability in microhabitats and therefore offers the con- 
ditions for the coexistence of antagonistic species (Rojas 
and Ojeda, 2010). Two main microhabitats are available 
to rockpool fishes (Griffiths et al., 2006): the substra- 
tum of the pool where fish can hide and the complex 
macroalgal cover used by midwater or pelagic species. 
Rojas and Ojeda (2010) demonstrated that small fishes 
prefer pools of low structural complexity where there 
are fewer ambush areas for predators. On the other 
Stress: 0 
® © 
Nonmetric multidimensional scaling plot for a compari- 
son of the centroids of (A) morphometric characteristics, 
and (B) physicochemical parameters from intertidal 
rockpools at Praia dos Castelhanos, Espirito Santo, 
Brazil. Numbers (1-6) are the numbers assigned to the 
tide pools (see Fig. 1). 
hand, fish distribution on a rocky shore is also affected 
by inter-intra specific interactions such as competition 
and predation (Gibson and Yoshiyama, 1999; Zander 
et al., 1999) that may lead to microhabitat segregation 
(Faria and Almada, 2001). Thus, many different physi- 
cal and ecological factors regulate the distribution and 
structure of fish community in rockpools (e.g., Gibson, 
1972, 1982), and the principal difficulty is to determine 
the respective contribution of each of these. 
There are about three “vertical” ecological pool-zones 
on a flat, fringing, intertidal rocky reef. The first is lo- 
cated at and just below the upper edge of the reef. This 
is an area under stronger atmospheric and terrestrial 
influence and that is subject to shifting conditions from 
reef to nonreef environment. Permanent residents and 
a few opportunistic fish that can tolerate the physi- 
ologic stress caused by physicochemical changes dur- 
ing the exposure of the reef to the air (Evans et al., 
1999), normally dominate this zone. There, B-pattern 
species, such as the eurythermic and euryhaline B. 
soporator and C. boleosoma, probably find resources 
available and lower predation risk because at low tide 
