268 



Fishery Bulletin 93(2), 1995 



Analysis of habitat preferences 



Cape Hatteras — Analysis of species' habitat prefer- 

 ences produced three factors (Table 4). Factor 1 was 

 related to species preferences for different types of 

 epifaunal assemblages. Factor 2 was related to dif- 

 ferent amounts of mounds and crinoids and presence 

 or absence of sargassum detritus. Factor 3 was related 

 to density of holes and presence of grass detritus. 



Factor 1 had high positive loadings for numbers of 

 small anemones, tubes, and gastropods, and a nega- 

 tive loading for numbers of large anemones. This 

 factor differentiated nekton species found more of- 

 ten in intervals with large numbers of small anemo- 

 nes, gastropods, and tubes, from forms more com- 

 mon in areas with few small anemones, tubes, and 

 gastropods (but where large anemones might be 

 found). Species' scores showed that the scorpaenid, 

 rattail, and shrimp were characteristic of areas with 

 large numbers of small anemones and tubes, whereas 

 cancroid crabs, lizardfish, and Species A were not 

 usually found in such areas (Fig. 1). 



Factor 2 had high positive loadings for density of 

 mounds and number of crinoids. Sargassum detri- 

 tus had a high negative loading on this factor. This 

 factor differentiated between forms found in asso- 

 ciation with crinoids and mounds, and forms found 

 away from such areas, more in association with sar- 

 gassum detritus. Species found more in areas with 

 crinoids and mounds included the eel and lizardfish. 

 Forms found in areas with few crinoids and mounds 

 (but with sargassum detritus) included Species A, 

 squid, octopod, and scorpaenid (Fig. 1). 



Factor 3 had high loadings for density of holes and 

 presence of grass detritus. Gastropods and crinoids 

 also contributed positively to this factor. Species 

 found more in areas where holes were dense included 

 the lizardfish, flounder, and scorpaenid, while spe- 

 cies not characteristic of such areas included the 

 squid, cancroid crab, and shrimp (Fig. 1). 



We compared species' and location variances on 

 each of the three factors to determine which species 

 appeared to be selecting subsets of the environment 

 represented by each factor. Figure 2 illustrates the 

 distributions on factor 1 of location scores, squids (a 

 significant variance comparison), and scorpaenids 

 (variance comparison not significant). 



Habitat selection according to type of epifaunal 

 assemblage (factor 1) was shown for the squid, which 

 preferred areas where neither small anemones nor 

 large anemones were found (Fig. 2). Habitat selec- 

 tion related to mounds and crinoids (factor 2) was 

 demonstrated for the hake. Hakes tended to be found 

 in areas with many mounds and crinoids. Habitat 

 selection with respect to density of holes and grass 



detritus (factor 3) was demonstrated for the hake, 

 sergestid shrimp, and lizardfish. These three forms 

 all tended to be seen in areas with intermediate num- 

 bers of holes. 



Cape Lookout — Only five species were observed in 

 enough intervals to include in a factor analysis of 

 their means. These included the rattail, eel, hake, 

 sergestid shrimp, and red deepsea crab. Factor analy- 

 sis of these species produced two factors (Table 4). 

 Factor 1 related to density of holes and mounds, and 

 factor 2 related to types of epifaunal assemblages. 



Factor 1 had high positive loadings for density of 

 holes, mounds, and tubes. Numbers of gastropods, 

 presence of invertebrate tracks, and presence of sar- 

 gassum debris contributed negatively to this factor. 

 This factor differentiated between species found more 

 often in areas with many holes and mounds and spe- 

 cies not usually found in such areas. The hake and 

 the red deepsea crab were characteristic of areas with 

 many holes, whereas the sergestid shrimp tended to 

 be found where holes, mounds, and tubes were less 

 densely distributed. None of the comparisons be- 

 tween environmental and species variances were sig- 

 nificant. 



Factor 2 had high positive loadings for number of 

 small anemones and number of large anemones, and 

 for presence of detritus (both grass and sargassum). 

 The hake was the species most characteristic of ar- 

 eas with many anemones and much detritus, whereas 

 the red deepsea crab was most characteristic of ar- 

 eas devoid of epifauna and detritus. None of the com- 

 parisons between environmental and species vari- 

 ances was significant. 



Discussion 



Mobile species respond to environmental variables, 

 seeking certain conditions and avoiding others. The 

 assumption that individuals select their environment 

 is central to this analysis, as it is to multivariate 

 analyses of habitat selection in general (James and 

 McCulloch, 1990). Environmental variables affect- 

 ing an individual's habitat choice may include abi- 

 otic variables, the presence and density of other spe- 

 cies, and the presence and density of members of its 

 own species. The individual's response to a variable 

 of importance may be negative (avoidance) or posi- 

 tive (attraction). Together, the habitat choices of all 

 individuals in an area produce the distributional 

 patterns observed by the investigator. 



We used patterns of species distribution and asso- 

 ciations between species and environmental vari- 

 ables as a guide to understanding the structure of a 



