diets was most even in kelp perch. Distributions of 

 all measures (s, b, 6/s) did not differ significantly 

 from normal for the kelp perch, but were strongly 

 skewed for the others. In fact, median dietary 

 evenness (6/.s) equalled the maximum possible 

 (1.0) for the others, because most foreguts con- 

 tained either a single item or, occasionally, two 

 items in equal amounts: 55% of the white seaperch 

 and 47% of the senoritas had but a single item, as 

 compared with only 24% of the kelp perch. 

 Moreover, one-item diets of white seaperch and 

 senoritas were more predictable in composition: 

 73% (of 30 foreguts) and 88% (of 25), respectively, 

 contained the same item, plant-encrusting 

 bryozoans. Other one-item foreguts of seaperch 

 contained either plants (4 foreguts), shrimps (2), or 

 polychaete worms (2). Other one-item foreguts of 

 senoritas had either small copepods (1), gnathiid 

 isopod larvae presumably from cleaning (1), or 

 serpulid worms (1). The 12 one-item kelp-perch 

 foreguts contained either small copepods (6), 

 gammarid amphipods (4), bryozoans (1), or 

 caprellid amphipods (1). 



Activity 



Activity breadths measured by feeding rates in 

 the field (Table 5, B) correspond to breadths 

 measured by swimming rates observed 

 experimentally (Table 6, B), even though these 

 two independent measures vary inversely among 

 species. The senorita had the smallest feeding- 



Table 5.- Activity breadths of tlie three fishes, as measured by 

 their feeding bites per minute in the field. Sample size is the total 

 fish observed in each species; S is the number of 2-h intervals 

 (maximum of 6) in which fish were observed to make the biting 

 motions. See Table 4 for further explanation (note that the na- 

 ture of feeding-bite breadth precludes samples estimates). 



Species 



Sample size 



8 



e/s 



Kelp perch 

 White seaperch 

 Senorita 



114 

 111 

 139 



5.29 

 5.60 

 3.62 



0.88 

 0.93 

 0.91 



FISHERY BULLETIN: VOL. 73, NO. 4 



rate breadth, because bites were observed over the 

 fewest 2-h intervals (Table 5, S). From mid-after- 

 noon on, the particular individuals followed during 

 the test were not seen to bite, even though they 

 swam actively about in the kelp canopy and below. 

 (Recall that a few other individuals were feeding, 

 but the general impression was of curtailment of 

 feeding activity then.) Yet the senorita had the 

 largest swimming-rate breadth because it swam 

 actively and continuously during all six daytime 

 intervals (Table 6, S). Thus both scaled measures 

 (B/S) were relatively large for the senorita: the 

 distribution of counts, whether of bites per minute 

 or swimming movements per 5-min period, was 

 relatively even among all intervals in which the 

 action occurred {S). But it may be misleading to 

 conclude that senoritas were then most consistent 

 in their daytime feeding activity. During the first 

 four daytime intervals when bites were observed 

 for all species, 40-96% of the kelp perch and 46-69% 

 of the white seaperch were recorded as biting. 

 However, only 13-26% of the senoritas were so 

 recorded. Therefore, senoritas show more 

 variability in feeding activity; i.e., individuals may 

 bite very rapidly for a few minutes then stop for 

 extended periods. On the other hand, individual 

 senoritas were the most consistent in their swim- 

 ming activity. Senoritas led the others in breadth 

 (6) and evenness (b/s), although the inter-species 

 difference in b values was not significant (Table 6). 



Habitat 



By all measures, the senorita had the greatest 

 breadth of habitat within the area of reef and kelp 

 where all three species co-occur (Table 7). The kelp 

 perch, which was more or less restricted to the 

 canopy, was most specialized in distribution. 



Overlap 



The white seaperch and senorita overlapped 



Table 6.-Activity breadths of the three fishes, as measured by their daytime swimming movements in a large 

 outdoor tank. S is the number of 2-h intervals (maximum of 6) in which fish swam across one or more grid lines 

 during a 5-min observation period. See Table 4 for further explanation. 



'Significant at P<0.005 



824 



