FISHERY BULLETIN: VOL. 73, NO. 4 



show more generalized behavior and are less 

 specialized for picking than either kelp perch or 

 senoritas. Like senoritas, white seaperch ate 

 mostly plant-encrusting bryozoans, but their 

 foraging behavior is quite different. White 

 seaperch typically feed alone or in very small and 

 loose aggregations. Feeding individuals often 

 hover head down within 1 m of the substrate and, 

 judging from their eye movements, search 

 carefully for food. Even so, the substantial 

 amounts of sand and other debris mingled with 

 the more select items in white seaperch guts in- 

 dicate that once the fish find their sedentary bot- 

 tom prey, they engulf it in relatively large and 

 indiscriminate mouthfuls. 



Underwater disturbances attract white 

 seaperch and senoritas. For example, the two 

 fishes commonly aggregate and feed where bat 

 ray, Myliobatis californica, are stirring up the 

 bottom with their wings. They are also quick to 

 follow and assemble about actively working scuba 

 divers. This seems to be an adaptation to forage 

 opportunistically in the wake and disturbance left 

 by others, a strategy which is commonly used by 

 tropical wrasses (Hobson 1974). In contrast, the 

 kelp perch appears to be much less aware of such 

 disturbances and often seems oblivious of an ob- 

 server at close range. 



Indirect evidence suggests that the plant 

 material ingested with the bryozoans is not a 

 primarily source of food for the fish. Only 10'^ of 

 the ingested material was bare of bryozoans, in- 

 dicating that white seaperch and senoritas select 

 the heavily encrusted bits. Also, their relative gut 

 lengths are less than expected for herbivores and 

 many omnivores. Odum (1970) noted that the ratio 

 of gut to fish length is usually less than unity in 

 carnivores, one to three in omnivores, and greater 

 than three in herbivores. Mean ratios from 74 

 white seaperch and 65 senoritas are only 

 0.82 + 0.028 (95% confidence interval) and 0.75 + 

 0.036, respectively. They do not differ significantly 

 from the mean ratio of 0.76 ± 0.024 from 95 kelp 

 perch, which ingest comparatively little plant 

 material. Likewise, Chao (1973) found no evidence 

 that the cunner, TautogolahrKs adspersu.s, a 

 temperate labrid from off the Atlantic Coast, as- 

 similates the algae it ingests. Small undigested 

 amounts from the intestine of the cunner are 

 usually associated with digested epiphytic 

 animals, including bryozoans. Primarily a shellfish 

 eater, the cunner also has a gut ratio that is less 

 than unity. 



Individual diets of kelp perch and senoritas vary 

 considerably from fish to fish, but this is not likely 

 attributable to facultative cleaning. Diets of white 

 seaperch, which do not commonly clean, were no 

 more concordant than those of the other two 

 species. Instead, opportunistic feeding in general 

 may account for most of the variability. Kelp perch 

 may switch from one patch of plankton to another, 

 or feed on the kelp surface as the opportunity 

 arises. Individuals were seen to dart back and 

 forth between open areas and the kelp surface, 

 selecting prey from either source. Although most 

 sefioritas eat large amounts of bryozoans, many 

 select small crustaceans, especially amphipods. 

 Hobson (1971) noted that senoritas not only eat 

 mid-water plankton, but are occasionally seen 

 picking about on the bottom. We observed that 

 they are usually among the first to arrive at un- 

 derwater chumming stations where sea urchins 

 are broken open. 



Yet cleaning contributes to the food breadth of 

 kelp perch and senoritas by adding items that can 

 be taken only by that process. And this points out 

 a major problem in measuring food breadth by the 

 "richness" or number-of-items measure, S. The 

 categories of food items cannot be defined objec- 

 tively, from the fish's point of view at least. For 

 example, the total number of items recorded for 

 the white seaperch would obviously increase if we 

 further diversified the benthic categories (which 

 are not exploited by the cleaners) by— say— distin- 

 guishing gastropods from bivalves within the cate- 

 gory of "crushed shells." Even though cleaning 

 increases S, its total nutritional importance to the 

 cleaner species may be negligible. 



Likewise, it is diflRcult to conclude whether or 

 not cleaners have specialized diets. The total items 

 eaten by either cleaner exceeded that eaten by the 

 supposedly more generalized white seaperch, and 

 the unsealed food breadth of the kelp perch was 

 greatest of all three species. But the kelp perch 

 and, to a lesser extent, the senorita are in fact 

 limited to smaller items because they have smaller 

 mouths. The 25 subjectively determined food cate- 

 gories included some 15 "small items" (usually <3 

 mm in diameter) but only 10 "large" (usually >3 

 mm). Therefore, the diet of the kelp perch ap- 

 peared to be relatively broad because it includes all 

 of the small items, several of which are exclusively 

 planktonic. On the other hand, the diet of the 

 white seaperch, which rarely visits the canopy, 

 appeared to be more narrow because it includes 

 relatively few of these small prey. Yet having a 



826 



