WATSON and DAVIS: LARVAL FISH DIETS IN SHALLOW COASTAL WATERS 



small items as well, but in addition consumed 

 larger items such as larger copepod species and 

 mysids (Table 9). Cross-shelf patterns in dietary 

 composition were apparent; for example, tintin- 

 nids in the Pr diet shifted from Stenosomella 

 spp. near shore to predominantly Condonana 

 spp. faithest from shore (Table 8). Pr larvae also 

 tended to consume more bivalve veligers more 

 frequently in the blocks farthest from shore, but 

 more gastropod veligers nearer shore (Table 8). 

 The FP larvae hkewise consumed most gastro- 

 pod veligers nearer shore, but did not display 

 cleai' evidence of a cross-shelf pattern in the con- 

 sumption of bivalve veligers (Table 9). The aver- 

 age number of prey items consumed by the Pr 

 larvae ranged from 2.9 to 8.8 per feeding individ- 

 ual (Table 8), while for the FP larvae the number 

 of items consumed ranged from 2.0 to 8.3 (Table 

 9). Cross-shelf patterns in the number of items 

 consumed were not apparent. The FP larvae 

 contained more prey items per feeding individual 

 in midwater and epibenthos than in the neuston. 

 There were no consistent differences between 

 strata for the Pr larvae or between midwater 

 and epibenthos samples for the older larvae 

 (Tables 8, 9). 



curred in blocks A (76.6 ± 7.0%) and B (77.1 ± 

 9.67c) where larval abundance was highest, and 

 the highest feeding incidences occuiTed in blocks 

 C (91.1 ± 9.4%) and D (85.7 ± 33.1%) where 

 larval abundance was low. However, the confi- 

 dence bounds about these estimates overlapped 

 in all cases. 



Larval queenfish consumed a wide variety of 

 prey types (Tables 10, 11). Pr larvae consumed 

 mainly small items, especially bivalve vehgers 

 and small copepods such as Pamcalanus panms 

 and Euteiyina acutifrons (Table 10). The FP 

 larvae also consumed these small items in addi- 

 tion to larger items, especially mysids and gam- 

 marid amphipods (Table 11). Cross-shelf pat- 

 terns in dietary composition and number of items 

 consumed were not clear. For the Pr larvae, 

 feeding specimens contained fewer prey items 

 per individual in block E than elsewhere (Table 

 10), but for the FP larvae no pattern was appar- 

 ent (Table 11). Both the Pr and FP larvae tended 

 to consume slightly more prey items per feeding 

 individual in midwater than in epibenthos. 

 Smaller prey contributed larger fractions of the 

 diet in midwater than in the epibenthos for both 

 larval stage classes. 



Seriphus politus 



A total of 742 larval queenfish (61% Pr stage) 

 were examined from samples collected between 

 September and November 1978 and between 

 March and September 1979 (Table 1). The over- 

 all 67% feeding incidence was comparable to the 

 72% night incidence noted in the day/night sam- 

 ple set. Feeding incidence differed httle between 

 the Pr larvae (73.5 ± 4.2%) and the FP larvae 

 (79.2 ± 4.8%). Relatively few larvae were avail- 

 able for dissection from the neuston samples, 

 and overall these larvae displayed the lowest 

 feeding incidence (48.5 ± 12.6% for the Pr lar- 

 vae; no FP larvae occurred in the neuston). 

 Feeding incidence differed little between mid- 

 water (76.4 ± 6.1%) and epibenthos (79.4 ± 

 3.8%). The feeding incidence for the Pr larvae 

 was highest in block A (81.5 ± 6.1%) where lar- 

 val abundance was highest, and ranged between 

 about 62 and 68% in the remaining blocks. Confi- 

 dence limits about the means for blocks B (63.1 

 ± 9.4%) and D (61.7 ± 11.2%) did not overlap the 

 confidence limits about the mean for block A, but 

 those about the means for blocks C (66.7 ± 

 12.4%) and E (65.9 ± 15.1%) did overlap the 

 confidence limits about the block A mean. The 

 lowest feeding incidences for the FP larvae oc- 



DISCUSSION 



Both the limited day/night sample series and 

 the much larger night-only sample set indicated 

 that larvae contained food well into the night. A 

 nonzero feeding incidence does not necessarily 

 imply recent feeding, however, but only indi- 

 cates the presence of food in the gut. For ex- 

 ample, a slow digestion and evacuation rate 

 might result in the appearance of noctm-nal feed- 

 ing even if the larvae in fact w'ere not feeding. 

 On the other hand, a low feeding incidence does 

 not necessarily imply nonfeeding, especially for 

 taxa that have a sti'aight gut, since these larvae 

 frequently void their gut contents during cap- 

 ture and fixation (June and Carlson 1971; Hay 

 1981). Hunter (1981) noted that fish larvae are 

 visual feeders lacking rods and retinomotor pig- 

 ment migration (e.g., Blaxter 1968) and are 

 probably largely restricted to feeding in daylight 

 hours. The low feeding incidences noted for 

 larval Paralabrax spp. in the night samples sug- 

 gests that they do feed only during the day. 

 Similarly, Paralichthys californicus larvae may 

 feed only during the day, although theii- day- 

 night differences in feeding incidence were 

 smaller and less convincing owing to the broad 

 confidence hmits about the %FI values. Both 



579 



