Sedberry and Cuellar Planktonic and benthic feeding by Rhomboplites aurorubens 



701 



plies no selectivity by the predator, that is to say the 

 fish is feeding on the prey in proportion to the prey's 

 relative abundance in samples taken in the habitat. 



The electivity index was calculated for species that 

 were numerically dominant in prey environment 

 samples or in fish stomach samples. Dominant species 

 included those that ranked among the five most abun- 

 dant species within stomach or environmental samples 

 pooled by depth zone. Prey samples and stomach col- 

 lections were pooled by depth zone (inner, middle, and 

 outer shelf) for comparison. Samples of benthic prey 

 were obtained at the 11 reef sites during 1980 and 

 1981 with a diver-operated suction sampler (inner and 

 middle shelf) or a grab sampler (outer shelf). Details 

 of benthic sampling are provided elsewhere (Wenner 

 et al., 1983; Wenner et al., 1984) and are only summa- 

 rized here. Briefly, divers obtained five replicate suc- 

 tion samples during the day at each inner and middle 

 shelf reef site by scraping the hard reef substrate en- 

 closed by a 0.1-m 2 quadrat box while simultaneously 

 sucking with an airlift device similar to that described 

 by Chess (1979). Suction samples were collected in 

 1.0-mm mesh bags. 



At the outer shelf stations, where water depth pre- 

 cluded the use of the suction device operated by divers, 

 quantitative 0.1 m'--samples were collected in reef habi- 

 tat with a modified Smith-Mclntyre grab. After re- 

 trieval, each sample was sieved through a 1.0-mm sieve 

 and retained organisms were identified and counted. 

 Most (75.0%) grab samples were taken during the day. 



Additional sampling was conducted to compare stom- 

 ach contents with potential prey species occurring in 

 the near-bottom water column above the reef. To de- 

 termine extent of feeding on the near-bottom plank- 

 ton, electivity values were calculated for samples ob- 

 tained with an epibenthic plankton sled and compared 

 with those from the suction and grab samples. The 

 sled was similar to that described by Brattegard and 

 Fossa (1991) but had a mouth opening of 0.5 m 2 and 

 runners that permitted it to sample 0.5 m off the bot- 

 tom. The sled also had a mouth-opening mechanism 

 designed to fish only when it was in contact with the 

 bottom. A 0.947-mm mesh net was attached to the 

 sled and two ten-minute tows were made per station 

 at night to minimize net avoidance. Sled collections 

 for invertebrates were made only in 1981. These 

 samples were analyzed for decapods, stomatopods, 

 cumaceans, and mysids, and comparisons with fish 

 stomachs were limited to these taxa. Underwater tele- 

 vision and diver observations were used to direct in- 

 vertebrate sampling to reef habitat. Further details 

 of all fish and invertebrate sampling can be found in 

 Sedberry and Van Dolah (1984) and Wenner et. al. 

 (1984). ' 



Daytime submersible observations of vermilion snap- 

 per behavior were made during dives aimed at visu- 

 ally censusing (transects and point counts) reef fishes 

 of the outer shelf on 14-15 July 1985. The submers- 

 ible used was the Sea Link /, which provided a pan- 

 oramic view (Askew, 1985) of shelf edge reefs and their 

 fishes. 



Results 



Vermilion snapper was very abundant in trawl catches. 

 Although it was relatively infrequent at inner shelf 

 stations (mean catch per tow of 12.4), vermilion snap- 

 per was very abundant at middle (243.7 per tow) and 

 outer shelf (140.1 per tow) stations. 



Approximately 115 species of prey were identified in 

 255 stomachs that contained food. Fish with stomachs 

 containing food were found at all times of the day; 

 22% of stomachs with food were collected between 0001 

 and 0600 hours local time; 17%, between 0601 and 

 1200 hours; 18%, between 1201 and 1800 hours; and 

 43%, between 1801 and 2400 hours. Although no at- 

 tempt was made in the field to quantify proportions of 

 stomachs with food, there were more stomachs with 

 obvious contents at night ( 1801-0600 hr). 



Most prey items found in vermilion snapper stom- 

 achs were planktonic or nektonic organisms (Table 1). 

 Amphipods, mainly planktonic hyperiids and caprellids 

 (e.g., Lestrigonus bengalensis, Phtisica marina), cope- 

 pods, and decapods (e.g., larval forms, Lucifer faxoni) 

 were the most frequently consumed taxa. These small 

 crustaceans were eaten in large numbers by smaller 

 vermilion snapper, but, with the exception of deca- 

 pods, contributed little to prey volume (Table 2). Mysids, 

 cumaceans, and doliolids were also frequently con- 

 sumed. The overall diet was dominated volumetrically 

 by squids and fishes. 



Significant feeding differences between size classes 

 were detected by the Tyler chi-square feeding hetero- 

 geneity index (Table 2). Several groups of crustaceans 

 and fishes demonstrated significant differences in fre- 

 quency between size classes of vermilion snapper. De- 

 capods were important prey for larger fish but domi- 

 nated the diet of fish between 50 and 100 mm SL. 

 Small crustaceans, such as copepods, stomatopods and 

 amphipods decreased in relative number and volume 

 in the diet with increasing size of predator, although 

 amphipods remained a frequent food item in all size 

 classes examined. Barnacles and cumaceans varied in 

 relative volume of prey for different size classes, but 

 cumaceans increased in frequency in larger fish. Al- 

 though squids and fishes were eaten by most size 

 classes, they were volumetrically most important in 



