NOTE Franks et al Stomach contents of juvenile Rachycentron canadum 



375 



^4 sn° 



'in" MS i V . L S 



Figure 1 



Collection sites and sample sizes of juvenile cobia, Rachycentron canadum. from the northern 

 Gulf of Mexico. The MS/AL collection area comprised 17 collection sites. 



a 13-m otter trawl with 4-cm codend mesh. Towing 

 speed was 2.5 knots, and the duration of tows ranged 

 from 0.25 to 1 h. Time and location were recorded for 

 each tow. Juveniles were collected in 25 net hauls 

 ( 18 by day; 7 at night). Whole specimens were frozen 

 aboard the research vessel, generally <0.5 h follow- 

 ing capture. In our laboratory, specimens were 

 thawed, measured to the nearest mm fork length 

 (FL), and weighed to the nearest 0.1 g total weight 

 (TW). Stomachs were removed, opened, and desig- 

 nated either as containing food or as empty. Stom- 

 ach contents were fixed in 10% buffered formalin and 

 then placed in 507c ethanol. A cursory examination 

 revealed that intestinal contents were too digested 

 to be identified; thus, intestines were separated from 

 stomachs at their confluence and discarded. 



Stomach contents were placed onto a 0.840-mm 

 mesh screen sieve and gently washed with fresh 

 water. Prey were identified to the lowest taxa pos- 

 sible, counted, weighed (blotted wet weight) to the 

 nearest 0.1 g, and measured volumetrically to the 

 nearest 0.1 mL by water displacement in a gradu- 

 ated cylinder. Prey too digested for unequivocal identi- 

 fication were assigned to one of the appropriate princi- 

 pal prey categories (fish, crustaceans, or cephalopods) 

 as "remains" and included in the determination of the 

 contributions of those principal categories to the diet. 



Stomach-content data were pooled for all stomachs 

 and analyzed for percent numeric abundance C7cN), 

 percent of total volume (%V), and percent frequency 

 of occurrence (%F) of prey items to describe quanti- 

 tatively the diet, where %N = (number of individu- 

 als of one prey taxon divided by total number of all 

 prey individuals) x 100; %V = (volume of one prey 

 taxon divided by total volume of all prey) x 100; and 

 %F = (number of stomachs containing prey of one taxon 



divided by total number of stomachs that contained 

 any prey items) x 100. These methods have been re- 

 viewed by Windell (1971) and Hyslop (1980). These 

 values were used to calculate an index of relative im- 

 portance (IRI) (Pinkas, 1971 ), where the importance of 

 an item is directly related to the size of the value: 



IRI = {%N + %V)x%F. 



Percent IRI ( %IRI) was also calculated and consisted 

 of the IRI value of each prey taxon divided by the 

 sum of the IRI values (Eggleston and Bochenek, 

 1989). Empty stomachs were excluded from the above 

 computations. 



A small number of other items found in stomachs 

 (e.g. mollusc shells and both larval and postlarval 

 crustaceans and parasitic worms) were considered 

 miscellaneous or nonfood items probably ingested 

 incidentally in normal feeding. These were not used 

 in calculating percentages or indices describing re- 

 lationships of food items in the diet. 



A stomach fullness index was calculated by divid- 

 ing the total prey weight of each cobia by its total 

 body weight x 100 (Sullivan and Gillman, 1993; Rooker, 

 1995). Stomach fullness was tested between daytime 

 and nighttime captures, and to further examine diel 

 feeding periodicity based on stomach fullness, cobia 

 were grouped by time of capture into 6, 4-h periods: 

 0000-0400 (tz=2); 0400-0800 (including dawn) (ra=5); 

 0800-1200 (n = 15); 1200-1600 (7? = 10); 1600-2000 (in- 

 cluding dusk) (77=8); and 2000-2400 (77=9). A one-way 

 analysis of variance ( ANOVA) was used to test for dif- 

 ferences in mean fullness between time intervals. 



Chi-square (x 2 ) 2 x 2 contingency table analyses 

 and Fisher's exact test were used to test for differ- 

 ences in frequency occurrence of principle prey cat- 



