336 
Fishery Bulletin 95(2), 1997 
size data to test the hypothesis that dolphins of all 
size classes eat prey of the same size. To do this, we 
fitted a linear regression to prey size versus the to- 
tal body length of the dolphins (Norris, 1961). We 
also tested the null hypothesis of no correlation be- 
tween size class of prey consumed and number of 
items consumed by using a Pearson Correlation 
Matrix (SYSTAT , 1992; unless otherwise noted, all 
statistical tests are interpreted with a=0.05). 
Geographic and seasonal variability 
w i - initial weight of the forestomach 
with contents (g); and 
w c = weight of the forestomach contents 
(g). 
Using a x 2 , we tested the hypothesis of no difference 
in the SFI during the course of the day (all stomachs 
were collected between 0600 and 1800 h). The data 
were stratified by time-of-day collected: 0600-0900, 
0901-1200, 1201-1500, and 1501-1800 h and by SFI 
categories: 0-30%, 31-60%, and 61-100% full. 
To test for variability in diet, we stratified the sample 
by season and area. For season, we used the two 
oceanographic seasons characteristic of the ETP: 
winter (January-June) and summer (July-Decem- 
ber; Reilly, 1990). For area, we stratified the sample 
by the two recognized management stocks: northeast- 
ern and western-southern (Perrin et al., 1994). How- 
ever, we divided the western-southern stock into a 
western and a southern section at the equator (Fig. 
1) because biological differences in pantropical spot- 
ted dolphins have been noted between the western 
and southern sections of the western-southern stock 
(Perrin et al., 1976, 1979; Barlow, 1985; Hohn and 
Hammond, 1985; Myrick, et al., 1986; Chivers and 
Myrick, 1993; Bright and Chivers 6 ). With the 10 most 
numerous species of fish and squid, we used % 2 to 
test the null hypothesis that there was no difference 
in prey consumed by season or area. 
Stomach fullness index (SFI) 
A relative index of stomach fullness (SFI ) was calcu- 
lated for each stomach with the method of Bernard 
and Hohn ( 1989) to estimate when pantropical spot- 
ted dolphins were feeding. With this method, the SFI 
can never equal 100% because the initial weight of 
the contents includes the weight of the forestomach. 
To compensate for this method, we adjusted the scale 
of the SFI to range from 0% to 100% by dividing the 
index for each stomach by the maximum index value 
in our sample (Eq. 1): 
= (W c ,W i ) x 100 
adj SFI nr 
( 1 ) 
where SFI ndj = adjusted stomach fullness index; 
SFI max = maximum stomach fullness index; 
6 Bright, A. M., and S. J. Chivers. 1991. Post-natal growth 
rates: a compariosn of northern and southern stocks of the off- 
shore spotted dolphins. Southwest Fish. Sci. Center, Natl. Mar. 
Fish. Serv., NOAA, P.O. Box 271, La Jolla, CA 92038. Admin. 
Rep. LJ-91-30, 24 p. 
Reproductive condition 
The reproductive condition of female pantropical 
spotted dolphins was determined by microscopic ex- 
amination of the ovaries and macroscopic examina- 
tion of the uteri and mammary glands (Perrin et al., 
1976; Akin et al., 1993). Using the mean number of 
fish and squid consumed by lactating (n= 57) and 
pregnant (n= 37) females, we used Student’s Atest to 
test the null hypothesis that there was no difference 
in consumption of fish and squid between the two 
groups. We also compared the SFI of lactating and 
pregnant females by time-of-day as described in the 
SFI analysis section. 
Results 
Our sample of 428 stomachs contained 49,798 prey 
items, representing 56 fish species and 36 cephalo- 
pod species (Table 1). Thirty-eight (38) of the species 
identified had not been previously reported as prey 
of pantropical spotted dolphins (Shomura and Hida, 
1965; Fitch and Brownell, 1968; Perrin et al., 1973; 
Bernard and Hohn, 1989; Roberts, 1994). Some spe- 
cies identifications could not be positively confirmed 
and were designated as species 1, 2, etc. of the low- 
est identifiable taxon. One crustacean, Pleuroncodes 
planipes, accounted for 2.4% of the crustacean re- 
mains (Table 1). The parasitic nematode Anisakis 
simplex was found in 38% of the stomachs. 
Cephalopods occurred in 354 (82.7%) of the stom- 
achs and fish occurred in 270, or 63.1%, of the stom- 
achs. However, the percent number of fish (66.6%) 
was higher than that for cephalopods (32.6%). Of the 
56 species of fish identified, 27 belonged to the family 
Myctophidae (lantemfish). As a family, myctophids had 
the highest percent by number of all prey (49.7%) and 
accounted for 74.6% of all fish prey. Myctophids occurred 
in 40.2% of all stomachs (Table 1). 
The cephalopod species identified belonged to two 
orders: 1) Teuthoidea (squids; 32 species) and 2) Octo- 
poda (octopuses; 4 species). Squids from the family 
