McKinnon: Feeding habits of Lagenorhynchus obscurus 



573 



By weight, anchoveta accounted 

 for more than half of all prey con- 

 sumed in every collection period 

 and its lowest frequency of occur- 

 rence was 93.3% (Table 3). The 

 percent weight of T. symmetricus 

 in dusky dolphin diets was always 

 low, ranging from 09c in the win- 

 ter of 1985 to a maximum of 14.5% 

 in the summer of 1986, but fre- 

 quency of occurrence was more 

 variable, ranging from 0% to 

 53.3% (Table 3). In a log-linear 

 analysis, the three-way interac- 

 tion between year, season, and fre- 

 quency of occurrence of T. 

 symmetricus in stomachs was not 

 statistically significant (G=1.91, 

 df=l, P>0.10), nor was the inter- 

 action between season and fre- 

 quency of occurrence significant 

 (G=2.26, df=l, P>0.10). The large 

 increase in frequency of occurrence 

 of this prey from 1985 to 1986, 

 however, resulted in a highly sig- 

 nificant interaction between fre- 

 quency of occurrence and year 

 (G=37.76,df=l,P<0.001). The only 

 other prey found sufficiently often 

 in stomachs for statistical testing 

 of frequency, but not so often that 

 insufficient variation was present 

 (as for anchoveta), wasL. gahi, for 

 which log-linear analysis yielded 

 a significant three-way interaction 

 between frequency of occurrence, 

 year, and season of collection 

 (G=13.44, df=l, P<0.001). Thus 

 neither season nor year exerted a 

 clear, independent effect on con- 

 sumption of L, gahi, although 

 there was considerable variation 

 among collecting periods ( Table 3 ). 



Reproductive status did not have 

 any obvious effect on dusky dolphin 

 feeding habits (Table 4). The fre- 

 quencies of occurrence of T. symmet- 

 ricus and L. gahi did not differ sig- 

 nificantly between either lactat- 

 ing females and pooled nonlactating mature females 

 (G=0.39, df=l,P>0.5; G=1.06, df=l, P>0.3, respectively 

 for each prey species ) or between lactating females and 

 all other individuals pooled (G=0.29, df=l, P>0.5; 

 G=0.17, df=l, P>0.5, respectively). Anchoveta were con- 

 sumed by both sexes almost without exception (Table 4). 



Table 2 



Relative importance of prey species of the dusky dolphin, Lagenorhyiuhus 

 obscurus, from the coastal waters of central Peru in the summers and win- 

 ters of 1985 and 1986 (n=136 stomachs, 9,137 individual fish and squid). 

 Percent weight and percent gross energy were calculated by using only the 

 six most important prey species. 



Percent total numbers values were sometimes very low, e.g. 0.01% or only 0.09r to an accu- 

 racy of one decimal place, even when a prey species was present in more than one stomach 

 and its percent frequency of occurrence was greater than 1%. 



There appeared to be at least eight species offish represented among the otoliths which could 

 not be identified. 



Prey size 



All dusky dolphin prey species for which lengths and 

 weights could be estimated averaged less than 30 

 cm (mantle length for the squids, fork length for M. 

 gayi, and total length for all others) and 300 g. The 



