570 



Fishery Bulletin 92(3), 1994 



Figure 1 



Map of Peruvian coastline, showing study sites. The 

 inset shows the area in relation to the remainder of 

 South America (Read et al., 1988). 



study (Read et al., 1988). Pucusana, a small fishing 

 town approximately 50 km south of Lima (Fig. 1) was 

 the principal collecting site. Stomach content samples 

 were also collected from the nearby ports of Cerro 

 Azul and Ancon, about 70 km south and 90 km north 

 of Pucusana, respectively (Fig. 1). Dolphin landings 

 at any given port were highly variable (Read et al., 

 1988), therefore collecting efforts were concentrated 

 at the port or ports where the most dolphins were 

 being caught. The collecting periods were the aus- 

 tral summers (1 January through 31 March) and 

 winters (1 July through 30 September) of 1985 and 

 1986. 



Field procedures 



Standard length, measured in a straight line along 

 the main axis of the body from the tip of the upper 

 jaw to the notch of the flukes, was taken for each 

 specimen upon being landed. Females were pressed 

 above the nipples and expression of milk, indicating 

 lactation, was noted. When dolphins were eviscer- 

 ated, female reproductive tracts were removed and 

 checked for the presence of a fetus. Ovaries were in- 



spected for corpora lutea and albicantia, then pre- 

 served in 10% formalin. 



Collection of stomach contents also began upon 

 evisceration, between approximately 6 and 48 hours 

 after capture. Each of the fore-, main, and pyloric 

 stomachs was separately rinsed through a series of 

 three brass sieves (Treacy and Crawford, 1981; Murie 

 and Lavigne, 1985) of mesh diameters 4.75, 1.40, and 

 0.425 mm. The sieved contents were then placed in 

 deep, water-filled plastic trays so that any remain- 

 ing flesh could be skimmed off. Fish otoliths and clean 

 squid beaks were retrieved and stored in 5-10% al- 

 cohol; squid beaks with flesh still attached were 

 stored in either 70% alcohol or 10% formalin. The 

 forestomach consistently contained the least digested 

 contents and the greatest volume, so only material 

 from that chamber was later quantified and analyzed 

 (see also Perrin et al., 1973; Bernard and Hohn, 

 1989). 



Reproductive status 



Reproductive status of females was classified after 

 macroscopic examination of gonads and accessory 

 reproductive tissues. Males were classified only as 

 to sex. Females were defined as 1 ) immature, if their 

 ovaries lacked corpora lutea and albicantia; 2) preg- 

 nant, if a fetus was visible in the uterus; 3) lactat- 

 ing; 4) simultaneously pregnant and lactating; or 5) 

 resting, if corpora lutea or albicantia were present 

 but there was neither a fetus visible in the uterus 

 nor evidence of lactation ( Perrin and Donovan, 1 984 ). 

 The "resting" category may have included females 

 with small embryos not detected during field inspec- 

 tions, in addition to individuals actually between 

 reproductive cycles ( Perrin and Donovan, 1984 ). Data 

 were incomplete for several females; therefore they 

 were classified as "unknown females." No informa- 

 tion on gender was available for several additional 

 samples; their sex and reproductive status were clas- 

 sified as "unknown." 



Identification of prey and calculation of 

 measures of relative importance 



A reference collection of otoliths from common Peru- 

 vian marine fishes was made from specimens pur- 

 chased at local markets and identified with 

 Chirichigno's ( 1978) key. Otoliths from stomach con- 

 tents were identified by comparison with this collec- 

 tion and collections belonging to the Instituto del Mar 

 del Peru (IMARPE) and to P. Majluf (University of 

 Cambridge). Squid species were identified from their 

 beaks by using published keys (Wolff, 1984; Clarke, 

 1986), through reference to beaks from local squids 

 identified and supplied by F. Cardoso (Museo de 



