KISHKRY [BULLETIN: VOL 86, NO 1 



ameter was measured to the nearest 0.1 mm with 

 a steel dial caliper. 



Stomach contents, if not larger than approxi- 

 mately 200 cc, were sampled whole; otherwise, 

 the bolus was made homogeneous by stirring, and 

 about 200 cc were preserved. Stomach content 

 samples were fixed and preserved in 10% buffered 

 formalin. The stomach content of each sample 

 was thoroughly agitated, separated with a plank- 

 ton strainer (mesh size 0.5 mm), rinsed of forma- 

 lin in deionized water, and blotted for 30 seconds 

 on blotting paper. The lump was then molded into 

 a cylindrical shape, and a portion of one end was 

 separated to make up 1 g of wet weight, measured 

 to the nearest 0.1 g. The subsample was then 

 placed with water in a gridded tray, and exam- 

 ined under a dissecting microscope. Contents of 

 the spiral intestine were discarded, because the 

 small crustacean prey was rapidly digested. 



Feeding habits were analyzed quantitatively 

 by computing the Index of Relative Importance 

 (%IRI) (Pinkas et al. 1971; Hyslop 1980) for each 

 prey species. The IRI combines percentage by 

 number iN), mass (M), and frequency of occur- 

 rence (F) in the formula: 



mi = (%N + %M) X %F 



Prey items were identified, when possible, to 

 lowest taxa or species, then the %N of all prey 

 species within each subsample was calculated. 

 When more than one species was present, all 

 items were individually counted. To obtain the 

 %M term of the equation, mean mass was calcu- 

 lated for each species by measuring the length of 

 each item contained in five randomly selected 

 squares on a tray, calculating the average length 

 of each prey species, and obtaining mass values 

 from Miller's (1966) Plankton Conversion Tables, 

 where mass is related to length for all main 

 planktonic taxa. The %N from all subsamples 

 were summed, and the percent from the new sum 

 was calculated, to calculate %IRI for each prey 

 species. The same procedure was applied to %M 

 and 7cF. The total 7(N was then added to the total 

 7cM, and that sum was multiplied by the total 

 %F, to obtain total IRI for each prey species, from 

 which the %IRI was calculated. When few items 

 (e.g., copepods) were found among a large amount 

 of partially digested euphausiid or mysid 

 shrimps, the possibility of reconstructing the 

 shrimp number within the subsample by count- 

 ing the digestion-resistant eye pairs was dis- 

 carded to avoid bias in favor of the shrimp frac- 



tion. The following method was adopted instead: 

 all odd prey items were counted and measured, 

 and their total mass was obtained from Miller's 

 tables; this was subtracted from the total weight 

 (1 g) of the subsample. The remaining weight was 

 divided by the mean weight of each individual 

 item, calculated by averaging the lengths of all 

 available intact specimens, and obtaining the cor- 

 responding weight in Miller's tables. A potential 

 biasing factor existed, when only a few prey re- 

 mains were found (e.g., when a relatively uncom- 

 mon item occurred alone in a stomach, therefore 

 contributing a value of 100 7(N and 7<M to the 

 total IRI). This factor was avoided by considering, 

 for quantitative treatment, only those stomachs 

 which contained more than 1 g (wet weight) of 

 recognizable food. All stomachs that had <1 g of 

 contents were in fact virtually empty, and the few 

 items found in them were treated only qualita- 

 tively. 



Remoras were fixed in 10% formalin, preserved 

 in 50% isopropanol, and deposited in the Marine 

 Vertebrate Collection, Scripps Institution of 

 Oceanography. Parasitic copepods were fixed and 

 preserved in 50% isopropanol and sent to the 

 Long Beach State University for taxonomic iden- 

 tification and study. 



RESULTS AND DISCUSSION 



A total of 262 mobulid rays, belonging to four 

 species iMobula thurstoni, M. japanica, M. 

 miinkiana, and M. tarapocana) were examined 

 between 1981 and 1984. Of these, M. thurstom 

 was the most abundant species, constituting 58% 

 of the total catch, followed by M. japanica (30%), 

 M. munkiana (9%), and M. tarapacana (3%). 



KEY TO THE SPECIES OF MORULA IN 

 THE GULF OF CALIFORNIA 



Mobula can be distinguished from Manta by the 

 mouth on the lower surface of the head rather 

 than being subterminal, and by the presence of 

 toothbands in both jaws. Moreover, Manta grows 

 to a greater size, and the size of its head, relative 

 to the body, is much greater than in Mobula. 



la. Branchial filter plates fused 



M. tarapacana 



(Spiracle in an elongated longitudinal 

 slit, dorsal to plane of pectoral fins. Teeth 

 tessellated, surface of crown pitted, buc- 

 cal edge comblike. Thick cover of acumi- 



48 



