SU 61662, Blanca Bay, lat. 24°46'30"N, long. 

 112°15'W), and occasionally to Cabo San Lucas at 

 the tip of Baja California (lat. 22°20'N, long. 

 112°27'W). There is one specimen (Ricker 1959b; 

 UBC 57-149) from Socorro Island in the 

 Revillagigedos Islands 250 miles south of Cabo 

 San Lucas. It is not common north of Point Con- 

 ception, Calif., but there are several scattered 

 records: off the Farallon Islands and off Eureka, 

 Calif. (Radovich 1961); Puget Sound (Starks 1910, 

 Patten et al. 1965); east coast of Vancouver Island, 

 British Columbia (Manzer 1965; UBC 64-394); and 

 coastal Alaska (Quast 1964) in Clarence Strait 

 northwest of Ketchikan (USNM 213494) and off 

 the Copper River at lat. 60°16'N, long. 145°32'W 

 (USNM 213495). 



Geographic Variation.- Although the temperate 

 northeast (c. lineolata) and southeast (c. 

 chiliensis) Pacific populations of Sarda chiliensis 

 are completely separated from each other by the 

 tropical eastern Pacific population of S. orientalis 

 (Figure 70), there are few differences between the 

 two populations; anatomically they are virtually 

 identical. The caecal mass extends posteriorly for 

 more than half the length of the body cavity in 

 ventral view in the northeast population; less than 

 half this distance in the southeast population. 



Walford (1936) used the posterior extent of the 

 corselet to distinguish the northeast Sarda 

 lineolata (pectoral fin extends scarcely beyond 

 corselet) from the southeast S. chiliensis (pectoral 

 fin extends an eye diameter further). Hildebrand 

 (1946) considered this character entirely unreli- 

 able in separating the two populations. Kuo (1970) 

 measured the extent of the corselet in 380 

 specimens from southern California and 49 from 

 Callao, Peru. He found that the corselet does ex- 

 tend further posteriorly in the southern popula- 

 tion but that the interpopulational differences 

 were not significant at the 5% level. He concluded 

 that the character was not reliable even for sub- 

 specific separation of the two populations. 



Morphometrically, there are some differences in 

 the ranges and means between the two popula- 

 tions (Table 19), but these are smaller than those 

 between some populations of S. sarda and S. 

 orientalis. Kuo (1970) found, through covariance 

 analyses and scatter diagrams, significant 

 differences at the 1% level between the two 

 populations in such morphometric characters as 

 head length, predorsal length, prepelvic length. 



and the distance from the origin of the second 

 dorsal and anal fins to the caudal base. 



Among meristic characters, the total number of 

 vertebrae is the best distinguishing character 

 between the southeast and northeast populations 

 of Sarda chiliensis. The northeast population has 

 more vertebrae (Table 9) with means of 44.9 for 

 our data, 44.8 for Kuo (1970), and 45 ± 0.47 for a 

 sample of 19 reported by Vildoso (1963b). Our data 

 for 23 southeast specimens {x 44.2) agrees with 

 that reported by Vildoso (1963b) for two samples, 

 44.15 ± 0.40 for 38 specimens, and 44.21 ± 0.40 for 

 a sample of 100. Kuo (1970), however, found even 

 fewer: a range of 42-46 with a mode of 43 and a 

 mean of 43.4 compared to our range of 43-46, mode 

 44, mean 44.2. The reasons for the discrepancies 

 between our data and that of Vildoso on the one 

 hand, and that of Kuo on the other are not known. 

 We have reexamined seven of Kuo's southeastern 

 specimens and find totals of 43 (1 specimen), 44 (5), 

 and 45 (1). 



There are also differences in the highly variable 

 number of teeth in the upper and lower jaws (Ta- 

 bles 5, 6) with the northeast Pacific population 

 having slightly more: upper jaw 20-30, x 23.7 vs. 

 18-28, X 23.0; lower jaw 15-25, x 19.4 vs. 14-23, x 18.9. 



The available data does not convince us that the 

 northeast and southeast Pacific populations are 

 subspecies. However, as the populations are genet- 

 ically isolated from each other and there are some 

 significant differences, there is practical value in 

 using the available subspecific names, and there is 

 ample historical precedent for the name lineolata 

 for the northeast population. Therefore, we have 

 emphasized comparisons between the two 

 geographic populations by labelling the popula- 

 tions northeast and southeast Pacific in the tables 

 but have retained the subspecies designations in 

 the formal synonomies. 



Sarda orientalis (Temminck and Schlegel) 

 Indo-West Pacific 



Pelamys orientalis Temminck and S/shlegel 

 1844:99 (original description, Japan), pi. 52. 

 Richardson 1846:268 (Sea of Japan). Gunther 

 1860:368 (description). Tirant 1885:46 (Cam- 

 bodia). 



Pelamys chilensis (not of Cuvier 1831). Day 

 1878:253-254 (description; Bombay), pi. 56, fig. 

 1. 



Sarda chilensis var. orientalis. Steindachner and 

 Doderlein 1884:179 (description; Japan). 



603 



