768 



Fishery Bulletin 91(4), 1993 



SW Africa similarity pairs. The New Zealand form is 

 set apart from all others by minimal tooth size and is 

 further distinguished from the SW African form by the 

 significantly higher number of teeth. 



A commonly used criterion of subspecific level dif- 

 ference between populations is whether or not 90% of 

 the specimens can be unequivocally assigned to the 

 proper group (Mayr, 1970). Perrin (1975a, 1990 1. for 

 instance, described three subspecies of spinner dol- 

 phin, Stenella longirostris, based on this principle. 

 Discriminant analysis allows correct identification of 



a group (96.5%, 91.7%, and 90.9%) of L. obscurus 

 skulls from respectively the Southeast Pacific, New 

 Zealand, and SW Africa, which warrants allocation of 

 subspecific status. It seems premature, however, to 

 formally describe subspecies and assign trinomial 

 names, because no skulls from Argentina have been 

 studied and the Chilean sample was too small to per- 

 mit definitive conclusions on the degree of heteroge- 

 neity in the Southeast Pacific group. In particular, no 

 skulls have been examined from dusky dolphins oc- 

 curring off southern Chile (south of 46°S), which are 

 thought to belong to the Argentinian stock (Van 

 Waerebeek, 1992a). Considering the persisting mor- 

 tality levels in gillnet and harpoon fisheries off Peru 

 and northern Chile and in purse-seine operations off 

 Argentina (Van Waerebeek and Guerra, 1986 s ; Guerra 

 et al. 1987; Crespo and Corcuera, 1990"; Van 

 Waerebeek and Reyes, 1993; Van Waerebeek, unpubl. 

 data), the question of separate subspecies or popula- 

 tion status is of major importance with respect to 

 conservation. 



The MANOVA results, suggesting some seasonal 

 morphological variation in L. obscurus caught off cen- 

 tral Peru, were not confirmed by other tests. However, 

 the picture may be blurred by multi-year pooling or 

 other sampling bias. High-resolution molecular tech- 

 niques, such as DNA-fingerprinting, should preferably 

 be used to explore this issue. 



Although results were complementary, the resolv- 

 ing power of non-metrical cranial characters (not in- 

 cluding tooth counts) was inferior to that of measure- 

 ments (i.e., fewer characters differed); therefore 

 non-metrical characters are not recommended as sole 

 method for subspecific discrimination. Based on stud- 

 ies of the common dolphin, Delphinus delphis, Perrin 

 et al. (1989) independently concluded that non-metri- 

 cal characters be used only in combination with met- 

 rical characters and that special attention be paid to 

 controlling conditions to maximize comparability of 

 data sets. 



Although little information is available on parasite 

 prevalences in dusky dolphins other than these from 

 Peru, preliminary indications of differential infesta- 

 tions add weight to the morphological findings. 

 Crassicaudid nematodes were practically absent in 

 the cranial sinuses of many hundreds of Peruvian 

 dusky dolphins but were retrieved from the cranial 

 sinuses of one out of two specimens from SW Africa. 



"Crespo, E. A., and J. F. Corcuera. 1990. Interactions between ma- 

 rine mammals and fisheries in some fishing areas of the coast of 

 Argentina and Uruguay. Doc. SC/090/G2 presented to the IWC, La 

 Jolla, CA, 20-21 October 1990. 



