756 



Fishery Bulletin 91(4), 1993 



Departamento de Ciencias del Mar, Universidad Arturo 

 Prat, Iquique, Chile; Museo Nacional de Historia Natu- 

 ral, Santiago de Chile; National Museum of New 

 Zealand, Wellington; Otago Museum and Otago Uni- 

 versity, Dunedin, New Zealand; Port Elizabeth Mu- 

 seum, Port Elizabeth, and South African Museum, Cape 

 Town, South Africa; U.S. National Museum of Natural 

 History, Washington D.C.; and the Zoologisch Museum, 

 University of Amsterdam. Unfortunately, only a very 

 few skeletal specimens from the SW Atlantic are avail- 

 able in Argentinian collections 12 and, for reasons of 

 economy, it was impractical to examine these. Metri- 

 cal and non-metrical data sets were collected by the 

 author only, in order to reduce error and observer bias 

 as a source of variation. 



Cranial maturity 



Juvenile specimens were excluded from the sample to 

 reduce variability due to age. As many skull speci- 

 mens did not have sexual maturity data, mature skulls 

 were selected partly by using indirect criteria. To iden- 

 tify one or more cranial indicator(s) of sexual maturity 

 in L. obscurus, selected sutures were inspected in 85 

 skulls (48 females, 37 males) from sexually mature 

 specimens from Peru and in 31 skulls from subadults 

 (large sexually immature specimens) from Peru, SW 

 Africa, and New Zealand. Skulls which had been boiled 

 (moderate heating for maceration excepted) or chemi- 

 cally treated were not included. 



The degree of suture fusion between cranial elements 

 was scored as: 0= no fusion, elements can be moved 

 freely; 1= limited fusion, suture line clearly visible at 

 all points; or 2= advanced fusion with partial oblitera- 

 tion of suture line. Occlusion of the pulp cavity in 

 teeth (OCCL) was scored as: 0= wide open; 1= par- 

 tially occluded; or 2= occlusion complete. The follow- 

 ing seven cranial sutures were defined as follows: 



SYM symphysis mandibularis (between left and 



right ramus) 

 MXDIST distal maxilla-premaxilla suture (rostral 



distal fusion) 

 PTPAL pterygoid-palatine suture 

 PTBAS pterygoid-basioccipital suture 

 LAC lacrimal-maxilla/frontale suture 



ZYG zygomatic-parietal/exoccipital suture 



FROC frontal-supraoccipital suture. 



The selection criteria for adult skulls were, in de- 

 creasing order of priority: 



'-'R. N. P. Goodall. Centro Austral de Investigaciones Cientificas, 

 Ushuaia. Pers. commun. July 1988. 



1 Skulls of sexual mature specimens as evidenced 

 in males by presence of seminal fluid upon gross ex- 

 amination of the epididymis (Sergeant, 1962); in fe- 

 males by presence of at least one corpus luteum or 

 corpus albicans in the ovaries or if lactating or preg- 

 nant, or both; 



2 Advanced fusion (score 2) in the frontal- 

 supraoccipital suture or advanced fusion in at least 

 two of the other six defined cranial sutures. 



The complete fusion of thoracic vertebral epiphyses, 

 indicating physical maturity, was used to confirm cra- 

 nial maturity of some specimens. 



Metrical characters 



Thirty-three standard metrical variables measured 

 with vernier calipers to the nearest mm and four 

 tooth counts, adapted from Perrin (1975a) and 

 Schnell et al. (1982 K ) (see Table 3) were recorded for 

 381 dusky dolphin skulls. Standardized skewness and 

 kurtosis were computed for all sample variables; if 

 significant departure from normality was suspected, 

 a Kolmogorov-Smirnov (K/S) one-sample test of the 

 fit was applied. A ratio of variances test with 95% 

 confidence limits was used to verify homogeneity of 

 variance in all sample pairs. STATGRAPHICS ver- 

 sion 4.2 programme (STSC Inc., 1989) was used on 

 an AT286 personal computer for most statistical ap- 

 plications. The a=0.05 level was the criterion chosen 

 for significance. All ^-tests were two-sided. Owing to 

 the large series of £-tests, a few significant results 

 (false positives) should be expected to occur by chance 

 alone. 



Sexual dimorphism in cranial measurements and 

 tooth counts was studied by means of <-tests and analy- 

 sis of covariance (ANCOVA, with condylobasal length 

 as covariate) in skulls of sexually mature females 

 (A r =59) and males (A/=50) from central Peru. To permit 

 pooling of sexes and utilization of cranially mature 

 skulls of unknown sex, the few sexually dimorphic char- 

 acters were not further considered in the analysis of 

 geographic variation. 



Seasonal variation was examined for in all variables 

 by using a multifactor analysis of covariance 

 (MANCOVA) with three-monthly seasons (January- 

 March, April-June, etc.) and sex as classification 

 variables. 



Geographic variation was analysed for skulls from 

 Peru (59 females, 50 males and 80 of unknown sex), 

 New Zealand (JV=47>, SW Africa (iV=40) and Chile 

 (iV=22). Both a one-way analysis of variance (ANOVA) 

 and an ANCOVA with skull CBL as covariate and with 

 geographic region as classification variable were uti- 

 lized to assess among-group variation in skull size and 

 shape. Next, most morphologically similar group pairs. 



