Van Waerebeek: Variation in skull morphology of Lagenorhynchus obscurus 



757 



namely Peru/Chile and SW Africa/New Zealand were 

 compared by means of /-tests to identify discriminat- 

 ing variables. 



Discriminant analyses (DISCRIM) were performed 

 on samples limited to specimens with a complete set of 

 measurements. This sample included 135 skulls from 

 Peru, 24 from New Zealand, 22 from SW Africa and 7 

 from Chile. A first DISCRIM with 21 characters maxi- 

 mized discrimination between geographic groups. Sub- 

 sequently, a series of discriminant analyses was ex- 

 ecuted by considering various combinations of a small 

 number of highly variable cranial characters so that 

 an optimum segregation between groups was obtained. 

 These discriminant functions can be used tentatively 

 to assign undocumented museum skulls to respective 

 populations. It must be cautioned however that no 

 specimens from Argentina have been examined and 

 the discriminant locus of that population remains 

 unknown. 



Non-metrical characters 



Initially, 35 non-metrical cranial characters (NMC), fol- 

 lowing Perrin et al. (1982 111 ), were determined, but se- 

 rious problems with score repeatability over time re- 

 duced the number of usable NMC to 27 (Table 1 and 

 Fig. 1). Similar problems have been reported for the 

 common dolphin by Perrin et al. ( 1988 11 , 1989). 



Analysis of sexual dimorphism was limited to skulls 

 of sexually mature specimens from central Peru 

 (A r =109), eliminating influences of growth and geogra- 

 phy. Sexually monomorphic characters (the full set, 

 see Results) were then tested for dependence on cra- 

 nial maturity status in 173 mature and 48 immature 

 skulls. Geographic variation in non-metrical traits 

 among the four geographic areas was studied after 

 eliminating variables influenced by sex and maturity. 

 All significance testing of variation in NMC was done 

 by means of chi-square contingency analyses. Tooth 

 counts were compared by /-tests. 



Table 1 



Non-metrical cranial characters utilized in present study (from Perrin et al., 1982"); see Figure 1. 



Character 1L/1R (left/right). 



Character 2L/2R. 

 Character 3. 

 Character 5L/5R. 

 Character 6L/6R. 



Character 9L/9R. 



Character 10. 



Character 12L/12R. 

 Character 13. 



Character 16. 

 Character 17. 

 Character 18L/18R. 



Character 19. 



Character 23. 

 Character 24. 

 Character 25. 

 Character 30. 

 Character 33L/33R. 

 Character 34L/34R. 



Position of the foramen (A in Figure 1) in the premaxillary at the base of the prenarial triangle, 

 relative to the level of the anteriormost projection of the antorbital processes. Codes: l=at; 2=ahead; 

 3=behind the level. 



Position of the anteriormost large foramina (B) in the maxillary relative to the level of the anteriormost 

 projection of the antorbital processes. Codes: l=at; 2=ahead; 3=behind the level. 



Asymmetry of position of the two anteriormost large maxillary foramina 1B1. Codes: l=symmetrical; 

 2=left foramen more anteriorly placed; 3=right foramen more anterior. 



Number of small foramina (C) in the maxillary anterior to the anteriormost of the large foramina. 

 Codes: l=one foramen; 2=two foramina;3 etc.; 6=no foramina. 



Number of foramina ID) in the maxillary behind a line at the level of the anterior edge of the external 

 nares and perpendicular to the long axis of the skull. Codes: 1=1 foramen; 2=2 foramina; 3 etc.; 6= no 

 foramina. 



Contact between maxillary and occipital, at point where occipital crest intersects margin of temporal 

 fossa (E). Codes: l=contact lor space of <lmml; 2=no contact. 



Contact between premaxillary and nasal, on right side (Fl. Codes: l=contact (or space of < lmm); 2=no 

 contact . 



Presence of an antorbital notch (G). Codes: l=notch present; 2=not present. 



Composition of the antorbital process, in lateral view, on left side (H). Codes: l=formed by lacrimal; 

 2=formed by lacrimal and frontal; 3=formed by lacrimal and maxillary. 

 Accessory foramen above foramen magnum (I). Codes: l=present; 2=absent. 

 Clear notch in upper margin of foramen magnum (J). Codes: l=present; 2=absent. 

 Fenestrations in occipitals, near foramen magnum and in exoccipital region (K). Codes: l=one fenes- 

 tration; 2=two; 3 etc.; 6=no fenestrations. 



Fenestrations in region of occipital swellings (L). Codes: l=one fenestration; 2=two etc.; 6= no fenes- 

 tration. 



Posterior contact between pterygoid hamuli (Mi. Codes: l=open igap >lmml; 2=closed. 

 Anterior contact between hamuli (Nl. Codes: l=open (gap>lmml; 2=closed. 

 Spine of palatines visible in anterior gap between hamuli (Ol. Codes: l=visible; 2= not visible. 

 Mesial fenestration in basisphenoid (Rl. Codes: l=foramen present; 2=absent. 

 Number of mental foramina. Codes: 1=1; 2=2; etc. 



Arrangement of mental foramina. Codes: l=foramina lie along a simple curve; 2=foramina do not lie 

 along a simple curve. 



