We examine differences in standard body length and skull morphology of male (n = 65) and female (n = 18) South African (Cape) fur seals, Arctocephalus pusillus pusillus, from the coast of southern Africa with the aim to develop an objective method for determining the sex of fur seal skulls. Males were found to be significantly larger than females in standard body length, with K-means cluster analysis successfully identifying 2 relatively homogeneous groups. Principal component analysis (covariance matrix) showed that the underlying data structure for male and female skull variables was different, and that most of this variation was expressed in overall skull size rather than shape. Males were significantly larger than females in 30 of the 31 skull variables. Breadth of brain case was significantly different for the genders. Relative to condylobasal length, males were significantly larger than females in 13 of the 31 skull variables used in the present study. These were gnathion to posterior end of nasals, breadth at preorbital processes, least interorbital constriction, breadth at supraorbital processes, greatest bicanine breadth, breadth of palate at postcanine 1 and 3, calvarial breadth, mastoid breadth, gnathion to anterior of foramen infraorbital, gnathion to posterior border of preorbital process, height of skull at base of mastoid and height of mandible at meatus. In males, these variables were associated with the acquisition and defense of territory (e.g., large head size and mass; increased structural strength of the skull; increased bite capacity). Two skull ratio parameters, breadth of braincase/condylobasal length and length of upper postcanine row/condylobasal length were significantly higher in females compared to males. Based solely on the skull data, mature males can be reliably distinguished from immature males and females using both (a) Classification and Regression Tree (CART) and (b) Hierarchical Cluster Analysis. Both approaches had difficulty in reliably distinguishing immature males from females. The Classification and Regression Tree method was the more successful in correctly distinguishing immature males from females.
|Number of pages||22|
|Journal||Proceedings of the Linnean Society of New South Wales|
|Publication status||Published - 2010|