ECOLOGY: PORT JACKSON SHARKS 541 



The deviation from Hardy-Weinberg equilibrium of the Newcastle popu- 

 lation, and the poor agreement of the Western Australian population, could 

 be due to the smaller sample size or to sampling error (Wright 1964), but 

 other factors cannot be ruled out. The Western Australian population con- 

 sisted largely of juvenile animals, while about 30% of the Newcastle sample 

 were juveniles. Hashimoto and Matsuura (1960) and Vanstone et al. (1964) 

 have described developmental changes in the concentration and number 

 of types of haemoglobins in Oncorhynchus sp. Manwell (1963) showed 

 that foetal and adult shark haemoglobin differed in primary structure. 

 It is conceivable, then, that phenotypic patterns in juvenile and adult 

 H. portusjacksoni are sufficiently different to admit the possibility of 

 miscounting. This could explain the discrepancies in the two populations. 



A further possibility, in the case of the Western Australian population, is 

 that the animals sampled were not members of a single breeding population, 

 but a random sample drawn from several subpopulations. Under these con- 

 ditions, and invoking Wahlund's Principle (Li 1955), there would be an 

 apparent shortage of heterozygotes and an excess of homozygotes in the 

 sample. These discrepancies are apparent in Table 3. As the sharks in the 

 Western Australian population were captured several kilometers from shore, 

 it is entirely possible that they were from the general population of the 

 areas, and not from a specific breeding site. 



Unfortunately, the distribution of phenotypes in the Newcastle popula- 

 tion cannot be explained in this manner. Possibly the genetic hypothesis, 

 obviously adequate for the other populations, is untenable in northern 

 waters. Newcastle is quite close to the northern limits of the range of 

 H. portusjacksoni on the east coast and, although the identity and func- 

 tion of the polymorphic protein is unknown, selection against the homo- 

 zygous condition may be occurring. 



Returning to the distribution of gene frequencies (Table 3), it is obvious 

 that there is considerable variation among populations. The reasons for the 

 differences are not so obvious. Lewontin (1974) stated that, "spatial varia- 

 tion in allelic frequencies, unlike temporal changes, cannot provide any in- 

 formation on the intensity of selection, because, in the absence of histori- 

 cal evidence, it must be assumed that the spatial pattern is an equilibrium 

 state. ..." Selection can of course be detected if there is a geographical 

 cline in the gene frequencies of several spatially separated populations. 

 Thus Sick (1961, 1965) and O'Gower and Nicol (1968) were able to dem- 

 onstrate a latitudinal cline in the frequencies of the genes coding for 

 haemoglobin in Gaddus sp. and Anadara sp. respectively. In the latter 

 case the environmental variable responsible for the alteration in gene fre- 

 quencies has been identified (Nicol, Collette and O'Gower, unpublished 

 data). Detecting temporal changes in gene frequencies requires data from at 

 least two generations. Considering the habitat, behaviour, and generation 

 time (approximately 10 to 12 years) of H. portusjacksoni, such data are 

 virtually impossible to obtain. 



One must therefore use the data available. These data (Tables 3 and 4) 

 indicate no regular pattern, but rather a mosaic of gene frequencies. 

 Kimura and Weiss (1964) have shown mathematically that in the absence 



