Thus the combined forces of founder effect, genetic drift, migration and mutation have undoubtedly shaped the genetic profiles seen in contemporary Austronesian speakers. This conclusion based on HLA data is supported by studies of the distributions of the nine base-pair (bp) mitochondrial DNA deletion and of the 3.7 kb deletion in the α globin genes. The nine bp mitochondrial DNA deletion is another example of the role of chance in determining genetic profiles; since mtDNA is maternally-inherited, the number of copies in a colonizing group is equivalent to the number of females in that group. Thus we see the gradual loss of the wild-type or common non-deleted mtDNA as the early Austronesian-speakers moved through the Pacific. The deleted form has a frequency of 16 per cent in the Moluccas, 10-40 per cent in coastal Melanesia, 77 per cent in Tonga, 87 per cent in the Cook Islands, but reaches near fixation (frequency of 100 per cent) in Samoans and the New Zealand Maori (Hertzberg et al. 1989).

The common form of mtDNA was not the only genetic material lost to the pre-Polynesians as they moved eastwards through the Pacific. There are other examples of clines in gene frequencies from west to east, with final and irretrievable loss of the allele. At the HLA-B locus, for instance, HLA-B27 is not represented in eastern Polynesia, although this is a common antigen in island Melanesia (11 per cent in New Caledonia) and occurs in Mauke Island (6 per cent) (Serjeantson 1989). Similarly, HLA-B13 is found in western but not eastern Polynesia.

Another example of chance effects is the 3.7 kb deletion in the α globin genes (α^-3.7), which has a frequency of 15 per cent in Maoris. This deletion is clearly carried on a chromosome of Melanesian origin, because there are particular features in the DNA flanking the deletion (restriction enzyme sites) that are otherwise seen only in Melanesians. This has given rise to claims of substantial Melanesian genetic input into the contemporary Polynesian genome (Hill et al. 1989). However, the gene can be seen to increase in a cline from western to eastern Polynesia, increasing from 1 per cent in Tongans to 12 per cent in Cook Islanders to 15 per cent in Maoris (Hill et al. 1989). Thus the α^-3.7 marker does not provide evidence that Polynesians evolved out of Melanesia; rather, it suggests that pre-Polynesians recruited at least one Melanesian into the mating pool.

It is clear that the recent PCR-based technological revolution in molecular genetics (Scharf et al. 1986), permitting rapid screening for the newly-sequenced alleles, will result in an explosion in knowledge of the genetic profiles of contemporary populations. PCR amplification of DNA from aged material has attracted much attention, but application of the PCR technique in reconstruction of genetic prehistory by analysis of contemporary populations is potentially very informative indeed.

How can the scholar of Austronesian prehistory best make use of the emerging PCR-based data? First, there will be an important role for the prehistorian who is prepared to examine the genetic tabulations and reinterpret them in language understood by his colleagues, wresting the data from the geneticists who with great abandon muddle geographic, anthropological and linguistic terms and muddle the reader (Houghton 1991)! Second, for the data to be relevant to archaeology they will need to be interpreted with due attention to the main thrust of this paper — that Homo sapiens is an evolving species. We conclude, from this small sample of the human genome, that founder effects, bottlenecks and mutations have resulted in a unique genetic profile in contemporary Polynesians, so that their origins are forever blurred. The Polynesian genetic repertoire at HLA-DR,DQ can largely be drawn from the East Asian gene pool, but has been irrevocably altered by evolutionary forces to generate the unique genetic repertoire that is distinctively Polynesian.