The linguistic dichotomy of populations in Melanesia into Austronesian and non-Austronesian-speaking peoples has tempted some scholars to over-simplify the genetic characteristics of these populations as malaria-resistant and malaria-susceptible, on the basis of gamma-globulin genetics (Clark and Kelly 1993). Kelly (1992) argues that if malaria in Melanesia did indeed have an antiquity greater than about 3600 BP, the malarious coastal lowlands would have been open for settlement by Austronesian-speaking immigrants because the non-Austronesian-speakers did not have the genetic capability to live permanently in malarious environments.
The data do not support this scenario. First, it is not clear how malaria could have been sustained without some minimal human population density in the low-lying regions. Second, continuous occupation of a malarious region induces some immunity to malaria among the inhabitants. The sporadic forays of non-Austronesian-speakers into the lowlands, as envisaged by Kelly, are much more likely to be deleterious than continuous occupation. This is evidenced by the finding that in New Guinea, hyperreactive malarious splenomegaly is confined to the Watut people, resident at an altitude of about 1000 m where malarial transmission is intermittent.
The third factor to be taken into consideration is the genetic profile of the contemporary population of coastal lowland Melanesia. There are several genetic markers, widespread in non-Austronesian-speaking lowlanders, which are thought to provide some protection against malaria. A description of these genes together with their population distribution is given by Serjeantson et al. (1992) and includes the thalassaemias, glucose-6-phosphate dehydrogenase deficiency, hereditary ovalocytosis and the Gerbich negative blood group. The population distributions of these genes provide some insights into genetic prehistory in Melanesia.
For instance, the gene causing hereditary ovalocytosis, a slight distortion of the red blood cell shape, has a frequency of about 10 per cent throughout lowland Papua New Guinea. The distortion in the red cell membrane arises from a 27 base pair deletion in the DNA encoding Band 3 protein, first shown in a Southeast Asian donor and confirmed as the same gene as that in some Melanesians. The Band 3 defect distorts a number of receptors on the red cell surface and may inhibit malarial invasion. Further evidence that hereditary ovalocytosis provides some protection from high density, lethal parasitemias is provided by epidemiological data and by the apparent absence from the population of people homozygous for the mutant Band 3 gene. That is, despite the apparent lethality of the mutant Band 3 gene when it is inherited from both parents (i.e. a double dose), a high population gene frequency is maintained, presumably through a selective advantage in malarious areas for those with a single dose.
Was this gene introduced to lowland coastal Melanesia, together with malaria, by early Austronesian speakers? The gene’s antecedents lie in the Aboriginal populations of Southeast Asia, but it is otherwise rare in contemporary populations of this region. It is found in appreciable frequencies in the Orang Asli populations of Malaysia, has been reported in the Austronesian-speaking Land Dayaks and Iban of Borneo, is prevalent in the hinterland of north and south New Guinea, is absent from New Guinea Highlanders, and is rare in populations east of the New Guinea mainland. This gene thus indicates the likely arrival in Melanesia of a non-Austronesian-speaking wave of immigrants before the early Austronesian-speakers.
The non-Austronesian-speaking Melanesians in malarious areas of New Guinea have other characteristic genetic mutations, such as the Gerbich-negative blood group that is a consequence of a large deletion in the glycophorin C gene. This mutation may have arisen in one of the populations of the Torricelli Mountains, where it is particularly prevalent, as the gene has a more limited distribution than hereditary ovalocytosis in New Guinea and is not found in the Gulf and Western Provinces. We have detected, using DNA analyses, the glycophorin C gene deletion in low frequency in Micronesia but not in Polynesia.
Among other genetic markers thought to be related to malaria, the most frequent in non-Austronesian-speaking peoples in lowland New Guinea is a 3.7 kb deletion in the alpha-globin gene, a mutation that has occurred on an unmistakably Melanesian-area chromosome as attested by the flanking DNA markers. This marker was carried into Polynesia, where it has sporadic occurrence, although it has attained appreciable frequency in New Zealand Maoris, presumably through genetic drift.