Volume 100 1991 > Volume 100, No. 2 > The early human biology of the Pacific: some considerations, by Philip Houghton, p 167-196
Back to Search Results                                                                                              Previous | Next   

- 167

This paper examines the thesis that the distinctive body form of the people of the wider Pacific is an adaptation to their unique environment. By the term “wider Pacific” I mean the region where original settlement and subsequent existence haverequired mastery of a maritime environment, and where land masses are relatively inconsequential in size. Effectively, this is the region defined by Pawley and Green (1973) as Remote Oceania, and in standard geographic terms includes all of Polynesia and Micronesia and the more eastern part of Island Melanesia. The thesis itself assumes the validity in this environment of a statement by Damon (1977:221): “Climate . . . does indeed seem to be the major regulatory factor for human body size and proportion”. In this statement are subsumed the classical biological rules of Bergmann (1848) and Allen (1877).

Bergmann's rule says that, for closely related mammals or birds, those living in cold regions tend to have greater body mass than those living in warm regions. Allen's rule says that animals living in cold climates tend to have shorter extremities than those living in hot climates. Some animals with a wide geographic distribution, such as rabbits and foxes, illustrate these rules well. In human terms, larger and more muscular people should be found in colder climates, where large muscle mass can produce more body heat, and the relatively smaller surface area of a larger body lessens heat loss. Smaller-bodied, or at least more linear, people should be found in hotter climates, where endogenous heat production is less necessary, and a relatively larger surface area allows for more efficient cooling. The geometric basis is that the volume and mass increase as the cube of the linear dimensions, whereas surface area only increases as the square of the linear dimensions.


I shall discuss first the body form of the people of Polynesia, their environment, and some relevant physiology. Micronesia is then more briefly considered in the light of the Polynesian situation. Comment on the human morphology in eastern Melanesia is considered in later sections of the paper.

- 168

A picture of Polynesian physique in prehistory can be obtained from at least three sources: records of earliest contact, anthropometric studies on relatively un-Westernised groups, and the skeletal evidence. In the historical record a picture of a large-bodied, strongly muscled people distributed over a wide expanse of the Pacific is repeated almost monotonously. Of the Samoans: “The men were a remarkably fine-looking set of people, and among them were several above six feet high, with Herculean proportions” (Erskine 1967:41). Of the Maori: “The Zealanders are generally tall and well made. Men of six feet high are by no means uncommon” (du Clesmeur 1914:471). Cook (1968:1:123) noted of the Tahitians: “the men in general are tall, strong-limbed and well shaped, one of the tallest we saw measured Six feet 3 inches and a half”. Of the Marquesans: “The inhabitants of these islands taken collectively are without exception the finest race of people in the Sea; for a fine shape and regular features they, perhaps, surpass all other Natives . . . The men are in general tall, that is about 5 feet 10 inches or six feet” (Cook 1968:2:374). The Lau is landers, of known Polynesian affinity, were observed to be “ a stalwart, well-proportioned and heavy-boned stock, slightly taller than the natives of Fiji proper. The men are muscular, with particularly well-developed legs . . . [they] do not grow fat as they age” (Thompson 1940:7). The muscularity of the legs often receives particular mention; in New Zealand, “The inhabitants are of a fine stature but their legs are so thick that they appear to be swollen”(Monneron 1914:279).

Anthropometric data are available from studies on relatively un-Westernised groups where traditional modes of living and diet have still dominated, the obesity of Westernised groups has not been in evidence, and where little genetic admixture with outsiders has taken place. Earlier studies (Buck 1923, Dunn and Tozer 1928) are confined to simple parameters such as stature, weight and body proportions but allow derivation of others such as body surface area. More recent studies (Prior et al. 1981, Finau et al. 1983) provide skinfold thickness data which allow estimation of body fat content, body density and lean mass. Data from a study of the Lau islanders of eastern Fiji (Lourie 1972) are included. The focus here is on male data, purely because more are available.

The anthropometric data in Table 1 confirm the historical record. There is no reason to suppose that the basic Polynesian phenotype has changed significantly since European contact, as there is an abundant confirmatory skeletal record. The Polynesians clearly are at the upper end of the range for stature for Homo sapiens (Wells 1969). The sitting height ratios indicate relatively short lower limbs. The upper arm and calf circumferences, a reasonable guide to muscularity, are large in comparison with other groups (Stini 1979) and lean body mass is large. As the skinfold thicknesses and body fat content are less than those generally obtained on recent Western groups (Lourie 1972), the substantial weight may be ascribed - 169 to the obvious Polynesian muscularity. Given the wide geographic spread, the consistency of the basic data is striking.

Table 1
Group Stature Weight S.H. ratic Limb Circumference Upper Arm Calf C.I. Skin fold Triceps subscap %Body Fat Surface Area DuBois Lean Mass
  (cm) (kg)   (cm)   (mm)   (sq metres) (kg)
Lau (74) 170.8 76.1 51.7 30.9 38.9 81.2 7.8 13.3 13.0 1.88 66.2
Tonga (198) 171.3 75.2       8.6   12.2 1.87 66.1
Hawaii (85) 171.3 77.3 52.6     83.4       1.90  
Maori (426) 170.6 75.2 53.8 29.2 37.9 77.7       1.85  
Pukapuka (39) 168.8 70.6           10.4 12.1 1.82 61.4
Tokelau (60) 170.8 76.1           11.8 13.4 1.87 66.4
mean 170.6 75.08 52.7 30.05 38.4 80.77 8.2 11.8 12.7 1.86 65.0

Table 1: Male Polynesian morphological data. Sources: Lau, Lourie 1972; Tonga, Finau et al. 1983; Hawaii, Dunn and Tozer, 1928; Maori, Buck 1923; Pukapuka and Tokelau, Prior et al. 1981. C.I. = cranial index. S.H. = sitting height. Body fat and lean mass derived from the equations of Damon and Goldman (1964). Body surface area derived from the equations of Dubois and Dubois (1915).

With the exception of New Zealand and one or two minutiae such as Rapa, Polynesia lies firmly within the tropics. Basic climatic data for the April-September period, when the trade winds are most constant, and outside the cyclone season, are given in Figures 1 and 2. More detailed climatic data are available, but, as we are concerned with events of past millenia, a broad picture is appropriate to the discussion. Water isotherms for July are above 21 degrees Celsius for the entire geographic range and above 27 degrees Celsius close to the equator (Fig.1). Surface air isotherms closely follow the water isotherms. Average surface wind speed for the June-August period is above 15 kilometres per hour over most of the region (Fig.2). Average annual cloud cover exceeds 40 per cent for most of the region. Humidity generally ranges between 70 and 90 per cent. Rainfall nearly everywhere exceeds 1250 mm annually, with a fairly even distribution through the year. The slightly warmer temperature and less consistent wind of the November-March period would not affect the conclusions reached in this paper.

The fact that emerges is that, by the rules of Bergmann and Allen, these tall and impressively muscled people seem quite out of place in their environment. One expression of these rules is the stature/weight ratio, which ranges from under 2.6 for individuals from cold climate regions such as Turkestan, Finland, Iceland and England, through to values above 3.2 for Vietnam, Burma and India (Molnar 1983). The Polynesian values of between 2.21 and 2.34 derived from the data in Table 1 are below those of any of the recorded cold climate groups. A related

- 170
Figure 1.
Pacific surface water isotherms for July.

ratio is that for weight/surface area, derived by Schreider (1950:286) for several groups. He observed there to be “obviously a decline in value of the ratio as we proceed from temperate to tropical regions”, from 38 for a French group to 32.5 for Andamese. The Polynesian value of about 40.5, again derived from the weight and stature data in Table 1, is higher than that of any of Schreider's temperature climate peoples.

Allen's rule predicts that limbs in Homo sapiens will be relatively shorter and stouter in the colder parts of the species range. The short-legged Eskimo, or Lapps (sitting height ratios above 52.5) may be contrasted with some long-limbed African Negro groups or Australian Aborigines, in whom the ratio is less than 48. The Polynesian sitting height ratios greater than 51 indicate that the considerable stature is derived from a long axial length and relatively short lower limbs.

A relationship between the form of the head and climate has been postulated. Beals et al. (1983) suggest that brachycephaly, around head with relatively small surface area, is a cold climate adaptation; dolicocephaly, a long head with

- 171
Figure 2.
Pacific average wind speeds in kilometres per hour for the June-August period.

relatively large surface area, is a warm climate feature. While there is no doubt that the head can be an important source of heat loss (Froese and Burton 1957), a more refined assessment of surface area than just the cranial index seems desirable. Yet it is worth nothing that the peoples of central Polynesia are brachycephalic — that is, the purported cold-climate form.

Into this scenario of body form and climate, then, the muscular, large-bodied, short-limbed, round-headed Polynesians seem to fit uneasily. The resolution of the paradox lies in an appreciation of the real oceanic environment, which is not immediately evident from the bland climatic data.

Firth (1963:30) described on Tikopia the period of the trade winds that blow steadily in the north-east to south-east quadrant from April till September. At this time “the sky is frequently overcast for several days at a time and the weather is often wet and even chilly”. This is a land-based comment. But it is particularly at sea in small craft, even a modern yacht, that the tropics can prove cold, the more so if it is wet, dark, or the sky clouded. Near Samoa, 15 degrees south latitude: “Surprising how chilly it is without the sun. Good to have on oilskins” (Roth - 172 1972:100). Whatever the merits of the voyaging canoes of Oceania, such as the Fijian dura or the great Polynesian double-canoes, there was no possibility of their being dry (Lewis 1972). A perpetual wetness would have been inevitable, particularly when working to windward, and of this there is an abundance of anecdotal evidence.

Finney (1979:106) sailed a reconstruction of a Polynesian double-canoe from Hawaii to Tahiti. “May 4 . . . It is cold and wet. We might be in the tropics, but the 15-knot trade wind blowing across the deck plus the nearly constant spray from the head seas makes it downright cold, especially at night when there is no warming sun. I still feel chilly, even when wearing a jogging suit under my foul weather gear”. And “Even when wearing foul weather gear as pyjamas we could not shut out the drips, dribbles and spurts of water that found their way into ears, down necks and into pants” (Finney 1979:188). On the Kon-Tiki raft Heyerdahl (1952:156) “. . . stumbled about the deck bent double, naked and frozen”. And the journal of Bligh (1961:152-63) during his small-boat voyage to the west following the Bounty mutiny, in the latitudes of Tonga, Fiji and Vanuatu, repeats again and again “miserably wet and cold”. May 5, 1789: “among the hardships we were to undergo, that of being constantly wet was not the least: the night was very cold, and at day light our limbs were so benumbed, that we could scarce find the use of them”. May 7: “Heavy rain came on at four o'clock . . . being extremely wet and having no dry things to shift or cover us, we experienced cold and shiverings scarce to be conceived”. On other days; “Our situation was miserable, always wet, and suffering extreme cold in the night . . . We suffered extreme cold and everyone dreaded the approach of night . . . At noon it was almost calm, no sun to be seen, and some of us shivering with cold”.

The high humidity of the oceanic environment exacerbates any discomfort. Siple (1968:410) comments: “From a physiological standpoint, even at 100 percent relative humidity, the vapour pressure gradient between warm moist skin and the atmosphere is so steep that evaporation may be equivalent to the rate found in the driest desert regions. Any moisture accumulation in clothing from exterior or interior sources will cause marked cooling. At high relative humidities, the body seems to react more intensely to chill in this type of weather than when the temperatures are much lower and still drier”.

Physiological considerations

These factors of persistent wetness and wind totally alter the conditions. The oceanic environment is potentially and frequently very cold, whether for voyaging, for the more mundane but routine activities of reef and coastal fishing in a small-island existence, or at times even for life ashore. To quantify the possible advantages of a large and muscular body in these wet-cold conditions I shall compare potential body heat loss and heat production for Polynesians with - 173 that for an un-Westernised group from coastal New Guinea (Harvey 1974). The basic comparative data are set out in Table 2. Essentially in this comparison a typical Polynesian male is being compared with a typical male coastal dweller in a semicontinental tropical climate in the western part of the Pacific. This heat-balance analysis draws extensively on an earlier study (Houghton 1990)

Table 2
  Surface area (sq. metres) %Body Fat Lean Mass (kg) Heat production  
        Shivering (kj/hour) moderate exercise (kj/hour) Heat Loss (wet-cold) (kj/hour)
Polynesian 1.86 12.7 65.0 1750 2250 3692
Melanesian 1.58 10.2 50.7 1320 1660 3136
M/P ratio 0.85 0.8 0.78 0.75 0.75 0.85

Table 2. Comparative morphological and heat balance data for Polynesian and coastal New Guinean phenotypes. Estimates of surface area, body fat and lean mass as for Table 1.

Regarding heat loss, the thermal conductivity of water is some 25 times greater than that of air, and Beckmann and Reeves (1966) observed that, even in water at 24 degrees Celsius, none of their fit young adult subjects could tolerate more than 12 hours immersion. Of course, the voyagers were not immersed, but Pugh (1966) concludes that a person in wet clothing, especially in the presence of wind, is in much the same situation as a person in water at a slightly higher temperature. In addition to the wetness, the oceanic colonisers were exposed to the wind and without protective clothing — in conditions at sea they effectively were naked — the wind-chill factor is significant (Court 1948, Siple 1968).

In this exposed situation, heat loss figures for each body phenotype may be derived from Siple's (1968) figure of 1356 kJ/square metre/hour with a 23 kph wind and air temperature of 21 degrees Celsius, which are climatic figures compatible with the Pacific data. This value is approaching the subjective “cool” level but takes no account of evaporative heat loss. When the factor of dampness or wetness is added, the rate of heat loss readily doubles (Pugh 1967, Maclean and Emslie-Smith 1977, Kaufmann and Bothe 1986) to enter the subjective “very cool” to “very cold” range, which is in harmony with the written accounts. The combined insulative value of clothing and air in these conditions is only about 0.3 clo (Burton and Edholm 1955). To quantify this heat loss in the wet-cold situation the figure of 1985 kJ/square metre/hour derived by Iampietro et al. (1958) for nude men at temperature 14.5 degrees Celsius, 91 per cent humidity and 16 kph wind is used, as approximating the frequent wet-cold environment; if anything, it is on the conservative side, underestimating the - 174 possible effective coldness. In these conditions the smaller phenotype with a surface area of about 1.6 square metres has a heat loss from the body of 3100-3200 kJ/hour while the larger phenotype with surface area of about 1.8 square metres has a loss of 3500-3700 kJ/hour (Table 2).

Body heat production above basal level is related direct to skeletal muscle mass. At basal metabolic rates, skeletal muscle contributes only 20 per cent of body heat production, but any increase above basal level comes almost entirely from this tissue. A level of metabolic activity some five times basal level is achievable by shivering (Hemingway 1963) corresponding in a 75 kg body to an oxygen consumption of about 1.5 litres/minute, and in a 56 kg body to about 1.1 litres/minute. With an energy value of 20 kJ/litre of oxygen at a respiratory quotient of 4.825, this consumption equates with a heat production of about 1700 kJ/hour in the larger body (Webster 1952; Bazett 1968). Similarly, Iampietro et al. (1960) suggest about 1750 kJ/hour as the maximum heat production achievable by shivering over several hours. Relating this to body weight: if a muscular 75 kg individual can produce 1750 kJ/hour then a 56 kg individual can produce about 1320 kJ/hour. Hayward and Keatinge (1981) comment on this advantage of muscle mass in cold response.

These heat loss and heat gain figures (Table 2) show that in wet-cold conditions at sea — relatively inactive, shivering and exposed —the smaller body is able to produce only 42-43 per cent of body heat lost, with an hourly deficit of 1800-1900 kJ. The larger body can supply 47-50 per cent of body heat loss with an hourly deficit of 1750-1950 kJ. The larger body is better off, but, for any individual, such exposure is unsustainable for more than an hour or so. For example, Pugh (1967) noted a drop of core temperature from 37.5 degrees Celsius to 36.4 in 25 minutes in an inactive subject in wet-cold conditions of five degrees Celsius and 14 kph wind.

If shelter is largely obtained from the wind — say to about 4 kph —the heat loss is reduced to about half the exposed loss (Pugh 1966, Clarke and Edholm 1985). Then, with a heat production of about 1750 kJ/hour, the large body can just maintain heat balance. The small body remains in deficit of about 300kJ/hour. The change in mean body temperature in one hour resulting from this deficit is given by the equation: T = D/WS, where D is the heat loss, W is weight in kg and S is the specific heat of the body tissues (Spealman 1968). In this example T is 1.1 degrees Celsius/hour. With such a deficit a core temperature of about 32 degrees Celsius, a condition of moderate hypothermia, is reached within eight hours. The smaller individual could avoid this deficit only by increasing oxygen intake by 250 ml/minute, 23 per cent over the intake of 1.1 litres. If the maximum heat output from shivering has already been reached, such increase would have to arise from deliberate muscular activity.

During moderate exercise, body oxygen consumption is about 25 ml/kg body - 175 weight/minute, which equates with body heat production of 30 kJ/kg/hour (Pugh 1966; Bazett 1968). The 75 kg body will then produce about 2250 kJ/hour in body heat, and the slighter body 1660 kJ/hour (Table 2). This level of moderate exercise is a reasonable estimate for much of canoe existence, whether voyaging or local fishing. Voyaging canoes were largely sailing craft, but could be paddled. Horvath and Finney (1976) investigated the physiology of physically fit canoe paddlers of Polynesian ancestry and determined a mean oxygen uptake of 2.2 litres/minute for mean speeds of 3.16 knots sustained over eight-hour periods of paddling. During periods of sailing the level of activity would be less, but still above basal levels.

During such moderate exercise both phenotypes would be still in considerable heat imbalance under the conditions defined above of temperature 14.5 degrees Celsius, 91 per cent humidity and 16 kph wind. The larger body is producing about 62 per cent of the heat being lost, the smaller body about 53 per cent. However, these figures can be related to the realities of canoe existence. During severe wet-cold exposure conditions, progress under sail would be possible and physical activity for most of the crew restricted; they would be huddled down out of the wind and protected by whatever (damp) coverings were available. Heat balance of the large muscular body could be maintained for long periods by shivering. Smaller individuals would range in condition from chilled to hypothermic or worse, depending on body size and the time the severe wet wind-chill conditions persisted. As the wind dropped and some ancillary paddling was possible, heat loss from the body would rapidly reduce, heat production increase, and, for the larger body, any deficit in body heat would readily be restored. (There would be no need to persuade the crew to paddle!). Quite quickly in calmer conditions, with wind speed below about 8 kph and with moderate exercise, the effective environment would be warm. It is clear that, in these conditions, seemingly minor changes in wind exposure may have substantial influence on body heat balance and, as noted by Pugh (1966:1285), “apparently trivial differences in metabolism or total insulation can have disastrous consequences”.

The natural insulation provided by body fat has been suggested as having had survival value in Polynesian colonisation (Beaven 1977, Baker 1984). Yet, while a great many studies have shown the value of subcutaneous fat in conserving body heat, particularly in cold water immersion, the subjects of these studies invariably have had a higher body fat content than non-Westernised groups, and frequently have tended to the obese or grossly obese (Pugh and Edholm 1955). This fat advantage, thus, “is probably to be regarded as little more than a fortunate benefit provided by a food store” (Keatinge 1969:20) and such levels of food storage are probably readily attainable only in modern Westernised conditions. Burton and Edholm (1955) also question the importance of fat as - 176 insulation in Homo sapiens, calculating that a thickness of 2.5 cm is required to provide one clo of insulation. There is little evidence that indigenous people in cold environments have a particular tendency to accumulate subcutaneous fat. Skinfold thicknesses in Eskimos are not large (Ducros and Ducros 1979), while Korean diving women actually have skinfold thicknesses less than the Korean female population in general (Rennie et al. 1962). Budd (1965) recorded a loss of subcutaneous fat in a group exposed over several weeks to a wet, cold environment, findings confirmed by O'Hara et al. (1979). Pugh (1963) noted a remarkable example of high-altitude cold acclimatisation to derive from metabolic response and not fat insulation. There is no doubt that Polynesians have a strong tendency to obesity in Westernised conditions but there are several possible explanations for this tendency. The limited historical comment on obese individuals in Polynesia is confined to those of high rank and apparent great appetite and indolence (Cook 1968, Freycinet 1978), and the latter record is 50 years postcontact. The body fat content derived from the Polynesian skinfold data of the relatively traditional groups of Pukapuka, Tokelau and Foa (Tonga) (Table 1) are not suggestive of even a subtle obesity, being lower than levels obtained for contemporary Western athletes (Damon and Goldman 1964). The New Guinean figure is even less, being some 80 per cent of the Polynesian value. Hayward and Keatinge (1981) derived equations for the determination of mean fat thickness over the entire body. The Pacific data do not allow precise use of their equations, but an approximation from the available skinfold thicknesses suggests a difference between the Polynesian and New Guinean of 0.5 mm. Even allowing for the significance of “apparently trivial differences”, on Burton and Edholm's (1955) estimate this is about 0.04 clo in insulation value, which would seem to be inconsequential.

Another reason for arguing against pronounced subcutaneous fat deposition as a selective feature in Polynesians is the lability of the oceanic environment. The coldness has been stressed, but, of course, it is frequently very hot indeed, particularly on land. Life was not all voyaging and fishing, or always cold at sea, and a substantial cloak of fat would often have been an uncomfortable garment. Muscle tissue, on the other hand, allows for rapid variation in heat production, and, in water-immersion studies, increased metabolic output is recorded as being of equal significance with fat in maintaining body temperature (Hayward and Keatinge 1981). All these considerations indicate that, in non-Westernised groups, the metabolic response to cold is of primary importance.

The thickness of Polynesian limbs, commented on in the historical record and confirmed by anthropometry, has advantage both in heat production and heat conservation. In wet-cold conditions the often-rapid onset of physical impairment corresponds to the effect of local muscle cooling of the extremities, and not to a developing general hypothermia (Vanggaard 1975). Pugh (1966) investi- - 177 gated muscle temperature gradients in the lower limb in rather severe cold-wet conditions, finding, after 40 minutes at five degrees Celsius, temperatures of 27 degrees at 2 cm depth in both leg (extensor compartment) and thigh (quadriceps), rising at 4 cm to 31 degrees in the leg and 37 degrees in the thigh. Function can continue adequately with decreasing muscle temperatures to about 27 degrees Celsius; below this level the outer and cooler layers of muscle will be inactive but retain an insulative function (Clarke et al. 1958). The insulative effect of muscle is at least half that of fat (Clark and Edholm 1985) and the thick muscular limb can remain warm and functional, at least in its deeper part, when circulation and metabolic function have become severely restricted in the thin limb. In survival conditions efficient limb function may be vital.

These figures for heat production and heat loss show that, in the oceanic environment with only neolithic technology, a larger-bodied individual is at a quantifiable and crucial advantage in maintaining body temperature. In addition, a thick muscular limb maintains its warmth and function better. In these wet-cold oceanic conditions only an individual approaching Polynesian proportions and muscularity could generally sustain body heat and limb function. For individuals of lesser build the consequences would range from extreme discomfort to hypothermia and death, depending on the persistence of the windy, wet-cold conditions.

Yet, despite these advantages of the large body, heat imbalance must have occurred at times for any individual, and the fluctuation between effective high and low temperatures in this environment needs emphasis. In the tropical Pacific there is always the chance of a bitter night being followed by a very warm day, or of the wind dropping and vigorous exercise becoming possible, with recovery of heat balance. “The different things that we suffered from during the eight days and seven nights that we drifted were . . . the heat of the sun and the cold of the night” (Biggs 1974:363).

This lability of effective temperature is the great distinguishing feature of the tropical oceanic climate, and neolithic technology provided no effective insulation against the pervading wind and wet. By contrast, such technology may be highly effective in cold, dry continental regions. I suggest that, whatever the mean air and water temperature, the environment of Remote Oceania is effectively the coldest to which Homo sapiens has adapted, and, at the time of Western contact, the people of the region displayed the supreme cold-climate body form. This influence of environment is pervasive, going beyond simple consideration of infracranial proportions and muscularity to shape the distinctive Polynesian head and dentition (Houghton and Kean 1987).

In the light of the importance of body size, it seems possible that a constraint in settlement of Remote Oceania may have been female survival. Demographic simulations of small-island settlement (McArthur et al. 1976) usually assume - 178 equal numbers of male and female in founding populations, but it is likely that this was not the usual situation and that females may have been at a premium. Speculatively, in this may lie the origin of the widespread Polynesian constraint on females taking part in fishing expeditions. Children, of course, would not have fared well at sea.


The concept of a large-bodied oceanic phenotype must be applicable to Micronesia. The seafaring comments have a familiar ring. O'Connell (1972:104) described events in a lifeboat aftershipwreck in the Carolines in 1826:

Even in a latitude which must have been within fifteen degrees of the equator, a night passed without sleep or food, in an open boat, washed by the continual breakings of the sea over it, chilled our whole frames . . . Broiling heat succeeded the chills of the night . . . through the night the wet chills, and the same heat and calm upon the next day. After two days and three nights of exposure the daughter died about ten o'clock on the third day . . . The mother, in her weak state . . . in a few hours followed her daughter”.

For much of Micronesia and, particularly, the Marianas, the later historical record and recent anthropometric data are at best suspect as a reflection of the human biology of the past, because of the impact of Spanish rule. Between 1668 and 1710 the native Chamorro people were reduced in numbers from an estimated 50,000-100,000, to the 3439 counted in the first census (Cordy 1983). Thereafter, disease, transportation, and colonisation from the Philippines effectively obliterated them as a physical group. However, the contact record does offer some gleanings. The men were tall, robust, well-built and of great strength (Blair and Roberston 1903). “The Marianos are in colour a somewhat lighter shade than the Filipinos, larger in stature, more corpulent and robust than Europeans” (Garcia 1937:21). According to Mendoza, the natives were light-complexioned, like Europeans, “although in their bodies they do not resemble the latter for they are as large as giants and of such great strength, that it has actually happened that one of them, while standing on the ground, has laid hold of two Spaniards of good stature seizing each of them by one foot with his hands and lifting them as easily as if they were children” (Blair and Robertson 1903:6:138). Oliver (1961:335) summarises the early descriptions of the Chamorros: “— in other words, much like Polynesians”.

Howells (1973:36) describes modern Micronesians as “undersized Polynesians”, with a mean stature between 160 and 165 cm. Yet, in a detailed study examining the validity of Bergmann's rule, Roberts (1953:553) established - 179 an inverse relationship between mean body weight and mean environmental temperature for Homo sapiens in general, but found Polynesians and Micronesians to be anomalous and distinctive in possessing “greater weights for their stature and age than any other tropical dwellers”.

The archaeological record is slight, but increasing. Seven males from the prehistoric Haifa Dai site on Saipan in the Marianas have a mean stature of 174.4 cm. Corrected (+6 mm) mean biepicondylar width for this group is 61.5 for nine females and 67 for seven males — that is, medium frame by modern standards, but large by the standards of prehistory. (This dimension provides a simple and accurate assessment of body frame size, and predicts weight more accurately than stature estimates: Frisancho and Flegel 1983.) From the Latte House site on Saipan, five females were estimated to have a mean stature of 158.4 cm, and a solitary male 175 cm (Roy and Tayles 1989).

There are limited data from less afflicted parts of Micronesia. Mean stature of 23 living males from Kiribati is recorded as 169.5 cm (Shapiro 1937) and an estimate for a single prehistoric male is 174.5 cm (Katayama 1985). On the southern fringes of the Carolines living statures of 171.4 cm and 172.3 cm have been recorded from Nukuora and Kapingamarangi (Schlaginhaufen 1929, Buck 1950). Further south still, a substantial skeletal series excavated on Taumako by B. F. Leach and J. Davidson, and dated to about 700B.P., have a mean male stature of 1751 mm (n=52) and female of 1643 mm (n=19) (Houghton 1981). These last examples are classified as Polynesian outliers on ethnographic and linguistic evidence, but, for the present discussion, it is best to ignore labels and consider the evidence in a purely biological context.

In essence, the available data for Micronesia, whether anthropometric from more peripheral and unscathed parts, or from the early historical record, or the increasing archaeological evidence, are appropriate to the model of a large-bodied oceanic phenotype.


The heat balance analysis indicates that selection for a large muscular body must have taken place in the very early stages of the exploration of Remote Oceania. The region in which this selection began is, logically, Island Melanesia, for, by the rules of Bergmann and Allen, a large physique is biologically inappropriate for the land-based tropical environment further west, and nowhere else on the Pacific rim is the geography right for a gradual biological and technological adaptation to the sea. In Island Melanesia neolithic people reached a convenient nursery for voyagers — large islands, in sight one from another, gradually attained, but diminishing in size and number — until, mature in experience of the ocean, technologically equipped, and now physically evolved - 180 for survival, they could confidently and freely move on to the ultimate expanse of water.

While selection for larger body form must have occurred, plasticity of the genotype can cope to some extent with changing environment (Cavalli-Sforza and Bodmer 1971). Just when genotypic plasticity is left behind and actual selection is occurring, is probably impossible to define. Rhoads (1984) offers an interesting statistical analysis of anthropometric data from several Solomon Islands groups, which make apparent that two populations, the Lau of northern Malaita and the Ontong Javanese, are particularly heavy. For these same two groups the functionally related nasal height and chest breadth also are particularly large. It is just these two groups whose environment and economy are, or were, essentially maritime, in contrast withthe land-based economies of all the other groups, save one. The exception is Ulawa, which registered above the 75th percentile on the three parameters. Ulawa has an essentially land-based economy but with some marine exploitation. Particularly in Ontong Java we may be seeing evidence of the transition to a fully fledged oceanic phenotype.

In my interpretation of these Solomon Islands data I may differ from Friedlaender (1987:357), who concludes that, for the Pacific, “Natural selection has not been a major determinant of summary population relationships in recent times”. I am uncertain of his terms “summary” and “recent” in this context, but my own view is that selection has been of major importance in shaping the physical form of the peoples of the Pacific.

The time that this selection for large body size took is uncertain, but there is a big gap between the earliest date for human occupation of New Ireland, 33,000 B.P. (Allen et al. 1988), and the first appearance of Lapita ware with its voyaging associations about3600 B.P. (Green 1982). While it is likely that humanity took many millennia to come to terms, biologically and technologically, with the oceanic environment, there is growing evidence that phenotypic change can take place quite quickly. Lister (1989) suggests a reduction in size to one-sixth of original body weight in less than 6000 years in an island isolate of red deer. An analysis of metabolic adaptation to high altitude by Peruvian Indians suggests that, in these people, true selective genetic change has occurred (Hochachka 1989), which, in light of the probable settlement time of the Andean region, has taken less than 5000 years. In a study of stature change in Italy over a single generation, Cavalli-Sforza and Bodmer (1971) determined a selective genetic component of 0.8 cm in the overall generational change of 3 cm. The stature heritability estimate used in their study (0.62) is lower than that determined in most studies and a more usual heritability estimate of about 0. 80 raises the selective genetic component to 1 cm. It seems most unlikely that this recent European group was exposed to the extreme selective pressure for body size that existed for the early colonisers in the Pacific. The analysis of directed selection - 181 is fairly straightforward (Cavalli-Sforza and Bodmer 1971) and, in an evolving oceanic group under considerable selective pressure, to look for an increase in male mean body parameters from 160 to 170 cm and 60 to 70 kg for stature and mass within less than 1000 years (40-50 generations) seems theoretically unexceptionable.

The view is taken here that the prehistory of Island Melanesia was distinguished by the gradual evolution within the region of larger-bodied, increasingly maritime peoples, antedating the Lapita culture but ancestral to its creators. At the same time and perhaps from a common gene pool, others of Homo sapiens remained essentially with the land and retained the tropical body phenotype.

The Lapita culture is prominent in current discussion on the emergence of pre-Polynesians in Island Melanesia. “What both the exchange system and the huge distribution of the Lapita cultural complex do imply is a great deal of skill in local and long-distance voyaging” Green (1982:16). Human skeletal material from secure Lapita contexts is still sparse, but, nevertheless, offers significant information regarding the evolution of the phenotype of Remote Oceania. In Table 3 are data derived from some Lapita-related individuals (Houghton 1989a, b and c; Kirch et al. 1989). It is helpful that they derive from the poles of known Lapita distribution. The series of eight adults from the Reber-Rekival site on Watom Island in the west is much the most significant in numbers. For this group the large body build dominates the findings. While the limitations of the sample must be continually born in mind, the mean male stature of 1784 mm is exceptional for a prehistoric group, and the female also is tall. These estimates are derived from Polynesian stature equations (Houghton et al. 1975) and are preferred because, for the Lapita remains, they give the least range of stature for a single individual from the different long bones. Use of the White American equations of Trotter (1970) still gives a tall mean male stature of 1750 mm while the stature of the single female is actually increased. All show bony evidence of a well-developed musculature.

The tallest individual (1809 mm) has a biepicondylar (lower humeral) breadth, corrected (+6 mm) for soft tissues and bone shrinkage of 67 mm. This falls into the “medium frame” category for contemporary American males, with a mean weight of 79.2 kg. The Watom individual, being at the lower end of the category, would be below this mean weight, but, as discussed above, a contemporary individual will be fatter than an individual existing under neolithic conditions, and it can be postulated that this was a muscular individual of some 72 kg weight. Indeed, the lean body mass of the neolithic individual may well have exceeded that of most individuals in the contemporary “medium frame” category. It is also to be noted that the contemporary data embrace many individuals of a weight and frame size that, on dietary and general environmental grounds, would have been extremely rare anywhere in prehistory. For a Neolithic - 182 group, the Watom people were, indeed, large and muscular. This phenotype, inappropriate to the land-dominated tropical environment to the west, places the Watom people firmly in the ancestry of the Polynesians.

Table 3
individual sex age(years) stature(mm)
Watom 2 female 20 1659
Watom 3 male 25-30 1798
Watom 5 male 30+ 1767
Watom 6 male mature adult 1800
Pea, Tonga male mature adult 1710
Lakeba male mature adult 1715
Mussau male mature adult 1705
mean male structure     1749

Table 3: Basic data on several Lapita-associated adults. All from Houghton (1989 a, b, c), except for Mussau (Kirch et al. 1989).

The individual from the Lakeba site in the Lau Group of Fiji (Houghton 1989b) is an adult male, probably in the fourth decade of life. The long bones are robust (though a defined robusticity index was not obtainable) and with physiologically bowed shafts to accommodate a considerable musculature. From the maximum lengths of the intact right forearm bones a stature of 1715 mm is suggested. The upper femoral morphology resembles the extreme eastern Polynesian form. The mandible, though influenced in morphology by advanced dental caries, verges on the rocker form, the right side displaying no antegonial notch and the left a slight notch. This array of skeletal features, and the stature estimate, are distinctively Polynesian.

The fragmentary human material from Pea, Tonga, Site To-1 (Houghton 1989c) is largely of a mature robust male whose skeletal form is compatible with the Polynesian phenotype. The long bones are robust and bowed with strong muscle markings. Stature is about 1710 mm, with Polynesian stature equations giving the most consistent estimate. Corrected biepicondylar diameter of 68 mm falls into the “medium frame” body category, with an estimated weight in excess of 72 kg.

In general, all these Lapita-associated human remains support the idea of the evolution within Island Melanesia, and later eastward spread, of a larger-bodied people adapted to the maritime environment.

The only early human material from the more eastern reaches of Island Melanesia of which I have knowledge is from the Taplins site on Efate, Vanuatu - 183 (Houghton 1975). Fragmentary remains of several individuals, associated with pottery of the Marigoasi tradition, and dated to about 2080 B.P. (human collagen, INS R5497) were excavated by G. K. Ward in 1974. (This date is currently under revision, but is not less than 2000 B.P.). Stature of a 16-year-old male is about 1650 mm, and — with rather less accuracy — a mature male has a stature of 1670 mm. A mean Nordin's score of 53 on second metacarpals of two individuals indicated reasonable bone structure and an adequate diet for the group — that is, there is no reason to suppose that the modest stature was a consequence of an unfavourable environment. Corrected biepicondylar diameter on an adult female was 57 mm which is in the “small frame” body category of Frisancho and Flegel (1983), with a mean body weight of 57 kg.

The mandibles of the Taplins group are robust, but the rami are low, reflecting a small mid-facial region and airway, and a rather slight physique. The malar (cheek) bones are small, particularly vertically, again indicating a small mid-face and airway. The shoulder girdle skeletons are gracile. Only one adult showed a rather minor first rib groove on a (right) clavicle, a feature that in its pronounced form has been associated with the consistent action of paddling canoes (Houghton 1980). The only femora preserved are most un-Polynesian; the Taplins specimens have unusually round upper shafts with a platymeric index of unity and large round foveae on the heads. The biological basis for these femoral differences is yet to be established, but will be related to the degree of muscularity around hip and thigh, the extent of which, in turn, reflects the extent of adaptation to the wet-cold oceanic environment.

This limited Taplins series suggests the presence of a rather smaller, non-oceanic morphology remaining confined to Island Melanesia and probably ancestral to many of the groups there today. This does not preclude the original sharing, by individuals exemplified by the Taplins phenotype, of a common gene pool with the people of the larger ocean.

If the biological concept of a large-bodied oceanic phenotype is valid, and Island Melanesia is the logical place for its evolution, then one is led inexorably to the conclusion that in Island Melanesia are the origins of all the colonisers of the wider Pacific in prehistory — Micronesia west and east, as well as Polynesia. The proximity to the Philippines of Palau and Yap and the string of islands towards Guam, make some human movement along this route a possibility. However the absence from the Philippines of a large-bodied phenotype, either recently or in the historic or prehistoric record, is against significant biological contact with Micronesia, whatever the linguistic connection. And, on biological grounds, the discovery of significant numbers of large-bodied individuals is not to be anticipated in the Philippines or further west. The hypothesis supported here is that the settlement of Micronesia, like that of Polynesia, was out of Island Melanesia by those biologically and technologically equipped to succeed on - 184 such voyages. Particularly in the expansion across Polynesia was there the opportunity for genetic drift to act on the already established large-body phenotype, with development of a most distinctive morphology. For Fiji, much has been made of Polynesian and “Melanesian” admixture (these terms are discussed below), but the likelihood is that we are seeing here the descendants of the early, still-evolving oceanic phenotype as it was at the time of settlement of Fiji. The “tail” of the body form distribution, created by the directed selection influence (Cavalli-Sforza and Bodmer 1971), is composed of the smaller, so-called “Melanesian” members of the population. A rather different matter — the environmental influence on inland and coastal peoples over the settlement period of this substantial island — must also be considered.

A common difference between many of the smaller-bodied, large-island people of much of geographic Melanesia, and the large-bodied, small-island people of the greater ocean is the lighter skin colour of the latter. An understanding of skin pigmentation in relation to climate is still elusive, but the Pacific distribution does fit with the general global distribution of lighter skin with cooler climates. It can be postulated that this is another selective adaptation, for, while light skin reflects more of the visible light, it allows the infrared part of the spectrum to penetrate more deeply. In the oceanic environment such a warming effect would be advantageous.


In the light of this biological perspective on the settlement of the Pacific, I think it is desirable to comment on three related matters. These are, firstly, some terminology frequently used in discussion of the origins and relationships of the people of Remote Oceania; secondly, the use of osteological data; and thirdly, the significance of recent DNA analyses on living Pacific groups.

Some standard terminology.

The terms Polynesian, Micronesian, Melanesian, Mongoloid and Australoid are commented on here.

The distinctive Polynesian somatotype has been described in the first part of this paper and is remarkably homogeneous considering its distribution over one-sixth of the globe. There is not space here to mention a considerable and consistent array of distinctive biochemical and physiological parameters. Indeed, if any group of Homo sapiens tends to compliance with the outdated concept of a racial type, it is the Polynesian, particularly East Polynesians. It is probably reasonable to continue to use the term “Polynesian” for a particular biologically confined group, though, with DNA and other studies, an increasing number of subtle differences between subgroups will, doubless, be displayed.

- 185

Until we have much more biological data we must withold judgment as to whether Micronesia was sufficiently homogeneous as to justify a single epithet along the lines of Polynesia, or initially, in genetic terms, has been one with Polynesia. The history of western Micronesia is such that prehistoric material will be crucial in this assessment.

By contrast, the hallmark of Homo sapiens in Melanesia is biological diversity; anthropometric variability goes hand in hand with serological polymorphisms of confounding profusion (e.g., Friedlaender 1975, Lai and Bloom 1982). This is part of the common knowledge of anthropology, yet the surprising thing is that the term “Melanesian” is still used with the implication of some sort of biological unity or homogeneity, and frequently as a foil to the term “Polynesian”. “People will inevitably persist in the naming of racial groups based on simple physical and even social attributes, but, at the very least, they should be made aware how grossly simplified any such taxonomic system has to be, and the diversity which a name such as ‘Melanesian’ masks, (Friedlaender 1975:215). “Melanesian” represents a geographic statement, indicating someone living within the confines of Melanesia. No phenotypic or genotypic uniformity can be implied from the term.

On a broader scale, the term “Mongoloid” often appears as a label for the “racial” affiliations or origins of Polynesians (and Micronesians; but the limited data prevent much discussion). Leaving aside the problem of defining “Mongoloid” in any sensible way, a scrutiny of the phenotypic characters purported to evidence this Polynesian association (Bellwood 1978, Coon 1966) is some-what underwhelming. It is a comment on the arbitrariness of these racial classifications that Montague (1960:451) saw the Polynesians as “a far-flung branch of the Mediterranean stock”. In form of head and form of the body overall, Polynesians are distinctive among Homo sapiens, and I can see no significant resemblance to any recent Asian group nor justification for assigning to them the epithet “Mongoloid”. Indeed, a major point of this paper is that adaptation to the environment of the wider Pacific has substantially changed body and head form, and many biochemical and physiological parameters, so that the settlers of Remote Oceania little resembled their ancestors out of Asia.

The term “Australoid” is rather peripheral to the sphere of this discussion, but “Australoids” are sometimes contrasted with “Mongoloids” and/or Polynesians. The unity of the “Australoids” seems to be the concept or stereotype of a rather slight, often rather short, and usually dark-skinned people, generally equated with the aborigines of Australia, Highland New Guinea, and much of coastal New Guinea and Island Melanesia. Bulbeck (1982) observes that the criteria are defined on recent material and then the divisions are extrapolated to the small number of older (albeit insecurely dated) specimens. Much of the recent variation, such as mandibular form and the extent of prognathism and differences in - 186 tooth size, is interpretable on environmental and functional grounds. Whatever one's views on the evidence used to derive modern “Australoid” and “Mongoloid” groups from Homo erectus in China or from Java and China respectively (Weidenreich 1947, Wolpoff et al. 1984), these attempts are rendered obsolete by recent analyses of mitochondrial DNA (Stoneking and Wilson 1989). These studies not only deny the regional descent of Homo sapiens from Homo erectus but also have Australia and New Guinea settled by different lineages of Homo sapiens out of Asia. The epithet “Australoid” does not seem to have any coherent biological basis.

It may be claimed that these terms are part of the essential vocabulary of the subject; that, like the term “race” itself, they do serve a useful purpose and that the topic cannot sensibly be discussed without them. Bellwood (1985:69) feels that “for an intelligible narrative of prehistory a concept of race is necessary”. I think this is an unfortunate and unnecessary view, for it makes all too easy the continuation of thinking in outdated racial terms (I should make it completely clear that I am talking strictly in a taxonomic context). Because these various terms, with the possible exceptions of Polynesian and Micronesian, are biologically undefinable, they are just cliches that form a real impediment to progress towards an understanding of the early human biology of this part of the world. I have no illusions as to the difficulty in discarding such impedimenta, which we all carry. Even in molecular biology, the very discipline that offers the chance to shake free of unbiological racial scenarios, discussion is frequently couched in typological terms. Melanesia is contrasted with Polynesia as if they were distinct and opposing genetic entities. It is startling to come to the opening sentence of the discussion in a recent paper on nuclear DNA polymorphisms and read “The three principal races in the South Pacific are Melanesians, Micronesians and Polynesians” (Trent et al. 1986:355). One must appeal to the molecular biologists to keep their minds uncluttered and not to read the anthropology texts.

As Jones (1981:190) has pointed out, the genetic differences among human beings within a local population are far larger than the genetic differences among the classically described races of humanity, or those which exist among nations within a “racial” group. He concludes: “The idea of racial ‘type’ — and, some would argue, of ‘race’ itself — is no longer a very useful one in human biology”. Certainly, in the Pacific arena, much of the terminology is typological and obsolete.

A matter related in its potential to spread confusion is the tendency of some writers to flip between a biological and a linguistic definition even in the course of a single sentence. Thus, Australoids are said to be displaced by Austronesians, or “coastal Melanesia has about 16 per cent Austronesian admixture” (Serjeantson et al. 1982:904). I think this ready equating of linguistic and genetic relationships - 187 is unwise. Stoneking and Wilson (1989) comment on the homogeneity of the mitochondrial DNA patterns in coastal New Guinea in the face of linguistic diversity. That is, the biological homogeneity extends across the divisions into Austronesian and non-Austronesian languages. For later prehistory in the Pacific it is obviously reasonable often to make the association between linguistic and biological categories (although not in terms of percentage admixture); but to tie the misty Asian origins of the Austronesian language family to a particular genotype or phenotype is speculation.

Analyses of osteological data

Statistical analyses of skeletal data, usually cranial, occupy considerable space in some studies of Pacific peoples. Comment is necessary on how the biological and environmental considerations discussed here might influence such studies. Their intentions are presumably taxonomic, although “clustering is not the same as phylogeny, however well the two seemingly or actually correspond. Clustering is simply an arrangement of data” (Howells 1984:6). I think there are grave problems with these studies beyond the common difficulty in complying with the rather strict demands of multivariate theory, though this is in itself a major problem and renders invalid many presentations. Methodologically, most take an inductive approach — a traditional and comprehensive battery of skull measurements is subjected to a multivariate analysis (Mahalonobis D2 is a favourite) and the results are offered for interpretation. The derived clusters are presented in various ways. It is usual for slight variation in the basic data, such as dropping of two or three measurements in order to increase sample size, to result in different, even contradictory, results. Within any large cluster diagram at least one bizarre juxtaposition can be assured, such as Hawaii with the Gulf District of New Guinea, or the Trobriands with Easter Island (Pietrusewsky 1976). Or the females of a group show a totally different kinship from that of the males (Howells n.d.). While a result is assured with such methodology, no judgment is possible as to which presentation, if any, is valid, and I suspect that, by and large, and wisely, prehistorians ignore these exercises. The hiccups are sometimes said to indicate the need for greater refinement of the statistics.

Davis and Hersh (1986:61–2) comment “ . . . it is important to state publicly that among professional mathematicians the skepticism . . . about mathematical biology is much stronger than it is among nonmathematical . . . biologists. This skepticism is rarely stated in print”. They define “rhetorical mathematics” as distinct from pure and applied mathematics. No practical consequences issue from rhetorical mathematics which “ . . . presents itself as applied maths, but it is easy to tell them apart . . . Either initially or ultimately work in applied maths - 188 leads back to the phenomenon being modelled. Rhetorical maths is often incapable of being tested against reality”. That seems a reasonable description of much use of statistics in the anthropological literature.

Warnings on the misuse of multivariate statistics have been sounded often enough (e.g., Kowalski 1972, Corruccini 1975, Rhoads 1984:248). Rhoads comments,

It is a misuse of the statistical procedures, and a form of neoPythagoreanism, to suppose that nature conforms to inductively discoverable arithmetical ideals, and that statistical methods are active agents in an inductive search. We are the agents and must acknowledge the responsibility. The application of a particular method is, or ought to be, a thoughtful act of substantive scholarship.

This leads to the core of the problem, which is fundamentally not one of statistical method but of biological validity. It is the problem of the fundamental data put into the statistical system. Unless there are clear and valid biological reasons for supposing the measurements to carry a reasonably uncluttered and common genetic content between compared groups, then these exercises are scientifically untenable. There is nothing original in this view. Statisticians, often rather unversed in biology, echo it; “ . . . it is necessary for the anthropologist to give a great deal of careful thought at the commencement of the project to his choice of measurements, rather than to measure everything possible” (Wilson 1984:263).

The problem is that there is scarcely a part of the body that may not be substantially influenced by the environment. Earlier I indicated that this influence is not merely on total body size and muscularity, but extends to that popular source of data, the skull. Here it does seem that the only region where the genetic message may possibly be relatively unobscured, at least after maturation of the nervous system, is the central component of the cranial base (Houghton and Kean 1984, 1987, Kean and Houghton 1982, 1987). The rest, vault, mid-face and airway, extent of prognathism, dentition and mandible, may all be profoundly shaped by environmental influences. This makes a nonsense of attempts to relate peoples across large expanses of prehistoric time and space on the basis of unconsidered phenotypic data. What is being achieved in these studies is a display of degrees of association of metric data obtained from skulls. It would be inaccurate to call the data “biological” for biological considerations have not gone into their selection. The associations tend often to look plausible, for they are fundamentally geographic (that is, environmental) but could be obtained more simply just by using an atlas.

Incidentally, the use of nonmetric traits does not offer an easy way out. Some - 189 features, such as those defined on the Sinjanthropus mandibles by Weidenreich (1936) and subsequently used by Larnach and Macintosh (1971), almost in their entirety owe their variability to functional influences. The genetic bases of most of the commonly used battery of discrete cranial traits are unknown, and heritability estimates are low and, of course, are group-specific (Sjovold 1984, Rosing 1984:319). The latter finds the analysis of such traits to be plagued with serious difficulties, particularly a virtually unknown genetic background, and concludes that the commonly used traits appear to be “unsuited for most populational investigations”. The same caveats apply to the use of variations in crown morphology of teeth. The increased understanding of phenotypic expression arising from molecular genetics means that such terms as “variable penetrance” are obsolete and that varying degrees of expression of a morphological trait may have different genetic bases. In arguing a case for the use of nonmetric traits in palaeoanthropological studies, Wolpoff et al. (1984:426) comment, “Comparisons are likely to be misleading or invalid unless they are between populations that are fairly closely related”. Exactly, and to use the traits to determine this relationship is hardly logical. Within an isolate, certain markers, particularly neural foramina, may prove to be useful, but at present there is no theoretical basis for comparisons across large expanses of prehistoric space and time.

When undertaken without consideration of the biology of the situation, statistical analyses, with taxonomic intent, of data obtained from prehistoric human remains in the Pacific and its environs, may simply be unbiological, unscientific and uninterpretable.

Phenotype and Genotype

Finally, I wish to look very briefly at the molecular evidence as it relates to the adaptive model presented in this paper. As a preliminary, it is worth commenting that the young science of molecular genetics should provide the fine-grained information on relationships within the wider Pacific that are attainable from no other source, and the prospect of obtaining DNA from archaeological bone is exciting. In these subtler matters I do not think that gross skeletal studies can ever provide firm answers. However, it would also be a mistake to believe that the patterns of origins and journeyings and relationships of different groups of Homo sapiens are writ crystal-clear on the genome. For example, Chen et al. (1987) give four dendrograms for genetic distances between Oceanic groups derived from gene and haplotype frequencies at the betaglobin and ALB-Gc gene clusters (the latter a non-DNA analysis). All dendrograms place the Polynesian (Rarotonga and Niue) and Micronesian (Kiribati) groups close together. However, the reshuffling of other groups in the different dendrograms brings a sense of dejà-vu to anyone accustomed to viewing the results of analyses of phenotypic cranial data. At least, in the case - 190 of the gene studies, these problems may resolve with more detailed analyses on larger numbers.

On the small, maternally transmitted and rapidly mutating mitochondria' DNA molecule a unique 9-base pair deletion has been identified as confined to “Asian” groups. Analysis of the distribution is still at a very early stage and the sample numbers are still very small, but it has been identified in frequencies of 5-16 per cent in Chinese and Japanese groups, and at 42 per cent in coastal Papua New Guinea (Stoneking and Wilson 1989). Subsequently, Herzberg et al. (1989a) have established an incidence for this mutation of 93 per cent in Polynesia, 82 per cent in Fiji, 8 per cent in the Tolai of New Britain, and 14 per cent in coastal New Guineans. The deletion was absent in 30 New Guinea Highlanders and 31 of 32 Australian Aborigines. Some nuclear DNA studies also show the link with Asia (Herzberg et al. 1989b)

In interpreting these findings one must resist the temptation to picture Mother rapidly coursing the northern coast of New Guinea, paddled along by a few stalwart relatives, and heading for Polynesia. The task is to marry the molecular and statistical data with the grosser biological and environmental picture of selection for body size and the realities of ocean voyaging and small- island settlement. As the evidence points to small group origins and the founder effect repeated again and again in the settlement of the wider Pacific, it would not be surprising if, in the known time scale, a single maternal origin out of Asia should be traceable. As Stoneking and Wilson (1989) point out, random walk theory demonstrates that, for a stable population of initial size n, then after n generations all individuals are likely to trace their origin back to a single female. It is inevitable that a maternal origin for Polynesians should be traceable to Asia. But this was not a migration out of Asia of Polynesians or pre-Polynesians (sensu Lapita) — indeed, as indicated earlier in this paper, the Polynesians are not more distinctive for their Asian origin than are the Highland New Guineans or the Australians (Stoneking and Wilson 1989), though they probably do share more of their genome with recent Asian peoples than do the others. Where the mitochondria' DNA actually came from is rather immaterial in the physical (phenotypic) evolution of the pre-Polynesians and pre-Micronesians in Island Melanesia.

Whatever the physical appearance of the first people into Island Melanesia, the pause while the oceanic phenotype evolved was long enough for the appearance of some nuclear DNA polymorphisms unique to Island Melanesia and Polynesia. These include a specific alpha-thalassaemia deletion and a restriction-enzyme-site polymorphism associated with a triplicated-zeta-gene chromosome (Hill et al. 1985, 1987). On present carbon-14 dates we know that some 25,000 years' worth of mutations were available. Finer-grained DNA studies may clarify these matters of origins and time.

- 191

For Homo sapiens, the wider Pacific, with its fluctuation between hot and very cold conditions, is a unique environment. Neolithic technology provided no useful protection against the wet-cold conditions, and among the adaptive responses was selection for a large muscular body. This change began in Island Melanesia.

As this new environment significantly altered human morphology, many of the taxonomic labels and methods applied to the external and skeletal form of Homo sapiens of the region are of dubious value. DNA analyses will increasingly be useful, particularly if archaeological bone proves a satisfactory source, but need to be carefully interpreted against the phenotypic and environmental situation.

  • Allen, J. A., 1877. The influences of physical conditions on the genesis of species. Radical Review, 1:108–40.
  • Allen, J., C. Gosden, R. Jones and J. P. White, 1988. Pleistocene dates for the human occupation of New Ireland, northern Melanesia. Nature, 331:707–9.
  • Baker, P. T., 1984. Migrations, genetics and the degenerative diseases of South Pacific Islanders, in A. J. Boyce (ed.), Migration and Mobility. London: Taylor and Francis, pp.209–39.
  • Bazett, H. C., 1968. The regulation of body temperature, in L. H. Newburgh (ed.), Physiology of Heat Regulation. New York: Hafner, pp.109–85.
  • Beals, K. L., C. L. Smith and S. M. Dodd, 1983. Climate and the evolution of brachycephalization. American Journal of Physical Anthropology, 62:425–38.
  • Beaven, D. W., 1977. Maori health. New Zealand Medical Journal, 85: 540.
  • Beckmann, E. L. and E. Reeves, 1966. Physiological implications as to survival during immersion in water to 75 degrees F. Aerospace Medicine, 37:1136–42.
  • Bellwood, P., 1978. Man's Conquest of the Pacific. Auckland: Collins.
  • ——1985. Prehistory of the Indo-Malaysian Archipelago. Sydney: Academic Press.
  • Bergmann, C., 1848. Uber die Verhaltuisseder Warmeokonomie der Thiere zu ihrer Grosse. Gottinger Studien, No. 8.
  • Biggs, B., 1974. A drift voyage from Futuna to Cikobia. Journal of the Polynesian Society, 83:361–5.
  • Blair, E. and J. A. Robertson, 1903. The Philippine Islands 1493–1898. Cleveland: Clark.
  • Bligh, W., 1961. The Mutiny on Board H. M. S. Bounty. New York: Signet.
  • Buck, P., 1923. Maori somatology. Journal of the Polynesian Society, 32:189–99.
  • ——1950, Material Culture of Kapingamarangi. Bulletin 200. Honolulu: Bernice P. Bishop Museum.
  • Budd, G. M., 1965. Effects of Cold Exposure and Exercise in a Wet, Cold Antarctic Climate. Journal of Applied Physiology, 20:417–22.
  • Bulbeck, D., 1982. A Re–evaluation of Possible Evolutionary Processes in Southeast
- 192
  • Asia since the Pleistocene. Bulletin of the Indo-Pacific Prehistory Association, 3: 1–21.
  • Burton, A. C. and O. G. Edholm, 1955. Man in a Cold Environment. London: Arnold.
  • Cavalli-Sforza, L. L. and W. F. Bodmer, 1971. The Genetics of Human Populations. San Francisco: Freeman.
  • Chen, L. Z. , P. G. Board and R. L. Kirk, 1987. DNA Polymorphisms in Oceania: a Study of Haplotypes within the Beta-globin and the Albumin-Gc gene Clusters. Man and Culture in Oceania, Special Issue, 3:19–32.
  • Clark, R. P. and O. G. Edholm, 1985. Man and his Thermal Environment. London: Arnold.
  • Clark, R. S. J., R. F. Hellon and A. R. Lind, 1958. The duration of sustained contraction of the human forearm at different muscle temperatures. Journal of Physiology, 143:454–73.
  • Cook. J., 1968. Endeavour Journal. J. C. Beaglehole (ed.). Cambridge: Hakluyt Society.
  • Coon, C. S., 1966. The Living Races of Man. London: Cape.
  • Cordy, R., 1983. Social stratification in the Mariana Islands. Oceania, 53:272–6.
  • Corruccini, R. S., 1975. Multivariate analysis in biological anthropology: some considerations. Journal of Human Evolution, 4:1–19.
  • Court, A., 1948. Wind chill. Bulletin of the American Meteorological Society, 29: 487–93.
  • Damon, A., 1977. Human Biology and Ecology. New York: Norton.
  • —— and R. F. Goldman, 1964. Predicting fat from body measurements. Human Biology, 36: 32–44.
  • Davis, P. J. and R. Hersh, 1986. Descartes Dream: the World According to Mathematics. London: Pelican.
  • DuBois, D. and E. F. DuBois, 1915. The measurement of the surface area of man. Archives of Internal Medicine, 15:858–81.
  • du Clesmeur, A., 1914. Journal of the Marquis de Castries, in R. N. McNab (ed.), Historical Records of New Zealand. Wellington: Government Printer, pp 445–81.
  • Ducros, A. and J. Ducros, 1979. Recent anthropological data on the Ammassalimiut Eskimo of East Greenland: comparative studies, in W. A. Stini (ed.), Physiological and Morphological Adaptation and Evolution. The Hague: Mouton, pp. 55–67.
  • Dunn, L. C. and A. M. Tozer, 1928. An anthropometric study of Hawaiians of pure and mixed blood. Papers of the Peabody Museum, 9:89–211.
  • Erskine, J. G., 1967. Journal of a Cruise among the Islands of the Western Pacific. London: Dawson.
  • Finau, S. A., J. M. Stanhope, I. A. M. Prior, J. G. Joseph, S. T. Puloka and P. N. Leslie, 1983. The Tonga cardiovascular and metabolic study. Community Health Studies, 7:57–77.
  • Finney, B. R., 1979. Hokule'a: the way to Tahiti. New York: Dodd, Mead.
  • Firth, R., 1963. We, the Tikopia. Boston: Beacon Press.
  • Freycinet, L. C.de S., 1978. Hawaii in 1819. Pacific Anthropological Records No.26.
  • Friedlaender, J. S., 1975. Patterns of Human Variation. Cambridge: Harvard University Press.
  • ——1987. Conclusion, in J. S. Friedlaender (ed.), The Solomon Islands Project. Clarendon: Oxford, pp.351–62.
- 193
  • Frisancho, A. R. and P. N. Flegel, 1983. Elbow breadth as a measure of frame size for US males and females. American Journal of Human Nutrition, 37:311–4.
  • Froese, G. and A. C. Burton, 1957. Heat losses from the human head. Journal of Applied Physiology, 10:235–41.
  • Garcia, F., 1937. Life and Matyrdom of the venerable Padre Diego Luis de Sanvitores, translated by M. Higgins. Guam Recorder, April 1937.
  • Green, R. C., 1982. Models for the Lapita cultural complex: an evaluation of some current proposals. New Zealand Journal of Archaeology, 4:7–19.
  • Harvey, R. G., 1974. An anthropometric survey of growth and physique of the populations of Karkar Island and Lufa subdistrict, New Guinea. Philosophical Transactions of the Royal Society, B 268:179–292.
  • Hayward, M. G. and W. R. Keatinge, 1981. Roles of subcutaneous fat and thermoregulatory reflexes in determining ability to stabilise body temperature in water. Journal of Physiology, 320:225–51.
  • Hemingway, A., 1963. Shivering. Physiological Reviews, 43:397–422.
  • Herzberg, M., K .N. P. Mickleson, S. W. Serjeantson, J. F. Prior and R. J. Trent, 1989a. An Asian-specific 9 base pair deletion of mitochondrial DNA is frequently found in Polynesians. American Journal of Human Genetics, 44:504–10.
  • ——K. Jahromi, V. Ferguson, H. H. M. Dahl, J. Mercer, K. N. P. Mickleson and R. J. Trent, 1989b. Phenylalanine hydroxylase gene haplotypes in Polynesians: evolutionary origins and absence of alleles associated with severe phenylketonuria. American Journal of Human Genetics, 44:382–7.
  • Heyerdahl, T., 1952. The Kon-Tiki Expedition. London: Reprint Society.
  • Hill, A. V. S., D. K. Bowden, R. J. Trent, D. R. Higgs, S. J. Oppenheimer, S. H. Thein, K. N. P. Mickleson, D. J. Weatherall and J. B. Clegg, 1985. Melanesians and Polynesians share a unique alpha-thalassemia mutation. American Journal of Human Genetics, 37:571–80.
  • ——B. Gentile, M. Bonnardot, J. Roux, D. J. Weatherall and J. B. Clegg, 1987. Polynesian origins and affinities: globin gene variants in eastern Polynesia. American Journal of Human Genetics, 40:453–63.
  • Hochachka, P., 1989. How Peruvian Indians adapt to altitudes. New Scientist, 122:14.
  • Horvath, S. M. and B. R. Finney, 1976. Paddling experiments and the question of Polynesian voyaging, in B. R. Finney (ed.), Pacific Navigation and Voyaging. Wellington: Polynesian Society, pp. 47–54.
  • Houghton, P., MS 1975. The Human Material from the Taplins Site, Efate, Vanuatu. Unpublished typescript, University of Otago.
  • ——1980. The First New Zealanders. Auckland: Hodder and Stoughton.
  • ——MS 1981 The People of Namu. Unpublished typescript, University of Otago.
  • ——1989a. Watom: the people. Records of the Australian Museum, 41:223–33.
  • ——1989b. The Lapita-associated human material from Lakeba, Fiji. Records of the Australian Museum, 41:327–9.
  • ——1989c. Comment on the human skeletal material from Pea, Tonga, Site To. 1. Records of the Australian Museum, 41:331–2.
  • ——1990. The adaptive significance of Polynesian body form. Annals of Human Biology, 17:19–32.
- 194
  • ——B. F. Leach and D. G. Sutton, 1975. The estimation of stature of prehistoric Polynesians in New Zealand. Journal of the Polynesian Society, 84:325–6.
  • —— and M. R. Kean, 1984. Cheshire cat craniology, in G. K. Ward (ed.), Archaeology at ANZAAS Canberra. Canberra: Canberra Archaeological Society, pp.80–5.
  • —— and M. R. Kean, 1987. The Polynesian head: a biological model for Homo sapiens. Journal of the Polynesian Society, 96:223–42.
  • Howells, W. W., 1973. The Pacific Islanders. Wellington: Reed.
  • ——1984. Introduction, in G. N. Van Vark and W. W. Howells (eds.), Multivariate Statistical Methods in Physical Anthropology. Dordrecht: Reidel, pp.1–11.
  • —— n.d. Skull Shapes and the Map: Craniometric Analyses in the Dispersion of Modern Homo. Cambridge: Peabody Museum.
  • Iampietro, P. F., D. E. Bass and E. R. Buskirk, 1958. Heat exchanges of nude men in the cold; effects of humidity temperature and windspeed. Journal of Applied Physiology, 12:351–6.
  • ——J. A. Vaughan, R. F. Goldman, M. B. Kreider, F. Masucci, and D. E. Bass, 1960. Heat production from shivering. Journal of Applied Physiology, 15:632–4.
  • Jones, J. S., 1981. How different are human races? Nature, 293:188–90.
  • Katayama, K., 1985. Human skeletal remains from Gilbert Islands. Man and Culture in Oceania, 2: 81–120.
  • Kaufmann, M. S. and D. J. Bothe, 1986. Windchill reconsidered. Aeronautical, Space and Environmental Medicine, 57:23–6.
  • Kean, M. R. and P. Houghton, 1982. The Polynesian head: growth and form. Journal of Anatomy, 135:423–35.
  • —— and P. Houghton, 1987. The role of function in the development of human craniofacial form. The Anatomical Record, 218:107–10.
  • Keatinge, W. R., 1969. Survival in Cold Water. Oxford: Blackwell.
  • Kirch, P. V., D. R. Swindler and C. G. Turner, 1989. Human skeletal and dental remains from Lapita sites (1600–500 B.C.) in the Mussau Islands, Melanesia. American Journal of Physical Anthropology, 79:63–76.
  • Kowalski, C. J., 1972. A commentary on the use of multivariate statistical methods in anthropometric research. American Journal of Physical Anthropology, 36:119–32.
  • Lai, L. Y. C. and J. Bloom, 1982. Genetic variation in Bougainville and Solomon Islands populations. American Journal of Physical Anthropology, 58:369–82.
  • Larnach, S. L. and N. W. G. Macintosh, 1971. The Mandible in Eastern Australian Aborigines. Oceania Monographs No. 17. Sydney: University of Sydney.
  • Lewis, D., 1972. We, the Navigators. Wellington: Reed.
  • Lister, A. M., 1989. Rapid dwarfing of red deer on Jersey in the last interglacial. Nature, 342:539–42.
  • Lourie, J. A., 1972. Anthropometry of Lau islanders, Fiji. Human Biology in Oceania, 1:273–7.
  • Maclean, D. and D. Emslie-Smith, 1977. Accidental Hypothermia. Oxford: Blackwell.
  • McArthur, N., I. W. Saunders and R. L. Tweedie, 1976. Small population isolates: a microsimulation study. Journal of the Polynesian Society, 85:307–26.
  • Molnar, S., 1983. Human Variation. Englewood Cliffs: Prentice-Hall.
- 195
  • Monneron, P. 1914. Journal of a voyage made on the vessel Le St Jean Baptiste., in R. N. McNab (ed.), Historical Records of New Zealand. Wellington: Government Printer, Vol. 2:263–95.
  • Montague, M. F. A. 1960. An Introduction to Physical Anthropology. Springfield: Thomas.
  • Naval Intelligence Division, 1945. Geographical Handbook Series Pacific Islands. Volume 4. London: Admiralty.
  • O'Connell, J. F., 1972. A Residence of Eleven Years in New Holland and the Caroline Islands. Pacific History Series No. 4. Canberra: Australian National University Press.
  • O'Hara, W. J., C. Allen, R. J. Shepard and G. Allen, 1979. Fat losses in the cold — a controlled study. Journal of Applied Physiology, 46: 872–7.
  • Oliver, D. L., 1961. The Pacific Islands. New York: Doubleday.
  • Pawley, A. K. and R. C. Green, 1973. Dating the dispersal of the Oceanic languages. Oceanic Linguistics, 12:1–67.
  • Pietrusewsky, M. P., 1976. Prehistoric human remains from Papua New Guinea and the Marquesas. Asian and Pacific Archaeology Series, No. 7.
  • Prior, I. A. M., F. Davidson, C. E. Salmond and Z. Czochanska, 1981. Cholesterol, coconuts and diet on Polynesian atolls. American Journal of Clinical Nutrition, 34:1552–61.
  • Pugh, L. G. C., 1963. Tolerance to extreme cold at altitude in a Nepalese pilgrim. Journal of Applied Physiology, 18:1234–8.
  • —— 1966. Accidental hypothermia in walkers, climbers and campers. British Medical Journal, 1:23–129.
  • —— 1967. Cold stress and muscular exercise, with special reference to accidental hypothermia. British Medical Journal, 2:333–7.
  • —— and O. G. Edholm, 1955. The physiology of channel swimmers. Lancet, 2:761–8.
  • Rennie, D. W., B. G. Coving, B. J. Howell, S. H. Song, B. S. Kang and S. K. Hong, 1962. Physical insulation of Korean diving women. Journal of Applied Physiology, 17:961–6.
  • Rhoads, J. G., 1984. Improving the sensitivity, specificity, and appositeness of morphometric analyses, in G. N. van Vark and W. W. Howells (eds), Multivariate Statistical Methods in Physical Anthropology. Reidel: Dordrecht, pp.247–59.
  • Roberts, D. F., 1953. Body weight, race and climate. American Journal of Physical Anthropology, 11: 533–58.
  • Rosing, F. W., 1984. Discreta of the human skeleton: a critical review. Journal of Human Evolution, 13:319–23.
  • Roth, H., 1972. Two on a Big Ocean. New York: Macmillan.
  • Roy, K. and N. Tayles, MS, 1989. The People of Afetna. Unpublished manuscript, University of Otago.
  • Schlaginhaufen, O., 1929. Zur Anthropologie der Micronesischen Inselgruppe Kapingamarangi (Greenwich-Inseln). Archiv der Julius Klaus-Stiftung für Verebungsforschung, Sozialanthropologie und Rassenhygiene, 4:219–87.
  • Schreider, E., 1950. Geographical distribution of the body-weight/body-surface ratio.
- 196
  • Nature, 165: 286.
  • Serjeantson, S. W., D. P. Ryan and A. R. Thompson, 1982. The colonization of the Pacific: the story according to human leucocyte antigens. American Journal of Human Genetics, 34:904–18.
  • Shapiro, H. L., 1937. The physical characteristics of the Ontong Javanese. Anthropological Papers of the American Museum of Natural History, 23:225–78.
  • Siple, P. A., 1968. Clothing and climate, in L. H. Newburgh (ed.), Physiology of Heat Regulation. New York: Hafner, pp.389–442.
  • Sjovold, T., 1984. A report on the heritability of some cranial measurements and nonmetric traits, in G. N. Van Vark and W. W. Howells (eds), Multivariate Statistical Methods in Physical Anthropology. Dordrecht: Reidel, pp.223–46.
  • Spealman, C. R., 1968. Physiological adjustments to cold, in L. H. Newburgh (ed.), Physiology of Heat Regulation. New York: Hafner, pp.232–39.
  • Stini, W. A., 1979. Adaptive strategies of human populations under nutritional stress, in W. A. Stini (ed.), Physiological and Morphological Adaptation and Evolution. The Hague: Mouton, pp.387–407.
  • Stoneking, M. and A. C. Wilson, 1989. Mitochondrial DNA, in A. Hill and S. Serjeantson (eds), The Colonization of the Pacific: a Genetic Trail. Oxford: Clarendon Press, pp.215–45.
  • Thompson, L., 1940. Southern Lau; an Ethnography. Bulletin 162. Honolulu: Bernice P. Bishop Museum.
  • Trent, R. J., K. N. P. Mickleson, T. Wilkinson, J. Yakas, M. W. Dixon, P. J. Hill and H. Kronenberg, 1986. Globin genes in Polynesians have many rearrangements including a recently described yyyy. American Journal of Human Genetics, 39:350–60.
  • Trotter, M., 1970. Estimation of stature from intact long bones, in T. D. Stewart (ed.), Personal Identification in Mass Disasters. Washington: Smithsonion Institution, pp.71–83.
  • Vanggaard, L., 1975. Physiological reactions to wet-cold. Aviation Space and Environmental Medicine, 46:33–6.
  • Webster, A. P., 1952. Caloric requirement of man in cold climates. Journal of Applied Physiology, 5:134–42.
  • Weidenreich, F., 1936. The mandibles of Sinanthropus pekinensis. A comparative study. Polaeontologica Sinica, Series D, 7:1–132.
  • —— 1947. Facts and speculations concerning the origin of Homo sapiens. American Anthropologist, 49:187–203.
  • Wells, L. H., 1969. Stature in earlier races of Mankind, in D. Brothwell and E. Higgs (eds), Science in Archaeology. London: Thames and Hudson, pp.453–67.
  • Wilson, S. R., 1984. Towards an understanding of data in physical anthropology, in G. N. Van Vark and W. W. Howells (eds), Multivariate Statistical Methods in Physical Anthropology. Dordrecht: Reidel, pp.261–82.
  • Wolpoff, M. H., X. Z. Wu and A. G. Thorne, 1984. Modern Homo sapiens origins: a general theory of hominid evolution involving the fossil evidence from East Asia, in F. H. Smith and F. Spencer (eds), The Origins of Modern Humans. New York: Alan R. Liss, pp.411–83.