Volume 71 1962 > Volume 71, No. 4 > Some bases for ecological inferences about the aboriginal population of the Hanapepe Valley, Kauai, by Richard Pearson, p 379-385
                                                                                             Previous | Next   

- 379

The research for this paper was done in 1961 at the Bernice P. Bishop Museum, and the University of Hawaii, where the author was a graduate student in Anthropology, as a Woodrow Wilson Fellow.

ECOLOGICAL RECONSTRUCTION through archaeology is in its infancy in Oceania. With the exception of New Zealand, 2 problems of this nature have not been dealt with in any island groups in Polynesia, Micronesia, or Melanesia. In this paper we evaluate historical reconstruction, archaeology and pollen analysis—three techniques for eco-historical integration being developed in the current archaeological programme in Hawaii. We feel that our methods are of interest not only to scholars of the Pacific World, but also to those of other tropical areas.

The Hanapepe Valley on the island of Kauai, Hawaii, was chosen for field work because it was of appropriate size for the time and resources at hand, and it had not been the scene of recent drastic changes. In addition, it had not been used for unusual agricultural ventures, which might have altered the terrain, such as those of the Koloa Valley, 10 miles to the east of Hanapepe, where many introduced crops were tested in the last century. In the Hanapepe Valley, the growth over the previously cultivated areas was not as dense as in other cultivated areas.

The key problem to be investigated was the extent of pre-contact cultivation and its location.

The problem was not simply the location of taro plot ruins in the valley, for it is full of such ruins from back to front and side to side. Rice production was introduced into the area in the late nineteenth century and, if the inhabitants followed the general pattern of contemporary Hawaii, Chinese began to grow taro shortly after that time. Therefore, the problem was to sort out which remains could be attributed to the Hawaiians, even in instances in which they were used by the immigrant population at a later date. It appears that the Chinese adapted their food production techniques to the local situation with such perfection that there is almost no break in the continuity of agricultural practices between the aboriginal and immigrant populations with respect to valley bottom-lands.

- 380

Wendell Bennett 3 stated that the entire valley floor of Hanapepe was under wet-land taro plot (lo'i) cultivation, wherever it was possible to irrigate, in pre-contact times. It is certainly true that, if we consider the long auwai (ditch) system used to bring water to the taro, in relation to the scattered areas which are known by land-claim records to have been Hawaiian taro plots, we can only conclude that the intervening areas, in which there are remains of many terraces, must all have been used for taro, or else the system would not work. The Hawaiian ahupua'a, or land division, which ran from the mountains to the sea, operated so as to assure the utilization of various environments by a single community; therefore, we should expect entire portions of the valley to be under cultivation and some settlement. In Tahiti and Samoa, communities did exist inland, and the same holds true for Hawaii, in the more favourable localities.

Our most complete demographic information comes from population figures. In 1838, the population of Hanapepe and the adjacent area of Waimea was recorded at 3,372. 4According to missionaries at that time, there were several reasons for considering this figure too low. The general population of the Hawaiian Islands decreased by two-thirds between 1832 and 1878, and Fornander 5 wrote that the depopulation continued at the same rate between 1778 and 1832. Although we cannot assume that the depopulation was constant on all islands and in all areas, we can establish that the population of Hanapepe, the smaller portion of the population of Waimea and Hanapepe, was several times as large in aboriginal times as it was at the time of the first census. (New figures show that the population of Hawaii was 300,000 in 1778, and that in 1835 and 1836 it was 107,000 6). Allowing for the depopulation, we should be able to state that the population in pre-contact was two or three times as large as that of the first census.

One may wonder why we did not simply resort to a list of water rights awarded to the Hawaiians for their taro plots after the Great Mahele (the land division in 1853, in which many Hawaiians secured titles to their taro lands from Kamehameha III). In Ancient Hawaiian Civilization 7 it is stated that although 1,000,000 acres were set aside for the common people, only 8,000 to 11,000 claims (each of which was termed a kuleana) were staked, embracing some 28,000 acres. In many cases the konohiki (land supervisor in the employ of the ali'i or nobility) discouraged people from claiming land for themselves. Chinen 8 points out that, in the Hawaiian Islands, 1,500,000 acres were reserved for the chiefs, 1,000,000 acres were retained by Kamehameha III, and 1,500,000 were given by the king to the government and the people. While the 28,000 acres claimed were supposed to be the - 381 most suitable for taro production, the chiefs and the king must have retained a considerable acreage of productive land. It seems certain that all the state land in the valley at present is suitable for taro; this land, reserved from the time of the Mahele, with later additions, was not distributed during the Mahele, and is not considered in land-claim records to be taro-land.

Those kuleanas which were claimed in the Hanapepe Valley after the Mahele were for the subsistence cultivation of indivdual families. They would have grown very little more than what was required to feed the families, since the political system had collapsed and there was no market for taro on Kauai. 9 The aboriginal population, on the other hand, would have required not only enough taro-producing land to feed itself, but also enough to support the king and the ali'i.

In consideration of the above material, we begin with a datum line consisting of the land awarded water rights because it was claimed at the time of the Mahele. The acreage must be multiplied by a coefficient to compensate for several factors. These are (1) the small proportion of aboriginal farmers making claims at the time of the Mahele, as compared with those living at the time, (2) the reduced population to which the claimants belonged, and (3) the reduced amount of land worked by the historic people when surpluses were not required by the political hierarchy. In addition to this, the state land at the head of the valley should also be added. Although our quantification is not to be taken too seriously, we submit that of the approximately 300 acres of deeded land below the junction of the tributaries of the Hanapepe River, the total (as a minimum) was utilized in aboriginal times; whereas, there are only 97.87 acres claimed on available records.

A century ago, an observer stated that the immigrants to Hawaii would re-use the lo'is already established by the Hawaiians; “The land is prepared and ready and only wants the same process of clearing as a kalo patch prepared for kalo planting. In opposition to the heavy expenses which the preparation of rice fields requires in South Carolina, I would mention only the hundreds and perhaps thousands of uncultivated kalo patches whose owners and planters have long since gone.

“. . . there they lie with their rich clayey soil ready at hand, possessing all the most desirable conveniences, level embankments, and streams of the richest (I mean chemically) water, fresh from the mould and decaying rocks of the mountains.” 10


Historical sources state that the walls of the ancient Hawaiian taro plots were wide: “In the spaces between the fields, which are from three to six feet broad, there are very pleasant shady avenues and on - 382 both sides bananas and sugar cane are planted.” 11 All the early writers agree on the width of the walls except Vancouver, 12 who stated that the walls in the Waimea Valley were wide enough for only one person at a time. It seems likely that the was referring to the low separation walls in the level areas. It is commonly held by the local inhabitants that the walls in Hawaiian times were much wider than in the time of the Chinese growers, and that often the Chinese purposely narrowed the walls to produce more growing space.

One wall which we cross-sectioned, in an area known to be Hawaiian, was between 13 and 15 feet wide. Probably the collapse of the upper portion of it and the subsequent flooding of the river exaggerated its proportions.

We also investigated archaeologically whether it was possible to distinguish between the wall composition of the Hawaiian and non-Hawaiian taro plots. Historic sources state that the Hawaiians often put heavy leaves in the edges of walls in order to secure them and make them more water-tight. In our excavations we could find no traces of such vegetable material, although under microscopic analysis the soil of the valley in general contains a great deal of vegetable matter. Modern growers in the valley, with whom we discussed taro cultivation, use only earth in building up the walls, while those of ancient times seem to have used stones or earth depending upon the availability of stone and the inefficiency caused by leaving it in the fields. On the basis of trenching, we concluded that it was not possible to distinguish structural differences in the walls which could be correlated with Hawaiian or immigrant land use.


It was suggested that we attempt pollen analysis of soil samples, to see if, by any chance, taro pollen could be found in an antecedent relationship to the pollen and/or foliage of rice and other plants which might have been grown by the Japanese or Chinese in deserted taro plots. Although cultivation techniques for both rice and taro crops included turning over the bottoms of the plots, it was hoped that the continuous supply of water and the silting of the bottoms of the lo‘is would make a stratigraphic relationship evident, if it occurred. With success, we could then state whether or not a particular area of land had been under Hawaiian taro cultivation even if its superstructure of walls and terraces had been modified recently.

In the analyses planned the aim was to recover traces of crop plants, and as most could not easily be identified even to genus (as for rice), the emphasis fell on a search for taro pollen for which there is at present only one very dubious published record—from delta deposits in India.

Dr. Lucy M. Cranwell, a Research Associate at the University of Arizona, at Tucson, offered to do a pilot study of some 30 soil samples, mainly from excavations in areas in which the history of the taro plots - 383 was not known: three of the five localities were not mentioned in the very incomplete Mahele records referred to earlier. From all localities specimens were taken at 6-inch intervals from the surface to the bottoms of the plots: the samples were treated with glycerine to keep them moist and were shipped at once to Tucson. In addition samples were selected from a taro plot whose cultivation had been abandoned within the last 25 years. The investigator approached the work with little optimism, as many difficulties had to be faced:

  • (1) Analyses of non-peaty surface samples from other lowland subtropical areas are not available for comparison.
  • (2) In both subtropical and tropical lowland regions heat combined with alternate wetting and drying had always been considered inimical to the preservation of waxy coats of pollen grains and spores—the most enduring parts of plants—and poor results could thus be expected in Hawaii where running irrigation water from the river would greatly increase aeration and thus destruction of these membranes.
  • (3) Few of the pollen types could be identified with certainty as little had been done with them, except by Dr. O. Selling, 13 who found no traces of cultivated plants, of course, in the Hawaiian montane bogs he studied. A collection of comparative material was thus needed and it proved very difficult to get taro pollen, in particular, since most herbaria had sterile plants only. However, fresh collections of rice, sugar cane and coconut flowers were made for Dr. Cranwell and foliage as well as taro was forwarded for comparative anatomical purposes.
  • (4) A special difficulty lay in the separation of taro pollen from that of other spiny-walled monocotyledonous types which might also have grown in Hawaii.
  • (5) In ancient practice the taro farmer plucked the infrequent scapes (flower-heads) of taro to conserve the energy of the plant. This could result here and there in an aggregation of immature grains, if any were preserved, but it is clear that the scapes might not be buried deeply enough for preservation under the conditions prevailing in a plot during any one season, or decade.

On the credit side it should be mentioned that since pollen grains are now being found with increasing success in surface soils of temperate climates, it was hoped that survival potentialities in Hawaii might be greater than many had anticipated.

The preliminary studies showed a uniform condition—the presence of a great bulk of organic debris, often with scrape of monocot tissue intact. The density of this material promises well for further work of an intensive nature, but tends to defeat the search for microspores. Much more work would be needed to segregate pollen and spores which are now hidden amongst debris. Some microspores were recognized in clear view, but their frequencies suggested that the whole pollen and spore content was very low, and limited to the more resistant types such as fern spores.

- 384

Unfortunately the fern spores tell us little about local conditions. They may be drawn from some distance and are thus indicators of regional rather than local conditions. The recovery of taro pollen, on the other hand, would be purely local, as the grains are insect-pollinated and thus never carried far.

Taro pollen was found only in the recently cultivated plots. Further research might well locate it in other areas, however, as many of the dominants of the forested surroundings were not represented at all by their pollen. This is an unusual state of affairs to be explained only by the eroding nature of a puggy matrix in a hot climate.

The results are thus not disappointing as they prove for the first time that taro pollen can survive. It is likely that higher frequencies of the grains will be found in more favourable localities in Hawaii, and especially in the cooler areas at the limits of migration of the Polynesian peoples.

Altogether the rich organic content points to long cultivation under wetland conditions. Whether taro alone was cultivated, or whether rice followed, is not clear. Grass pollen was so rare that the presence of rice (a member of the grass family) seems doubtful. (See Appendix I.)


As the problems of establishing actual chronologies are met, our interest may shift from these to equally pressing ecological and demographic questions. Armed with factual material—the most useful of which is historical—researchers may produce testable hypotheses which can be verified through specific field projects. We have shown three possible techniques for the substantiation of such hypotheses. Although the methods are rough and often ambiguous, they are useful, and await further refinement.


Fungal bodies: Not common.

Fern spores: Comparatively abundant.

  • 1. Polypodiaceae
  • 2. Gleicheniaceae
  • 3. Tree-fern types

Lycopod spores: Occasional.

Pollen grains:


  • Grasses: Very rare in all preparations. One may be of cereal type.
  • Sedges: One grain seen.
  • Others: Varied and fairly abundant.
  • Taro (Colocasia): Three grains recovered. They are very small (about 25µ).
- 385


  • Chenopods: Rare.
  • Myrtaceae: Occasional.
  • Others: Various, not identified.

Tissue: Fragments reminiscent of taro.

Wood fragments: Present but probably not identifiable.

  • BENNETT, W. C., 1931. The Archaeology of Kauai. Honolulu. Bernice Bishop Museum, Bulletin 80.
  • CHINEN. J., 1957. The Great Mahele. Honolulu. University of Hawaii Press.
  • FORNANDER, A., 1878. An Account of the Polynesian Race—Its Origin and Migrations. London. Trubner & Co.
  • HANDY, W. C., EMORY, K. P., BRYAN, E., BUCK, P., et al., 1933. Ancient Hawaiian Civilization. Honolulu, Kamehameha Schools.
  • KOTZEBUE, Otto von, 1821. A Voyage of Discovery into the South Sea and the Bering Straits. London.
  • Pacific Commercial Advertiser. Sept. 9, 1858. Honolulu.
  • PEARSON, R. J., 1960. The Extent of Hawaiian Taro Patches in the Hanapepe Valley, Kauai. (Manuscript in the B.P. Bishop Museum, Honolulu.)
  • SELLING, O., 1947. Studies in Hawaiian Pollen Statistics. Honolulu. Bernice P. Bishop Museum, Special Publication No. 38. 3 volumes.
  • STATE OF HAWAII, 1961. Population Trends in Hawaii, 1778-1960. Honolulu. Research Report No. 3, April 11, 1961. Dept. of Planning and Research. (Mimeographed copy.)
  • TAYLOR, N. H., 1958. “Soil Science and New Zealand Prehistory.” New Zealand Science Review, 16:71-79. New Zealand.
  • VANCOUVER, G., 1801. A Voyage of Discovery to the North Pacific and Around the World. London, Stockdale.
  • WILKES, C., 1845. The Narrative of the United States Expedition During the Years 1838, 1839, 1840, 1841, and 1842. Philadelphia. Lean and Blanchard.
1   I wish to thank Dr. K. P. Emory of the Bishop Museum, and Dr. L. Cranwell of the University of Arizona, for their advice and inspiration. The material on pollen analysis in this paper is the work of Dr. Cranwell.
2   Taylor 1958.
3   Bennett 1931:114.
4   Wilkes 1845, 4:165.
5   Fornander 1878, 2:165.
6   State of Hawaii 1961.
7   Handy, Emory, Bryan and Buck 1933.
8   Chinen 1957:31.
9   Vancouver (1801, 2:198) states that during his second visit he could see the effects of war in the neglect of the irrigated taro plots on Maui. This shows that the disruptive effects of European contact were evident very early.
10   Pacific Commercial Advertiser, Sept. 9, 1858.
11   Kotzebue 1821, 1:340.
12   Vanvouver 1801, 2:164.
13   Selling 1947.