Acknowledgments

The period spent by me on Mangaia extended from December 1929 to April 1930, the months alluded to as the hurricane season. It was the slack time of the year when no vessels visited and the fruit trade was suspended. As a return for the assistance of the Cook Islands Administration, I agreed to act in the place of the Resident Agent, Mr. J. McGruther, during his absence on leave, and was officially gazetted a Commissioner of the High Court of the Cook Islands to try all offences committed during my term of office. The Hon. Sir Apirana Ngata, Minister for the Cook Islands in the New Zealand Cabinet, and Judge Hugh Ayson, Resident Commissioner, who were both interested in the ethnological survey, appointed me acting Resident Agent for the months of my stay on Mangaia as much to facilitate the survey as to assist the administration.

Except for morning sick parades, the weekly court, which occupied less than an hour, and the monthly paying of civil servants, my official duties made little demand on my time. In my official position I was authorized to use the resources of the government in such way as I deemed expedient. The government, therefore, devoted its attention to the ethnological survey. The police in the villages acted as assistants in gathering the people together for the purpose of making head and body measurements. The district and subdistrict chiefs were called upon to conduct official visits of inspection through their districts. The maraes and old battlefields were visited and described by local experts. Prisoners utilized to clear the growth off the maraes took as much interest in practical archaeology as did the Resident Agent. To Sir Apirana and Judge. Ayson my thanks are due for giving an ethnologist such a unique opportunity for carrying out field work. To the ariki Matekeiti and the district and subdistrict chiefs who form the Island Council of Mangaia my personal thanks are rendered for their hearty coöperation. My thanks are due to all my informants, but particularly to Aiteina and Akaeakore for the Ngati-Vara genealogies and history and to Daniela Tangitoru for information on certain crafts. I thank Mr. and Mrs. J. McGruther for the use of their residence and equipment. My wife assisted page 4me as recorder in the anthropological work and took charge of the reciprocal hospitality so necessary in the campaign of a field worker in Polynesia.

Geography

Mangaia, the island farthest south in the Cook archipelago, lies no miles southeast of Rarotonga. Marshall (16, p. 1)^* records its position, its circumference as 17.62 miles. Its area is estimated as 17,500 acres (26, p. 361).

In the absence of a complete official map, Marshall (16, pl. 1) made a reconnaissance of the whole island in 1923. He was assisted by his native guide, Mama Akaui, who knew the localities on the western side of the island, but some of the information given by him for other parts is inaccurate. His greatest lapse is the omission of the district of Karanga, which lies between the districts of Tavaenga and Ivirua, and completes what are known as the "six districts" (puna) of Mangaia. This subject was thrashed out with me on the slopes of Karanga by a group of old men with Marshall's map before them. They indicated the boundaries of Tavaenga and Karanga as shown on the readjusted map (fig. 1).

As the physical features of the island have had a marked influence on the culture of the people, Marshall's physiographic outline (16, p. 7) is quoted:

The island is surrounded by a coral reef, which in most places has a width of not more than 300 feet; at one place its width is nearly 1,000 feet. From the inner side of the reef platform the coast rises abruptly to a height of 10 to 30 feet and continues as a gentle and more or less uniform slope, which rises 20 feet in a distance of about 600 feet. From the inner edge of the slope a cliff 65 feet high rises abruptly, forming a wall which may be climbed at very few places. Extending inward from the top of the cliff is a broad surface, from which project innumerable sharp limestone pinnacles, when travel is impracticable except where paths have been made. This raised platform of eroded coral rock is called the Makatea; a little more than a half mile from the coast its surface stands about 200 feet above sea level. At its inner edge, the Makatea is terminated by a steep cliff which descends 150 feet or more—in places almost to sea level—and which is unscalable for the greater part of the circumference of the island. At the foot of this inward-facing escarpment lie extensive swamps which receive all the run-off from the central part of the island. The swamps are drained by subterranean channels passing to the sea through the Makatea at depths of more than 150 feet. The central part of the island is a dissected mass of volcanic rock which rises gradually to a height of 554 feet above sea level. Generally this igneous core is separated from the Makatea by the moat which contains the swamp, but in a few places projecting spurs extend downward to the Makatea platform. The summit of the island is a flat surface three quarters of a mile long and less than a quarter of a mile wide. The central point of this flat is called Rangimotia, the "Crown of Mangaia."

The fringing reef which surrounds Mangaia has, according to Marshall (16, p. 10), a raised "Lithothamnion ridge" 60 to 90 feet wide. The seaward edge, exposed to a height of about 2 feet at low water, is explored by women-folk for shellfish and echini. Deep chasms, varying in width and depth, page break page 5penetrate the rim. Where they cross the rim, they form channels (ava), which are useful for launching small canoes. The channels, however, do not penetrate far, and none of them reach the shore. An attempt has been made at Oneroa to enlarge and lengthen one of these ava to provide a boat passage to the shore. The lack of suitable natural passages has affected the method of launching and beaching canoes.

From the inner side of the raised outer border of the reef there is a sudden downward slope of 2 or 3 feet. The reef flat, which extends from here to the shore, ranges from 80 to 300 yards wide; its surface is always below low-water level. The flat is covered at high tide. The stretch of water between the raised rim and the shore is usually alluded to as the "lagoon." Depressions pitted in the uneven flat form pools when the tide recedes. Some pools are fairly large and are frequented by fish which come in and out with the tides.

The reef flat is bordered inland by a "terrace slope" 300 to 600 feet wide. Before the advent of Christianity the terrace slope was not permanently occupied, but the national marae of Orongo is located on the terrace and at times the Shore High Priest of Rongo took up his residence near the marae. Two divisions of the village of Oneroa are now located on the terrace. Coconuts have been planted all along the coast.

The makatea is penetrated by many crevices and caves which once gave refuge to fugitives after battles. It is probable that some tribes would have been exterminated but for the concealment thus provided. Access to the coast was by steps up the inner and outer faces of the makatea connected by narrow winding paths that avoided the pinnacles and crevices. Plants grow in the red friable soil between the pinnacles. Taro flats form the puna lands over which so much fighting took place. Marshall (16, p. 35) states:

The taro flats are about 40 feet above sea level, where the streams that supply the water to them issue from the valleys. They descend by a series of artificial terraces and end about 20 feet above sea level at the base of the makatea wall. Here the water enters a sink, which may be as much as 10 feet below the general surface, and passes in channels through the makatea, finally issuing on the inner side of the reef flat in the form of springs.

The volcanic slopes in the center of the island are covered with fern (Gleichenia) and support groves of toa (Casuarina). The stream valleys have dense thickets of small trees and shrubs. The whole central region, referred to as "the mountain" (maunga), gave cover to fugitives belonging to defeated tribes.

The People

The Mangaians are brown-skinned Polynesians. The anthropometrical measurements of 200 males and 100 females show them to be of medium page 6stature; the average is below that of the neighboring Rarotongans. The men are good workers and much in demand for labor.

The population in 1823 was estimated by Williams (27) as between 2,000 and 3,000. The government census in 1926 gave a total of 1,241, consisting of 636 males and 605 females. Previous census reports show that the population has decreased steadily since 1823 and reached its lowest number (1,230) in 1921. The village of Titikaveka in Rarotonga contains a large proportion of Mangaians. Some of the decrease was due to emigration to Rarotonga in search of employment. The decrease from introduced and endemic diseases was severe, though the missionaries did their best with the remedies they had. The Health Service at Rarotonga, which now makes occasional visits to Mangaia, has done much to check the excessive mortality through campaigns directed against venereal disease, hookworm, and yaws. Mrs. McGruther, a trained nurse, attends regularly to all sickness and has a well-equipped dispensary.

The Language

The Mangaian dialect closely resembles that of the other islands in the Cook archipelago, but contains a number of local words. The alphabet inaugurated by the missionaries contains 5 vowels, a, e, i, o, and u, and 8 consonants, k, m, n, ng, p, r, t, and v. The wh of Manihiki and New Zealand and the f of Tahiti are absent.

The letter h was omitted by the missionaries from the alphabet compiled for Rarotonga. Mr. Stephen Savage and I once held that this letter should have been included, as we thought that the h sound was represented in the spoken language (22, p. 22). I am now of the opinion that the sound is, perhaps, more truly a glottal closure, and would be represented better by the hamzah. Both K. P. Emory and J. F. Stimson of Bernice P. Bishop Museum were already of this opinion, and they have made me doubt the accuracy of my original statement. I have always objected to hearing the Taranaki and Whanganui tribes called the "cockneys of New Zealand," because I could always hear the h sound in their speech. It is probable from my Cook Islands experience that what I heard was the glottal closure and not the fully aspirated h sound. The use of the glottal closure gives the speech a jerkiness which is a marked feature of the Whanganui subdialect of Maori and is sometimes humorously imitated by other tribes. This jerkiness probably results from the application of the glottal closure to the h sound. I had long noticed that the Cook Islands use of the h resembled that of the Taranaki and Whanganui tribes, and now feel that the similarity lies in the use of the glottal closure. In the native text of this paper the h is represented by the hamzah.

If the letter h had been introduced into the alphabet, however, there would have been much less confusion. For example, ua may represent ua (rain), page 7uha (female) or huha (thigh). Similarly aa is evidently meant for aha (what), and there is nothing to indicate that it may be meant for haha (to feel). Sometimes the problem can be solved only by hearing a word pronounced. In Manihiki, Rakahanga, and Tongareva the aspirated h is used.

Though the k sound is used, it has been dropped, as in Tahiti, in a number of words: aore (kaore), ia (kia), o (ko), and uete (kumete). In the native text in this work the hamzah has not been used for a dropped k.

The m sound has been dropped in two words, kuara for kumara (sweet potato), and uete for kumete (wooden bowl), but there is no glottal closure.

Whether or not some words written with v are sounded as though written with w is doubtful in Mangaia as in Aitutaki (23, p. 24).

Christian (3) has compiled a Mangaian vocabulary.

Sources

The native culture of Mangaia has been profoundly affected by Western culture for more than a century. The inferiority of many of the old customs and institutions was so impressed upon the minds of the Mangaians by their new teachers that they not only gave them up, but even tried to forget they had ever existed. When I asked an old man what the old people (ai metua) thought of a certain native institution, he replied, "Pe'ea taua e kite i te manako a te 'etene?" (How can you and I understand the thoughts of the heathen?)

The field worker is forced to seek for details in the printed pages of early observers, many of whom present biased pictures. The Rev. Wyatt Gill, a resident missionary who was able, because the old men alive in his time had grown up before the advent of the first missionaries in 1823, to record much of the history of Mangaia and many of her songs, wrote largely to interest the British public in the work of the London Missionary Society. In his eight published books (5, 6, 7, 8, 9, 10, 11, 12) there are more detailed stories about murders, human sacrifices, and cannibalism than about the more constructive institutions of Mangaian culture. Yet Gill has covered Mangaian history from the first native settlements to the advent of Christianity. I have used Gill's material freely in order to piece together the details scattered through his different books to form a continuous historical narrative which serves as the basis for my study of the rise of military power in Mangaian social organization.

In estimating the duration of occupation of Mangaia, Gill used lists of priests, though his informant, Mamae, had full pedigrees of the Ngati-Vara tribe. These pedigrees, supplied by Aiteina, are here used for the first time. Much valuable information compiled by the Ngati-Vara bards is contained in the songs given by Gill in connection with historical events. I have not only page 8introduced the hamzah into the texts, but for most of the songs quoted I have used my own more literal translations. The native text regarding the mythological origin from Vatea is given in full in this paper, both to show the points of difference from Gill's English record and to supply a native text for the use of students of linguistics.

I have also had access to a manuscript by Mamae, an informant of Gill and former pastor of the church at Mangaia.

A source book drawn principally from government records was compiled for my use under the direction of Sir Apirana Ngata. The information concerning the periods immediately preceding and following the proclaiming of the British Protectorate has been of invaluable assistance.