Volume 39 1930 > Volume 39, No. 153 > An evaluation of early genealogies used for Polynesian history, by John F. G. Stokes, p 1-42
                                                                                             Previous | Next   

- 1

IN this paper I attempt to clarify what, to my mind, have been certain misconceptions of native information on the part of two noted authorities on Polynesian traditions, namely Abraham Fornander (6) of Hawaii and S. Percy Smith (22) of New Zealand. The contribution of these men are most valuable and will, I hope, stand as monuments in Polynesian research and continue to guide many students. They will stand better, I believe, if the weak places be pointed out. These weak places are certain genealogies or traditions which affect in particular the chronologies suggested by Fornander and Smith, and certain views of Fornander. I shall follow especially the matter of chronology, and shall include such side-lights as the study may warrant.

Fornander and Smith worked along similar lines, rationalizing the legends, utilizing the traditions, calculating dates by genealogies, and blending the whole into works of real historical value. They also reached similar conclusions in regard to early Polynesian migrations. Fornander brought his proto-Polynesians through India, while Smith was content to originate his within India. Both had them resident in Indonesia at the beginning of the Christian era. Fornander moved his Polynesians into Polynesia near the end of the first century or the beginning of the second century A.D.; Smith, at some indefinite time between A.D. 50 and A.D. 450. When it is known that Fornander worked primarily on Hawaiian traditions, and Smith on southern-Polynesian traditions, their views would appear to be mutually confirmatory. The probability of chronological accuracy was strengthened by apparent agreements in genealogies of New Zealand, Rarotonga, Tahiti, and Hawaii, - 2 twenty-six and twenty-eight generations before 1900 (22, pp. 20-1), and the assumption was made that the earlier dates were also of probable acceptance.

However, a weakness in the chronology will at once be suspected when it is remembered that Fornander based his chronology on a generation of thirty years, while Smith preferred twenty-five years. Certain portions of the genealogies, accepted as of human beings, are or seem to be parts of cosmogonies. Both authors assumed an approximate unity of race within Polynesia, and made insufficient allowance for the possible existence within the region of other races or peoples without traditions, or of peoples whose traditions may have been eliminated by those whose records have been preserved.


Confidence in the results is not increased by a knowledge of the wide disagreements among the genealogies preserved in the different island-groups. It would be expected that since Fornander and Smith were writing sketch-histories of the Polynesian people, there would be, in order to assure a tolerable accuracy for a chronology, a fair agreement in 1, the number of generations; 2, cosmogonic and mythical human names; and 3, human names down to the traditional separations through migrations.

The agreements are negligible:—

1. In the accompanying Table 1 of generational counts prior to 1900, we may note a range of 30 to 154 generations.

TABLE 1. Range of Polynesian genealogies in generations prior to 1900 A.D.
  Tregear (28, p. 667) Smith (20 and 22)
Maori 39-139 52
Moriori 154 105
Rarotonga 30 38-138
Society 36 40
Marquesas 146 74-140
Hawaii 78 93
Samoa 57 55
Tonga 37 35

The divergences, it will be observed, are not merely great between the island-groups, but also between the genealogies - 3 of each island. Later I shall show that Smith's main authority in Rarotonga is responsible for recording two genealogies which differ in their counts of generations by about one-third. Except for one of these genealogies, Smith (22, pp. 29 ff.) attached little value to portions more than forty or fifty generations prior to 1900.

2. The cosmogonic and mythical human names are seldom to be found either in close agreement, or in such order as to establish their identity in the different genealogies.

3. The human names in agreement either in genealogical order or in identity of succession are so few as to be exceptions and not the rule in the portions preceding the migrational epochs.


Fornander and Smith were both handicapped in the genealogical comparisons through the Polynesian habit of name-changing. This was such that a chief might be referred to by many different names in as many genealogies. However, it is difficult to believe that this is the only explanation. Probably not as many as five per cent. of the names prior to the latest migrational epoch are in agreement in fair identity, succession, and genealogical position. There is a possibility of course that the chiefs received posthumous names, as did the Japanese emperors, (4) but no note has yet appeared concerning such as a Polynesian custom. This might be worth while following up in New Zealand, where much of the early information concerning Polynesia has been preserved. At the same time, the matter of name-changing would not seriously affect the results of this paper, which is directed towards the separation of cosmogonic and human names in genealogies.


As a matter of fact, disagreements are to be anticipated. The opening address of the priest and teacher in the Whare-wananga (Maori college of sacred lore) was as follows:—“There was no one universal system of teaching in the Whare-wananga. Each tribe had its own priests, its own college, and its own methods. From tribe to tribe this was so. The teaching was diverted from the true - 4 teaching by the self-conceit of the priests which allowed of departure from their own doctrines to those of other Whare-wananga. My word to you is: ‘Hold steadfastly to our teaching; leave out of consideration that of other tribes. Let their descendants adhere to their teaching, and you to ours.’” (20, p. 84).

What more enlightening statement has ever come to those who have struggled to reconcile the early Polynesian cosmogonies and human genealogies? That there were many sects in New Zealand alone is evident from Tregear's comparisons (28, Kore, Papa, etc.). And the existence of sects is to be expected among an intelligent people like the Polynesians.


We shall be better able to evaluate the accuracy of the genealogies if we know their purpose and the means of preserving them. Despite the halo of sanctity or superstition which is supposed to surround the Polynesian genealogy, I believe that its greatest value to the Polynesian and the greatest cause of its preservation was its material use.

My information for the following is mostly from tribal New Zealand and monarchial Hawaii—the limits of political range in Polynesia.

In New Zealand, the genealogy with its accompanying tradition establishing the tribe, served to identify the individual with the tribe, and the tribe with the land. The most important point undoubtedly was that it served as a land title. In olden times, apparently, each member knew his own genealogy back to the tribal ancestor (Cf. Best, 3). However, land titles and tribes in New Zealand do not go back further than A.D. 1350 (22 generations), the time of the last migration, or possibly A.D. 1150 (30 generations) on the basis of the Toi-kai-rakau traditions. Of what practical value, then, are genealogies prior to the establishment of the land title, namely, the arrival of the particular ancestor? It becomes certain, then, that each migration and long settlement has had the effect of wiping clean the pre-migrational slate. They serve no useful purpose. That such a thought must have controlled is evident when it is found that what should have been the most authentic of the genealogies, namely the one given out by the house of - 5 sacred traditions (20, p. 143) allows only fifty-two generations or 1300 years from the creation of man to the year A.D. 1900.

In monarchical Hawaii, where title to land was vested in the king, genealogies were mainly of importance to the aristocracy. They served (1) to keep pure the royal line of descent (17, p. 80) and (2) to ascertain for courtiers desiring political preferment the degree of their consanguinity to the king (17, p. 254). In neither case were long genealogies necessary. For the first, close blood-relationship was the most desired. For the second, a knowledge by the applicant of ten generations of ancestors was demanded. As in New Zealand, the portions of the genealogies preceding the latest traditional arrivals are decidedly hazy.

What then about the portions of the genealogies antecedent to the latest migrations? They have, of course, some philosophical interest for the teachers, and, as shall be discussed later, had an incantational value, such as might be used in instilling ancestral mana into a divine chief (see p. 12 below). They were, however, known to few, and one may safely conclude that they, of any portions of the genealogy, would be the first to be forgotten, as of the least practical value, and least likely to be challenged by others.


The methods of preserving genealogies in Polynesia had, without doubt, certain similarities. In Hawaii and New Zealand there were professional genealogists, who transmitted their knowledge to succeeding generations entirely in secret and under rigidly prescribed rules. Acolytes were selected from among children or young people who showed aptitude and ability. Details of the Maori methods have been published (20, p. i), but unfortunately no other account, which might serve as confirmation, has reached print under the authority of the native college of learning. In this one, the precautions taken (special building, ritual, etc.) when the recitals were written down (about 1860) are a guarantee of the genuineness of the document. But its value for confirmation of other Polynesian recitals is minimized by the opening statement, as given above, of the old priest and teacher.

- 6

While, as he stated, there were different sects and different teachings from tribe to tribe, the teachers of one sect had access to the secret sessions of another. The divisions were not between tribe and tribe, but between the initiated within the colleges, and the uninitiated or the outsiders. The sanctity of the knowledge was thought to require almost absolute concealment. It might seem that at times there was a willingness to mislead the uninitiated. There were occasions, however, when genealogies or other recitals were necessary outside the Whare-wananga. A difficult birth would be eased by the recital of the father's genealogy (3, p. 284) or the groom's pedigree might be recited at his marriage (3, p. 469), but as Smith (20, p. 146) shows, the portions prior to Maui, the noted ancestor, were supposed never to be so revealed—“they were too sacred, for they deal with the gods and their more immediate descendants, the first of mankind.” However (20, p. vi) later pedigrees were the property of all.

This guardedness as between initiated and uninitiated extended to other recitals, and required the intentional use of obsolete or esoteric terms. The frequency of these unknown words was pointed out (20, p. iii) by the natives to their teachers, three of whom were present, “and their reply was that it was not proper to use ordinary words for matters referred to in their teaching.” There are other references of this kind (20, p. vi) “….. a recitation [in esoteric terms] intended to be delivered to the common people, whilst its true meaning was known only to the priests of old …. the larger part of the teaching of the Whare-wananga was never known to the common people—it was too sacred.”

With this intent to conceal from, and possibly to deceive, their own tribesmen, did the foreign antiquarian fare any better?


In the early portions of most Polynesian genealogies, we find the following names or personifications in this order:—Tumu or a derivative, Atea or Rangi, and Tiki. There are seldom less than two attributed to an island group, and frequently the three of them (see Table 2). When examined for significance, such will be found as “origin,” - 7 “sky father,” and “first man” respectively, although seldom so recorded in the published genealogies. In some genealogies, Tumu or Atea may appear as the “first man,” even though the name Tiki may later appear in the same list. In between these epochal names, and also following that of Tiki, may be found many others, which it is at times difficult to recognize as human, and others which are unquestionably cosmogonic. They vary apparently according to the requirements of the occasion and the taste and ability of the genealogist. There is not the slightest doubt that, for incantational purposes, a genealogy is required to be as full and to go as far back as the genealogist can supply it. We may rely on the later portions as correct, because known to so many. There was no check on the earlier parts—which perhaps were not much in use. The thought has often crossed my mind that given the salient features or epochal names in a cosmogonic genealogy, the intermediate portions, so unimportant and so easy to forget, might occasionally become merely compositional to a genealogist with ability to make quick mental associations.


Fortunately, a remarkable example showing the mechanics of such composition has been preserved in the Kumulipo, or Hawaiian chant of the Creation, published by H. M. Kalakaua (12) in 1889 and translated and published by H. M. Liliuokalani, Queen of Hawaii and Patron of the Polynesian Society. (15) This chant, it is stated, was a prayer for the dedication of an ancestor to the gods, and was also “the special property of the latest ruling family of the Hawaiian Islands, being nothing less than the genealogy in remote times of the late King Kalakaua” and his sister, Queen Liliuokalani. According to the Queen's statement to her companion (29) of later days, the chant was committed to writing by her brother in 1856, when a national celebration brought old people to Honolulu from all parts of the islands. These old people supplied the information, but it is not clear if the chant was supposed to represent the recital of one native authority, or a combination of the recitals of many. Much of it is undoubtedly mere filling, as one who examines it may readily see. It is important to note, however, that it includes several of the genealogies referred to but not - 8 given by Malo (17, pp. 20 ff), that it is a composition recorded and believed in by Hawaiians, and that it is the only Hawaiian document of its kind which has reached the public with such weight of authority behind it. Much of it resembles the creation chant of Tahiti (11, p. 336). It appeals to me as the class of material said to be founded on fact.

The Kumulipo chant is of particular importance at the present. It reveals the mechanism of composition, and the incantational purpose of the genealogy.


As may be recognized on examination, the mechanism is largely a matter of name and term-associations expressed through identities and antitheses, and a casual analysis of the chant will explain the composition of the earlier portions of many of the so-called genealogies from Hawaii, and probably from other parts of Polynesia.

The associations are apparently limited only by the experience of the people. They are found in forms, biological and verbal, in meanings or ideas, and in topographical situations. With the names visualized as printed in two parallel columns (as indeed most of them are) with the “husbands” in the left column and the “wives” in the right column, the associations are found to run vertically, horizontally, diagonally or parallel, and in blocks or groups, or only in pairs. Occasionally the last word or syllable of a block serves to suggest the name which heads or controls the new block. Frequently a pair each of identities and antitheses occurs in the same passage.

As a matter of fact, we are getting simultaneously examples of Hawaiian music and poetry—music in the rhythm of repetition, with the theme varying in a well organized manner, and poetry in the rhyme implied by the associations.

Word and syllable-forms are those most frequently employed in the associations, chiefly in the terminal or initial syllables. When such are not recognizable, the similarities are frequently to be found in the association of ideas. These phases of association were regarded very highly in Hawaii as an intellectual accomplishment, and took a very important part in the profession of hoopapa (contests of wit)—(8, pp. 574, 582).

- 9

In the latter portion of the chant, another mechanism may be observed—the cumulative phase indicated by counting. When this is recognized, it soon becomes apparent that the so-called genealogy is an incantation intended to instil divine mana into a chief (p. 12 below).


The mechanism of the composition may be best illustrated by examples given below, each of which is one of many. The page number refers to that in the Queen's book. 15 The left column indicates husbands, and the right, wives. The Queen's translations are repeated in parentheses and my own are in brackets. Common antithetical pairs are found in suffixes, representing: light—dark, day—night, above—below, land—sea, skies—earth, skies—land, before—behind, etc. These run intermittently through the chant and need no special illustration.

Similarity of biological forms:—

p. 8—The Papaua (mother-of-pearl) was born, the Olepe (oyster) was its offspring.

Similarity of word-forms, but antithetical situations:—

  • p. 9—The Akiaki was born and lived in the sea, Guarded by the Manienie Akiaki that grew in the forest.

Similarity of biological and word-forms, and of situations:—

  • p. 12—The Nana (fish) was born, the Mana (fish) was born in the sea and swam.

Similarity of word-forms, with internal letter-change:

  • p. 18—The Naio (water-worm) was born and became parent Its offspring was a Nalo (fly) and flew.


  • p 25—The Hooipo was born, and became hooipoipo (loving).

Natural gradation of significance associations:—

p. 34—Pua (flower, glow) Ena [red hot]
Puaena (bright glow) Enaena (great heat)
Wela (warm, heat, burnt) Ahi (fire)

The next is a similar illustration, but with antithetical termination for the group:—

p. 34—Kawala (dash it away) Mahuli (turn over and over)
Huli (turn over or seek) Imi (search)
Loaa (found) Olioli (joy)
Huhu (anger) Leawale (ecstasy)
- 10

Euphonic associations:—

p. 35—Kaiwi Iwia
Kulea Kulia
Makou Koula

A group of variable associations:—

p. 35—Puili (embrace) Apomai (embrace me)
Puiliili (frequent embraces) Liilii (tiny bits)
Puiliaku (catch it) Heleihea (go where?)
Mokukapewa (cut the fish tail) Naalo (the fronts)
Mokukaia (the fish is cut) Naele (the blacks, miry)
Piala (sprinkle water on rocks) Heleua (walk in the rain)
Kiamo (material for filling cracks) Komo (enter)
Koikua (falseness) Keaho (the breath)
Koiele (swing) Kauhi (the coverlet, a yam)
Paele (daubed with black) Peleiomo (Pele's sudden appearance)
Keomo (plug) Omoomo (suck)

An interesting group of three pairs, showing similarity of ideas only, then such similarity and reduplication of root-forms, and finally reduplication only:—

p. 39—Mahele (to divide, share) Puunaue (divide)
Kaohi (reserved, hold back) Kaohiohi (hold back)
Kona (his or hers, a district) Konakona (disgusted)

The following pair may carry the idea of skin-chafing through wearing a rough garment:—

p. 43—Ukinala (coarse grass plaited) Ilimaka (raw skin)

An example illustrating the cumulative phase is now met with, and is followed by other groups counting five or more, but more regular and even than the first. I have omitted the “wives” column:—

p. 46—Pa (hit)

  • Pana (bow and arrow, snapped)
  • Panakahi (one snap)
  • Paikekalua (hit in the hole) [probably the second stroke]
  • Puukolukolu (three heaps)
  • Napuueha (four heaps)
  • Palimakahana (work by fives)

Then follow long files of “husband” names with similar beginnings, and “wife” names in a variability of association.

- 11

Of the males, there are:—

pp. 48 to 64-54 beginning with Kupo-, of which 36 begin with Kupolo-
71 beginning with Polo- (including some with such syllables)
73 beginning with Lii-, most of which open with Liili-
62 beginning with Alii-
60 beginning with Mua-
62 beginning with Loi-

They have much the appearance of a telegraph code. When a change of theme is due, it is sometimes indicated by the presence of the new syllables controlling the next file. For instance, Pololii precedes Lii-. At the end of the Lii- file, the next male is given as A, and his wife as Lii, and the succeeding male file is Alii-. The Queen was unable to translate many of the names. The use of the diagonal transition of sounds from female to succeeding male is often met with in these files. For instance (hyphens are my insertion):—

p. 51—Polo-hili Kau
Polo-kau Uli
Polo-uli Polo
Polo-polo Hamu
Polo-hamu Nini
Polo-nini Haihai
Polo-haihai Hei
Polo-heihei Hanuai
Polo-hanuai Ewa (uneven)

In the latter part of the book are published versions of nearly all the genealogies used by Fornander, but they differ greatly. Fornander's version of the Kumuhonua tradition states that Kumuhonua (namely, “world-origin”) was the mythological “first man” created by Kane, Ku, and Lono, with Kanaloa (an evil god) looking on. The Queen's version follows the more commonly accepted Hawaiian account that the first man was Wakea, and shows Kumuhonua as the father of the gods Kane and Kanaloa. This is more in conformity with the wider Polynesian concept.

The great climax in this genealogical cosmogony is of course the introduction of the “first man,” or, as it was no doubt formerly, the birth of the god responsible for the creation of the first man. It is a very important epoch, because it must of necessity be the connecting link between human beings and the ancestral gods. The cumulative - 12 phase, as in the creation-chants in Tahiti (11, p. 342) and, I observed, in a Rimatara account, becomes intensified by counting kahi (one), lua (two) to iwa (nine) in horizontal and diagonal succession, until some unknown term is met. When the compound name is properly divided, there is a strong indication of a sex-phase if kii be translated “male” and hina, “female,” as explained in pp. 16-18 below.

p. 72—Keakenui (great longing) Laheamanu (stench of birds)
Kahi-anakii-akea Lua-anahinakii-papa
Kolu-ana-hina-kii-akea Ha-ana-hina-kii-papa
Lima-ana-hina-kii-akea Ono-ana-hina-kii-papa
Hiku-ana-hina-kii-akea Walu-ana-hina-kii-papa
Iwa-ana-hina-kii-akea Lo-hana-hana-hina-kii-papa
Welaahilauinui (burning heat of her beauty) Owe (grating sound)
Kahiko-lua-mea Kupulanakehau
Wakea Papa

Wakea or Akea and Papa having been introduced with this powerful incantation, the genealogy runs down the Ulu line to Ka-i-imamao or Lonoikamakahiki, of very sacred and high birth—the ancestor for whom the chant was sung.


We may now consider the incantational phase of the chant. Kalakaua 12 entitles it “he pule hoolaa ali'i”—“An incantation for royal dedication or consecration.” It was used to dedicate the very high chief Lono—direct descendant of the gods—and also to deify Captain Cook, who was believed to be Lono resurrected. Why, then, the consecration and deification, since Lono was already divine? It might seem that such mortal divinities represented suitable empty shells into which the mana would need to be placed before the divine phases could be operative. It is like an electric battery which, no matter how excellent the quality, is of little use until charged. The charging of the divine chief with mana is made possible through the belief in the magic omnipotence of names. We find in the Kumulipo chant the name of every form or being apparently known to Hawaiians, arranged in progression from the state of chaos through all known and many other stages of growth, creation, evolution or generation, through gods and human beings, right down the line to the new-born chiefly babe. There is nothing which may be regarded as a prayer. The mere recitation of names forms a chain along which the accumulative mana of ages untold may be moved into the

- i Page is blank

- ii
- 13

recipient shell. I have observed fears among Hawaiians that mana may not go straight. The chain then becomes a verbal tube, the leaks of which are closed by the repetition of the innumerable name variations.

These names, of course, may only be used by and for the proper person—a fact which explains the Queen's reference to the chant as being her “special family property.”

This charging of the divine chief with divinity is paralleled in other ways in Hawaii and elsewhere (cf. Best's references to karakia and “magic”—3). It is interesting to note that in the Society Islands, Ta'aroa, the Creator, did not create through his omnipotence alone, but “conjured forth” gods (11, pp. 338, 356-7) and “conjured up” man (11, p. 402).

The following Hawaiian curse to kill a sorcerer follows the general lines of the Kumulipo completeness. As will be observed, nothing remains necessary to picture all the processes of entire annihilation of the body after death by the magic fire. And, of course, such a symmetrically complete curse is bound to have its effect!

“The fire burned fiercely, fire of the night of Lanipili.
Burned whither, the fire, fire of the night of Lanipili?
Burned to the heavens,
Death (reaching) to the heavens,
Rotting, to the heavens,
Maggots, to the heavens,
Mould, to the heavens,
Made ashes, to the heavens,
The dying of the sorcerer
With his accomplice, O Kane ….”

We may, then, expect to find in a composite Polynesian genealogy, not only a human genealogy, but every phase present in the local cosmogony.


The priest of the Whare-wananga referred to different sects and different teachings—“departures from the truth,” etc. Variations will be found frequently enough as one dips into the cosmogonies and early genealogies. And yet, these variations are not so diverse but that they may be recognized as branches of a limited number of stems. I shall select, and attempt to follow back one of these stems, which I shall refer to as “orthodox,” on the basis of distributional extent.

- 14

A good starting-point is the “first man” myth, of which there are several versions in Polynesia. I shall select as orthodox that of Tiki (dialectical variations, Ti'i and Ki'i), which I shall endeavour to show runs in a broad crescent right from Hawaii to New Zealand (Table 2). Samoa and Tonga, on the one side of the crescent, and Easter Island on the other, have not maintained the creed. This distribution, however, is in part a matter of clear record, and in part the result of deduction. Tiki as the name of the first man is established in the islands lying within the crescent, except in Mangaia and in Hawaii. In Mangaia, Tiki was a woman, “sister to Veetini, the first who died a natural death” (10, p. 170). Obviously, the myth of Tiki, the “first man” was formerly present and has been submerged. In Rarotonga, the next island, I have been able to find but one direct reference to the first man Tiki, although the inferences are numerous. The records are perhaps incomplete. Tiki (Hawaiian Ki'i) as the first man is not known in Hawaii, but all indications are that he was formerly present and had been displaced. We may arrive at this conclusion through several routes.


Distribution of sculptured human images, and the term tiki for the same, show remarkable co-ordination, as given in Table 2. Except for Hawaii, where images termed ki'i are present, Ki'i, Ti'i or Tiki is the “first man.” In Rapa, images are absent, but Tiki, the first man, is present in myth. In the islands outside the crescent, there is variability. In Easter Island, images are strongly established, but the term used is mohai. In Tonga, images are rare and are termed tamapua. Images are reported as absent from Samoa. In these last three localities, Tiki, as first man, is not reported. Obviously the crescent indicates the former distribution of a cult maintaining the Tiki creed, which has been overlaid in Hawaii. Ki'i, however, is present fairly early in the alleged genealogies of the group.


If any doubt remains that Ki'i as “first man” was known in Hawaii or to the Hawaiian stock, it will be removed after examining into the esoteric meaning of the term and - 15 name. Best (2, p. 55) states that within the Maori colleges of sacred learning “the true sacerdotal term for the male organ was tiki, and Tiki is the personified form thereof.” It was not a term known to laymen, who merely knew tiki as a carved image, or Tiki as the first man. But as Best goes on to show, Tiki is more specifically the personification of the male organ of Tane, the god or personification of procreation. In Polynesian, tane is “male.” However, Tane and Tiki were somewhat interchangeable in New Zealand (28, p. 510). It is not improbable that Tane in this connection is a later and euphemistic name for the original Tiki. I suggest this because in Rapa, where the god Tane was unknown, Tiki is known as the “first man.”


With this esoteric understanding of Tiki, I may introduce some observations I made in Rapa and the next island, Raivavae. In Rapa, where graven images were absent, there are at the place where the mythical Tiki magically seduced his two daughters, two tall and slender unhewn megaliths. The modern natives call them te ka'o, “house rafters” (namely, ground-rafters) no doubt because they slope a little—which explanation is of course inapplicable. Perhaps in place of te ka'o, they should have been tiki.

In the next island, Raivavae, I found in more or less chronological order as deduced, megaliths, phallicized slab, phallus with human face, phallus with human face and limbs, and transitions from some intermediate point to pregnant females. In other words, in Raivavae there are transitions partly in form and fully in association from the plain unhewn column to the well-sculptured image, all under the control of phallicism. In Raivavae the sculptured images are known as ti'i. Ti'i as the “first man” was once present as a great god, as was Tane.

This progress of image development is, as I observed, partly paralleled in the Society Islands—an observation further confirmed by photographs shown me by Mr. K. P. Emory and Dr. E. H. C. Handy of the Bishop Museum. It is indicated in some other parts of Polynesia.

Have we not evidence of an ancient and well-established cult, following a creed of Tiki as the “first man” and using megaliths termed tiki as symbols? Tiki, I suggest, was succeeded by Tane, “man,” not necessarily as a new - 16 creed, but as an euphemism, and with this euphemistic tendency came the impetus to mask the phallus with a human form.

In Hawaii, where sculptured images are known as Ki'i, and where Kane (Tane) held ancient sway, we find a high development in wooden sculpture, and a comparatively poor one in stone. The latter may or may not be phallic. There is a head and face carved on the end of a natural stone column. However, phalli in the form of columns, large and small, natural or carved naturalistically or conventionally, were numerous, but not under the term tiki. They were known as Pohaku-o-Kane according to Fornander (6, p. 47). Other terms variously applied were Kane-hoa-lani or Nanahoa (see p. 19 below) and pohaku eho. 1 Of these, practically no information is available.

We thus have phallic forms preserved in Hawaii in stone with but little knowledge concerning them but that they were connected with the name of Kane. The material being stone, the objects may have been very ancient. We have also well-developed images in wood, which is perishable. They, therefore, are recent. All images are termed ki'i. On the basis of euphemism, have we not the overlay of Tane or Kane on Tiki or Ki'i in the matter of the stone columns or phalli, the existence of tiki in Hawaii as the phallus being established by the term being used for sculptured images? If not, then some other derivation must be furnished for the wide Polynesian term tiki, ti'i, or ki'i for images.

The concepts of Tiki and Tane are so close that I do not believe they can be dissociated. I believe from the foregoing that we may associate the use of the term tiki for images where so found, the secular use of the name Tiki as the “first man” and the worship of the god Tane.


For the next comparisons, we may take up the myth of the “first woman,” of which there are more than one. For the orthodox, Hina (Hine or 'Ina) may be selected—Hina, the wife of Tiki. As shown in the distributions (Table 2) Hina, the mythical first woman, is generally the wife of Tiki, the first man. The main exception to Hina belonging to the ‘first woman’ myth is again Hawaii, - 17 where, however, Hina appears as the wife of Ki'i in the genealogy above referred to (p. 16). It might be noticed that the weakness or absence of the Tiki myth in Rarotonga, Mangaia and Hawaii, so far as information is available, is paralleled by that of Hina, the first woman.


The earth-formed woman or man is another myth which offers interesting comparisons. Where it occurs, it is either Hina or Tiki, except in Hawaii. There, Kumu-honua, the neo-mythic first man, is given the honours. The orthodox favours Hina, the woman, when the distribution and sources of information are evaluated.


The mythical father-daughter incest theme may also be orthodox. It is not reported from Rarotonga, Austral Islands, Society Islands, or the Tuamotus, due possibly to censorship. It belongs to Tiki as the personification of Tane in New Zealand, and to Tiki in Mangareva and Rapa, but is attributed to Rongo, the son of Vatea, in Mangaia, to Atea in the Marquesas, and to Wakea or Akea in Hawaii. As a primitive rationalization in myth, it is closer to the Tane-Tiki cult than to the Atea cult, if we adopt the evolutionary phase of the latter (see p. 24 below). Here again we find the irregularities in the Cook and Hawaiian Islands, but the father-daughter incest myth was present in Hawaii.


There is another comparison which I originally included, and then omitted as not primary enough, namely, Hina as the moon- or sky-maiden. It has been restored on account of the significance Best (2, p. 58) attaches to this Hina, “the personified form of the moon, who is the tutelary being of women.” This application, it seems to me, is a decidedly secondary elaboration. Hine or Hina, as the “first woman,” merely indicates the “female element” personified. “Man was born of woman.” Hine is “female” and in some dialects is “woman” (28, p. 71), and is a root of the widely-spread Polynesian word vahine (woman). Hina and Hine are very commonly used for women's names in Polynesia. Hina, as the first woman, was obviously impersonal, as was Tiki, explained above (p. 16). In other - 18 words, Tiki and his wife Hina were never persons, but are personifications of the male and female elements or functions which lay-members of certain sects among the Polynesians have handed down as persons. At the same time, they have equal value for comparative purposes as would the names Adam and Eve.


In the Society Islands cosmogony (11, p. 403) we find that the son of Ti'i and Hina, first man and woman, was Uru [Growth]. In Moriori genealogy (28, p. 669) we find among the divine sons of Rangi, the following names in order:—Tiki, Uru, and Nganana. These in the Hawaiian dialect are Ki'i Ulu, and Nanana.

It has been indicated above that on account of certain analogies, Tiki and Hina, “first man and woman,” should be present in Hawaii. If now we find early in Hawaiian genealogy a Ki'i with a wife Hina, and possibly a son Ulu or Nanana, we may be assured that such names are merely cosmogonic and not human. Such do occur in the royal Hawaiian genealogy, and are shown at No. 13 in Table 3, where Ki'i and Hina-koula are listed as the parents of Ulu and Nana-ulu—sons from whom all the Hawaiian kings are said to have descended. Ulu is also the father and grand-father of Nana and others with names beginning with Nana. In other words, Hawaiian genealogy has been carrying names, as human, which are found to be those of the first man, woman, and child of the Society Islands, or sons of the Sky-Father among the Moriori.


There are other suggested cosmogonic parallels through the names mentioned. In one New Zealand myth, several wind personifications named Ngana are immediately ancestral to Tiki (28, p. 276). In Hawaii, the Nana names are also close to Tiki, but in succession. In another Maori myth, (20, p. 118), Uru-te-Ngangana heads the list of divine sons, while in Hawaii, Nana-ulu heads one of the royal genealogies. It is difficult to escape the conviction that this portion of the genealogy represents cosmogonic fragments which have been either misunderstood by the native narrator or foreign transcriber, or have been deliberately confused for the purpose of concealment from the outsider.

- 19

The names beginning with Nana (Table 3), a succession of which are found after several genealogical epochs, may have been merely personifications of a human male—a parallel of which is found in the name Tiki. Kane (tane) is “male,” and we find that Kane-hoa-lani and Nana-hoa in Hawaii are synonymous terms for pillar- or finger-like stones or projections on ridge sky-lines, and, as names, are present as a Priapian personification used in an erotic manner. Hoa in this connection is not a dictionary word, and should probably be ho'a (N.Z. hoka). Nana then, in Hawaii, is synonymous with kane, and may have been an ancient or esoteric term for kane, “man.” The Maori comparatives (28, p. 278) ngarara, ngaro, and ngata are of interest here, especially as the last is an obsolete term for “man.” Perhaps were the true etymology of the Hawaiian Nanahoa known, it would signify ‘serpent erect.’ Nana in the Hawaiian genealogies is sometimes given as a woman's name, but such with Tiki is present in New Zealand (28, p. 510) and has already been noted for Mangaia.

In Hawaiian legend, the father of Ki'i was Kahiko (tahito, ta'ito, or tawhito in other dialects), the wide meaning of which is “ancient.” In New Zealand it also indicated “organs of generation” (2, p. 58), no doubt as a sacerdotal term. It is obvious that Kahiko in Hawaiian genealogy is only a personification serving to introduce the Ki'i epoch. Ta'ito, in Rarotongan myth, is found to follow immediately after Tiki (Table 7).


We may follow down, for a little way, the names of the Nanaulu and Ulu genealogies (Table 3) concerning which Fornander (6, pp. 197-207) constantly deplored the absence of historical data, or intermarriages between the contemporary descendants. I have hyphenated some of the names on the Nanaulu line, between Nos. 14 and 23, to illustrate the similarity of their composition to those illustrated in the Kumulipo chant above (p. 9). Since these names were originally recited, they should be read aloud in order that the compositional effect may be appreciated. They are obviously not human. The names between Nos. 24 and 29 appear to be human, but it is not until No. 29,

- 20
- 21

Maweke, is reached that the human element is certain. The Ulu genealogy is much longer and more complicated, and while some of the early names are cosmogonic or compositional, it will be better to postpone further discussion on the Ulu line until the recognizable human names may be considered (p. 29 below).


We may now turn on our tracks and follow back past Ki'i to the beginning of this genealogy, which is headed by Wakea and his wife, Papa. The commonly accepted Hawaiian myth is that these two were the first man and woman (15, p. 72 and 17, p. 23). These two are also said to have produced, as children, the Hawaiian Islands (17, p. 21). Fornander (6, pp. 170 ff rejects the myths, and sees in Wakea and Papa human ancestors of the Polynesians, and Wakea a chief in Indonesia.


A reference to Table 2 will indicate that Papa is the orthodox “Mother Earth,” so far as the distribution may be traced. Her full name in Hawaii, Papa-nui-hanau-moku, “Great Papa giving birth to islands,” is sufficient to so identify her. As the mother of gods or men, the distribution is less wide. Specifically as a “first woman,” she is found only in Hawaii, although such position is inferred in the Cook Islands. In Hawaii (15, p. 79) Papa is said to have married her male descendants as far as Kio (No. 7, Table 3) who rejected her. This descendant-marrying phase is present in a Rimatara cosmogony. Papa indicates “stratum,” “horizontal,” “broad, level surface,” as the earth was conceived in the native mind. With this obvious recognition of Papa as Earth-personification, it becomes evident that Papa, as the first woman, has displaced Hina in Hawaii.


Concerning Wakea or Akea, the husband of Papa and the “first man” in Hawaiian myth, there is no question that he had displaced Ki'i. The father of both these “first men” is Kahiko—explained above as an introductory epochal personification. The greater antiquity of Ki'i and Hina to Wakea and Papa is shown in the introductory incantational count introducing them (p. 12 above):—

Iwa-ana-hina-ki'i-akea Lo-hana-hana-hina-ki'i-papa
- 22

The general concept of Wakea or Atea in Polynesia is “expanse,” “light-expanse,” the “light-space between the sky shell and the flat earth,” and the personification of daylight. Occasionally, as in Tahitian cosmogony, Atea becomes the sky (11, pp. 406 ff). Awakea or Avatea is “noon.”

However, Atea is also known in variable distribution as a god, and as a primal ancestor with or without Papa. In cosmogony Atea is found in Marquesas, Society Islands, and New Zealand, but in New Zealand his place is otherwise taken by Rangi (sky).

We may now begin to find evidence of the sectarianism warned by the priest of the Whare-wananga (p. 3 above). In New Zealand (20) Rangi (Sky-Father) husband of Papa (Earth-Mother) is attacked by his divine sons, among whom Tane is leader, and is against his will thrust upward to the present situation. Tane as pointed out by Tregear (28) is widely worshipped in Polynesia “either as the male principle in Nature, or as the god of Light.” In Ra'iatea (11, p. 364) Tane is the son of Atea and Papa. In Tahiti (11, p. 411) Tane raises the sky, Atea, against the protest of Atea. In another Society Islands legend (11, p. 455) Tane attacks the sky, Atea, his parent, but prevails not. In the Tuamotus (11, p. 349) Atea is attacked twice and then destroyed by Tane. In the Marquesas, Atea displaces Tane as god of Light (28, p. 28).

While Tane is supposed to have become the god of Light in Hawaii, it was evidently a displacement of Atea. Awakea “noon” is noted by Andrews (1) as also the “name of the god who opened the gates of the sun.” The genealogy opening the Fourteenth Era in the Kumulipo Chant (15, p. 75) introduces Wakea, and then follows with the placing of the stars and the moon. In a version where he was personified, Wakea, the “first man” was attacked by his elder brother despite the warning that Wakea would prevail because it was “the time of sunlight” (17, p. 313). This is, of course, a native rationalization of Wakea as the personification of light or day. The account follows which introduces the Tane contest as, having conquered his brother and gained the kingdom, Wakea is attacked and defeated by Kane-ia-kumu-honua, but later turns defeat into victory (17, p. 313). - 23 Thus in the “Light personification” and “contest with Tane,” Wakea in Hawaii is not isolated from the rest of Polynesia.

As ancestor, Wakea (Atea) through Atanua, is the progenitor of gods and men in the Marquesas. Hawaii is in line with Rarotonga, Mangaia, Society Islands, and New Zealand, in the theme of Atea or Rangi, in company with Papa, being the parent of gods or heading genealogies. In Mangareva there is substitution of names; Tangaroa, alias Atea, is the “originator” with his wife, Haumea. Haumea in Hawaii is an alternate name of Papa.

In New Zealand, the stronghold of the Tane worship, there is “little distinct idea of Atea as a person; it is only as a vast abstraction that he has existence” (28, p. 28). Atea is known as the last of the eighteen Ages or Time-spaces, counting upward from Void or Nothingness. Curiously enough, even this phase seems to have been recalled by early Hawaiians. As Tregear (28, p. 672) has pointed out in this part of the Hawaiian genealogy preceding Ki'i (Table 3), five of the twelve names present are among the names of the cosmogonic time-spaces in this New Zealand myth. These are shown in column 5 of the Table, together with four others which have some similarity.

The origin of the Wakea incest-myth in Hawaii may perhaps be traced. As related, Wakea continued the human race by means of Hoo-hoku-ka-lani, his daughter by Papa. Hoohokukalani may be translated as “Be-star the heavens” and the account, Emerson (17, p. 317) has clearly shown, is part of a solar myth. The incest in Marquesas is attributed to Atea, with Atanua (Dawn). In Mangaia, it is the son of Vatea and Papa, Rongo, to whom the incest is ascribed. This Rongo is the great god of Mangaia, and Rongo is closely associated with Atea in the Marquesas. In the Society Islands the rest of the analogies are found. Tumu and Papa, upon order of the Creator, Ta'aroa, produce a child, Hotu-i-te-ra'i, as a pillar to prop up the sky. The name in Hawaiian is Hoku-i-ka-lani, practically the same as the daughter of the Hawaiian Papa. Atea in the Society Islands is sometimes the sky and sometimes the light-space. Atea's wife is given as Fa'a-hotu (Hawaiian, Hoo-hoku) and Hotu. In another legend, the wife or husband of Atea (who changes sex) was Papa.

- 24

It is apparent that Hawaii has had more or less of the myths in this connection found elsewhere in Polynesia but, the information not being imparted direct from the repositories of sacred learning in Hawaii, the accounts have been confused, or partly concealed, and an over-rationalization by uninitiated natives has done the rest.

In brief, Wakea and his immediate descendants recorded in Hawaii as human beings are found elsewhere in Polynesian cosmogonies, and when found in genealogies, such names are not suitable for use in chronologies.


I may here interpolate a suggestion arising from the present discussion and the distribution of names. We may regard the cosmogonic portions of the genealogies as evidence of progression of thought, whether local or introduced.

For instance, we have the wide distribution of the Tane-Tiki cult having as its basis an absence of philosophical enquiry and an acceptance of the theory of man's origin simply from “male” and “female.” In New Zealand (20, p. 117) native speculative enquiry has extended to the Sky (Rangi) as protean father and to Mother-Earth, whose seventy divine sons are personifications of natural forces and complete the visible universe. The leader of these sons is Tane. Behind all is a supreme but inactive being named Io, who endows Tane with power. Tane, in addition to being the procreator, is also the personification of light. The mechanics behind the myth seem to be creational, namely appointive. It may be of significance to note that while practically all the Polynesian gods are among the seventy sons, Atea—a great god in south-central Polynesia—is not included.

Atea, however, is present in New Zealand as a concept of abstraction of Light, and is found as the last and highest of the eighteen time-spaces beginning with Kore “Nothingness,” as already referred to with the Atea of Hawaii. It has much the suggestion of an evolutional cosmogony, as noted for Atea in the Marquesas, and in the Kumulipo chant leading up to Wakea in Hawaii. It is possible that the Mangaian myth (10, p. 3) of the creation of Vatea and of a succession of others from the sides of Vari-ma-te-takere “The Very Beginning” was originally another such. It is - 25 highly probable then that the New Zealand version, although the briefest, is also truest to the original.

Reduced to their elements, we find Atea the Light-concept and Tane the Procreator. But in the wide distribution through Polynesia, either Atea or Tane may be found as the procreator or the light-concept. We also find mythical contests between Atea and Tane. Do we here find evidence of the sectarianism warned against by the Maori priest-teacher? Apparently we do. It is interesting to note that the teacher, himself a follower of Tane, did not once mention Atea, so far as the index shows. The Tane and Atea cults, apparently, although in opposition, did not hesitate to borrow from one another so far as was necessary (20).

The Tane worshippers, as already stated for New Zealand, were apparently content with an assumed always-existing earth and sky as a cosmogonic beginning. Hence there was no room for further thought on pre-mundane origins. The names Tumu-tumu-whenua and Tupu-tupu-whenua, “originating or sprouting world,” are present, but not placeable in cosmogony.

The Atea philosophy, on the other hand, being evolutionary, could reach from one beginning to another—it was never content. In Hawaii, it reaches back first to Kumuhonua (world origin) and earlier to Kumulipo (origin of deep darkness, namely chaos) and to indefinable earlier stages. With this tendency to reach back in the cosmogonies, we find the tendency to correspondingly set back the supposed first man. Thus Tumu (origin) displaces Atea, as Atea had displaced Tiki. We find an illustration of this in Hawaii in what follows.


We may now take up the genealogies, regarded by Fornander (6 and 9) and others as human, and which precede Wakea. Thus in Fornander we find several, each leading back to a “first man.” The names of these alleged “first men,” Kumuhonua (world origin); Hulihonua (seek world) in the Kumu-uli (darkness beginning) genealogy; Opu-ka-honua (the world buds and expands) etc., will serve to identify the genealogies as cosmogonies, as will some of the names which follow them. Take, for instance, those numbered 16 to 22 in column 2, Table 4, where we find the - 26 smooth progression from intense darkness to light. These are supposed to be names of the “wives.” Numbers 16 to 20 in columns 3 and 4 serve similar illustrations.

Fornander's version of the Kumuhonua genealogy, as he points out (6, pp. 87, 97 etc.), follows in a most remarkable way the order of arrangement of the Biblical genealogies, while their genealogical counts and accompanying episodes are similar. These are given with fair completeness in Fornander's notes on the Legend of Hawaii-loa (9, p. 266) and are indicated in part in Table 4. In this, No. 13 Nuu represents Noah and No. 23, Lua-nuu, Abraham. Such a coincidence or close similarity between “Polynesian and Hebrew-Chaldean legends” (6, p. 101) formed Fornander's main argument for his theory of partial Cushite origin of the Polynesians.

I have sought long for Fornander's authority for the version he used, and so far have found the list of names fairly well confirmed down to No. 13. Down to this point, Fornander used two lists which are well established. The first is part of his Kumuuli genealogy (Table 4) taken from the Kualii chant (7, p. 386). This is also found in the Kumulipo chant opening up the Fourteenth Era (15, p. 75) in which, after 29 generations, Wakea is introduced, followed by the stars and moon as mentioned on p. 22 above (cf. 7, p. 386 and 15, p. 76).

The second list follows another genealogy in the Kumulipo chant, ending as shown on p. 12 above. In this, Kumuhonua precedes Wakea by 20 generations (15, p. 72). This Kumulipo version also appears in a MS. by Bishop Maigret (16) differing only enough to show that it was not a transcript.

Neither of these lists confirm Fornander's Kumuhonua genealogy below No. 13, and it is at this point that the Biblical analogies begin to be so pronounced.

Fornander states (6, p. xii):—“Mr. J. Kepelino has furnished some valuable chants, and the groundwork of the ‘Kumuhonua’ legends, most of which was confirmed by the late Mr. Kamakau.” The resumé of this material apparently is what is given as the “Legend of Hawaii-loa” (9, p. 266), “Compiled and condensed in English from Kepelino and S. M. Kamakau.”

- iii

- iv Page is blank

- 27

Most of this data must have been obtained through personal interviews. Kamakau's publications (1866-1871) listed by Thrum (27, p. 53) omit confirmatory references, except where Kamakau is attempting to prove that the Hawaiian and Catholic religions were identical. As is known, Kamakau was born while the native religion held sway. He became a pronounced Protestant, persecuted the Catholics and then in 1859 embraced Catholicism and became its exponent, and in his last years appeared to revert to a type of mixed native faith which had survived (27, p. 46 and 30, p. 181).

What is known to-day as Kepelino's history (14) is dated June 20, 1868, and is included in a MS. book with other notes, all written down by the late Bishop Maigret. It covers but a part of the material noted down by Fornander (9, p. 266). However, as Kepelino was a school teacher and Fornander was once Inspector-General of Schools (1865-1871), there were doubtless frequent opportunities for the two to consult on Hawaiian antiquities. The information in Kepelino's history and in Fornander's notes is in part identical and the rest is of the same class, and we may, I believe, accept Kepelino as Fornander's chief authority in this connection.

Kepelino's written history opens up with discourses on the existence of a Hawaiian Trinity composed of Kane “God,” Ku “Son of God” and Lono “The Spirit.” Kanaloa is a “slave” of Kane, and later appears in Fornander's notes as “The Devil.” Many brief litanies and chants are included as though for confirmation of Kepelino's statements. The following is given as the code of Kane:—

  • 1. Do not take the name (Kane, Ku, Lono) in vain.
  • 2. Do not steal.
  • 3. Do not fornicate.
  • 4. Do not murder.

Kepelino's history also gives, adapted to the local horizon and with Hawaiian descriptive or cosmogonic names, the Biblical accounts of the Creation, the “first man,” his wife from his side, the deluge and the ark, the escape from King Wahanui with the aid of a younger and an elder brother who divide the sea, after which King Wahanui meets his death. The same younger brother, later in the wilderness, obtains water for the people by striking a rock with - 28 his rod, after which the people continue to the promised land. Under the name of Hawaii-nui, we find the earlier episodes related by Fornander with regard to Hawaii-loa, and there are no genealogies in the body of the account. The genealogy of Kumuhonua attributed above to Maigret (p. 27), follows the MS. index at the end of the history, and may have come from Kepelino, although there is nothing to indicate this.

The marked Biblical influence upon Kepelino in his ideas on Hawaiian antiquities will have been noticed. However, Kepelino was perfectly candid about his methods and, I believe, sincere in his efforts. In his foreword 14 he states:

“In Hawaiian history there are many authorities [versions], one differing from the other. This was so with the history relating to God, with the Creation of the World, with Kumuhonua [first man] and so forth.

“On this account, all the versions related by the historians have not the same authoritative value, but I have united [reconciled] the correct portions of those accounts which were partly correct.

“The history relating to God was not entirely lost to the [original] Hawaiians. They had the true religion, but their descendants departed greatly from it and worshipped artifacts.”

What was Kepelino's reconciliation of historical and religious facts has been partly shown, and will be more readily understood when I state on good authority 29 that among Hawaiians Kepelino was well known for his militant Christianity. His zeal apparently led him to read into Hawaiian religion a non-existent Hawaiian Trinity, with chants adapted to confirm its existence, and also as shown in part in Table 4, to apply Hawaiian cosmogonic or other terms as names for the characters in Genesis, and domicile them in Hawaiian tradition with their appropriate episodes. Finally, these names are understood, as previously pointed out, to represent human characters who were ancestors of Papa-nui-hanau-moku, or old “Mother Earth.”

This kind of information did not appear before the decade of 1860-70, when there was a kind of renaissance in Hawaiian history. Many educated natives, with Kamakau as a central figure, were writing and publishing on early native life, and there were frequent and occasionally acrimonious disputes between authorities (27, p. 49). The type - 29 of information was mainly traditional, and formed valuable historical material. The cosmogonic and theogonic phases were not dealt with very much, nor did they compare with the clearness of native concept revealed from the sacred colleges in New Zealand. 20 I doubt if Fornander received any of the sacred traditions in a pure state, a distant approach to which is shown in the Kumulipo. It is highly probably that the zeal of Kepelino, which caused him to deceive himself, has also mislead Fornander, and until other information comes to light, it would seem wise to withhold consideration of Fornander's conclusions based on his Kumuhonua or Hawaii-loa legend.


We may now consider names in Hawaiian genealogies which may, with a reasonable degree of certainty, be regarded as human and therefore suitable for chronologies. The unacknowledged references below are from Fornander (6 and 7).

Fornander depended largely on the genealogies following Maweke (Nanaulu line) for his later historical dates and his checks on the other genealogies. The Maweke lines were well chosen for the purpose, if we accept his lists, as, following through the descendants of the second and third sons, there is an even count of 30 generations from A.D. 1900 to Maweke (see Table 5). Traditionally, these lines remained undisturbed through foreign voyaging. Another line, through Maweke's eldest son Mulielealii and the latter's third son Moikeha, gives results similar to those preceding, namely 28 and 29 generations on different branches. However, it is indicated in the traditions that Moikeha was a traveller, and the line may have been shortened on that account. Moikeha left Hawaii Island and settled in Tahiti; abandoned Tahiti and settled on Kauai Island, married there and received the kingdom. He was then probably middle-aged. The genealogy is carried on through the Kauai marriage. We may thus add one generation to Moikeha's line for comparative purposes, making the count to Maweke on this line 29 or 30, as against 30 on the other two. Maweke's count may thus be established at 30 generations prior to A.D. 1900.

- 30
- 31

The Ulu line divides at the brothers Puna and Hema. The Hema line branches at Hanalaa. The later checks on all the lines are very satisfactory, but there is none in the earlier parts.

The earliest check on the Maweke and Puna lines is at Maelo and Lauli-a-Laa, 22 and 23 generations on the Moikeho line and 22 on the Puna line. This check may be criticised when the traditions are partly examined, as Maelo was the great-great-granddaughter of Moikeha through his eldest son, Hookamalii. Lauli-a-Laa was the son of Laamaikahiki, the adopted son of Moikeha. Diagrammed, it seems improbable. However, Laamaikahiki when young was taken to Tahiti by Moikeha, and left with the latter's older brother Olopana. In Moikeha's old age, he sent his third son to Tahiti to invite a visit from Laamaikahiki. The latter remained in Tahiti until Olopana died, and then came to Kauai. Later, moving to Oahu, Laamaikahiki married three wives, who each on the same day gave birth to a son—the wonder of which is oft repeated in the traditions. At this time, Laamaikahiki must have been well on in years and there would be no clash in the counts if his son Lauli also married late, or Maelo was very young, or Hookamalii were an early son of Moikeha. There is some evidence concerning the last suggestion. Hookamalii was Moikeha's eldest son, but evidently was not of Kauai birth, because he did not inherit the kingdom of Kauai. He settled on Oahu, where Moikeha's father had held land (7, p. 54).

The genealogical count for Laamaikahiki on the Puna line is 23, but in view of the circumstances narrated, it should be advanced at least to 24.

There is but one more genealogy which may be checked at about this point, namely that on the Hema—Hanalaanui line. We get two checks, one direct through the marriage of Kanipahu, 23, of the Hema line to Hualani, 25, of the Maweke line. The other is indirect. Kanipahu's other wife was Alaikauakoko, who also married Lakona, the fourth in descent from Kumuhonua, Moikeha's elder brother. The indirect count makes Lakona's genealogical age 24—the average between Hualani and Kanipahu.

Acording to Fornander, all the earlier portions of the Ulu lines have been tampered with by bards or genealogists of contending political factions. His claim that the line - 32 beginning with Pili is a grafting on a local genealogy seems well taken, as Malo (17, pp. 25, 312) states that Paao, who brought Pili, arrived in the time of Lonokawai (Lanakawai). The latter in the genealogy precedes Pili by two names.

Fornander believed that Pili was contemporary with Maweke's sons, and in order to make him so, dropped two names between Pili and Kanipahu above mentioned. I have restored the omissions and the count is still only 28. Regarding Kanipahu as at 24 (the average of him and his contemporaries on the Maweke lines), Pili's five generations prior to Kanipahu would make the count 29. However, if elisions are made, there should have been at least four. Paao brings Pili from some foreign country and secures for him the kingdom. Six generations later, Paao is instrumental in placing on the throne Kalapana, the son of Kanipahu (17, p. 327).

Paumakua, on the Puna line eight generations before Lauli-a-Laa, is counted as 30 in the Table. Fornander believed that Maweke and Paumakua were contemporaries, but placed the latter a generation after Maweke in the reconciled table of genealogies (6, p. 249), after dropping three names. Fornander did not begin his generational account from the last person in the line. As insufficient evidence has been brought forward for dropping the names, I have restored them and added one generation on account of Laamaikahiki above mentioned. I would tentatively set the genealogical count for Paumakua on the Puna line at 31.

Fornander has deducted another three from the rest of the Puna line, discarded the Hanalaanui line above Pili, retained the Hanalaaiki line with nine elisions, and separated it entirely from the Hema descendants. Fornander's arguments for his decisions would no doubt be satisfactory if all his references were available. His main checks, however, are indirect ones through Niheu-kolohe (6, p. 201 and 7, pp. 31 ff) the son of Hakalanileo, the local Menelaus in the Hawaiian Iliad. Fornander (7, p. 33) notes that he has found no genealogical references to Hakalanileo or his five children. The legend as related is clearly a myth.

It seems better to restore the counts, as in Table 5, and let them rest until other information be found. I do this largely because of the position of Hema, who, with his descendants Kaha'i (Tafa'i, Ta'aki, Tawhaki), Wahieloa - 33 (Vahieroa) and Laka (Rata) are well known in Polynesian tradition. The genealogical counts for Hema in Hawaii are 44, 46 or 47. According to Smith (22, p. 33) the count in Rarotonga is 49 and in New Zealand 51. If these be correct, then the Hawaiian counts should be lengthened, not shortened.

Best (2, p. 115) regards Tawhaki as a mere personification. He is such a widely distributed culture-hero that it is difficult to think otherwise. At the same time, the identity of immediate ancestry and descendants and other relatives in genealogies in New Zealand, Rarotonga, and Hawaii, together with the fact that Hawaiians and Maoris claim descent from Tawhaki and Rarotongans from Kari'i, his brother, might serve to identify them as human. In Hawaiian legend are given the places of their birth, death and burial (17, p. 323).

Before closing the Hawaiian section, I may point to another check from the New Zealand genealogies. The Hawaiian chief ‘Olopana and his wife Lu'ukia settled in the Society Islands (as the locality seems best placed) 26 to 27 generations before A.D. 1900. Olopana is mentioned as the brother between Kumuhonua and Moikeha mentioned above. As Smith points out (22, pp. 21 ff) Olopana and Lu'ukia in the Maori dialect become Koropanga and Rukutia, and husband and wife of these names are noted in Maori genealogy as living 27 generations before 1900. Tahiti as a former home of the Maori is now generally accepted through Smith's investigations.

We may, then, regard the following genealogical dates as acceptable from Hawaiian sources until corrected by other information:—Paumakua of the Puna line, 31; Maweke, 30; Pili and Paao, 29; Olopana and Moikeha, 27 and Laamaikahiki, 24. These do not represent the first arrivals in the Hawaiian Islands, as there are undated references to others previous.


We may now take up the Rarotongan genealogy on which Smith based his chronology. Turning to Table 6, it will at once be observed that as the first of the line, Tumu is in process of displacing Atea. In one genealogy (column 1) Atea is the husband of Papa. In another (column 3) Te Tumu is the husband and Atea the son. However, as we - 34 have already found these names in Hawaiian cosmogonies misread as human genealogies, we are better able to recognize them (Cf. Table 2). Our Hawaiian discussions will help to abbreviate those on Rarotonga.

Smith's chronology of the Polynesian race was practically based on only one genealogy, namely, that of Tangiia as related by Te Ariki-tara-are (26, p. 135) one of the last high priests of Rarotonga. It is given in part in Table 6, column 3. It was checked for total count (22, p. 27) by five others in Rarotonga, two of which (Table 6, columns 1 and 4) have been published. They agree very poorly with each other and with that of Tangiia in names and in order of succession. Another genealogy recited by Te Arikitara-are (26, p. 64) and known as that of Tamarua (column 2) has a count a third as long again as that of Tangiia, while in the earlier portion it is shorter. This is indicated by the contemporaneity (26, p. 67) of Tamarua-metua and Ui-te-rangiora, who are No. 34 on the Tamarua line and No. 44 on that of Tangiia.

Smith (22, p. 27) doubted the accuracy of the count of the Tamarua genealogy on account of the inclusion of three groups of names preceding No. 88, Te Itonga. There were 14 names beginning with Avaiki-, 21 beginning with Te Angai- and 15 beginning with Mau-o-. This is characteristic of the repetitional pattern in the Kumulipo chant mentioned on p. 11 above. Smith dropped them out, and was, I believe, justified in so doing. At the same time, after the existing Tamarua name repetitions began, there were 28 of them in sequence, so that no fixed rules for repetitional names may be laid down.


The Tangiia tradition of Te Ariki-tara-are (26, p. 55) is to the effect that Te Tumu and his wife Papa were chiefs in an ancient fatherland and that Papa bore first:—Te Uira, Te Aa, and Te Kinakina. These names I translate as “lightning,” “storm,” and “anger” or “disturbance” respectively. They suggest nature-forces as born to Rangi and Papa in New Zealand myth. 20 Next were born the gods Rongo, Tane, Tu, Ruanuku, and Tangaroa. After that, the chiefs were born—Te Nga-taito-ariki, Atea, Te Tupua-nui-o-Avaiki, Te Pupu, and Kaukura. Then six priests with

- v
Family Tree. (1) Iro in the Genealogy of Pa., (2) Tamarua by Te Ariki-tara-are, (3) Tangiia by Te Ariki-tara-are, (4) Tangiia and Akimano by Matatia, 1. Atea=Papa Descendants—, 1. Te Tumu=Papa, 1. Te Tumu=Papa Te Uira, Te Aa, and Te Kinakina (personifications), Rongo, Tane, Ruanuku, Tu and Tangaroa (gods), 2. Te Nga-taito-ariki, Atea, Te Tupua-nui-o-Avaiki, Te Pupu and Ka'ukura (chiefs), The common people, 1. Tupua=Tonga-iti “The first chief of Upolu was Tupua; his wife Taito. These are Tangaroa and Tonga-iti.”, 2. Te Uira and Te A-a Descendants of Te-a—, 3. Mua, Eanga, 5. Unga, 6. Engi, 28. A-io, Peke-to-io, 30. Peke-to-ake, Peke-tea-tama, Ia-tea-tama, Ia-te-po, Ia-te-ao, 35. Ia-maina, Ia-te-ata, Ia-tupu-ranga, Ia-makaro, Ia-tangata, 40. Tangata-nui, Tangata-rai, Tangata-katoa, I-te-katoa-ranga, Ia-te-atu, 45. Tiki, Taito-rangi-ngunguru, Taito-rangi-ngangana, Toro-ki-matangi, Te-ira-panga, 50. Tu-tarangi, 2. Te Tupua-o-Avaiki, 3. Tu-te-rangi-marama, Mookura, 5. Rua-i-te-karii, Te Ake-kura, Te Ake, Moo-uri, Moo-tea, 10. Io-tini-i-Avaiki, Kura-a-moo, Tau-tonu, Amau-aitu, Te Ariki-ivi-roa, 15. Kake-tua-ariki, Tua-ariki, Te Ariki-oki-tini, 18. Te Kura-a-moo, 2. Te Nga-taito-ariki, 3. Mua, Eanga, 5. Maina, Maeata, Makaro, Te Marama, Te Eanga-ki-te-ao, 10. Tangata-kato, Te Atu-tanganga, Te Atu-te-ngatata, Te Atu-te-ki, Te Atu, 15. Tiki, Taito-rangi-ngunguru, Taito-rangi-ioio, Taito-kuru-angiangi, Taito-kuru-ma-rakamea, 20. Vaitakere, 21. Te Tarava (ct. Maui-tiki-tiki),36. Te-ira-panga, 37. Tu-tarangi, 2. Nga-taito-ariki, 3. Nga-tupua, Te-ate-a-nuku, 5. Te-ate-a-rangi, Te-uira-o-te-rangi, Te-aa-o-te-rangi, Te-mangungu-a-te-rangi, 9. Te-ngatata-o-te-rangi, 34. Tamarua-metua-o-Avaiki, 44. Ui-te-rangi-ora, 38-51. Avaiki-, 52-72. Te Angai-, 73-87. Mau-o-, 88. Te Itonga, 53. Tuna-ariki (ct. Tu-ei-puku), 56. Atonga, Te Arutanga-a-nuku, Te Arutanga-a-rangi, Te Amaru-ariki, 60. Te Amaru-enua, Te Uenga-ariki, Te Uenga-enua, Kau-tea, Kau-mango, 65. Vai-iti, Kau-kura, Kau-ngati, 67. Rauru, Atonga, Te-arutanga-nuku, 70. Te-arutanga-rangi, 71. Kau-kura, Maru, 70. Moe-te-ra-uri, 71. Iro (ct. Tangiia), 105. Tamarua-pai (ct. Tangiia), 68. Tangiia, 73. Tangiia, Manatu, 74. Akimano, [75. Iro], (1) Nicholas (18) and Smith (22)., (2) Te Arikitaraare (26), (3) Te Arikitaraare (26) and Smith (22)., (4) Stair (23).

- vi Page is blank

- 35

names beginning with Rua-kana- (and who had been born of Te Veka-o-te-Po—probably the Messenger from Eternity) investigated the body of Papa, who may now be clearly recognized as “Mother Earth,” and ascertained that the multitude of commoners would be born.

The human line is carried on through Te Nga-taito-ariki. Smith (20, p. 56) was intrigued with this name on account of the combination of Te and Nga, and at one time thought that it indicated a people or a race, and not an individual. I believe that he was then more correct. The Te may well serve to indicate a personification of a group—nga-ta'ito-ariki, “The royal ancients”—names thoroughly applicable and descriptive. Nga-ta'ito-ariki and one of the brothers, Te Tupua-nui recall the names Kahiko and Kupua coming together early in Hawaiian cosmogonies (6, p. 181).

It will be unnecessary to examine further the above-mentioned names for cosmogonic characteristics if we look closely at those which follow them in the Tangiia genealogy. Certain of the latter have been transferred from Table 6 to Table 7, for better comparison with others from the Pa

TABLE 7. Comparisons of Cosmogonic Names in Rarotongan Genealogies, in Table 6.
3. Mua 3. Mua, first
4. Eanga 4. Eanga, arising, floating, breathing
35. Ia-maina  
36. Ia-te-ata 5. Maina, moon
38. Ia-makaro 6. Maeata (?)
39. Ia-tangata 7. Makaro, dimly visible
40. Tangata-nui 8. Te Marama, the light
41. Tangata-rai 9. Te Eanga-ki-te-ao, the rising to the world
42. Tangata-katoa  
43. I-te-katoa-ranga 10. Tangata-kato (probably katoa, see No. 42 opposite) all people
44. Ia-te-Atu  
45. Tiki  
46. Taito-rangi-ngunguru 11. Te Atu-tanganga
47. Taito-rangi-ngangana 12. Te Atu-te-ngatata
  13. Te Atu-te-ki
  14. Te Atu
  15. Tiki
  16. Taito-rangi-ngunguru
  17. Taito-rangi-ioio
  18. Taito-kuru-angiangi
  19. Taito-kuru-ma-rakamea
- 36

genealogy (column 1)—published by Smith but not used by him on account of displacements. These have been rearranged.

If we now follow down the Tangiia line (Table 7, column 3) after allowing for errors through repetition or transscription, we find a succession of names indicating cosmogonic changes. The first five, Nos. 3 to 7, agree with the first five in the Pa genealogy as rearranged in column 1. When we reach No. 10 in column 3, Tangata-kato, we meet with an untranslatable name, and find it a probable typographical error for Tangata-katoa, as found in No. 42, column 1. As we continue down, we find Tiki, in both genealogies, followed by several of the name of Ta'ito, but we have also found that neither genealogy is complete, gaps in each being filled from the other. One portion perhaps omitted from the Tangiia line is that preceding Tiki on the Pa line. It is given more fully in Table 6, beginning with No. 28, A-io. If these names are read over audibly, the compositional similarity to the Kumulipo chant will be observed.

However, while cosmogonic and compositional names are abundant in these sections, it is unnecessary to prove them so because, as in Hawaii, we have found Tiki and several Ta'ito in genealogies headed by Atea or Ta'ito or both. I believe that there cannot be the slightest question that this Tiki represents the mythical “first man” in Rarotonga, the name of which has been overlaid as in Hawaii. Still later, we find No. 20, Vaitakere, on the Tangiia line, as a contemporary of Tangaroa the god—so recognized from the episode accompanying the name (25, p. 68). It is thus certain that all the names on the Tangiia genealogy down to No. 20 should be eliminated as non-human.


It is possible that seven names should be dropped from the Tangiia genealogy near the generation of Tangiia. Matatia's version of the Tangiia genealogy (23, p. 130) given in Table 6, column 4, carried at No. 67 the name Rauru, followed by Atonga, Te Aratanga-nuku, Te Aratangarangi, and Kaukura. The version of Te Ariki-tara-are confirms Atonga and the two Aratanga, but inserts seven names before Kaukura. Smith (23, p. 130) at one time - 37 confirmed this portion of Matatia's version with data from New Zealand, but later changed his views. He was perhaps correct at first, as the manner of pairing the names inserted is suggestive of composition. I have followed Smith's first decision.

With the elimination of the seven names mentioned, the Tangiia genealogy is found to check approximately at the names of some of the voyagers. In Table 6 we find the name of Te Irapanga, father of Tu-Tarangi, at No. 49 in column 1 and No. 36 in column 3. These are 22 and 25 generations respectively back from the time of Tangiia. The contemporaries, Tamarua-metua and Ui-te-rangiora (26, p. 67) are No. 34 in column 2 and No. 44 in column 3. After making the adjustments mentioned, they are found 21 and 17 generations prior to Tangiia's time. While neither checks are very close, they about average possible errors and confirm the Tangiia line as modified as possibly correct.

Through Tamarua-metua and Ui-te-rangi-ora, we may be able to approximate another date on the Tamarua line. Smith (22, p. 283) had placed the period of Te Kura-a-moo (No. 18, column 2) at 65 B.C., apparently running back first along the Tangiia line to Te Nga-taito-ariki, and then forward along the line from Tu-te-rangi-marama. Counting backwards from Tamarua-metua, the date for Te Kura-a-moo would be A.D. 400, or 60 generations back from A.D. 1900. The change will bring Te Kura-a-moo more within the range of dates of other probable human beings. These appear as Te Irapanga, 51, Ui-te-rangiora, 43, and Te Aratanganuku, 30 generations back from A.D. 1900. At the same time none may be considered trustworthy until confirmed by the many genealogies in MS. said to be recorded in New Zealand.

Later genealogical dates set by Smith should be accepted without question. When confirmed by genealogies from two widely separated islands, or by numerous independent genealogies within one island, there can be no room left for doubt.


Smith's adoption of twenty-five years for a Polynesian generation is preferable to Fornander's thirty years. However, the best comparison available to me is the Japanese royal line between A.D. 400 and A.D. 1900, following the - 38 detail in Brinkley's History of the Japanese People. 4 After eliminating the lateral successions and counting down the direct line, I arrive at 57 generations averaging 26 1/3 years. The line was not infrequently carried through the younger sons—one of them the sixth. In view of this, Smith's estimate seems very fair.

On this basis I suggest in Table 8 certain modifications or confirmations of the principal dates offered by Smith (22, p. 283).


We have been discussing the presence in Polynesia of a people who were certainly present in A.D. 1200, and probably much earlier. The earliest possible date whether within or outside Polynesia to which traditions refer is A.D. 400, and this is not securely fixed. However, all these dates are arrived at through native traditions, the family resemblance of which indicate that they belonged fairly well to the same group of people. There were, however, other tradition-bearing people, as in Samoa, Tonga, and Easter Island whose records were different.

The traditions frequently refer to the migrants finding previous inhabitants in the islands occupied. In New Zealand, the earlier people are described as of a different race, and a people who did not preserve genealogies (21, p. 74). In Hawaii, Paao found a people who did not maintain the distinctions between sacred chiefs and commoners. 14 In Tahiti, there was once a similar state to that of Hawaii, if we may judge from the references to the Manahune. Rarotonga was occupied, apparently, before Apopo arrived. Easter Island has the legend of the long-eared and the short-eared people. The Tuamotus also have accounts of earlier inhabitants. While these may be nothing more than migration of myth, the New Zealand account is too circumstantial to be ignored.

Sullivan (24, p. 236) and Dixon (5, p. 384) have both indicated through detailed analytical work, that the Polynesians are comprised of several racial types, the proportionate distribution of which suggest racial migations.

Thus, when we attempt to deal with Polynesian chronologies, it is important to know if they are based on the traditions of all the people, or of only a part.

- 39 TABLE 8. Polynesian Dates (cf. Smith, 22, p. 283)
Individual Authorities Generation from 1900 Year, A.D. Evaluation
Te Kura-a-moo Rarotongan 60 400 Tentative
Te Irapanga Rarotongan 51 625 Tentative
Tawhaki Rarotongan and Maori 48 700 Tentative
Ui-te-rangiora Rarotongan 43 825 Tentative
Whatonga or Atonga Maori and Rarotongan 31 1125 Tentative
Paumakua, son of Pau Hawaiian 31 1125 Tentative
Te Aratanganuku Rarotongan 30 1150 Tentative
Maweke Hawaiian 30 1150 Tentative
Pili and Paao Hawaiian 29 1175 Tentative
Olopana and Moikeha Hawaiian and Maori 27 1225 Acceptable
Tangiia Rarotongan and Maori 26 1250 Acceptable
Hiro Rarotongan and Maori 25 1275 Acceptable
Laamaikahiki [last Hawaiian traditional contact with Hawaii] Hawaiian 24 1300 Acceptable
Last Polynesian migration to New Zealand Maori 22 1350 Acceptable
- 40

As a matter of fact we find the Polynesian area surrounded with land-masses of greater or less extent supporting a great racial variety of people. Most of them are people who maintain legends but no traditions. A few are historically minded.

Situated then as Polynesia is, since it does support different racial types, and there are traditions of earlier inhabitants than those recording the fact, we may assume with reason an early occupation of Polynesia by a people not maintaining traditions, and a later occupation by those who did maintain traditions. There may have been more than one occupation by each class of people. In any case, if there were an earlier occupation by a traditionless people, then of course the date of occupation is not reached through genealogies.

So far as our depending only on Polynesian genealogies for our chronology, it might perhaps be well to pay attention to St. Paul (19):—“Neither give heed to fables and endless genealogies, which minister questions, rather than goodly edifying which is in faith: so do.”

NOTE.—Full comment on this interesting and informative paper is not a task to be undertaken lightly; it covers such a wide field. It reminds us of another, of which it has been said, “it is a fair field full of fighting folk.” Two remarks may, however, be made: One is that caution should be observed in regarding the male and female elements or entities in early cosmogonies or divine genealogies as “husband” and “wife,” social concepts hardly applicable to cosmogonic evolution. The anthropomorphic interpretation of cosmogonic phenomena is perhaps unavoidable with uncivilized as with civilized man; our own concepts are nebulous enough; those of primitive man were at any rate as nebulous, if not more so. Another remark is that caution should also be exercised in the interpretation of names or words; giving them meanings according to the meanings of the component parts is only guesswork; true, the guess may often be correct, but we are not sure in all cases that the names or words are not corrupt, and the hazard of the guess is thereby doubled. In the corruption the parts of the corrupt form may yet have a logical meaning, when the interpretation, though apparently correct or apparently plausible, is actually quite wrong. Examples of this were noted regarding the publication, Ancient Tahiti, reviewed recently. An example in one of our own collections in New Zealand comes aptly. The gnomic phrase, Kia whakaoho koe i taku moe ko te whatu turei a Rua, has been translated, “When you startle my sleep let it be for the kernel that stands in the chest of Rua.” Now the Maori phrase baffles even the Maori; the - 41 pakeha has attempted to extract a kernel by splitting the word turei and translating tu rei—producing a European phrase that is as obscure as the Maori and so interpreting it not at all. It is not material food the Maori of the saying would be wakened for; according to the one old Maori who ventured in attempting an explanation the food was the satisfaction of the desire that is at the root of human existence. We walk in the midst of pitfalls, so must walk circumspectly.—EDS.


1. Andrews, Lorrin. 1865. A Dictionary of the Hawaiian Language Honolulu.

2. Best, Elsdon. 1923. “Maori Personifications,” Journal Polynesian Society, Vol. 32.

3. Best, Elsdon. 1924. The Maori, Vol. 1 (Mem. Pol. Soc., Vol. 5).

4. Brinkley, F. 1915. History of the Japanese People. New York.

5. Dixon, Roland B. 1923. History of the Japanese People. New York.

6. Fornander, Abraham. 1878. The Polynesian Race, Vol. 1. London.

7. Fornander, Abraham. 1880. The Polynesian Race, Vol. 2. London.

8. Fornander, Abraham. 1917. Fornander Collection, B. P.Bishop Museum Mem., Vol. 4. Edited by Thos. G. Thrum.

9. Fornander, Abraham. 1919. Fornander collection, B. P. Bishop Museum Mem., Vol. 6. Edited by Thos. G. Thrum.

10. Gill, W. W. 1876. Myths and Songs from the South Pacific. London.

11. Henry, Teuira. 1928. Ancient Tahiti (B. P. Bishop Museum, Bull. 48).

12. [Kalakaua]. 1889. He Kumulipo no Ka-I-imamao. Honolulu.

13. Kamakau, S. M. 1866-1871. Articles in Nupepa Kuokoa and Ke Au Okoa, Hawaiian newspapers. Honolulu.

14. Kepelino. 1868. MS. in Hawaiian. Catholic Mission Library, M. 20.

15. Liliuokalani. 1897. Hawaiian Tradition of the Creation. Boston.

16. [Maigret, Louis]. MS. Genealogy, Kumuhonua. Catholic Mission Library, M. 20.

17. Malo, David. 1903. Hawaiian Antiquities. Translated by N. B. Emerson. Honolulu.

18. Nicholas, Henry . 1891. “Genealogies and Historical Notes from Rarotonga,” Journal Polynesian Society, Vol. 1.

19. Paul, St. “First Epistle to Timothy, 1, 4.”

20. Smith, S. Percy. 1913. Lore of the Whare-wananga (Polynesian Society, Mem. 3).

21. Smith, S. Percy. 1915. Lore of the Whare-wananga (Polynesian Society, Mem. 4).

22. Smith, S. Percy. 1921. Hawaiki. Auckland, New Zealand.

23. Stair, John B. 1895. “Flotsam and Jetsam from the Great Ocean” (Journal Polynesian Society, Vol. 4).

- 42

24. Sullivan, Louis R. 1923. Marquesan Somatology (B. P. Bishop Museum, Mem. 9, 2).

25. Te Ariki-tara-are. 1899. “History and Traditions of Rarotonga,” Journal Polynesian Society, Vol. 8.

26. Te Ariki-tara-are. 1919. Journal Polynesian Society, Vol. 28.

27. Thrum, Thos. G. 1918. “Life and Labours of S. M. Kamakau,” Hawaiian Historical Society, 26th Annual Report.

28. Tregear, Edward. 1891. Maori-Polynesian Comparative Dictionary. Wellington.

29. Webb, E. Lahilahi. Personal communication.

30. Yzendoorn, Reginald. 1927. History of the Catholic Mission in Hawaii. Honolulu.

1   Cf. M. keho in Williams' Dictionary ii, 2.