interpretation of and reason for our own distant forefathers' thought and belief, should only provoke a smile. We may be unintentionally misled also by the native himself, who, unless exceptionally well informed, may give us his explanation, and be therein as thoroughly mistaken as we should be in some wild guess of our own. Take, for instance, cannibalism, and hear it explained by different observers and inquirers. One, after talking with native cannibals, explains that it arose from the desire for revenge; another because the courage, strength, &c., of the dead man can be inherited by the person eating the body. Still another teaches that it originated in scarcity of food. In fact, each native supplying the information believes that he is telling the truth, but he is absolutely ignorant of the reason why, perhaps ten thousand years before, his ancestors began their indulgence in human food. It may have arisen at funeral ceremonies, where, as in Hawaii, a dead chief was devoured by his own family only, or it may have been through some mysterious rite of communion, such as “eating the god,” known to different barbaric peoples. In any case the origin is wrapped in mystery and the dusty cloud caused by the procession of countless centuries. Customs, beliefs, usages of savages should all be carefully noted and recorded; but the deductions, the theories of origins, &c., arising therefrom should be very carefully and thoughtfully considered.

If this is true in the realm of anthropology, still more cautious should be the inquiry into the realm of etymology I believe that the man who gives an authoritative etymology of any word at all is a very bold man, and if he does so outside the lines of historic linguistics he is even a reckless man. Take up an English etymological dictionary written by a master and note how often he has to write “derivation un-known.” Then, again, how short are some of the etymo-logical pedigrees; they reach back two or three centuries and then are lost. Again, there are those words which, though not bearing the label “derivation unknown,” plainly show that they ought to have such a label, for their explanation is evidently pure guesswork. Even of those words which in the hands of men like Professor Skeat are the most reliable in the language, again and again there are instances where results have to be altered in the light of later knowledge, so that an etymological dictionary twenty years old is a thing to be handled with discretion. If, then, it can be said of the classical or the European languages that their etymologists are daring men, it is borne in upon the student that one who gives at present any Polynesian etymologies is simply a fool Of course, I do not allude to mere translations or explanations of words. It may be, for instance, quite true that the

translation of wai-rarapa is “flashing water”; but if we go further and try to trace wai and rapa to their source, giving a derivation for either of the words, we shall be, in our present state of knowledge, very devoid of wisdom. I say this as a preface to laying before you a sample of the difficulties besetting the pursuit of etymologies. I doubt the native exponent of the primary meanings of words just as I doubt the native explanation of some immeasurably ancient custom. The native will tell you something, but how did he know ? Unless his knowledge is guided by some tradition whose authenticity is indubitable (and they are few), he is simply guessing, misled by sound-resemblance—keen guessing perhaps, but without a substratum of fact, and perfectly incapable of proof. This applies to all languages, not to Polynesian only. Panini was, doubtless, a reliable as well as an ancient grammarian of the Sanscrit language; but when there is a question as to the real root of a word I should prefer the opinion of a modern scholar, with the light of comparative philology falling on his work, to the most authoritative dictum of the most learned Brahmin. I have shown in a former paper^* how it is almost certain that the Maori word ua (rain) was a worn-down form of surangi, and I have now a more curious series of transpositions to show. No one would be likely to consider from mere sound-resemblance that such Maori words as rau, pungawerewere, nape, ranga, and here were of common origin or closely connected with each other.

I was led to consideration of this subject when reading a Spanish book, “El Sanscrito en la lingua Tagalog,” by the well-known linguist Don T. H. Pardo de Tavera. One thing in the work particularly attracted my attention—viz., that the Tagal words given by him as allied to Sanscrit seemed to arrange themselves into two divisions, early and late Sanscrit. Of these, the long compound words have been apparently adopted by a late borrowing, but the shorter words are such as those found scattered through the Malay Archipelago. Crawfurd noted some of these, and they may fairly be classed as a stratum of language underlying the Malay and other tongues that were brought down in the Mongolian invasion from Asia. This stratum is composed of relics of that mother speech of Arya that preceded Sanscrit, as it preceded all European languages. Such words as Tagal acsaya, “to destroy, disperse,” is probably the Sanscrit kshaya, “slow destruction, ruin” (as the Kawi akshama = Sanscrit kshama), and Sanscrit bahala, “a hundred millions,” is the Tagal bahara, “a weight of 150 kils.,” through Malay bahara, also a measure

of weight. The Tagal castoli, “musk,” is the Sanscrit kastri, “musk”; the Tagal bangsi, “a flute,” is the Sanscrit vangi, “a flute.” With such words as these the Polynesian language has little affinity; they are too modern; but as soon as we get into the shorter and more radical words the Polynesian conveys the older form far better than the Tagal does. The Polynesian vaha, “to carry,” is far nearer the old root √VAH, “to carry,” than the Malay bawa is; while the Polynesian rapa, “to flash” (Aryan, √LAP, “to shine”), and kapa, “to flutter” (Aryan, √KAP, “to vibrate”), may serve as examples of closer affinity to the ancient root than any Tagal form can show except in the later groups of borrowed words.

It is not, however, my intention at the present moment to discuss this question, as the evidence is very voluminous, and it is beside the present point, which is, that one of Don Pedro's words, though late Sanscrit, leads to some very interesting disclosures in word-relationship. He points out that the Tagal lalaua, “a spider,” is the Sanscrit lalasrawa, “a spider,” and that it has taken this form because the letters sra in the middle of a word are very difficult for a Philippine native to pronounce. This may be so, and, if the connection is a fact, supplies us with a valuable, link in regard to the unsuspected relationship both of Malay and Polynesian with a Sanscrit word in this instance. The Malay has both lawa-lawa and labalaba as “spider,” and the cobweb is called sarang-labalaba in Malacca (sarang =“nest”), as in the Sulu Archipelago, where “spider” is lawalawa, the web is lawai-lawalawa, a double compound of the word. The Bisaya (another Philippine dialect) has lawa, “a cobweb”; but where the Tagal word is so valuable is in the first syllable, for the lawa requires the prefixed la to connect it properly with Sanscrit lalasrawa. Monier Williams, in his large Sanscrit dictionary, tells us that lalasrava means “distilling saliva,” and “spider,” from a root √LAL, “to play, sport, to loll the tongue,” and hence such words as lala, “saliva,” lalaya, “to emit saliva.” But if we admit that the Malay lawalawa, “the spider,” is a variant of the Tagal lalaua, “a spider,” or Tagal lawalawa, “a spider's web,” we have also to admit the Lampong lawah, “a spider,” the Sangi-Manganitic lelawah,” “a spider,” and the Timorese naba, “a spider,” because this latter makes a frequent I to n change with Malay. The Malay laba becomes Timorese naba by a rule, just as Malay lilin, “wax,” becomes Timorese nini. Malay lima, “five,” is Timorese nima, &c. This is also a Polynesian letter-change, as the Tongan neka, “joy,” is the Maori reka; the Tongan nima, “five,” the Polynesian lima; the Hawaiian ununa, “a pillow,” the Maori urunga, &c.

If we pursue the Malay words lawalawa and labalaba into Polynesia we arrive at some curious results. We do not find the word in its proper form and meaning, although we have it as a compound in the Fijian virita-lawalawa, “a spider's web.” The Hawaiian lawalawa does not either mean “spider” or “cobweb,” but its signification shows that it applies to actions resembling a spider's work. Thus it means “to stretch cords from one place to another; to fasten something”; (2) “to bind, as a grass house or anything in danger”; (3) “to bind round and make fast.” This dialect appears to have substituted n for v in direct words for spiders, thus we have (instead of lawalawa) lanalana, “the name of a large brown spider that stands high on its legs”; (2) “a rope with which the outrigger of a canoe is tied to the arched connecting-poles”; (3) “to cause to float.” Lana is “to float, to swim in the air or on the surface of the water; buoyant.” Also, to complete the resemblance in Hawaiian (and show the looseness of l and n), nanana, “the long-legged spider”; (2) “to swell up, as the abdomen”; (3) “a spider's web”: punanana, “a species of spider”; (2) “the spider's web.” But with this last we revert to the Malay form, for with the Hawaiian pulawalawa, “bound tightly and firmly, as a thatched house with cords from post to post,” shows the original idea of a web-like cordage. Note, however, that the Hawaiian lanalana, “the spider,” is the direct equal (according to the Polynesian ‘Grimm's Law’—viz., Hawaiian l and n represents Maori r and ng) of the Maori ranga and raranga, “to weave.”

I need not remind my readers of the connection always maintained in classical poetry and legend between the spider and the weaver, the spinner and the web. Even in our vernacular we speak of “the web” on the loom, and the fable of Arachne has blended itself with almost all thought on the subject. Not only in Maori but in other Polynesian dialects does the idea of weaving pertain to the word ranga. We have the Samoan lalaga, Tahitian raraa, Tongan lalaga, Paumotuan raraga, Mangarevan raraga, all meaning “to weave,” and also the Futuna lalaga, “to weave,” lagalaga “to weave a basket.” So that the words which in Hawaiian mean “spider” and “web” become in other dialects the words for “weaving.”

This is strengthened by the Timorese word naba, “a spider,” shown above as being a form of Malay labalaba or lawalawa, for this word strongly suggests the Maori word nape, “to weave,” arose from a root √LAP or NABH (a variant of √LAP or LAV), as in the old Sanscrit root √NABH, “to bind, to connect.” It appears again in the Tongan nabe, “a method of plaiting sinnet,” and in the Paumotuan

nape, “to weave,” “a tress or plait,” and in the Samoan nape, “to be entangled,” fa'a-nape, “to tie loosely.”^* If the final vowel be objected to, I would point out that the Samoan nape, “to be entangled,” has another form (lave or lavelave, “to be intertwined, intricate”), and that Lorrin Andrews, in his Hawaiian dictionary, shows that lawa is sometimes lawe, as in lawaia, “to take fish,” from lawe, “to take,” and ia (Maori ika) “fish.” In another Hawaiian example we have lawelua, “to tie up a second time,” which is from lawa, “to tie up,” and lua “twice.” Also the Tongan lalava, “to bind with sinnet,” is the equivalent of Samoan fa'a-lave, “to take a turn of a rope round something,” and Sunda (Java) lawe, “thread for weaving.”

It is, however, in metathesis that the strangest confirmation of the connection comes out. In the Motu (a Polynesian dialect in New Guinea) valavala, “a cobweb,” we have an exact transposition of lavalava, and this valavala varies in other places with velevele or werewere in a dozen curious ways. In the Maori pungawerewere, “a spider” (also puawerewere and puwerewere); in the Tahitian puaverevere, “a cobweb” and “gauze”; in Marquesan punaveevee, “spider” and “cobweb”; and Mangaian pungaverevere, “a cobweb,” we have one form. In the Samoan apugaleveleve, “spider” and “cobweb”; in the Tongan kaleveleve, “a large spider”; and Futuna kaleveleve, “a spider” and its “web,” we have the other. But to make assurance doubly sure we have also the Sinangolo (New Guinea) kavalavala, “a cobweb,” which is the original vowel again. In the Paumotuan pugaverevere means “cloth,” again showing the connection between the cobweb and weaving. The kaleveleve, “a spider,” above noted, has its Hawaiian affinity in kawelewele, “the names of certain short ropes about a canoe,” and “the beard”; but in this dialect punawelewele is the “spider” and its “web” as before, but with an addition well worth notice, for Lorrin Andrews gives us an etymology of a sister-word. He says, “Punawele, to be small in size, to be fine, as threads of spiders' webs. From pu and nawele, to be fine or small.” I cannot help thinking this derivation doubtful, because punapuna means “made fine, scattered, blown away,” and punawe “to divide,” which latter is a form of puu-nauwe, “to divide into parts or parcels,” evidently from puu, “to heap up,” which, in its causative form hoo-puu (Maori whaka-puku) means “to heap up, as stones, to cast lots, to divide a country