382 



NATURE 



[September 21, 191 1 



tnni of quality. When Mr. Ray has read again my 

 detailed statement of method he will be the iirst to thank 

 me for having introduced some principle of comparison 

 where none has hitherto existed. I am convinced that 

 this coefficient of quality amounts to more than a make- 

 shift ; its value will exist in its applicability as a constant 

 factor. Mr. Ray cites my determination of coefficient 93 

 for Nguna and 76 for Sesake, " the same language, 

 Sesake being a colony from Nguna." The lower coefficient 

 of Sesake is not a matter of mere opinion ; it is the 

 mathematical product of the presence in that language 

 of several consonant mutations expressive of a wider 

 divarication from the Proto-Samoan stem. If we were in 

 possession of practically complete dictionaries of Nguna 

 and Sesake, this coefficient of quality would, in my best 

 judgment, still be applicable. Suppose that such 

 dictionaries should exhibit Nguna as having Proto-Samoan 

 loan material quantitatively to the extent of 100 words 

 to the thousand, and Sesake 125 ; of course, it is under- 

 stood that I am using figures symbolically. We should 

 err in assigning to Nguna a 10 per cent. Proto-Samoan 

 element and to Sesake 12J per cent. We should have to 

 regard quality as well as quantity ; we should have to 

 employ my quality coefficient (subject to such recomputa- 

 tion as a better supply of data would make possible) as 

 an essential factor. Thus Nguna, as a result of the com- 

 bination of the two elements, would be indicated by 

 9-3 per cent, and Sesake by 9-5 per cent. Then and thus 

 only will it be possible to state positively that Nguna 

 and Sesake have received the same influence from the 

 Proto-Samoan, the question of whether qua Melanesian 

 they are the same language being a matter wholly 

 distinct. 



It was of set purpose that I have omitted the discussion 

 of the syntax of the grammar of Oceanic languages. Our 

 present material is all stated most inaccurately; our 

 authorities, without exception, have stated their con- 

 siderations of grammar in terms of the grammatical cate- 

 gories of inflected speech. The grammar of these isolating 

 languages must be newly written. It will form a large 

 part of the Samoan dictionary of Polynesian philology 

 upon which I am now engaged. In various publications 

 I have made preliminary announcement of some of the 

 principles of this grammar — e.g. in my monograph on the 

 Beach-la-mar jargon, just published, I have dealt with 

 the '-volution of the verb in the diffuse attributive ; but in 

 the present incomplete stage of the study it did not seem 

 advisable in the volume under review to seek to exceed 

 the phonetic examination of the material there assembled. 

 William Churchill. 



New York, August 28. 



Is my review of Mr. Churchill's book, "The Polynesian 

 Wanderings," I certainly intended my remark as to the 

 liability of error arising from deficient and imperfect 

 materia! to apply both to quantitative and qualitative 

 comparisons. With regard to the former, there can be 

 no dispute, and the impracticability of quantitative com- 

 parison was recognised by Mr. Churchill on p. 143 of his 

 book. But a qualitative comparison as given by Mr. 

 Churchill in his tables, and referred to in the letter, seems 

 In be equally liable to error through inaccuracy in the 

 material. Referring again to the tables for Nguna and 

 Sesake, which I took as typical cases, Mr. Churchill finds 

 that the lower coefficient of Sesake — i.e. 76 as opposed to 

 Nguna 93 (implying that Sesake was less like Polynesian 

 than Nguna) was due lo the presence in Sesake "of 

 .! consonanl mutations expressive of a wider divari- 

 cation from the Proto-Samoan stem." Mr. Churchill's 

 taken from the lists of Codrington and 

 Gabelei ource of his much shorter li-i in Nguna 



is not stated, Using a longer list by Bp. Patteson (from 

 which those of Codrington and Gabelentz were derived) 

 and a very long (MS.) vocabulary of Nguna, I find that 

 all Mr. Churchill's examples in Sesake are identical with 

 Nguna, and all the Nguna are identical with Se 

 The consonant mutations are the same. Thus the quality 

 of the likeness to Polynesian is the same for both 



language rid gi if differe in Mr. Churchill's 



results is entirel) due to the lists being defective and 

 appearing under different names. For the accurate 



I . 87] 



application of Mr. Churchill's comparisons the vocabu- 

 laries used mast be equal in extent and signification. 



My desire in the note on Mr. Churchill's omi 

 discuss grammar was to direct attention to the fact that 

 lie had not shown that an} distinctively Polynesian particH 

 (as, e.g., the article te, the verbal signs /.tin, no, the 

 so-called possessive words loku, toku, &c.) were used in 

 Melanesian languages. Sidney H. Ray. 



Habits of Dogs. 



1 have read with interest the letter of Miss Everett it. 

 your issue of August 31 on dogs eating wasps, as I have 

 a poodle which also eats them, with evident satisfaction. 

 Id generally catches them alive, but will also pick them 

 up from the floor when recently killed; he evidenth suffea 

 somewhat from the sting, but only for ten or fifteen 

 seconds. 



I have always attributed this liking for wasps to some 

 stimulating action of the poison similar to that produced 

 by formic acid on man ; this idea was suggested by the 

 following plan, learned in Switzerland, 



If a freshly peeled wand be plunged into an ants' nest, 

 so as to be bitten by the infuriated ants, and is then 

 passed between the lips, a sensation of refreshment is 

 experienced which appears to be out of proportion to the 

 effect which one would expect from the mere acidity. 



I believe that 1 have read somewhere that formic acid is 

 a stimulant. Robert Vina isles. 



8 rue du Sundgau, Mulhouse, Alsace, September 8. 



On p. 348 Dr. Kidd asks if it is known to be a common 

 thing for dogs to carry hedgehogs in their mouths. I can 

 only answer for my own dog, a fox-terrier. Last season 

 a hedgehog strayed into our garden, and appeared anxious 

 to stay ; but the dog carried it in his mouth repeatedly, and 

 so teased it in various ways, that we were not surprised 

 when the hedgehog beat a permanent retreat by abscond- 

 ing. The curious circumstance in the affair was that the 

 dog appeared to carry the hedgehog, rolled in a ball, with- 

 out causing his lips to bleed; in this particulai 

 seems to be cleverer than Dr. Kidd'- 'log ! 



Braewyn, Earlsfield Road, Wandsworth Common. 

 R. Hooper Pear- 



A Gilbert White Manuscript. 



Mention has been made in the Press of the reci nl saj 

 of a hitherto unpublished manuscript by Gilbert White. 

 It consists of a nature calendar which the author of " The 

 Natural History of Selborne " carefully drew up 

 h. wrote the first of the letters which are the basi- of his 

 famous book. To the latter, he tells us, he meant to add 

 an "Annus I listorico-Naturalis," and it seems that the 

 MS. in question was intended to be used in this connection. 

 I am pleased to say that the Selborne Society will shortly 

 produce the calendar (which is particularly interi - 

 facsimile, and print a limited edition on Italian hand-made 

 paper. 



I should be glad to give further particulars to any of 

 the many admirers of Gilbert White who would 

 have them. Wilfred Mark Webi 



(Honorary Secretary.) 



The Selborne Society, 42 Bloomsburv Square, 

 London, W.C., September 10. 



Miniature Rainbows. 



Tin: recent letters on miniature rainbows recalled to my 

 mind a rather interesting case which I obsen 

 years ago n Inversnaid, on Loch Lomond. Here a small 

 Stream makes a pretty waterfall; and while stand 1 

 the pool at (he base of the fall, and directly opposite the 

 fall itself, I noticed first a brilliant rainbow- reflected in 

 the pool. The actual bow was formed in the spray abolB 

 flic pool, and, unless my recollection is greatly at fault. 

 a appeared less brilliant than its reflection. But the bow 

 and its reflected image, viewed together across the pool. 

 formed an almost complete circle, broken only whi re 'ho 

 extremities of the real bow in the spray appeared to con 

 down towards the surface of the pool. 



A. I.. I'll 



1, ill a, Shrewsbury Lane. Shooters Hill. 

 September 16. 



