NA TURE 



November 7, 1907 



In a precisely similar way, but using equation 2 instead 

 of I, the value 13937 is obtained, which can be increased 

 to 13-948. 



This being the case, the question is, How can this value 

 arise? The experimental work upon which the figures 

 are based seems to have been carried out with every 

 possible precaution, and all the values agreed very closely. 

 The only possible weakness appears to lie in equation 5 ; 

 the result was deduced from only three experiments, and 

 the agreement was not so good as usual. Assuming for 

 the moment that this value is wrong, it is easy to calcu- 

 late by how much it is so. Taking N = i40io, we may 

 say 



242208 + 6'62 



If the values of s, w, and y are taken as before, this 

 gives 0-028 as the minimiim value of s, and it is hardly 

 to be expected that such a large error could have escaped 

 notice. Another possibility is that all the errors are about 

 five times as large as the values given, but even if this 

 were so it would be very improbable that they should all 

 be of such a nature as to raise the atomic weight. Con- 

 sequently, granted that the discrepancy is due to experi- 

 mental error, it is almost certain that the analysis of 

 potassium nitrate is at fault. This was carried out by 

 heating with silica, and if the nitrate was not completely 

 decomposed the number 87-232 would be too great. This 

 is the only explanation which seems reasonable ; and, 

 moreover, if the ratio obtained in this experiment be 

 changed to 100 : 87-203, all five equations become consistent, 

 and yield frequently accepted values for the atomic weights, 

 silver being 107-883. 



Whether this be the true explanation or not, it is obvious 

 that the method outlined above affords an excellent means 

 of checking atomic weight determinations, and is also 

 applicable to finding the absolute weights, since there is 

 no great accumulation of errors. 



H. E. W.4TSON. 



University College, Gower Street, October 26. 



The Fauna of Madagascar. 



Madagascar, with certain adjacent islands, has been re- 

 garded by some naturalists as forming a distinct 

 " region," the Malagasy, equivalent to the other main 

 regions of the world. On the other hand, Messrs. P. L. 

 and W. L. Sclater (" The Geography of Mammals," 

 London, 1899, p. 108) adhere to the earlier opinion of the 

 first-named of the two authors, as well as of many sub- 

 sequent writers, and place Madagascar in a subregion only 

 of the Ethiopian region. They remark that " Madagascar 

 appears to contain a sample of the ancient Ethiopian 

 fauna, which has been almost exterminated on the main- 

 land." 



The archaic nature of the Madagascar fauna has lately 

 (Zool. Jahrh., 1902) received further confirmation at the 

 hands of Miss A. Carlsson, who found that the peculiar 

 Viverrid genus Eupleres showed likenesses to both the 

 Viverrine and Herpestine sections of the Viverridje, and 

 was therefore probably an ancient type of Viverrid. 

 Having had lately the opportunity of making some 

 anatomical observations upon another Madagascar genus, 

 viz. Galidictis, I am able still further to support this view. 

 This interesting Viverrid has the external scent glands of 

 the Viverrine section, and a crecimi which is comparatively 

 long, like that of the Herpestinae. The brain, like that 

 of Eupleres, shows intermediate characters. Finally, the 

 archaic nature of this animal is demonstrated by the com- 

 pletely double uterus, a feature new to the Carnivora, 

 where a bicornuate uterus is at least the rule. 



It has been pointed out that Madagascar also shows 

 an unexpected likeness to the neotropical region in its 

 fauna, especially in the group of reptiles. As to 

 mammals, the late Dr. Dobson showed reasons for 

 believing that the alleged close resemblance between the 

 Cuban Soleonodon and the Mascarene Centetes had been 

 exaggerated ; but among the Reptilia there are genera 

 which are common to the two regions, e.g. the snakefe 

 Boa and Corallus. I have been able lately to compare 

 Corallus madagascariensis with a South American form, 

 C. cookii. In the former the bronchus extends a long 



NO. 1984, VOL. yyl 



way down the larger lung, the liver is prolonged by one 

 lobe nearly to the gall bladder, the umbilical vein of the 

 foetus does not persist, and the mode of distribution of the 

 intercostal arteries is as in the pythons. In the latter 

 species these characters are as in the Anaconda. 



The anatomical differences may possibly seem slight to 

 those not specially acquainted with the structure of 

 serpents ; but in the features mentioned there is, if any- 

 thing, rather a greater difference between the two species 

 of Corallus than between two admittedly distinct genera 

 such as Eryx and Python. It is very desirable that the 

 alleged close resemblance between other forms occurring 

 both in Madagascar and in the neotropical region should 

 be carefully scrutinised. Frank E. Beddard. 



Zoological Society's Gardens. 



The Interpretation of Kendelian Phenomena. 



Dr. Archdall Reid's letter in Nature of October 3 

 contains a very positive statement in reference to the rela- 

 tion of Mendelian phenomena to man, which I think 

 should be immediately answered. I delayed supplying an 

 answer because I wished to discuss his statement on a 

 tangible basis. I desired to analyse certain data which 

 I have been collecting, and which throw light upon the 

 problem of segregation in man. This analysis is as yet 

 incomplete, but it is sufficient to show that Dr. Archdall 

 Reid is too confident when he asserts that " there is no 

 segregation in man," and that, "with the exception of 

 eye-colour, and possibly one or two other traits, such as 

 the Mongolian eyelid, human hybrids appear to blend 

 every character as perfectly as skin-colour." 



The accounts which 1 have collected deal with various 

 marriages between Europeans (chiefly Scotch) and the 

 Canadian Red Indians. It is well known that many of 

 the early European settlers in Canada married Red Indian 

 women. The resulting half-breeds in their turn were in 

 some cases intermarried, and in others mated to 

 Europeans. 



The Canadian Red Indians can be marked off from 

 Europeans by six definite characters, which concern the 

 nature of the hair, eyes, skin, cheek-bones, nose, and 

 beard. The Indian hair is invariably black, long, glossy, 

 and lank, and cannot be confounded with European hair ; 

 the eyes are almost invariably black or, very seldom, dark 

 brown ; the skin is tawny brown-yellow (varying from pale 

 olive-yellow to dark brownish yellow) ; the cheek-bones 

 are high (there is no obliquity to the eyes, thus differing 

 from the Mongol) ; the nose is very prominent and broad 

 at the base, and is of the btisqiie type, that is, the profile 

 is m,ade up of two lines, which diverge widely from the 

 bridge towards the base ; and, lastly, there is either no 

 beard or a very scant one of straight hairs on the face of 

 the men. These characters, when well developed, are so 

 different from the corresponding features in Europeans that 

 they cannot easily be confused. No one, for instance, 

 would mistake the long, lank, black hair and black eyes 

 of the Indian for the thick red hair and blue eyes of 

 some of the Scotch persons concerned in the histories now 

 under review. 



We may therefore use these six characters as dififer- 

 entiating ones, and may tentatively regard the Indian 

 characters as being allelomorphic to the corresponding 

 European ones. For the sake of brevity I will use 

 symbols, which will have the following significance ; — 



I = Indian, E = European, H = hair, E = eyes, S = skin, 

 C = cheek-bones, N = nose. 



First, then, with regard to the matter of dominance. 

 We must, in this case, be quite sure that the European 

 concerned marries a full-blood Indian. In the cases which 

 I have so far collected, I have only one marriage of such 

 an Indian with a European, and there were only two 

 children of the marriage. The European was a Highland 

 Scot. His complexion was fair, and eyes blue. I have 

 no information of the colour of his hair, since it was 

 white with age when observed, but it was quite thick and 

 not lank. In all the features (with the exception of the 

 beard, of which I have no information) which mark off 

 the Indian from the European, the son and daughter of 

 this marriage were quite Indian. 



So far, then, as this one case will take us, these five 



