626 



NA TURE 



[October 27, 1904 



LETTERS TO THE EDITOR. 

 {The Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Further Discovery of Dodos' Bones. 



Since the astonishing discovery, in 1865, of innumerable 

 bones of the dodo in the peat of the Mare au.\ Songes by 

 Mr. George Clarli, of Mah(5bourg, in Mauritius {Ibis, 1866, 

 pp. 141-146), whereby Prof. Owen was enabled to describe 

 the greater part of the skeleton of that remarkable bird 

 (Trans. Zool. Soc, vi. pp. 49-80), and the subsequent re- 

 searches at the same place of Mr. Sauzier in 1889, the 

 results of which, when worked out by Sir Edward Newton 

 and Dr. Gadow (Trans. Zool. Soc, .xii'i. pp. 281-302), almost 

 wholly completed our knowledge of its osteology — besides 

 affording evidence of the former existence of other con- 

 temporary species now extinct— nothing more has been re- 

 corded on the subject.' It was therefore with great interest 

 that, just five years ago, October, 1899, I received a letter 

 from M. E. Thirioux informing me of his having found, in 

 the preceding month of .\ugust, some remains of at least 

 two dodos in a small, partly collapsed cave, about 800 feet 

 above the sea, and about two miles, and a half from Port 

 Louis. Encouraged by this success .M. Thirioux continued 

 his operations, a matter of some difificultv, not to sav 

 ■danger, from time to time, and was good enough to keep 

 me acquainted with many of the results, sending me photo- 

 graphs of the bones which he was fortunate in "disinterring 

 from the soil. They were not all dodos' bones, but some be- 

 longed to other extirpated forms of birds— as the brevipennate 

 parrot (Lophopsittacus), the " Poule Rouge " (Aphana- 

 pteryx), and the coot— and reptiles— as Didosaurus and 

 ■one or more of the land tortoises— all of which are verv 

 miperfertly known, while some of the small dodo bones are 

 of great rarity, and at least one of them (the pvgostvle) 

 had not been seen before. From that time until verv recentlv 

 M. Ihinoux has been continuing his researches ' and has 

 consequently formed a very considerable collection which 

 he now writes to me he has disposed of to the Museum of 

 Mauritius, and I can but express the fervent hope that some 

 competent person may be found to work it out and publish 

 a memoir on it which will be a worthv successor to those 

 that I have already mentioned. ' Alfred Newton. 



ibridge, October 20. 



Ca 



The Forest-pig of Central Africa. 



■ Tu"''",'; '■"''' ''^■° ^°°'^ niounted specimens of the forest-pig 

 m the Museum of the Congo Free State at Tervueren, near 

 Brussels where I had the pleasure of examining them in 

 jul% last. MA. Dubois, conservator of the Roval Museum 

 t ^ u "'^""■y St Brussels, told me that he intended 

 to describe the animal in conjunction with Dr. Matschie 

 01 berlin. but I am not aware that their description has 

 .^et been published, so that I hope the forest-pig mav re- 



fn!^'?. K "aT" ^Ju*"^ excellent name Hylochoerus, proposed 

 101 It by Mr. ihomas. 



F^;^! r^fj^^'*"" ," 't!''''' "iystf"ous animal " of the Congo 

 Forest a luded to by Sir Harry Johnston in his letter on this 

 ^ubject (Nature, p. 601), I have little doubt that it was 

 the f^ne antelope of the genus Tragelaphus, latelv described 

 V r w - Tf^ "^ Baeocephalus cnryceros isaacsoni (.inn. 

 rt«M7) V. p. 310, and Proc. Zool. Soc, 1902, ii. 

 p. 3I0)- Ihe first pair of horns of this species was obtained 

 i>\ Mr K J. Jackson in 1897 (see Proc. Zool. Soc. 1807 

 P- dp-), but It IS onlv recently that the perfect specimen 

 which now adorns the mammal gallery of the British 

 .Museum was procured. 



The " abnormally developed horns of the cow eland " 

 referred to by Sir Harry Johnston have nothing to do with 

 this antelope They will be found fullv described and figured 

 in the Book of Antelopes " (vol. iv. p. 209). 



P. L. SCLATER. 



Average Number of Kinsfolk in each Degree. 



The letter you forward to me from Prof. G. H. Bryan 

 gives an opportunity of discussing the question somewhat 

 more thoroughly than space allowed in mv brief memoir of 

 September 20. 



The writer says: — "Is Dr. Galton's deduction of d — i 

 correct? I should have thought that if a parent had d 

 male and d female children, each female child would have 

 d—i sisters and d brothers." 



The objection holds good only on the erroneous suppo- 

 sition that each and every family of 2d children consists of 

 d boys and d girls ; it does not hold good on my supposition 

 that each such family contains on the average d boys and 

 d girls. The inclusion of the omitted word introduces a n<"W 

 set of considerations. They depend on the variety of the 

 possible forms of combination of boys and girls in 2d 

 children, which are 2ci-|-i in number, and on the frequency 

 of each of these forms, which is given by the d+i terms of 

 the binomial expansion of (i-l-i)"''. The exact character 

 of the process concerned is clearly appreciated by thoroughly 

 working out some particular case, say that of d = 2i, where' 

 the number of children, 2d, in each family will be 5. There 

 are then 6 possible combinations of boys and girls, forming 

 6 different classes, shown in the first three lines of the table. 



1 Some reoiited dodos' bone-, 

 Zool. Soc, 1885, p. 719), turned i 



NO. 1826, VOL. 70] 



id to have been found in 

 to be turkeys' (ofi. ctf., i 



cavern (P^ 

 [90, p. 402). 



In line (4) is shown the number of sisters in any one family 

 of each of these classes (n.(n— i) sisters to n girls). 

 Thus in each famil)' in class vi. there are 5 girls, con- 

 sequently 5x4 = 20 sisters, in class v. there are 4 girls, and 

 4x3 = 12 sisters, and so on. The total number of combin- 

 ations of boys and girls in a family of 5 children =2^ = 32, 

 which are distributed into si.x classes according to the familiar 

 binomial fashion as above ; these are shown in line (5). 

 Multiplying each entry in (5) with that in the same column 

 in (3) we obtain line (6), which shows that the total number 

 of girls in the 32 families is 80 ( = 2ix 16, as it should be). 

 Multiplying similarly the entries in (5) by those in (4) we 

 obtain line (7), which shows that the 80 girls have between 

 them 160 sisters ; consequently each girl has on the average 

 2 sisters. This is identical with my rf — 5. 



I have made similar calculations for values of d=i, I5, 

 2, 25 (above), and 3. In each case the result is that a girl 

 has on the average d — 5 sisters. It may therefore be 

 assumed that the reasoning by which I originally arrived 

 at that deduction is correct. 



Before concluding, I should like to direct attention to a 

 slip of the pen in the last line but one of my memoir, which 

 somehow escaped correction; the term d = s should have 

 been 2^ = 5. The context corrects the mistake, which may 

 nevertheless puzzle the reader for a while. 



Fran'cis Galton. 



Mendel's Law. 



In his letter of last week detailing his most interesting 

 experiments on cross-bred maize, Mr. R. H. Lock makes 

 the following statement : — " I see from the published account 

 of a recent discussion at the Cambridge meeting of the 

 British Association that the facts of Mendelian segregation 

 are still disputed by the biometric school of evolutionists." 

 Now it is easy to make a general statement about some 

 vaguely defined group of men, and I have no right to speak 

 for biometricians as a body. But as inventor of the term 

 biometry, I may perhaps be allowed to say what I under- 

 stand by it as a science, and to restate what 1 said with 

 some emphasis at the Cambridge meeting. Biometry is 

 only the application of exact statistical methods to the 

 problems of biology. It is no more pledged to one hypo- 



