December i6, 1922] 



NA TURE 



827 



The enigma of the three parallel rivers is explained 

 as due to their valleys having been worn out along 

 clefts through which the drainage from south-eastern 

 Tibet was enabled to escape through the mountain 

 rim of Chinese Tibet. This rim had been formed by 

 the Himalayan movements which were due to the 

 intense compression of the crust ; on the relief from 



that pressure the mountain ranges were broken by 

 transverse clefts, and large blocks sank be1 

 network of fractures. The basins formed b; 

 subsidences gave the rivers great powers of enlarging 

 their channels and thus of excavating the deep steep- 

 sided valleys which are now the most conspicuous 

 features in the topography of south-western China. 



The Present Position of 

 \\ "HALING has been practised as an industry for 

 ^ » some centuries. The pursuit of the Atlantic 

 right whale was carried on in the Bay of Biscay at 

 an earlv date, and was active at least so long ago 

 as the twelfth century. The Greenland right whale 

 was hunted in three areas, at successive periods, at 

 first oft Spitsbergen from about 1610, when few 

 Atlantic right whales were left, then in Davis Straits 

 from about 1719, and finallv in the North Pacific 

 and Bering Sea from about 1840. The sperm whale, 

 which occurred in the whole of the tropical belt, 

 though by no means restricted to this area, was 

 hunted from about 1712. 



The successful introduction of the modern harpoon- 

 gun, with a harpoon carrying an explosive charge, 

 dates from 1865, and has revolutionised whaling, 

 by making it possible to capture the large and swift 

 rorquals or fin whales. Modern whaling is concerned 

 mainly with the humpback whale, the fin whale, 

 and the blue whale, all of which are widely dis- 

 tributed in nearly all seas, although it is not certain 

 whether each of these whalers' names indicates the 

 same species in all parts of the world. After rorquals 

 had been hunted in such localities as the Varanger 



1 Substance of a paper read before the Association of Economic Biologists 

 by Sir Sidney F. Harmer, F.R.S., on November 10. 



the Whaling Industry. 1 



Fjord, Newfoundland, Iceland, the British and 

 Norwegian coasts, and elsewhere, whaling on an 

 unprecedented scale commenced off the edge of the 

 Antarctic Continent in 1005, and is still being con- 

 ducted energetically. The total catch in this area 

 has exceeded 10,000 in a single year. 



The principal whale-products of economic im- 

 portance are : train-oil, sperm-oil, spermaceti, baleen, 

 ambergris, whale meat, and the various forms of 

 whale-meal or " guano." In a well-conducted factory 

 all parts of the carcass are utilised. 



With the exception of the Antarctic whaling, 

 vvhii li has had a career of less than twenty years, 

 whaling has been carried on consistently to' an ex- 

 cessive amount, leading to the most " serious re- 

 duction of the number of whales. The Atlantic and 

 Greenland right whales were decimated almost to 

 the point of extermination, the sperm whale in- 

 dustry has practically disappeared, and little re- 

 mains now but the Antarctic whaling grounds. 

 The efforts of all lovers of Nature should be directed 

 to the restriction of whaling to an amount which 

 is not inconsistent with the permanent preserva- 

 tion of these magnificent marine mammals and of 

 the industry which they are so unfortunate as to 

 support. 



Biometric 



T N the current issue of Biometrika (vol. xiv. 

 •^ pts. i. and ii.) Dr. Kirstine Smith discusses the 

 standard deviation of a coefficient of correlation com- 

 puted from data derived from classes, members of 

 which are mutually correlated, with special reference 

 to the case of fraternal and parental correlations 

 calculated from entries of siblings. She finds, inter 

 alia, that the best determination of a fraternal cor- 

 relation from a given number of observations is 

 obtained by taking (1 +i/r) offspring individuals 

 from each family, where r is the fraternal correlation. 

 Mr. Egon S. Pearson contributes an important 

 memoir on variations in personal equation. The 

 experimental basis of the research was a series of five 

 sets of measurements of different type ; the form of 

 sessional change, i.e. the resultant of factors operative 

 within each series, is separated from the secular, or 

 long period, change effective from one session to 

 another ; appropriate forms for the expression of each 

 are discussed. It is evident that in the determination 

 of the precise value of the correlation between succes- 

 sive judgments in a series, one has to reckon not only 

 with phvsiological or psychological common factors, 

 the organic basis of the correlation, but also with 

 accidental errors which blur the record — the observa- 

 tional errors of some writers — and reduce the numerical 

 value of the correlation. It is found that the correla- 

 tions between successive judgments decrease approxi- 

 matelv in geometrical progression with the number of 

 intervals, a finding consistent with the assumption that 

 there is little or no partial correlation between the 

 observers' true estimates at intervals greater than one. 

 The chief practical outcome of the work is to show 

 that although " experience and accuracy may be 

 gained by practice, it does not follow that the cor- 

 relation between successive judgments will disappear." 



NO. 27/2, VOL. I IO] 



Studies. 



The memoir is not only of practical interest to all 

 experimenters, but also contains several contributions 

 to statistical algebra. In connexion with the work on 

 pp. 37 et seq., a reference to the memoir of Anderson 

 {Biometrika, x. 260I would have been in place, but 

 no doubt Mr. Pearson will deal more fully with the 

 literature of the subject in a sequel. He is to be 

 congratulated on his first appearance in a field where 

 one bearing his name must be judged by the highest 

 possible standard. 



Dr. Ernest Warren's paper concludes the account 

 of work partly described in 1917 concerning inherit- 

 ance in the foxglove. Dr. Warren holds that " the 

 evidence of the present investigation is therefore 

 definitely against any general application of the theory 

 of pure lines and of genotypes of any appreciable 

 magnitude, and further it indicates that selective 

 breeding within self-fertilised generations of a homo- 

 geneous race is capable of modifving that race to a 

 marked degree." 



Prof. Karl Pearson and Mr. Egon Pearson show how 

 to find a general polychoric coefficient of correlation, 

 i.e. to fit the " best " normal surface to data subject 

 to the limitation that the marginal totals are exactly 

 reproduced. The arithmetical work is heavy, and the 

 suggestion is that a determination of the correlation 

 ratio from the array means — not a laborious task — 

 will usually suffii e 



Mr. James Henderson discusses the expansion of a 

 function in tetrachoric functions, a matter of some 

 importance to those who use the frequency svstems 

 favoured by Scandinavian mathematical statis- 

 ticians. 



It will be obvious that the fourteenth volume of 

 Biometrika is as valuable to statisticians as its pre- 

 decessors. 



