;o 



NATURE 



[March 9, 19 16 



1404 students. The three agricultural colleges had 

 938 students. Three schools of mining had 668 

 students, and 267 students were registered in the four 

 schools of forestry. During the first semester follow- 

 ing the beginning of the war, the total number of 

 matriculants fell to 64,700 in forty-seven of these insti- 

 tutions. The four schools of forestry were closed, and 

 the veterinary school in Munich became a part of the 

 University. During the winter of 1914-15 about 50,000 

 of these students were in the field or available for 

 service; that is, 7575 per cent, of the 66,000 German 

 male students registered at the beginning of the war. 

 Of the 66,000 German male students who were regis- 

 tered at the end of the summer of 1915, only 12,000 

 are still in attendance at the schools, so that about 

 54,000, or 8i-8i per cent., of German higher students 

 are now enrolled in the army. Of the 13,785 univer- 

 sity students registered during- the summer semester 

 of 1870, only 4400 (32 per cent.) were at the front, and 

 3200 of this number fell in the field. 



SOCIETIES AND ACADEMIES. 

 London. 

 Royal Society, February 24.— Sir J. J. Thomson, 

 president, in the chair.— Prof. Karl Pearson : Mathe- 

 matical contributions to the theory of evolution. 

 XIX. — Second supplement to a memoir on skew varia- 

 tion. This memoir adds certain additional types of 

 frequency curves to those published by the author in 

 memoirs in the Phil. Trans, of 1895 ^"d 1901. It 

 sums, up by aid of a diagram the old results and the 

 present additions. It further illustrates by an impor- 

 tant general case that frequency curves are distributed 

 over a wide area of the y8,, y8., plane, where /3,, p. 

 are fundamental statistical constants, and that only 

 evil can arise from inflating the Gaussian point (i8i = o, 

 ^82 = 3) to cover the whole of this area. The entire 

 subject is, In the author's opinion, of much import- 

 -ance, as significant differences are in many branches 

 of science determined by the so-called " probable 

 ■error" of the measured quantities, whether they be 

 means, standard deviations, or correlations. But such 

 "probable errors" have little, if any, meaning, if it 

 can be shown that the sample value is not even the 

 most probable value of the statistical constant In the 

 sampled population, and that the samples are not dis- 

 tributed in a form in the least approaching the Gauss- 

 ian distribution about the mean value of samples. 

 In every case it is needful to determine the actual 

 frequency distribution, and in nine cases out of ten in 

 samples such as are in common use in psychology, 

 astronomy, or physics — what the statistician terms 

 small samples — it is easy to demonstrate that the dis- 

 tribution Is very far from the Gaussian type, but may 

 be markedly skew to such an extent that the ordinary 

 "probable error" Is meaningless. — F. P. Burt and 

 E. C. Edgar : The relative combining volumes of 

 hydrogen and oxj'gen. The gases were measured suc- 

 cessively In a constant-pressure pipette at 0° C. and 

 760 mm. pressure. (I) In the first series hydrogen and 

 oxygen were prepared by electrolysis of barium 

 hydroxide solution. The hydrojjen was purified by 

 passag-e over charcoal cooled In liquid air ; the oxygen 

 by liquefaction and fractionation. Mean value for 

 ratio of combining' volumes from twelve experiments 

 was 2-00294. The figure 200288 Is adopted as final 

 value for ratio of combining volumes at 0° C. and 

 760 mm. pressure. This differs from the value of 

 Scott (200285) bv onlv 3 parts in 200,000. The result- 

 ing atomic weight for hydrogen (0 = 16) computed from 

 Morley's value for the densitv ratio (008987^/ 1-42000) 

 is 1-00772, very nearly the arithmetic mean of Morley's 

 and Noges's values (1-00762 and 100787). — W. Mason : 



NO. 2419, VOL. 97] 



Speed effect and recovery in slow-speed alternating 

 stress tests. Repeated cycles of equal direct and 

 reverse torque have been applied to mild steel speci- 

 mens of tubular form, and systematic measurements 

 made of the range of the corresponding torsional 

 strains. The author attempts to account for variatlDUs 

 of strain on the hypothesis of alternate production and 

 hardening of •'mobile material" in the steel. — W. M. 

 Thornton : The Ignition of gases by impulsive electrical 

 discharge. The Ignition of gases by impulsive dis- 

 charge is considered first as a function of sparking 

 distance. It is shown that tjie shorter the distance 

 the greater the spark, so that the volumes of the leasl 

 igniting- sparks are, in a typical case, the same foi 

 all spark lengths. Ignition may occur with intense 

 momentary brush discharge, generally with the trw 

 disruptive spark. The products of combustion ar€ 

 found to be ionised and to carry a positive charge. 

 The gases examined were mixtures in air of hydrogen, 

 methane, propane, and pentane; of ethylene and acetyl- 

 ene ; carbon monoxide and cyanogen ; coal gas and 2 

 mixture of equal volumes of hydrogen and methane, 

 Hydrogen, propane, pentane, and carbon monoxide 

 rise gradually in difficulty as the percentage of oxygen 

 is reduced ; methane is Ignited bv the same sparli 

 whatever the percentage of gas may be ; acetylene anc 

 cyanogen have the stepped atomic type of ignition ; 

 ethylene Is more inflammable in rich mixture, 

 Hydrogen and methane In equal volumes are ignitec 

 as methane in type, hydrogen in magnitude. 



March 2. — Sir J. J. Thomson, president, in the chair, 

 —J. B. Cohen, H. D. Dakin, M. Daufresne, and J, 

 Kenyon : The antiseptic action of substances of th< 

 chloroamine group. The probability that the forma- 

 tion from proteins of substances containing haloger 

 was an Intermediate agent In the germicidal action 0I 

 hj'pochlorltes made It desirable to investigate system- 

 atically a number of substances containing the (NCI 

 group. Among the substances investigated, the mosi 

 promising were the group of sulphochloroamides firsi 

 prepared by Chattaway. The following are the mair 

 results of this Investigation : — (i) almost all the sub- 

 stances examined containing the (NCI) group posscsse,c 

 very strong germicidal action. (2) The presence In the 

 molecule of more than one (NCI) group does no1 

 confer any marked increase in germicidal power. (3 

 The germicidal action of many of these chloroamim 

 compounds Is molecule for molecule greater than thai 

 of sodium hypochlorite. (4) Substitution in the nucleu! 

 of aromatic chloroamines by CI, Br, I, Ch^, C0H5, 01 

 NO, groups does not lead to any very great Increag( 

 in germicidal activity. More commonly there is £ 

 moderate diminution. (5) The chloroamine derivative! 

 of naphthalene and other dicycllc compounds of sulpho- 

 chloroamlde type closelv resemble simpler aromatic 

 chloroamines in germicidal action. (6) The few bromo- 

 amines examined show a slightly lower germicida 

 action than the corresponding chloroamines, bui 

 sodium sulphobromoamides are much more active thar 

 sodium hvpobromite. (7) Derivatives of proteins pre- 

 pared bv the action of sodium hypochlorite and contain- 

 ing (NCI) gfroups are strongly germicidal. Blooc 

 serum inhibits their germicidal action to much the 

 same extent as it does with sodium hypochlorite 01 

 the aromatic chloroamines. Among the above pro- 

 ducts /)-toluene sodium sulphochloroamlde was selectee 

 as being on the whole most suitable for practical use 

 It is easily and cheaply made ; it is relatively non- 

 irritating to wounds ; it is non-toxic and very solubk 

 in water, and may be kept unchanged, both in th< 

 solid state and in solution for a long period. — I. J. B 

 Sollas and Prof. W. T- Sollas : The structure of th< 

 Dicvnodont skull. This is an account of a skull o 

 Oudenodon studied in serial transverse sections. 1' 



