246 



NATURE 



[January 16, 1902 



and the hair-colour from seven tofourteen are but little, in England 

 al any rate, subject to environmental influence. The cephalic 

 index remains almost constant throughout life. Now the mean 

 value for the fraternal correlation for these three characters is 

 o'5i6l, and for stature, span, forearm and health, which might be 

 supposed to be largely influenced by environment, it is 0-5179. It 

 thus seemed to me that environment was not an important factor 

 in modifying the correlation of the //i_)'«Va/ characters bet.veen 

 brothers. In other words, environment does not injlneuce the con- 

 stant uniformly in one direction. This view was apparently 

 appreciated by Prof. Cockerell when he wrote " the treatment of 

 successive children is not the same." Any influence of environ- 

 ment, strange as it may seem, was thus found to be negligible. 

 Turning to the intellectual characters, my own a priori 

 conception was that I should for the first time be able to 

 distinguish between nurture and heredity. I expected much 

 higher correlations in the case of temperament and probity than 

 in that of physique. I was therefore somewhat surprised when 

 the values came out much the same as in the case of the physical 

 characters, say an average of 05. Of course it is open to I'rof. 

 Cockerell to say that o'3 of this only is due to heredity and 02 

 to environment, or whatever other division appears to him 

 probable, but he will then have to explain why the sum of the 

 two makes o'5, and why the influence of heredity is less in 

 the intellectual than it appears to be for the physical characters. 

 There is the obvious direct scientific interpretation which seems 

 to me the true one, environment does not act in one direction in 

 either case, and the mental and physical characters are inherited 

 precisely at the same rate. To those who have taken the 

 trouble, as I have done, to examine carefully the mental charac- 

 teristics of a family at intervals of a century apart, so that we 

 are not troubled by the co-environment peculiar to brothers, it is 

 needless, perhaps, to urge the very strong inheritance of mental 

 qualities. If Prof. Cockerell attributes it to his third factor, 

 " pre-existing soul," I should, indeed, be proud to have aided in 

 the demonstration of its reality, although I fail entirely to see 

 how it is to be done "by just such methods as Prof. Pearson 

 employs." Meanwhile most people will, I think, prefer to stick 

 to heredity. 



Of the last paragraph of Prof. Cockerell's letter I understand 

 not a word. Correlation is quite independent of variation, and 

 although skull capacity is highly variable as compared with 

 length of femur, I see no reason for supposing the former is 

 therefore less strongly inherited than the latter. 



I have not touched on the influence of " local races " on my 

 data, because that appears to be a factor which has escaped 

 Prof. Cockerell, and so I am not bound to state a doubt which 

 I have well considered in order to reject it. 



Karl Pearson. 



Magnetostriction of some Ferromagnetic Substances. 



Wf. avail ourselves of your valuable Journal to give a short 

 notice of some new results, obtained in our researches on 

 magnetostriction, being a continuation of our investigation on 

 the same subject, given in the " Rapports presentes au Congres 

 international de Physique," t. ii., by one of us. 



It was generally believed that ferromagnetic bodies show 

 change of length by magnetisation, but not of volume. 

 Minute as the effect generally is, it is now placed beyond 

 dispute that iron, nickel, cobalt and especially steel difler also 

 in bulk in the magneti.sed state from that in the unmagnetised. 

 In the course of our researches on the magnetostriction of different 

 ferromagnetic bodie.« in the form of ovoids we came across a 

 substance which shows a remarkably large increase of volume. 



Examining the magnetostriction of reversible nickel-steels of 

 different composition, which were kindly placed at our disposal 

 by Dr. Guillaume and M. Dumas, we find that alloys con- 

 t.aining 46 per cent., 36 per cent, and 29 per cent, of nickel all 

 show increase of length several times greater than that observed 

 in ordinary iron. But far more striking is the change of volume ; 

 of the three above-mentioned alloys, the etTect increases as the 

 percentage of nickel becomes less ; with 29 per cent, alloy we 

 observed a motion of 5 mm. of the liquid in the capillary tube 

 (diam. 0'4 mm.) attached to the volumenometer containing the 

 specimen (volume » = 10 c.cm.) under examination. Thus 

 the greatest change of volume by magnetisation amounts to 

 ivjv = 51-1 X 10-" in H = 1690C.G.S. With ordinary iron, 

 the same change 5 t'/t' = i '2 x 10 " in the same field, being 



NO. 1 68 I, VOL. 65] 



only about l/40th of that observed in nickel-steel. The 

 magnetisation of 25 per cent, nickel alloy is so feeble that it is 

 impossible for it to be magnetometrically measured, and the 

 change of length is inappreciably small, but the volu me change 

 is measurable and amounts to o'2 x io~' in H — 1790. It is 

 thus quite probable that there is an alloy containing somewhat 

 more or less than 29 per cent, of nickel that indicates largest 

 increase of volume by magnetisation. 



We at first thought it would be possible to trace some con- 

 nection between the thermal expansion and the change of length 

 by magnetisation. .No such relation seems to exist ; the 36 per 

 cent, alloy, which is the least expansible by heat, indicates 

 tolerably large elongation by magnetisation. 



As regards the Wiedemann effect, nickel-steels behave very 

 much like iron, showing the maximum amount of torsion in 

 moderate fields. 



As is well known, the behaviour of cast cobalt, as regards the 

 change of length by magnetisation, is opposite to that of iron, 

 but the volume change in the same metal is much smaller. By 

 annealing cobalt in a charcoal fire it assumes a pale ashy colour, 

 and the magnetic character is greatly changed. The metal 

 becomes less magnetisable, and shows constant decrease of 

 length accompanied by increase of volume. 



We have also found, by actual experiment, that the effect of 

 stress on magnetisation and the magnetostriction in cobalt and 

 in nickel-steel are reciprocally related to one another, as was 

 already established for iron and nickel. H. Nag.\ska, 



K. Honda. 



Physical Laboratory, Imperial University, 

 Tokyo, December 3, 1901. 



Results of International Magnetic Observations made 

 during the Total Solar Eclipse of May 17-18, 1901.' 



To test further the results obtained by the United States 

 Coast and Geodetic Survey magnetic parties during the total 

 solar eclipse of May 28, 1900, regarding a slight magnetic effect 

 that may be attributable directly to some change produced in 

 the electrification of the upper atmospheric strata by the 

 abstraction of the sun's rays due to the interposition of the moon 

 between the sun and the earth, an appeal was made for inter- 

 national cooperation in magnetic and allied observations during 

 the recent total solar eclipse. 



The repetition of the observations was doubly interesting 

 owing to the fact that the present eclipse occurred in the 

 opposite magnetic hemisphere to that of the year 1900, and 

 hence the opportunity was afl'orded for ascertaining whether the 

 magnetic effect was reversed in its general character to that of 

 1900, as is, for example, the case with the diurnal variation 

 in passing from one magnetic hemisphere to the other. The 

 conditions, however, for obtaining observations at a number of 

 stations distributed along the belt of totality, as was done in 

 1900, and thus testing whether the magnetic effect again followed 

 directly in the wake of the shadow cone, were not favourabl • 

 owing to the present location of the belt of totality. 



In response to the appeal, simultaneous magnetic observa 

 tions were made on May 17 from 14 to 21 o'clock Greenwich 

 mean astronomical time — an interval amply covering the time 

 of the eclipse^at a number of stations encircling the entire 

 globe, three of which were in the belt of totality. The prime 

 purpose of making the observations so as to cover the entire 

 globe was to furnish the possibility of separating a possible 

 eclipse magnetic effect from a contemporaneous magnetic storm 

 of the usual type. The eclipse effect, for instance, doubtless 

 would be confined to a very small belt, wher eas a customary 

 magnetic storm, in conformity with the usual experience, would 

 manifest itself at practically the same moment of time over a 

 very large area and thus be felt at stations far from the totality 

 belt. 



At none of the outside stations has a disturbance of any 

 appreciable size been thus far reported to me, the general con- 

 sensus of opinion of observers at these stations being that 

 " nothing unusual occurred." 



At the three stations within the belt of totality the majority 

 of opinion is that something unusual did occur during the time 

 of the eclipse. 



1 Presented before the meeting of the .Astronomical and Astrophysical' 

 Society at Washington. December 30, 1901. 



