.;io 



NATURE 



[June 14, 1917 



assigned to these types, and it is clear that if 

 Russell's densities are correct the sixth-root law 

 must be close to the truth. 



If a is the radius of a star the total radiation 

 will be proportional to a^T*, which varies as 

 go?-, i.e. as M. The total radiation thus depends 

 only on the mass, and not on the density or stage 

 of evolution. The absolute luminosity is a fairly 

 good measure of the total radiation for the range 

 of temperature here considered, though, of course, 

 the visibility of the radiation changes a little with 

 the temperature. We shall thus have the total 

 radiation constant as we pass through the series 

 of spectral types, and the luminosity roughly con- 

 stant (with deviations amounting to about i^ mag- 

 nitudes). This is just the feature which Russell 

 has pointed out in the luminosities of the giant 

 stars; they are practically the same whatever the 

 type of spectrum.'* 



It may be remarked that this theory avoids a 

 difficulty noticed by J. Perry ^, that when y is 

 less than A, the heat within the contracting star 

 IS greater than the energy set free by contraction, 

 leaving less than nothing for radiation into space ; 

 the difficulty is even more serious than Perry 

 considered, for he did not make any allowance for 

 the enormous store of ethereal energy necessary 

 for equilibrium with matter at high temperatures. 

 But we have seen that by taking account of radia- 

 tion-pressure the interior temperature is much 

 reduced ; less internal heat is therefore needed ; 

 and there is, in fact, an ample balance of energy 

 left for dissipation even when 7 is considerably 

 below 4. 



With a molecular weight smaller than 54 the 

 importance of radiation-pressure is reduced ; for 

 example, with molecular weight i8 radiation- 

 pressure is 6/7 of gravity, instead of 19/20. But 

 it still plays a ' predominant part until we come 

 down to molecular weight 2. Reasons have been 

 urged in favour of a low average molecular 

 weight — perhaps as low as 2. It is probable that 

 the atoms are highly ionised by the radiation of 

 short wave-length within the star ; and if most 

 of the electrons outside the nucleus are split off 

 from each atom we shall actually have an average 

 weight for the ultimate independent particles 

 nearly equal to 2, whatever the material (exclud- 

 ing hydrogen). Radiation-pressure is then less 

 than half gravity ; but the two principal laws, 

 which seem to be verified by observation, are 

 arrived at as before. Moreover, the order of 

 magnitude of fe is scarcely altered ; it is now 5 

 instead of 30 C.G.S. units. Nor is the internal 

 temperature much changed. In fact, the effect 

 of ionising the atoms is that the pressure of the 

 superincumbent layers is supported by a mixture 

 of cathode rays and X-rays, instead of by X-rays 

 alone ; our doubt as to the proportions in which 

 these occur and as to which will predominate is 

 no serious hindrance, because the main results 

 are nearly the same in any case. 



A. S. Eddington. 



* Loc. cit., p. 252, Figs. I, 2, and 3. 

 5 Nature, vol. Ix., p. 35c<. 



DR. W. H. BESANT, F.R.S. 



THE death of William Henry Besant on 

 June 2, in his eighty-ninth year, will be 

 mourned, in all sincerity, by a far greater num- 

 ber than he would have anticipated, supposing 

 that he ever wasted a thought on the subject. 

 Among these will be a legion of his old pupils, 

 who had the opportunity of learning to know 

 him in a peculiarly intimate way. Until 1880 

 or so Besant and Routh had almost a monopoly, 

 for many years, in coaching pupils for the 

 Mathematical Tripos. Besant 's method was 

 rather odd, but very effective with. the right sort 

 of man. At the cost of immense labour he had 

 written out, with his own hand, a set of "book- 

 work and rider " papers covering the whole 

 range of the examination. The pupil, on each 

 of his three weekly visits, found one of these 

 papers awaiting him in the outside room, and 

 proceeded to answer it as well as he could on 

 the backs of old examination scripts. If he 

 had not brought a pen of his own, he had to 

 search among a lot of ancient quills until he 

 could find one that was not hopelessly spoiled. 

 Presently, Mr. X would be politely summoned 

 to an inner parlour, where his last exercise would 

 be returned to him corrected and annotated, and 

 if hehad failed to answer any question he would 

 be either shown a solution or given a hint how 

 to proceed. 



Of course, it was not every pupil that was 

 taken separately like this ; some of them were 

 taken in small batches (not exceeding five or 

 six), but the general method was the same. It 

 should be added that once every week each pupil 

 took away with him a printed problem paper to 

 be done at leisure in his own rooms. The 

 results were marked, and the list was available 

 for inspection. 



As a member of St. John's College staff 

 Besant used to give "lectures" of a sort; but 

 (unlike Routh) he eschewed formal ' lectures on 

 bookwork. His -solutions of problems were 

 always original and elegant, and he had, the great 

 advantage (for a coach) of being equally good 

 in geometry, analysis, and dynamics. 



Besides being one of the par nohile fratrum 

 of coaches, Besant was a busy and trusted 

 examiner, and in this connection it may be 

 recorded that he used to say that ten minutes 

 of oral examination were worth any amount of 

 written ditto. 



Besant was too much engrossed by Jiis 

 proper work io add much to mathematical litera- 

 ture. His text-books on conies, dynamics, 

 hydrostatics, and hydrodynamics deserved their 

 popularity, and are still worth consulting, though 4 

 their point of view is now rather antiquated. His 1 

 one thoroughly original printed work, the tract 

 on roulettes and glissettes (first edition, 1869;- 

 second edition, enlarged, 1890), shows all his I 

 qualifications at their best. Besant had really | 

 studied Newton, and had an exceptional power ' 

 of estimating different orders of infinitesimals ; 

 from a figure. His invention of the term 



NO. 2485, VOL. 99] 



