12 



SCIENCE 



[N. S. Vol. XXV. No. 627 



duration of collision is comparable with 

 the period of vibration, the transfer of en- 

 ergy is 'infinitesimal on account of the ex- 

 treme rarity of these collisions.' 



This point of view has, I think, been sup- 

 ported by the experience of every one who 

 has attempted to obtain characteristic spec- 

 tra from gases under conditions in which 

 electrical and chemical processes were ex- 

 eluded. 



The assumptions back of Jeans 's discus- 

 sion are to be justified, if at all, by experi- 

 ment. Hence the importance of such work 

 as that which King,^* Hale^° and others 

 have recently been doing by means of the 

 electric oven, and of the fundamental ex- 

 periments of Wood upon optical resonance. 



As to the bearing of the Saturnian atom 

 upon this fact, one finds it in about the 

 same position as any other elastic atom, 

 except that for ordinary mechanical rigid- 

 ity one has to substitute the quasi-rigidity 

 which comes from rotation of the electrons 

 about the positively charged center. 



The effort to render a gas radiant by 

 means of high temperature alone has been 

 aptly characterized by J. J. Thomson as an 

 effort to boil a tea-kettle by burning down 

 the kitchen; the spectroscopic analogue of 

 Lamb's roast pig. 



In view of all the evidence, the conclu- 

 sion would appear to be that spectroscopists 

 have greatly exaggerated the role of tem- 

 perature in terrestrial sources. That the 

 behavior of an are or spark is determined 

 largely by the temperature of its electrodes 

 there can be no doubt ; but it seems almost 

 equally certain that the effect of changing 

 temperature upon the character of the 

 spectra is produced through the intermedi- 

 ation of changed electrical conditions in the 

 source. 



CONCLUSION 



In turning the pages of Kayser's great 

 ="King, Asiroph. Jour., 21, 236, 1905. 

 "Hale, Adams and Gale, ibid., 24, 213, 1906. 



compendium, which so adequately repre- 

 sents the present phase of spectroscopy, 

 theie is but one. tinge of disappointment; 

 and this is that, in the presence of such a 

 wealth of facts, there is so little in the way 

 of fundamental well-established unifying 

 principles. And yet the only remedy ap- 

 pears to be one of the homeopathic sort, 

 namely, more facts. 



There is, however, this comfort: things 

 are not as bad as they used to be. Our 

 condition is somewhat that of the old judge 

 who never liked to admit that he was not 

 in perfect health. On one occasion, when 

 he was just recovering from an illness, a 

 friend met him on the street and asked him 

 how he was feeling. "Well!" said the 

 judge in reply, ' ' I am not quite myself, but 

 I am a great deal better than I was at the 

 time when I was not so well as I am now." 

 This, too, is to be remembered, that we 

 can never hope for any solution which can 

 in any sense be called final— aU solutions 

 are merely passing phases— the problem is 

 not one either of mathematics or of history. 



Let us then continue our search for the 

 facts of the ease confident in the belief that 

 when this work has been property accom- 

 plished, the unifying principle will be at 

 hand. It should be to us a matter of no 

 small pride that it is to our fellow members 

 in this section that spectroscopy owes the 

 bolometer, the curved-grating, the echelon, 

 the spectroheliograph, and, therefore, in 

 large measure the beautiful results ob- 

 tained with these instruments. 



In the meantime we must, I believe, all 

 gladly admit two things: first, that the 

 atom which we associate so closely with the 

 Cavendish Laboratory more nearly sup- 

 plies the desired principle than anything 

 else which has been offered, and secondly, 

 that the emission of a line spectrum is a 

 very imperfectly comprehended phenom- 

 enon. Henry Crev? 



