462 



NATURE 



[March, 19, 1908 



observation of which I have latterly made use — which 

 deductive method I attempted in former times, and upon 

 which others have relied since — will, I am persuaded, con- 

 tinue to be incompetent to deal with this real problem of 

 jiature unless man's knowledge of molecular physics receive 

 such unhoped-for accessions as will enable him to trace the 

 history of single molecules. Meanwhile, what I advocate 

 is that we avail ourselves of the mixed method, which 

 introduces data established by observation to supplement 

 the deductive method at the point where the deductive 

 method fails. G. Johnstone Stoney. 



30 Chepstow Crescent, W., March 6. 



Postscript, added March 13. — Nature of yesterday's date 

 announces the last supposed spectroscopic detection of 

 water vapour upon Mars by one of Prof. Lowell's assistants. 

 Observations of a like kind had been recorded by Sir Wm. 

 Huggins and Prof. Vogel, and the wave-lengths of three 

 of the lines observed were measured by Vogel, two of 

 which may possibly be water lines recorded by Rowland, 

 but not the third. 



On the other hand, Campbell and Keeler in a better 

 climate did not see them. Now, however, they seem to 

 have appeared again. This would be the behaviour of a 

 very variable coloured vapour like NO^ ; and what 1 

 should desire is that an adequate study be made of the 

 absorption spectra of the several such vapours which are 

 unable to maintain themselves in our atmosphere on 

 account of the presence of water, but are presumably to 

 be found on Mars if water does not exist on Mars, and 

 which if present will account for the orange colour of 

 large tracts upon that planet, and for the variations of 

 its colour at different seasons which are conspicuous. 



It is to be regretted that the observers to whom we owe 

 so much — from Schiaparelli to Lowell — have kept in view 

 only one of the competing views as to the state of things 

 on Mars instead of- at each step considering them both, 

 especially as the one they have preferred is that which 

 some physicists have felt to be the least probable. 



G. Johnstone Stoney. 



The Isothermal Layer of the Atmosphere. 



Like Dr. Chree (p. 437) I have had experience of the 

 vagaries of self-recording instruments, but I have generally 

 been able to trace them to some remediable defect in the 

 instrument or to the ignorance or carelessness of those 

 who use them. I fancy that the man who constantly 

 uses a certain instrument, and uses it intelligently and 

 not by mere rule-of-thumb, has a fairly correct notion of 

 the magnitude of the errors to which it is liable. If not, 

 w-hat reliance are wc to place on any instrumental 

 observations? 



It is quite natural, however, to doubt the observations, 

 and when this investigation first commenced I confess that 

 I did the same. Now that hundreds of ascents have been 

 made with different instruments, in different countr^s and 

 in widelv different circumstances, and all the results 

 obtained are in striking agreement, such a view seems to 

 me to be quite untenable. It is true that different instru- 

 ments sent up with the same balloon have given widely 

 difTerent temperatures, but the results have been published, 

 not concealed, and the instruments improved. I ascribe 

 these discrepancies, which are the exception, not the rule, 

 to solar insolation, which we avoid in England by making 

 our observations after sunset. 



With regard to the general question, the difliculties of 

 registering a true temperature are two : — (i) stagnant and 

 unmixed air which may be at different temperatures in 

 different parts of the same garden ; (2) the proximity of 

 bodies of large thermal capacity, which by radiation and 

 convection mask the true air temperature. Kites and 

 balloons when they have left the earth are free from these 

 errors, excepting that No. i applies to a balloon which 

 does not burst when swimming at its highest point. Since, 

 however, stagnant air does not matter provided sufficient 

 time is allowed, and in this case time is allowed, I do 

 not see what source of error there can be save solar 

 insolation. 



My belief in the accuracy of the thermojnetric results 

 obtained in England is based on inference from the follow- 

 ing facts. If a good trace, together with the constants of 



NO. 2003. VOL. 7y] 



the instrument, is given to two persons, they, working 

 quite independently of each other, will get practically 

 identical results. If the trace and instrument only Ijie 

 given to two persons, they, each calibrating the instru- 

 ment for himself, will obtain similar results for tempera- 

 ture within the limits stated, but the agreement for height 

 may differ by a kilometre or more in the higher parts. 

 Hence I believe in the accuracy of the temperatures, but 

 do not claim any great accuracy for the heights. 



Now with reference to Dr. Chree's questions. 



(i) Each station is held responsible for the accuracy of 

 its own results, and I am not acquainted with the routine 

 pursued at each individual station, but the general practice 

 certainly is to test each instrument in spirit cooled by 

 solid CO, both before and after each ascent. 



(2) Answered above. 



(3) No. The instruments used on the Continent are 

 expensive, and being heavier require a more expensive 

 balloon, and we have no funds with which to meet the 

 expense, especially when it is remembered that balloons 

 and instruments in England are lost about three times out 

 of ten. We hope that this will be done on the Continent 

 before long. W. H. Dines. 



Classification of Secondary X-Radiators. 



In Natl'RE of Fcbruarv 13 there is a letter by Dr. C. G. 

 Barkla and Mr. C. A. Sadler in which the authors dividi- 

 the elements — according to the qualities of the secondary 

 X-rays emitted bv them — into four groups founded upon 

 the atomic weights, without consideration of any other 

 quality of the element. It may be of interest to mention 

 that practicallv the same classification was given by mi- 

 as early as iSq6 in the Naturwissenschaftlichc liiitidscliav 

 (vol. xi., p. 485), and that this classification was also 

 dealt with in a treatise published by Prof. Voller and 

 myself in the Atuialen der Physik und Clicinic (vol. Ixi., 

 p.' 88, 1897). To this treatise there is added a table printed 

 directly by the secondary rays of a great nuriiber of 

 elements, and this shows not only the great difference 

 between the elements of the different groups, but also the 

 agreement in the behaviour of the various elements of the 

 same group. B- ^^ alter. 



Hamburg, Physikalisches Staatslaboratorium, 

 March 2. 



Gods and Godlings. 



Lest some readers should infer from your obituar> 

 note on Sir Denzil Ibbctson (March 12, p. 443) that this 

 distinguished anthropologist invented the word "god- 

 lings " for the rural deities of India, it is worth noting 

 that " godling " was good English in the sixteenth cen- 

 tury, and has never been allowed to drop. The Philo- 

 logical Society's " New English Dictionary " quotes 

 Lambarde's " Perambulation of Kent " (i.';7o-6) on raising 

 altars "to this our newe found Godlyng " : and examples 

 from Drummond of Hawthornden, Dryden, and Peter 

 Pindar show the convenience of the word. Coleridge pre- 

 ferred " godkin " for a minor deity with masculine 

 attributes, but sanctioned " goddessling." Charles Colton 

 in 167s permitted a certain cult of "little Goddikins " ; 

 Coventry Patmore regarded " godlet " as the more dignified 

 appellative. Anthropologists have therefore had a fairly 

 ample choice ; but it should be added that in some of thi 

 above examples, at least. Dr. Mu '' '"'" -•■■■•— 



suspected a " jocular " intention. 



nd his coadjutors 

 David Patrick. 



Edinburgh, March 14. 



Tabulated Values of Certain Integrals. 

 In Nature, October 24, 1907 (p. 639), the integrals 



x = - I cos ii-du and y= f sin !jV« are given. I shall be 



2,' 2.1 



grateful if any of your readers can inform me where I 

 can obtain tables of the numerical values of these integrals, 

 or any other tables that will reduce the labour of the 

 numerical calculation of them. C. E. Adams. 



9 Telford Terrace, Oriental Bay, Wellington, 

 New Zealand, January 18. 



