322 Notices respecting New Books. 



flakes were freshly split, their thickness determined with the 

 spectroscope, and they were torn to pieces with ivory forceps 

 to avoid the possibility of metallic contamination, which 

 might result if scissors were used. In all cases I find 

 evidences that metallic conduction commences with one volt 

 of applied potential when the plates are between 20 and 30 

 wave-lengths apart, but this distance can be accurately 

 measured only when semitransparent metallic films are used, 

 so that my results have not been very consistent. 



I have worked with the speculum plates mounted in the 

 interferometer, and though I have obtained conduction before 

 the plates appeared to be in contact, it appears to me that 

 the separation can be much greater when mica flakes are 

 present than when the gap is quite free from all material. 

 With clean mica flakes pressed between the plates I have 

 obtained conduction when light is still freely transmitted 

 through the crevasse. With the plates mounted in the in- 

 terferometer, however, the crevasse transmits little or no 

 light when conduction commences. The light is of a deep 

 blue colour, which is an interesting observation, as showing 

 that a narrow crevasse refuses transmission to the long waves 

 first. The colour effect is doubtless the result of the cir- 

 cumstance that the light which gets through has suffered 

 multiple reflexions, though the same thing would doubtless 

 be observed, though to a less degree, with a metal having 

 no selective properties. 



Upon the whole I feel disposed to believe that the ex- 

 periments indicate that the effect is a real one, due possibly 

 to the presence of an atmosphere of electrons in the imme- 

 diate proximity of metallic surfaces. I realize, however, that 

 it will be necessary to carry out to a successful conclusion 

 the experiment with the semitransparent films before any 

 definite conclusion can be reached ; for there seems to be 

 evidence that the presence of a solid dielectric such as mica 

 or quartz dust increases the effect. 



XXVIII. Notices respecting New Books. 



The Structure of the Atmosphere in Clear Weather. A Study of 

 Sounding with Pilot Balloons. By C. J. P. Cave, M.A. 

 Cambridge, at the University Press, 1912. Pp. xii + 144. 



HE hook deals with observations made between January 1907 

 and March 1910 on 200 balloons sent up by the author, mostly 

 from Ditcham, Hampshire. A few were of the " ballon sonde " 

 type carrying instruments, but the large majority were true 



T 



