1920] 



on Low Temperature Studies 



267 



therefore easily obtained. When necessary a correction was made 

 for the quantity in the curved surface of contact between the liquid 

 and the ring. Screens of approximately O'l mm. thickness were 

 readily obtained by using a tambourine of silk-gauze chiffon instead 

 of a rubber membrane. This was dipped into the liquid to be 

 examined, and when lifted out retained a fairly uniform layer in 

 the tiny spaces between the crossed silk threads. Separate measures 

 of the proportionate obstruction of the stretched gauze alone showed 

 that only '60 per cent, was stopped : a result found to be in agree- 

 ment with microscopic measurements of the threads and spaces of 

 the material. The wetted gauze was then cooled as rapidly as 

 possible and weighed in the same way as the membranes, etc. The 

 chief objection is that the mixed liquid and silk surfaces are not 

 strictly plane. 



The results obtained are grouped later, but some aspects of their 

 interpretation should first be dealt with. An obvious correction to 

 such measures is the reflection loss from the plane surfaces. This, 



as given by Fresnel, is y = [—— ^) ,' where y is the proportion 



reflected at the surface, and n the index of refraction between the 

 media separated by this surface. 



The curve (Fig. 14) shows the growth of y (ordinate) with n the 

 index of refraction (abscissa). The upper curve is for substances in 

 air, and the lower curve gives the reduced values of y in liquid air 



for values of n, corrected by ^—^, the relative indices for air, gaseous 



and liquid. Table III. gives the values for four typical substances 

 with increasing n, corresponding to normally incident " D " line 

 radiation. 



Table III. 



This correction for the rubber membiane or rock-salt is within 

 the experimental error, but for silver chloride is quite considerable. 

 When this substance is used to obtain very thin capillary films of 

 different liquids, the results may be quite misleading unless carefully 

 analysed. So also for direct measurements of iodine or other 

 very refractive materials. This is, however, fairly simple on the 



