On the Diffusion of Liquids, 413 



Note. — Silver films might possibly be found useful in the 

 construction of illuminating apparatus for the microscope. If 

 well varnished, they might well replace the usual mercurial 

 reflecting surface used in illuminating-mirrors, as the mercurial 

 layer is very perishable, particularly in a damp climate, as in 

 India during the rainy season. 



LIV. On the Diffusion of Liquids. 

 By John H. Long, of Lawrence, Kansas, U. S. A. 



[Concluded from p. 329.] 



IN the above I have explained the general workings of the 

 method and discussed its chief sources of error. It 

 remains now to mention several other causes of error. From 

 the tables given it is seen that the amounts diffused undergo 

 slight fluctuations from time to time. Yarious causes can 

 be assigned for this. The temperature undergoing slight 

 changes has its effect, of course, in producing currents in the 

 solutions, besides changing for the time being their rates of 

 diffusion. Other disturbing currents are produced when, by 

 any cause, the diffusion-vessels are jarred, as when, for instance, 

 a door in their vicinity is rapidly closed. In order to avoid 

 another source of error, it is to be observed, of course, that the 

 table upon which the vessels stand be horizontal, in order that 

 g h be parallel to ef. 



I turn now to the experiments relating to the comparison 

 of different salts as to their rates of diffusion — the real object of 

 the investigation. This necessitated the performance of a 

 great number of single experiments ; and in order that the 

 results of these be comparable, it was necessary that the con- 

 ditions be in all cases the same. This end could be attained, 

 of course, by diffusing all the salts, one after the other, from 

 the same tube, observing the above-mentioned points regard- 

 ing concentration &c. But to carry the investigation out in 

 this way would require a very long time. The only alterna- 

 tive was to make use of a larger number of diffusion-tubes ; 

 and this I did. Twelve tubes and beakers, having as nearly 

 as possible the same dimensions and of the form represented 

 in fig. 1, were obtained, and compared by filling them with 

 2n NaCl solution and observing the amounts diffused in equal 

 times, the velocity of the water being in all cases 40 cubic 

 centims. Some of these tubes were almost immediately re- 

 jected : in one the section of i had been left elliptical instead 

 of circular by the glassblower ; in another the point b was too 

 small and did not allow the water to pass through fast enough ; 



