414 Dr. J. H. Long on the 



and a third was rejected for still another cause. After a num- 

 ber of trials six were found which gave constant results, and 

 diffused amounts of salt in equal times proportional to 17, 18, 

 17, 17, 17, 17. Of course these rates Avere subject to slight 

 fluctuations ; but the averages for any length of time were 

 constant. In searching for the cause of the larger amount 

 diffused from the tube ~No. 2, it was found that in it i was 

 slightly shorter than in the others. A normal NaCl solution 

 stood in the six tubes at the height ef when the velocity of 

 the water was 40 cubic centims. The exact dimensions of the 

 tubes are as follows — the internal diameter of ar st 1*5 mil- 

 lim., the diameter at b 1 millim., the diameter of i 14*5 mil- 

 lims., the distance ge 11 millims., and the distance g k 15 

 millims. The beaker-glasses, when filled to the level a 6, con- 

 tained 240 cubic centims. All the experiments described 

 below were performed with these six tubes, and those with the 

 coloured solutions above also. The other experiments cited 

 above were performed with other tubes having slightly differ- 

 ent dimensions. In these the part i had a slightly greater 

 diameter (about 16 millims.) and was somewhat shorter. 



At the beginning of the series of experiments given below 

 each tube was filled with a different salt-solution ; but this 

 plan was soon abandoned, and two salts were then diffused 

 simultaneously — that is, in three of the glasses a solution of 

 one salt, and in the other three a solution of the other. In 

 some cases the amounts diffused from each of the three glasses 

 of each set were determined separately; and in other cases (with 

 the nitrates and sulphates always) the amounts from the three 

 tubes of each set were determined together. The values given 

 below are in all cases the averages reduced for one cell. 



The amount given by tube "No. 2 was first reduced to the 

 standard of the other five. 



The velocity of the water was maintained constant at 40 

 cubic centims.; and the temperature of the room was 14°-16° C. 

 As explained above, after filling the tubes, about 48 hours 

 were allowed to elapse before the determinations of the diffused 

 salt were commenced. The appearance of the diffusion-cells 

 as they stood when in operation is shown in fig. 3 (PI. VII.). 



The Mariotte bottles held in some cases 8 litres, and in 

 others 4. The latter are here represented. 



The salts made use of were in most cases obtained quite 

 pure ; in other cases it was necessary to purify them myself 

 (LiCl, LlN"0 3 for instance). Where possible, the solutions 

 were made by weighing out the required number of grammes 

 of dry or well-crystallized salt and dissolving this in a litre 

 of water. The strength of this solution was then afterwards 



