Diffusion of Liquids. 



423 



has been made for temperature, as the exact coefficients for 

 this factor are not yet known. Besides, the temperature was 

 nearly the same in all cases of both series of experiments, being 

 with 'Graham about 60° to 65° F., and with me 14° to 16° 0. 



Salt diffused from n solutions in 24 hours. 



Formula. 



KOI ... 

 NaOl ... 

 NH 4 CI 

 LiCl ... 



CaCl,... 

 SrCl 2 .., 



BaCl 2 .. 

 MgCl 2 .. 

 CoCl 2 .. 

 KiCl,,. 

 KBr .. 

 NaBr .. 

 NH 4 Br 

 KI 



From my 

 tubes. 



From 

 Graham's. 



gr. 



0-0599 

 •0351 

 •0369 

 •0233 

 •0476 

 •0686 



gr. 



00638 

 •0373 



•0378 



•0547 

 •0766 



•0936 



•0974 



•0373 



•0444 



•0397 





•0394 





•0965 



•1037 



•0525 



•0645 



•0617 





■1364 



•1416 



Formula. 



NaT ., 

 KCy ., 

 KIN O3 

 NH 4 fr 

 NaN0 3 

 LiNO, 



CuS0 4 

 ZnS0 4 

 MnS0 4 



MgS0 4 

 Na 2 S0 4 



(NH) 2 S0 4 | 



From my 

 tubes. 



gr. 



0-1008 

 •0499 

 •0614 

 •0544 

 •0446 

 •0353 



•0429 

 •0294 

 •0252 

 •0269 

 •0226 

 •0209 

 •0478 

 •0482 



From 

 Graham' i 



gr. 

 0-0959 



•0715 



•0596 

 •0531 



•0296 

 •0241 

 •0235 

 •0194 



•0150 

 •0300 

 •0369 



It will be seen from the above that the correspondence in 

 the results for the chlorides, bromides, iodides, and some of 

 the nitrates is as close as could .reasonably be expected from 

 methods so different as were the two here employed. But the 

 differences in the cases of BaN" 2 6 , SrN" 2 6 , and the five sul- 

 phates are far greater than could be allowed for errors of 

 observation or accidental disturbances. Most of Graham's 

 values represent the mean rates of diffusion for periods vary- 

 ing from 5 to 8 days. The values for BaN 2 6 and SrN 2 6 , 

 however, are the mean rates for 11*43 days ; and those given 

 for MgS0 4 and ZnS0 4 are the mean rates for the period of 

 16-16 days. The other sulphates were diffused for 8 days. 

 The discrepancies might be accounted for, in some cases, by 

 assuming that the periods chosen by Graham were too long to 

 admit of a fair comparison. Indeed it would seem from 

 my results obtained with the sulphates, that had they been 

 diffused longer and the mean values taken, they would have 

 been equally as small as those found by Graham. It will 

 be of course recognized that a chief advantage in the method 



