394 



EEPORT 1886. 



Fig. 3. 



Fig. 4. 



mimerary tubes and stopcocks, being sometimes employed to assist this 

 (fig. 3). 



To diminish evaporation a layer of parafiBn was sometimes put on each 



vessel, and my assistant, 

 Mr. Robinson, hit on the 

 happy device of connecting 

 the two layers of paraffin 

 by another and a much 

 shorter and stouter siphon 

 tube full of paraffin, so as 

 to equalise the levels and, 

 if possible, to keep them 

 equal, and yet not to tap off 

 any of the current by this 

 supernumerary but almost 

 infinitely resisting path (fig. 4) . 



We now got much more consistent results, and for a long time thought 

 things were pretty satisfactory, so we proceeded to make numerous obser- 

 vations, varying the character and 

 length of the tubes, the number of 

 volts applied, the strength of the 

 solution, &c. 



The difficulty of electric endos- 

 mose is an obvious one, and it may be 

 owing to this that the results of these 

 experiments are not very concordant. 

 Omitting minor corrections, however,, 

 and taking the measured ratio of the 

 distances of the ring precipitate, from 

 the BaCl2 vessel eni, and from the 

 Na2S04 vessel end of the tube, for all those experiments where a paraffin 

 levelling tube was used, we get the average i-esult that the ring forms 3'2 

 times as far from the BaClg end as from the NaaSO^ end (see p. 400). 



This may be recorded by saying that the Ba travels three times 

 the distance that the SO4 travels in the same time, or that Ba travels 

 three times as quickly as SO4 ; a result which, though simple a^d definite, 

 is probably incorrect. But it is possible to argue that what we are 

 measuring is not this simple ratio of speeds but a more complex ratio 

 depending on all the substances. Thus, consider the substances in action, 



Naa SO4 I HCl HCl HCl | Ba CI2, 

 and pass the current from right to left. 



On Hittorfian principles, H travels to meet the SO4, and journeys the 

 greater part of the distance between them ; CI also travels to meet the 

 Ba. Call the true velocities of these four ions, h, s, c, b respectively ; then 

 the position of the precipitate of the meeting Ba and SO4 may be thought 

 to really measure, not b : s, but 



b s 

 c ' h' 

 Now, on Kohlrausch's calculation, 7t=29,Z>=3'3, c^5'3, sis not certain, 

 but from Hittorf we may take it as roughly equal to the velocity of K, 

 viz., 51 ; in which case 



bh 3-3 X 29 OK 

 cs 6-3x5"l 



oil 

 SwCl 



Paretffin 



■mis ^"4 



