ON THE VISCOSITY OF SOLUTIONS. 405 



Fawsitt * was unable to confirm these results of Rudorf, and did not obtain any 

 evidence of " negative viscosity " at any concentration. Our measurements, while not 

 in close agreement with Fawsitt's results, also give no indication of " negative 

 viscosity" at 25° C, the values of 1/% being 1*005 for 0'125 m. solutions, and T002 

 for '03 125 m. solutions. 



When determinations were made at 8° C, the ratio became slightly less than 1, but 

 only in the case of the most dilute solution. The ratio n\n Q for a concentration of 

 0-03125 mol per litre is 0*9985. This value, and also all the other data for carbamide 

 solutions, have been repeatedly determined with independent solutions, and with two 

 different samples of carbamide, and though it is so nearly equal to 1, it is certainly less 

 than 1. The "negative viscosity" would probably be more pronounced at lower 

 temperatures. 



Hitherto the only exception to the general rule that aqueous solutions of electro- 

 lytes alone exhibit " negative viscosity " has been the fact, established by Muhlenbein 

 and Wagner, t that some organic substances in organic solvents, e.g. cyanobenzol in 

 ethyl alcohol, also exhibit it. Aqueous solutions of carbamide have now been shown to 

 do the same. Other substances may also do so at low temperatures, whilst, as already 

 mentioned, mercuric cyanide may be expected to do so at high temperatures. 



6. It seemed probable that a comparison of the viscosity-effect of adding 

 equivalent quantities of an electrolyte, and of a non-electrolyte, to water on the one 

 hand, and to an aqueous solution which at the temperature of experiment has the 

 same relative viscosity as water {nj% — 1), might yield interesting results. 



Accordingly, a large quantity of solution of potassium chlorate was made up 

 (approximate concentration "3 mol per litre) which at 30° C. had the relative viscosity 

 0'998. This substance was chosen because slight variation in the concentration has 

 very little effect on >i, nor does >//% change to any great extent with change of 

 temperature ; the necessary concentration, also, is not great. 



It was necessary to select a salt which has no ion in common with potassium 

 chlorate, so as to avoid alterations in ionisation, etc. ; ammonium iodide fulfils this 

 condition ; carbamide was used as a non- electrolyte which does not react with the 

 chlorate. 



The results obtained show that when the amount of solute added is small there 

 is no recognisable difference between the pure water and the chlorate solution. When 

 the amount of solute added reaches 1 mol per litre there is a slight but distinct 

 difference in both cases ; this is probably mainly due to the fact that the ratio of the 

 number of molecules of water to those of ammonium iodide or carbamide becomes more 

 widely different when the number of molecules of the solute is great. 



This difference is in the same direction and approximately to the same extent with 

 both solutes. These experiments do not throw much light on the question of the 



* Proc. Roy. Soc. Edin., 25, p. 52, 1904. 



t Ze.it. f. Phtjs. Ghem., 46, p. 872, 1903. 



TRANS. ROY. SOC. EDIN., VOL. XLV PART II. (NO. 15). 57 



