at Very Low Rates of Sheai 



G91 



The tube was then removed and very lightly silvered inside 

 by means of a silvering- solution. It was then replaced, and 

 the apparatus rilled with water from the same source as 

 previously and allowed to stand 44 hours, at the end of which 

 time the following results were obtained : — 



Time 



liquid was 



in tube. 



44 h. 



Cali- 

 Temp. bration 

 reading. 



Change 



of scale 



in 5 min. 



Value of; 



scale- 

 division. 



K 



Viscosity 



(low 

 shear). 



Viscosity 

 (high" 

 shear). 



Per cent. 



dif- 

 ference. 



1247 15-83 



7-62 



•001670 



•06506 



01222 



•01226 



-•3 



4-4 h. 



12*20 



7 60 





•06.312 



01226 



•01225 







45 h. 



1232 „ 



7-69 



» 



•06486 



01211 



01221 



-•s 



46 h. 



1233 



7-61 



•• , 



•06509 



•01224 



01220 



+ •3 



Thus the difference disappears when a silvered tube is used. 

 The silvering was too slight to affect the diameter of the 

 tube in any appreciable proportion, and besides, an appreci- 

 able decrease of diameter would have caused an increase in 

 the result for the coefficient of viscosity while the effect was 

 a decrease. The results stated seem to show conclusively 

 that small quantities of some constituent of glass, probably 

 silicates of the alkali metals, dissolved out by water, have a 

 very great effect on the viscosity of the water at very low 

 rates of shear. The maximum rate of solution of glass (of 

 glass vessels) in distilled water at 18° was found by 

 Kohlrausch* to be 0*48 nig. per square dcm. per day, the 

 rate for most forms of glass being much smaller than this. 

 At t\\\< maximum rate the amount dissolved in the water in 

 the tube used in these experiments would produce a concen- 

 tration of 0*000007 gm. per c.c.,and this (or probably a much 

 smaller amount) increased the viscosity at low rates of shear 

 by 3 per cent. The large effect produced may be due to the 

 tendency of >ilicic acid to separate out from the solution in a 

 gelatinous form. The result was of course not due (at least 

 directly) to impurity existing in the water before it entered 

 the tube : for this would leave the effect of the silvering 

 wholly unexplained. The distilled water used was that pre- 

 pared fur ordinary laboratory use by distillation in a block- 

 tin still. The glass tube was of soft German sodium-glass. 



Another tube gave similar high and somewhat variable 

 results for the viscosity of water and the ordinary value for 

 that of kerosene. Silvering the tube was not tried as it was 

 accidentally broken. The results obtained bv these two tubes 



* Ber. der deutsch. Chem. Ges. vol. xxiv 



.61. 



