1903.] and the Mechanical Friction of the Solvent. 



at moderate temperatures with the quadratic interpolation formula, 

 already applied by Poiseuille, has long been known to me. The 

 literary priority regarding this observation belongs, however, to 

 Messrs. Bousfield and Lowry. 



In my calculations I have made use of the values chosen by Mr. 

 Heydweiller,* as the most probable between the limits of 0° and 30 \ 

 These values were selected from different observers, especially Messrs. 

 Thorpe and Eodger. Since the constants were calculated for this 

 range of temperature, they have practically the same signification for 

 the fluidity of water as the former constants for the conductivity. 



The formula 



<p t = 55 -68 + 1 -981/ + 0-0105* 2 C.G.S. 



or 



fa = 94-74[l +0-0249(7- 18) + 0-000111(7-18) 2 ] C.G.S. 



represents the fluidity <f> within the limits of the table, 0° to 30°, with 

 a maximum error of 1/1000, that is, with about the same degree of 

 accuracy as Deguisne's formula for the conductivity. As far as 90° 

 the error would not exceed 1/100. 



The curve represented in the figure instead of having the factor 

 94*74 (which has no reference to the conductivity), was given the 

 arbitrary factor 67*0 in order to give the curve the desired position 

 close to the lowest curve of conductivity, that of sodium valerate. 



The coincidence of the two curves is striking, indeed the coefficients 

 0-0249 and 0-000111 differ little from those of sodium valerate 0*0244 

 and O'OOOlll.f The curves of conductivity are cut by the fluidity 

 curve approximately in the same region to which they converge. 

 The curve of fluidity passes through zero at - 34° G.J 



Messrs. Bousfield and Lowry calculated from the measurements of 

 Thorpe and Rodger the coefficients 0*0251 and 0*0001 15. This curve 

 differs very little from mine, especially in the neighbourhood of the 

 crossing point. 



* Mr. Heydweiller calculated at ray request the table for the ' Lehrbuch der 

 Praktischen Physik ' (Tab. 20a, 1901). 



t The same would apply for Deguisne's observations on Na 2 HP0 4 (0'0241 and 

 O-000105) and on NaHC 4 H 4 4 (0'0241 and 0'000109), which, however, on account 

 of the unknown constitution of these salts in solution, I have left out of account. 



Messrs. Bousfield and Lowry further called attention to the fact that the tem- 

 perature change which I have found for ordinary distilled water (practically a 

 very dilute solution of C0 2 ) corresponds with the temperature change of fluidity. 



X It is also to be mentioned that Messrs. Lyle and Hosking, from their interest- 

 ing observations on the viscosity and the electrical resistance of O'l to 4 normal 

 solutions of NaCl between 0° and 100° draw the conclusion : " The curves so 

 arrived at are remarkable, in that they indicate that for solutions of the strengths 

 experimented with, both the fluidity and the sp. mol. conductivity vanish at a 

 temperature of — 35°* 5 C." The manner of extrapolation is not dealt with. — 

 1 Phil. Mag.,' May, 1902, p. 496. 



