134 G. H. KNIBBS. 
loss at the entrance to the tube is sensibly the same for efflux at 
constant temperatures under any pressure,! for the times of efflux 
are very nearly equal. The results have been reduced to their 
values at 6° and 11° C. by formula (37), and the higher pressures 
to the value for 24 atmospheres. | 
Konig, 1885.—Using two tubes of nearly equal sectional area, 
Kénig made five efflux experiments between the temperatures 
155° and 18-1° C., correcting by Hagenbach’s formula. In my ~ 
reduction I have assumed g to be 980-95: its value was not 
mentioned in Kénig’s account of his work. a 
Couette, 1890.—In his interesting ‘Etudes sur le frottement des 
liquides,’ Couette gives a method of observation by which the 
corrections for fall in pressure at the entrances of tubes and fot | 
their end conditions—see (12a) and (13)—are eliminated. By 
means of selected capillary tubes of equal radii, but of vey ‘ 
different lengths, he connects three reservoirs, through which & 4 
‘flow from a still larger reservoir is established under pressure, the 
efflux through both tubes finally becoming equal. The pressures 4 
in the three reservoirs are then measured by a manometer of ae 4 
branches. Putting y =mpq?/(x? R*), / having the same significe 4 
tion as in § 9, P, denoting the fall in pressure betweet the 
reservoirs connected by Z,, and P, the corresponding fall betwee? 
the reservoirs connected by Z,, we have ; 
oe. Le eee ee? : 
hence if, as there is reason to believe, y, =y. and 1, =ls, and if 
also it were possible to employ two tubes of equal radius BW 
should have 
=-——, 1? (42 
8 7 L, -L, i ) z that 
Couette’s formula.? But more rigorously it may be assumed ae 
this is not possible ; and therefore we are involved in considering 
ea cinahmeettietatnleel hcl cree ai pa asia tg te pee 
1 The difference in the correction will probably always be les* ans 
-the unavoidable errors in the recorded times of efflux. 
? Annales de Chimie et de Physique, 6 sér. t. 21, p. 468- 472+ 
