442 
PROFESSOR G. G. STOKES OK MR. CROOKES’S EXPERIMENTS ON 
supposition of its truth, which is found to be in accordance with the results of 
experiment. 
Hitherto the ratios of the coefficients of viscosity have been deduced from a part of 
Mr. Crookes’s tables, in which the logarithmic decrements changed very evidently 
with the pressure. We may now deduce those ratios by what may almost be deemed 
an independent method, namely, by attending only to the part of the tables at which 
the logarithmic decrement is all but constant. If the condition of ideal simplicity 
supposed at the outset were quite attained, we might disregard the pressures in the 
comparison, which would entitle that method to be considered quite distinct from the 
former; but as that condition is not absolutely reached, it will be proper not wholly 
to neglect the condition of correspondence of pressure, though a rough determination 
of correspondence will suffice. 
Suppose, then, we take the air pressures from 120 to 2G millims. The ratio of the 
corresponding pressures in oxygen, &c., is given in the last line but two of the 
preceding table. The corresponding limits are for oxygen 132 to 29 ; for nitrogen 
and carbonic oxide the same (sensibly) as for air; for carbonic anhydride 65 to 14; 
for hydrogen the limits are not given in the table, but they are 864 and 187 nearly. 
In strictness each pressure should be considered separately ; but as the intervals were 
not intentionally divided in a different manner for the different gases, and as the 
logarithmic decrements are very nearly constant between the specified limits, it seems 
sufficient to take the mean for each gas of those corresponding to pressures that lie 
between the assigned limits. We thus get for air, ‘1002; oxygen, '1120 ; nitrogen, 
'971 ; carbonic anhydride, ‘822; carbonic oxide, ‘971 ; hydrogen, ‘500. Reducing to 
air =1, and writing down for comparison the numbers expressing the ratios of the 
coefficients of viscosity to that of air given in the previous table, we have for the 
ratios in question— 
From air pressures, 760 to 160 
O. 
1-121 
. m 
£ <?> 
o 
C0 2 . 
0-829 
CO. 
0-971 
H. 
„ „ 120 to 26 . 
1-118 
0-969 
0-820 
0-969 
0-499 
Values adopted. 
1-120 
0-970 
0-822 
0-970 
0-499 
We see that almost identically the same numbers are obtained whether they are 
deduced from the higher pressures, for which the logarithmic decrements notably 
diminish with the pressure, or from the part of the tables in which they are nearly 
independent of the pressure. The greatest difference is in the case of carbonic anhy¬ 
dride, where it is rather more than one per cent. This difference is in part accounted for 
by the omission of the correction for the time of vibration. If the times of vibration 
at corresponding pressures as determined by Mr. Crookes be taken, they will be found 
to be very nearly the same ; indeed, the differences are quite comparable with the 
errors of observation of those times. Perhaps the differences could be got with greater 
certainty from theory than from observation. According to theory the effect of the 
