442 Proceedings of Royal Society of Edinhurgh. [sess. 
oil and in air, and mercury was also dropped in air, so as to obtain 
three different conditions of parting. 
The following are the numbers for water in oil : — 
Growth Time 
of Drop. 
120 " 
27" 
11*3" 
Vol. of Drop. 
•4096 
•4607 
•5611 
Quantity passing through Neck 
after Uupture had begun. 
•0091 
•0570 
•1479 
Kesidue. 
•4005 
•4037 
•4132 
The quantity passing through the neck was calculated by multiply- 
ing the rate of flow by the time of rupture of the neck, — 3 seconds. 
By subtracting this quantity from the observed drop we ought to 
get a normal drop, so that the figures in the column headed “ residue ” 
ought to be exactly the same, and we see that is not the case. After 
some study the cause was found in the viscosity of the oil, which 
resists the downward movement of the growing drop ; so that on 
measuring the extreme width of the drops with a cathetometer at 
different rates, it was found that the higher the rate the larger was 
the drop, as shown by its greater diameter. 
In the case of water dropping in air, this effect had quite dis- 
appeared, while in the case of mercury in air an effect of an opposite 
character was manifested. 
In the case of water dropping in oil, the time taken to cut off the 
neck after the drop was ripe could be measured with ease, and was 
found to be independent of rate; but in the case of mercury 
dropping in air direct observation could not be made, so the time of 
rupture of the neck was calculated by finding how long the flow 
must have continued through the closing neck in order to account 
for the increase in weight. 
Rate of 
Calculated Time Occu- 
Addition to 
Normal 
Observed Drop 
Flow. 
pied by Rupture. 
Drop through. 
Drop. 
at given Rate. 
6-431 
X -036" 
•2315 
+ 
4119 
•6434 
4-594 
X *042" = 
•1884 
-i- 
4119 
•6003 
2-450 
X -0515" 
•1262 
+ 
4119 
•5381 
1-252 
X -0707" 
•0885 
+ 
4119 
•5004 
0-5788 
X 
o 
oo 
-4 
•0504 
+ 
4119 
•4623 
0-4598 
X -104" 
•0478 
-1- 
4119 
•4597 
From these numbers we see that the quicker the rate the shorter 
must have been the duration of the neck after rupture had com- 
menced. This apparently anomalous decrease of the time arises 
from the same cause as the opposite effect in the case of water in 
