at the total solar Eclipse in 1882. 
67 
to consider the rates observed at other times during the three 
days and the following table, which gives the amount of water 
distilled in each hour in the different days, may be used. 
If the experiments were to be repeated, I do not know in what 
particular the conditions of weather, as they were on the forenoon 
of the 18th, could be made better. Yet if we admit that they 
might be improved in the proportion that the conditions between 
10 and 11 a.m. on the 18th are better than those obtained 
between the same hours on the 16th, when they were the most 
unfavourable, the rate would have to be increased in the proportion 
58 8 : 82'4 and it would become 15 x T4 = 21 c.c. per minute. 
1 
Time 
Cubic centims. water distilled 
from to 
16th 
17th j 18th 
A.M. 
9 10 
78-1 
10 11 
58-8 
70-25 82-4 
11 12 
69-5 
69-7 82-8 
P.M. 
2 3 
73-2 
65-5 | 71-6 
3 4 
57-9 64-8 
This correction is certainly too great when considered as an allow- 
ance for faulty weather, and even if held to cover all instrumental 
deficiencies, such as imperfect equatorial adjustment and others, 
I believe it will be still much in excess of the truth ; moreover, 
I am convinced that if the calorimeter furnished steam at this 
rate it would be in such conditions that it would be impossible 
to stand by it and attend to it on account of the excessive heat. 
Having thus indicated the maximum correction which can 
be applicable to our observations we return to the consideration of 
the observations themselves, where we are on the sure ground of 
experiment. 
In the circumstances we may, without sensible error, take 
the cubic centimetre of water to weigh one gramme. In specify- 
ing quantities of heat we do so in gramme-degrees (Celsius) 
(gr.° C.), or kilogramme-degrees (kg.° C.), as the case may be. 
Similarly, quantities of work are expressed in kilogramme-metres 
(kgm.). We take the latent heat of one gramme of steam as 
535 gr.° C., and the specific heat of water as unity, and the 
mechanical equivalent of heat as 0 - 425 kilogramme-metres per 
gramme-degree. Therefore the heat required to transform T5 grs. 
5—2 
