BETWEEN THE FREEZING AND BOILING-POINTS. 
227 
It will be seen that over the range of flow from *67 gramme to ‘39, the value of 
the heat-loss remains constant to within the limits of error of the different measure¬ 
ments, and gives a mean value of '07128 in watt-seconds per degree rise. The 
largest variation from the mean is for a flow of '49 gramme per second, and amounts 
to '00019 watt on a total supply of 2T watts per degree, which is less than 1 in 
10,000. For the small flow, the difference from the mean of the other flows amounts 
to '00069 watt on I T watts. This shows an increase in the heat-loss of nearly 
7 parts in 10,000, and is much too large to be included within the limits of error. It 
is evident, then, that for flows below '3 gramme per second, the conduction effect 
commences to be measurable, and cannot be eliminated by the method of “ cold ” 
readings. 
For Calorimeter C, the measurements for the different flows are given under date 
February 20, 1899. These are for double intervals of 15 minutes each, and include 
the same limits of flow as for Calorimeter I). I have taken S = '00490, instead of 
— '00469 as given from these measurements, so as to give results comjmrable with 
the other sets of observations made at about this time. 
Calorimeter C. —Mean Temperature, 29 o- 09 C. February 20, 1899. 
§ = '00490. J = 4T794. 
(16. 
f 
Q. 
(EC - JQ d6)/dd. 
Difference from 
mean, '04997. 
8-2608 
•398498 
•04972 
- -00025 
8-2560 
•398540 
•04988 
- -00009 
8-2199 
•501957 
•05016 
+ -00019 
8-2301 
•501026 
•04998 
+ -00001 
7-9646 
•666042 
■ 05009 
+-00012 
7-9775 
■664388 
•04999 
+-00002 
8-2281 
•258114 
•05057 
+ -00060 
8-2284 
•257947 
•05070 
+ -00073 
The largest variation from the mean value of the heat-loss is -—'00025 watt, and 
amounts to a little more than 1 in i 0,000. For the small flow, the divergence from 
the mean of the other flows amount to '00067 watt, and shows that the heat-loss has 
been increased, which agrees very closely with the result obtained for Calorimeter D. 
The agreement of the results for the two calorimeters, the one with a 3-millim. bore 
tube and the other with a 2-millim., seems to prove fairly conclusively that the 
increase in the heat-loss taking place below a certain limit of flow, cannot be 
attributed to a change in the radiation loss from the fine bore tube, but can only be 
2 g 2 
