234 MESSRS. W. R. BOUSFIELP AND W. ERIC BOUSFIELD 



values of J ', special care was given to the determination of the first interval from 

 0* C. to 13 C. In Appendix F are given the data obtained from 17 experiments, 

 like that above described, in which this interval was measured. The table shows the 

 initial and final temperatures a and b, the final temperature b being diminished by 

 the rise of temperature due to obturator and stirrer-heating. In the fourth column 

 is given the difference, A0, which is rise of temperature due to electrical heating only. 

 The total joules expended are placed in the next column, the joules required for 

 heating the calorimeter in the next, and the difference, i.e., the joules expended in 

 heating the water, in the next. The weight of water, multiplied by A0, which gives 

 the mean calories for the interval due to electrical heating, is placed in the next 

 column, and the ratio joules/calories, which gives the mean value of J for the interval, 

 in the next. 



Taking the means, we get for the mean interval 1'30 C. to 13'09 C., the mean 

 value for J, which is 4 '1921. 



In the last column are shown the differences of the individual experiments from 

 the mean. We may correct for the interval from C. to 1'3 C. without sensible 

 error by taking the approximate value for J as 4 '2. The value at which we thus 



arrivei8 Jo 13 4-1937. 



The mean error in individual values works out at about 1 in 2000. We estimate 

 that in reading any interval, long or short, by means of thermometers such as we 

 were using, there may be a casual error of 0'01 C., so that on an interval of 12 C., 

 the mean error of 1 in 2000 on individual values is about what we might have 

 expected. The mean probable error in the result is less than 1 in 6000, so far as it 

 depends on casual errors of observation. 



In Appendix G are set out shortly the results of the other point-to-point experi- 

 ments, which cover the intervals 13 C. to 27 C., 27 C. to 40 C., 40 C. to 55 C., 

 55 C. to 73 C., and 55 C. to 80 C. These results are calculated in the same way 

 as for the interval C. to 13 C. When these results are corrected for small 

 differences of temperature at the beginning and end of the intervals, we obtain for 

 the mean value of J in each interval the figures set out in Table VIII 



TABLE VIIL Mean Values of J for Intervals. 

 Interval. 



6. J/. 



C. C. 



13 4-1937 



13 27 4-1752 



40 4-1756 



40 55 4-1935 



55 73 4'2024 



55 80 4-2056 



