ON THK SPKCIFIC HEAT OF WATER. 229 



11. Value. ofJ a K 'for Distilled Water. The value of JIB*** having been determined 

 for "tap water" and the capacity of the calorimeter having been thence deduced, 

 the next step was to determine what correction of the value was required for distilled 

 water. The experiments for this purpose followed precisely the same course as the 

 " tap- water " series described in Section 10. The distilled water used had a con- 

 ductivity of about 1 x 10~ 6 at laboratory temperature. Experiments were made to 

 determine whether the air contained in the distilled water affected the specific heat, 

 but the results were inconclusive, as the differences which appeared were of the 

 same order of magnitude as the possible experimental errors. 



But in nearly all the experiments made with distilled water, the water used was 

 Substantially air-freed. This precaution was necessary in later experiments in which 

 the water was heated up to 70 C. or 80" C., as at these temperatures dissolved gas 

 would otherwise come off freely, probably in sufficient quantity to render the results 

 inaccurate. For this reason we adopted the plan of keeping the water in bottles of 

 a capacity of 8 litres from which the air was exhausted. In Appendix E are set out 

 the data obtained from 15 experiments, the result of which is to show that the mean 

 value of J over this interval for distilled water exceeds that for tap water by about 

 0'003. The table sufficiently explains itself. In the table in Appendix E the rise of 

 temperature due to obturator, stirring, &c., is deducted from the final temperature 

 before entering in the table. The joules required for heating the calorimeter are 

 obtained by the multiplication of A0 by the mean capacity per degree of the 

 calorimeter, as determined in the preceding section. The average initial temperature 

 of the 15 experiments is 13 '11 C., and the average final temperature 54 '51 C. 

 Making a small correction to reduce to round temperature figures, we obtain as the 

 value of the mean calorie, from 13 C. to 55 C., 4*1821, or in round figures 



J 18 = 4*182. 



This may be regarded as our fundamental figure, and we believe it to be accurate 

 within 0'001. The deduction of the value for distilled water from that for tap 

 water is based on the mean of 15 experiments, of which the mean error of individual 

 observations is about 1 in 2600, whilst the mean probable error in the result is 

 O'OOOS, so far as it depends on casual errors of experiment. 



In confirmation of the result at which we arrive, regard may be had to the results 

 of the experiments for shorter intervals described later in Section 13, which were 

 made to determine the variation in the value of J from point to point for smaller 

 intervals. From the mean values of J for the intervals 13 C. to 27 C., 27 C. to 

 40 C., and 40 C. to 55 C., we can deduce the mean value for the interval 13 C. to 



55 C. The value so deduced is 



J, 3 M = 4*1819, 



which is in close agreement with the figure to which the experiments already 

 i It -scribed lead. 



