THE MECHANICAL EQUIVALENT OP HEAT. 
497 
investigation could bring our results into absolute agreement with those obtained by 
Rowland, since, owing to the difference in the expressions for the temperature coeffi¬ 
cients of the specific heat of water, it is inevitable that our conclusions should agree at 
some one temperature, but must necessarily differ when expressed in terms of a thermal 
unit at any other temperature, and, thus, changes in the values of the units would only 
alter the temperature of agreement. For example, Dr. Guillaume has pointed out 
to me that the experiments of Commandant Defeoeges lead to the conclusion that 
tlie value of g, at Greenwich, should be increased from 981‘17 centims. to 981'24 
centims. A similar correction would slightly increase Rowland’s value of J, and 
thus cause our point of agreement to be about 12° C. instead of 11°‘5 C. Again, the 
value of g, assumed by Lord Rayleigh (‘Phil. Trans.,’ A., 1884, p. 427), would have 
to be slightly increased, and the resulting values of the electro-chemical equivalent of 
silver, and of the absolute electromotive force of a Clark cell, would require modifi¬ 
cation, but the only result of any such change would, as before, be to shift the 
temperature of agreement. 
It is, therefore, evident that the difference is chiefly due to ermrs in thermometry. 
An error of 0°'01 C., in the value of the range 14° to 25° C., would suffice to explain 
the discrepancy between the results, and the close agreement in the range value, as 
obtained from our standard and the Paris nitrogen standard (see p. 430, supra), 
appears to me to warrant a brief criticism of Rowland’s methods and conclusions. 
The science of exact thermometry has made great strides during the past fourteen 
years, and, no doubt, much of this progress is due to Rowland’s initiative, for his 
work undoubtedly marks a distinct advance in this subject. Great, however, as his 
services have been, he would, I feel sure, be the first to admit that his investigations 
of 1878-79, by no means exhaust the possibilities of further progress in the science 
of thermometry. In his paper, he constantly indicates possible causes of error, and, 
in some cases, makes valuable suggestions as to methods by which they may be 
eliminated. 
In a previous paper (‘Phil. Trans.,’ A., 1891, p. 155) it was pointed out by 
Mr. Callendae and myself that “ the indications of the air-thermometer cannot be 
relied on beyond 0°‘01 C.... the limit of accuracy of a single reading of the barometer.” 
On p. 95 of his paper, PvOWLAND gives the data on which the following state¬ 
ments are based. A difference of ’03 millim. in the reading of his barometer gave a 
difference of 0°’01 C. in the air-temperature, and an error of 0°’3 C. in the estimation 
of the temperature of his barometer-column would produce the same effect. From 
his remarks, it is evident that he regarded his possible error from this cause alone 
as ± 0°-01 C. 
Again, his temperatures depended on the direct observation of the height of a 
mercury manometer column by means of a cathetometer, and everyone who has 
worked with such an instrument knows the difficulty of obtaining measurements 
accurate to ’02 millim. (It is true that small differences can be read to a much 
mdcccxciii.—A. 3 s 
