﻿Prof. It. Bunsen's Calorimetric Researches. 393 



The reason that I was not aware of HerschePs investigation, 

 and therefore did not mention it in my memoir, will probably be 

 understood when I state that it appeared in an appendix to an 

 astronomical work published in 1847, which not only escaped 

 my attention, but even appears not to have come into the hands 

 of physicists themselves ; at any rate, as far as I can learn, no 

 one who has been engaged on calorimetrical investigations during 

 the quarter of a century which has elapsed since the publication 

 of HerschePs volume has ever mentioned the proposal of the 

 illustrious astronomer, much less either tried or used the method ; 

 otherwise I certainly should not have omitted most distinctly to 

 state that the idea of measuring the amount of melted ice by the 

 diminution of volume which thereby occurs is due to Herschel. 

 On the other hand, I may be allowed to take advantage of my 

 friend Andrews's remark to point out, perhaps more precisely 

 than I have done in my memoir, one at least of the most impor- 

 tant respects in which the instrument proposed by Herschel 

 differs from that which I have described. 



Herschel absorbs the heat which is to be measured by the 

 alteration of volume produced by the ice melting, in a vessel 

 filled with water and floating ice. In my arrangement the 

 same purpose is effected in a vessel made out of a solid mass of 

 artificially congealed ice. The medium employed by Herschel, 

 under the circumstances of the experiment, allows heat to pass 

 through it; that which I use, on the contrary, cannot transmit 

 any of the heat. In regard to the quantity of heat to be mea- 

 sured, my instrument bears to that of Herschel the same relation 

 as a water-tight vessel does to a sieve as regards the measure- 

 ment of a volume of liquid. We should, however, hardly call 

 two methods the same, one of which measured a liquid with a 

 water-tight vessel and the other with a sieve. The loss of heat 

 occurring in a measurement made with a vessel filled with ice 

 and water is great enough to throw doubt upon any exact calo- 

 rimetrical determination, as may be seen by reference to expe- 

 riments detailed in my memoir, and as every one knows who has 

 tried to determine the freezing-point of thermometers in a mix- 

 ture of water and ice instead of in snow. The warmed water 

 which fills the spaces between the pieces of ice in HerschePs ap- 

 paratus, and which, according to its temperature, either rises or 

 sinks, comes in a few seconds, and without having completely 

 equalized its difference of temperature, into contact with the sur- 

 face of the glass vessel surrounding the medium, and gives up 

 this heat, which is thus withdrawn from measurement. 



The extent of uncertainty which is introduced by such a mode 

 of measurement may be estimated when we remember that, ac- 

 cording to experiments given in my paper, more than 30 minutes 



