the Thermal Unit. 445 



Rowland. — 1 have elsewhere expressed my admiration of 

 Rowland's work, and were it not for some doubts regarding the 

 thermometry I should accept his results as conclusive. The in- 

 vestigations by Professor Schuster above referred to " open out 

 the possibility that Rowland's value might have to be reduced 

 somewhat when referred to the Paris air-thermometer/' 



Professor Schuster's tables, however, indicate that any such 

 possible change would be very small, and the further com- 

 parison (already mentioned) by Messrs. Bartoli and Stracciati 

 between Rowland's thermometer and a Tonnelot thermometer 

 from the Bureau International shows the agreement between 

 the two standards to be very close. The results of these 

 investigations modify my previously expressed apprehensions 

 regarding Rowland's thermometry, although I join with 

 Professor Schuster in his desire that more light should be 

 thrown on the matter by a direct comparison of a Rowland 

 thermometer and the International Standard. Rowland's 

 value at 15° is 4*1895 x 10 7 , and at 10°, 4'2 x 10 7 . 



Miculescu. — The following is a quotation from Professor 

 Schuster and Mr. Grannon's paper : — " We now turn to an 

 investigation of Miculescu (Annales de Chimie et de Physique, 

 vol. xxvii. 1892), in which the mechanical equivalent of heat 

 is measured directly by what seems a very excellently devised 

 series of experiments. Its result is 4*1857 x 10 7 . He does 

 not state the exact temperature to which this applies, but all 

 his experiments seem always to have been made between 10^ 

 and 13°, so that we may assume 11°'5 to be the mean tempe- 

 rature of his experiments. Rowland's value at that tempe- 

 rature is 4*1999 x 10 7 *. We must draw attention to one 

 point in Miculescu's work which requires clearing up before 

 we can give to it any decisive value. Everything in the 

 experiments depends on the measurement of a couple, the arm 

 of the couple being the distance between two knife-edges ; one 

 of them had to support a weight of more than 43 kilograms. 

 The distance between these knife-edges is said to have been 

 28 cms. in all experiments. Very insufficient information is 

 given, however, as to how that distance was measured, and it 

 would almost seem as if the author had trusted to the maker 

 in adjusting the central knife-edge to the zero-point of that 

 scale. If the apparatus is still in existence, it might be well 

 to make sure that no error has been introduced through a 

 wrong estimate of the distance of the lever-arm. ' 



* I do not agree with Professor Schuster's interpretation of Rowland's 

 value at this temperature. The curve on Plate VII. is constructed 

 from the values given by Rowland in his final table (p. 196), and 4-197 

 would appear to be the more accurate value. 



Phil. Mag. S. 5. Vol. 40. No, 246. Nov. 1895. 2 I 



