12 0. Barns — Change of Heat Conductivity on 



chronometer beating half seconds was at hand for time measure- 

 ment, and temperatures (since relative values enter the formulae) 

 were directly given by the deflections of the dead beat galva- 

 nometer, reduced to arcs. For charges l cm thick (//), observa- 

 tions were taken every 20 sec , for charges -15 em thick every 30 sec , 

 and for charges *19 cm thick every 60 sec . Usually 24 readings 

 were made for each heating, which were then combined in 

 four batches of 6 observations each. By joining the 1st and 

 4th, 2d and 5th, etc., I therefore obtained 12 data for k, with 

 the mean temperature corresponding to each. 



Readings of the zero of the galvanometer, and of the tem- 

 perature of the environment were taken before and after the 

 time measurements; temperatures of the hot bath at the begin- 

 ning, temperatures of the cold bath at the end of the work. 



13. Exhibit. — To avoid prolixity I will only give an example 

 of the way in which the work was done. In addition to the 

 corresponding increments of time (of) and logarithmic tempera- 

 ture [d log u), table 3 contains the (actual) temperatures, m, to 

 u t , of the upper plate at the beginning and end of each batch 

 of six consecutive readings. Thus each value of k is a mean 

 of three pairs of sufficient data. Furthermore v is (henceforth*) 

 the actual temperature of the environment, U the initial tem- 

 perature of the upper and lower plate, d the temperature of 

 the cold bath; and finally is the mean temperature of the 

 charge, the bottom being constantly cold (d°) and the top hot 

 (u°). The table also contains the mean value of the corrective 



m = l/(l- Apch.F^l - ^j/A^cJcFAq^ 



of equation (7) §10. 



The two cases in which the environment has the final (cold) 

 or the initial (hot) temperature of the plates are distinguished 

 as method I and II, respectively. In case of solid thymol both 

 were tested and led to identical results. §17. 



To obviate discrepancies due to imperfect adhesion, expan- 

 sion, etc., of solid thymol, three different values of J are intro- 

 duced as a check on the method. 



14. Constants. — Supplementing §§ 8-10, 1 need only add that 

 an allowance (/') of 5 per cent of F.w&s made for the contrac- 

 tion of thymol on solidifying. 



F= 182-3 cm 2 A x — l-300 cm 



F x — 244-5 " p x — 8-94 " 



F—f— 173-4 " c x = -933 g. cat. 



* To avoid cumbersome notation. 



