OP ARTS AND SCIENCES. 269 



Kirchlioff and Hansemann (Wied. Annalen, Vol. IX. p. 6) con- 

 tained only 0.129 per cent of carbon, and was therefore not cast- 

 iron. I have as yet found no description of the iron used by Ang- 

 strom, but it is probable that he too used a tolerably pure specimen. 

 Reviewing the course of my experiments, I find no reason for sup- 

 posing my values to be, as a whole, much too low. The very great 

 interval of temperature, about 50°, between thermometers 8 and 7, 

 next the heated end of the bars, was no doubt too great for the 

 greatest precision of results, but a number of calculations of the 

 conductivity for points between thermometers 7 and 6 give a mean 

 very near those found for points between 8 and 7. The natural 

 conclusion is, that cast-iron has a much smaller conductivity than 

 wrought-iron. 



But whatever may be the value of my results as to iron alone, 

 there can be little doubt that the main object for which the experi- 

 ments were made, namely, to determine approximately the relative 

 conductivity of cast iron and cast nickel, has been attained. The 

 thermal conductivity of the nickel here used is very nearly the same 

 as that of the cast-iron bars with which it was compared. 



In spite of the very considerable amount of laborious study that 

 has been devoted to the thermal conductivity of iron, authorities 

 differ widely in their conclusions, especially upon this particular, 

 the change of conductivity with change of temperature, as the 

 numbers above given show. This difference comes in part from 

 different opinions as to the change of specific heat of iron with 

 change of temperature. The work of a number of investigators 

 within the past ten years has added greatly to our knowledge upon 

 this latter subject, and possibly recalculations of thermal conductiv- 

 ity from old observations might now be made with advantage. 



I hope soon to make trial of a method for estimating thermal 

 conductivity somewhat different from any heretofore used. 



January 7, 1893. 

 I have recently, with the help of students, made additional ob- 

 servations upon steady flow and the rate of cooling in the Southern 

 cast-iron bar, placed now in the main lecture-room of the Physical 

 Laboratory. The new observations on rate of cooling extended 

 about 46° only, — from +110° to +64°; but for this range they 

 indicated a rate about 10% greater than was found in the base- 

 ment in 1890. A rough calculation from this new study of the 

 bar indicates for end no. 1 a conductivity practically the same as 



