244 PROCEEDINGS OF THE AMERICAN ACADEMY 



Instrumental Errors. 



The barometer was one which had been compared directly with 

 another which had been studied somewhat by a cathetometer, and by 

 comparison with a Signal Service standard in Boston. Its constant 

 error must have been reduced to a fraction of a millimeter, which, so 

 far as it is constant, would not sensibly affect the temperature measure- 

 ments. Scales were by Brown and Sharpe, of Providence, R. I., and 

 had no errors sensible in this work as compared with a standard scale 

 bv Prof. William A. Rogers. Thermometers used were corrected. 

 Those for the more exact work were by Baudin of Paris, and were 

 read to fiftieths of a degree. Pernet's method of thermometry was 

 employed. Calibration and steam exposure corrections were applied. 

 A more extended discussion than has been published, so far as we 

 are aware, as to the precision necessary in the component measure- 

 ments entering into the air thermometry, was made with a view to best 

 proportioning of parts and their elimination of determinate constant 

 errors. 



Preparation of Substances. 



This was done by Mr. Gleason, under the direction of Prof. L. M. 

 Norton and Mr. C. W. Andrews of the Chemical Department of the 

 Institute. 



Naphthaline. — The pure product from Kahlbaum, of Berlin, was 

 used without subsequent treatment. Samples were fractionated, and 

 all distilled within 0°.3. These distillates were kept separate, each 

 being divided into as many parts as there were tenths rise in tem- 

 perature, and the melting points of all were found the same. The 

 range of melting and solidifying points of the naphthaline, as taken 

 from the original package, was 79°. 38 to 79°. 68 ; after use through the 

 entire series of observations in the boiling-point apparatus it was 

 79°.42 to 79°.84. 



Benzophenone. — The method of Friedel, Crafts, and Ador * was at 

 first selected, on account of its apparent simplicity and the ease of pro- 

 duction of considerable quantities of the substance in the pure state. 

 For reasons not known, the rate of production was too small ; and the 

 process was abandoned in favor of that of Chancel,| namely, the dry 

 distillation of benzoate of calcium. The benzoate was prepared by 

 neutralizing an aqueous solution of benzoic acid with milk of lime. 



* Comptes Rendus, xxxv. 673. 



t Liebig's Ann. d. Chemie u. Pharm., lxxii. 279; lxxx. 285. 



