104 proceedings: philosophical society 



eter and the resulting change in temperature are measured. A cor- 

 rection for the heat withdrawn in the expelled liquid is determined 

 by special experiments. By the use of data for the heat of pressure 

 variation of the liquid obtained from separate measurements, the 

 corrections for pressure variation are applied, the result being a second 

 determination of the specific heat of the saturated liquid. 



The greatest difference between the mean results of both methods 

 and the results of either method as represented by empirical equations 

 is less than 1 part in 1000. 



In the measurements of latent heat the jacket temperature was kept 

 constant while a measured amount of ammonia was evaporated, 

 slightly superheated, and withdrawn from the calorimeter. The 

 approximate amount of heat required to effect this change was added 

 electrically, the small balance being due to thermal leakage and change 

 in temperature of the system, both of which were kept small and were 

 measured. Analysis of the process occurring in the calorimeter during 

 an experiment leads to a method of calculation whereby data from 

 other sources than the direct observations enter only in the computation 

 of correction terms, which can be made small by careful manipulation. 



The result of each of the determinations agrees with the mean result 

 as expressed by means of an empirical equation within 1 part in 1000. 



Discussion: The paper was discussed by Messrs. White, Burgess, 

 SosMAN, and Bichowski. 



The 798th meeting was held at the Cosmos Club, January 5, 1918; 

 Vice-President Humphreys in the chair; 54 persons present. The 

 minutes of the 797th meeting were read in abstract and approved. 



Mr. 0. S. Adams presented a paper on Lambert's conformal conic 

 projection. The paper was illustrated by lantern slides. 



Since the spheroidal surface of the earth is nondevelopable, it is 

 impossible to make a perfect map of a section of any extent upon a 

 plane surface. The best that can be done then is to choose an approx- 

 imation that may preserve the features desired in the proposed map'. 

 A projection is called conformal or orthomorphic when any infinitesimal 

 element of the map is exactly similar to the element that it represents. 

 The Lambert conformal conic projection is admirably suited to the 

 mapping of any region that has no great extent in latitude. The 

 parallels become concentric circles and the meridians become radii of 

 this system of circles. The projection is for this reason very easily 

 constructed and thus fulfils one of the practical requirements for a 

 projection. With this method of projection, a map could be con- 

 structed of the United States that would not be in error of scale in 

 any part by more than 1.2 per cent. The fact that angles are pre- 

 served and that the error of scale is within the limits of scaling makes 

 its use in France of great service at the present time. The projection 

 is of interest historically because it was originated by Johann Heinrich 

 Lambert in 1772, and afterwards fully discussed by Gauss. Since 

 the projection is conformal, it is especially interesting to mathemati- 



