522 PROCEEDINGS OF THE AMERICAN ACADEMY. 



the jacket were brightly silvered to a distance of about five centimeters 

 above the vapor exit. The vaporizer was set so low that the level (F) 

 of the calorimeter water came up on its jacket to a height of a few 

 millimeters. This brought the boiling liquid in the vaporizer to within 

 one centimeter of the calorimeter water, largely protected the vapor 

 from condensation in this zone, and prevented most of the evaporation 

 from the surface of the calorimeter water at the line of immersion. 

 The vaporizer projected a short distance through the adiabatic cover, 

 and, since the projecting part was silvered, the amount of radiation 

 was not excessively large. The space around the vaporizer was closely 

 packed with cotton. With this apparatus, values were obtained which 

 were slightly higher than those previously obtained with the earlier 

 forms of apparatus, when made at the same rate ; but they w^ere not 

 as consistent as desired, partly because of the fact that the diameter 

 of the exit of the delivery tube was much too large. 



Another attempt to reach a better result led to the cutting off of 

 still more of the delivery tube below, the thermal protection being 

 provided only by a rubber cup, cut from a child's ball, placed 

 between the vaporizer and condenser. The rubber cup was packed 

 with cotton, as was also the space around the vaporizer. After a few 

 determinations, however, the apparatus was discarded, because the 

 correction for radiation and conduction was very large (0.040° per 

 minute), and also because of the difficulty in making a tight joint. 



The form of apparatus finally adopted combined all the advantages 

 of the preceding forms, and is shown in Figure 4. The boiling liquid 

 was surrounded by a vacuum jacket whose walls were about one centi- 

 meter firom the walls of the boiling compartment, save at the bottom 

 where the space was but five millimeters across. In addition to the 

 hood covering the upper end of the vapor delivery tube, the tube was 

 also provided with another trap to catch and retain any liquid that 

 might in any way gain access thereto. This trap was placed as low as 

 possible so that the distance between it and the condenser might be 

 reduced to a minimum, yet it was surrounded by the boiling liquid in 

 order to prevent condensation within it. The heating coil was placed 

 so low as to make sure that the liquid surrounding the trap was at the 

 boiling point. If prematurely condensed liquid now gained access to 

 the condenser, it must come entirely from condensation in the short dis- 

 tance between the trap and the calorimeter water, — because all previ- 

 ously condensed liquid was caught in the trap. After an experiment 

 the trap usually contained a few drops of liquid, sufficient to have in- 

 troduced appreciable error, had it been allowed to reach the condenser. 



The interior of the vacuum jacket was brightly silvered to a height 



