RICHARDS AND ROWE. — SPECIFIC HEATS OF LIQUIDS. 177 



is produced upon the temperature of the calorimeter by air thus 

 introduced. Accordingly all the tubes were loosely packed with 

 cotton wool which served to prcAcnt undue circulation.^ A further 

 safeguard against error from this cause was found in increasing the 

 depth of submergence of the "submarine." 



Two similar accurately standardized thermometers were used to 

 register the temperature in the calorimeter on the one hand and in 

 the outside bath on the other. For the sake of accuracy these ther- 

 mometers should be alike in scale and in thermometric lag. Those 

 employed were especially made, and were very well adapted for the 

 purpose. They covered a range from 15.4 to 20.8 degrees, and each 

 degree was 74 millimeters long. A small auxiliary thermometer (L) 

 suspended wholly in the air beside the stems of the larger instruments 

 gave the mean temperature of the exposed portions of these stems. 

 The thermometers were standardized with the greatest care by refer- 

 ence to standard instruments which had been carefully studied at the 

 Bureau Internationale des Poids et Mesures. Their standardization 

 will be described in another place. As the final results usually varied 

 no more from the mean value than 0.05 precent, it is clear that the 

 thermometers were usually to be depended upon to within 0.001°. 



In the outer pail of course the concentration of the alkali might vary 

 within wide limits without harm, and the sulphuric acid introduced 

 into it was always the most concentrated grade of pure acid of com- 

 merce,* which was delivered through a burette graduated in tenths 

 of a degree Centigrade, as already explained in previous papers. 



Within the platinum bottle, howe\'er, the acid needed very precise 

 determination in order that the heat yielded to the calorimetric system 

 by its neutralization by a slight excess of alkali might be definitely 

 apportioned in each experiment. 



The steps taken to prepare these solutions of definite strength are 

 described below. Other experimental details will be discussed later 

 when the individual experiments are described. 



With this apparatus, then, were made many series of experiments in 

 which first pure water M-as contained in the calorimeter-can, and after- 

 wards the various dilute solutions. In each case a definite amount 



3 The sketch is not strictly drawn to scale and the size of the rods is exag- 

 gerated for the sake of clearness. 



4 Crude acid would have answered, but the pure acid was found to be more 

 constant in composition than cheaper cruder grades. As the burettes were 

 graduated to deliver a constant volume of acid for a definite temperature rise, 

 this point was of importance. 



