196 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Of course, in each case where acid B was used, the higher value 

 must be used in the computation, whereas in each case where acid 

 A was used, the lower value is the proper one. 



In the following table each series of experiments used the average 

 quantity 114.863 grams of. sulphuric acid per experiment — at least 

 the deviation from this average amount was smaller than the probable 

 experimental error, so that the amount may be taken as a constant. 

 In none of the individual experiments except in the first series with 

 water (Table II), did the amount of acid exceed 114.87 grams or 

 sink below 114.85 grams. The averages for the several series were 

 much closer to 114.863 than this. The heat capacity of the solid 

 apparatus also was constant, being always 12.60 grams except in the 

 first four experiments with water already tabulated; and a constant 

 amount of caustic alkali was used as always. The other quantities 

 found in the experiments are given in the following table, the figures 

 there being (as has been said) total averages similar to those given 

 at the foot of the table just above. Because of the great similarity of 

 the experiments, it does not seem worth while to give each one in 

 detail. Anyone can calculate for himself the average results for the 

 specific heats from the data given below. The only essential figures 

 not immediately indicated by the table are the deviations among the 

 individual experiments, but these are sufficiently shown by the "pro- 

 bable errors" given in the last column, calculated according to the 

 theory of least squares. From this it will be seen that two of the 

 series were somewhat less concordant than that given above, but 

 most of them were fully as good as the sample series, or indeed even 

 better. In another respect, also, perhaps, this table of averages 

 does not quite do justice to the work: — it does not give a vivid repre- 

 sentation of the very large expenditure of labor, extending over many 

 months, which was needed to olitain the numerous results so concisely 

 tabulated. 



The main ol)ject in determining these specific heats was to pro^'ide 

 a basis for computing the heats of neutralization of the various 

 substances in question, but the values are ne\'ertheless not without 

 interest in themselves. 



The significance of results of this sort are of course most easily per- 

 ceived when the results are computed in terms of gram-molecules, and 

 after this has been done, two sets of comparisons are possible: — in 

 the first place, the effect of dilution upon the molecular heat of a 

 given substance, and in the next place the effect of varying the ions 

 concerned. 



