On the Steam Calorimeter. 



233 



of the copper box referred to in (1), that is, a cylinder 3*65 cm. in 

 diameter X 4'50 cm. in length. This will be about 405 grams of 

 copper, giving through a range of 90° a precipitation of 6*343 grams 

 of steam. The increment in 10 minutes, less that of the carrier, is 

 0'5 X 2 — 0*34 x 2 = 0"32 milligram. The error introduced by neglecting 

 this is 1 part in 20,000 about. 



These figures afford an idea of the extent to which this radiation 

 effect, if neglected entirely, affects the results obtained in these calori- 

 meters. And, in practice, even these are excessive ; they double the 

 error actually obtaining, for it will be found that half the interval 

 assumed for observation will be more than sufficient in order to be 

 quite sure of the weighing, and of the condition of the substance 

 when of such dimensions as I have assumed. 



What radiation error then can be detected with this method of 

 condensation is not alone in general extraordinarily small, but its 

 amount is easily ascertained, and a close approximation to its entire 

 elimination possible. To the question whether undetected error from 

 radiation or other causes enters into the quantity of steam precipitated 

 upon the substance in the first instance an answer, based on direct 

 experiment, cannot be given. Comparative experiments on sub- 

 stances, using other methods, would not afford a conclusive answer, as 

 the exact extent of the error entering into other methods is at least 

 equally open to surmise. It is to be said, however, that, a priori, no 

 grave error is to be expected, and what experimental tests have been 

 applied appear to show that such error if existent must be very small. 

 I have mentioned the general results of these tests. 



The Construction of the Steam Calorimeter. 



What is the best form to confer upon a steam calorimeter? There 

 are so many conditions to fulfil that the choice is really not very large, 

 It must permit of being filled rapidly with steam, which should pre- 

 ferably descend in the calorimeter, as it then mixes less with the air. 

 Arrangements must be made for allowing a slow current or circula- 

 tion of steam to contiuue all the time the weighing is being effected. 

 This must be such as will not interfere with the accurate deter- 

 mination of the weight. If this circulation of steam be stopped at 

 any time there is a minute but definite fall in temperature. A 

 sensitive thermometer will show this. It is about the one-thirtieth o,f 

 a degree. If the current be stopped when the body is accurately 

 poised, and after an interval be started again, a minute increase of 

 weight is at once apparent. The substance has cooled in the interval 

 and is reheated on re-establishing the current. 



Although this current of steam may be quite slow while weighing 

 is proceeding, there must be complete control over it, so that it may 



