234 



Mr. J. Joly. 



be made rapid at first when air and mist are being swept out, as 

 the object then is to let in the pure steam as quickly as possible 

 around the substance. 



The construction at the point where the suspending wire passes 

 through to top of the calorimeter, ascending to the balance, must be 

 such that no rubbing of the wire in its vibration up and down occurs. 

 That is, the wire should hang in the centre of the necessarily small 

 orifice provided for it ; as this is a troublesome adjustment, an auto- 

 matic arrangement should be provided. No condensation of steam 

 must occur on the wire where it passes out, or above that point, and 

 steam must be hindered from passing up along^ the wire into the 

 balance. 



To effect the accurate determination of t°, it is important that the 

 temperature in the interior of the calorimeter change slowly. This 

 necessitates that the walls of the calorimeter be fairly non-conducting. 

 They must withal be light, or they will remain hot an inconveniently 

 long while after an experiment,, and will take long to heat, which 

 hinders the rapid filling of the calorimeter with steam. 



The interior of the calorimeter must be easily got at for drying out 

 and cleaning, and for putting the carrier readily in its place. It 

 should fit together fairly steam tight, and be simple in construction, 

 and so cheaply made. 



A sectional elevation to a scale of one-fourth of a convenient form 

 of the calorimeter is given in PI. 6, fig. 1 ; I have worked a good deal 

 with it, and have found it fulfils the requisite conditions. 



I may observe that, as regards the- condition of preserving a steady 

 internal temperature, no form not very cumbersome will confer 

 perfect satisfaction if used in a room in which a rapid variation of 

 temperature is suffered to occur just before an experiment. In every 

 case it will be necessary to carefully screen off the boiler supplying 

 the steam, so that the waste steam and hot gases from the burner pass 

 up a flue or directly out of the room. It would be best of all to locate 

 the boiler in a neighbouring room, taking a steam pipe through the 

 wall. I have myself suffered more from defect in this part of 

 the arrangements than any other, and what discrepancies occur 

 from one experiment to another I attribute to unsteadiness of 

 the initial temperature. I generally find that in the fifteen or 

 twenty minutes during which the boiler is heating, the thermo- 

 meter in the calorimeter may show a variation amounting to 

 one-tenth, often to one-fifth, of a degree. My practice is to take 

 three readings during that interval, one just before lighting the 

 burner beneath the boiler, a second when steam is up, and a third 

 just before making an experiment, after the steam has been let flow 

 freely out of the boiler up a flue for eight or ten minutes to clear air 

 and mist out of the boiler. I assume the mean of these three as the 



