342 On Delicate Calorimetric Thermometers, 



portion of this temporary alteration is due, not to the bulb, 

 but to the mercury not entirely filling the bore of tube of its 

 own accord, owing to the extreme flatness of this latter at 

 the edges. The tube of which 63616 was made presented no 

 such imperfections. However, neither the permanent nor 

 temporary difference can be in any degree attributed to the 

 action of the tube in this latter instrument, but is wholly due 

 to the bulb itself; for the thermometer was first made up 

 with a small bulb *, containing only 3 grams of mercury, in 

 which state it showed no difference whatever in the readings 

 obtained under different conditions, Avith or without tapping. 

 The comparative smallness of this temporary alteration in 

 the case of 63616 renders this instrument far more conve- 

 nient and satisfactory than the other. Instead of requiring 

 two or three minutes continuous tapping in order to obtain 

 a correct reading, less than one minute is sufficient, and this 

 must very materially increase accuracy in calorimetric deter- 

 minations. I have not as yet used this thermometer in a 

 sufficient number of experiments to ascertain the experi- 

 mental error with it ; but the uniformity of its action and 

 the increased concordance of the rate of cooling registered in 

 successive intervals, leads me to feel confident that the error 

 with it will be considerably less than with 62839. 



From what has already been said as to the smallness of the 

 reading-error in this instrument, it may be stated that we 

 have pushed thermometric delicacy quite as far as is de- 

 sirable with regard to calorimetric work, and that to produce 

 a further degree of accuracy in this branch of study, the 

 methods employed, and not the thermometers, must be im- 

 proved. 



In conclusion, I wish to take this opportunity of expressing 

 my warmest thanks to Mr. Casella and to his assistants for 

 the scientific interest which they have shown in this investi- 

 gation, and the ungrudging manner in which they have 

 spared no pains to help me in attaining my object. 



* Where the bulb is to contain as much as 20 or 30 grams of mer- 

 cury, it is well to have the instrument made up temporarily with a small 

 bulb, for the purpose of calibration, as with fine tubes some rough hand- 

 ling is often necessary to pass a short thread of mercury along them. 



It would appear that 30 grams of mercury is not an uncommon amount 

 in the instruments used by Berthelot (Mec. Chim. i. 165), although they 

 are far less delicate than those described in the present communication. 

 He even refers to thermometers containing 250 grams ! (p. 156). 



