60 On the Heat and Cold produced by the 



caufmg the bulb to yield a little to the pre flu re of the air. 

 It has however been found upon trial that the fame effects 

 take place whether the thermometer is fealed or not. My 

 experiments accord with this, having made a thermometer 

 and left it unfealed for the fcjfprefs purpofe : in all the expe- 

 riments with condenfed and ravened air, there was no fenfible 

 difference obferved to arife from the inequality of p refill re on 

 the external and internal furfaces of the bulbs, the fealed and 

 open thermometers varying the fame in kind and alfo in de- 

 gree, except from circumftances to be noticed hereafter. 



It being certain then that a real change of temperature 

 takes place, it remained to determine the quantity and man- 

 ner of that change. Having chofen a fmall and consequently 

 fenfible thermometer, with a fcale of degrees fufficicntly large 

 to admit of diftinguifhing one-tenth of a degree, I proceeded 

 to afcertain feveral fa&s experimentally. 



Exp. I. Took a receiver, the capacity of which was about 

 120 cubic inches, and fufpended the thermometer with its 

 clear bulb in the central part of it; then letting the whole 

 require the temperature of the room, which was without a 

 fire, I cxhaufted the air and afterwards reftored it, marking 

 the effects upon the thermometer. The medium of feveral 

 trials nearly agreeing with each other was as under : 

 The thermometer in the air of the room flood at - 36"°. 8 



, f U nk upon exhaufticn to - 34 .7 



, ■ rofe when the air was reftored to 38 .9 



The fuddenntfs of the fall and rife puzzled me moft : after 

 reflecting upon it for fome time, I conjectured that the real 

 change of temperature of the air o- medium was much 

 greater than the thermometer indicated, but that the in- 

 equality exifted only for a few feconds of time, becaufe the re- 

 ceiver, See. immediately impart heat to, orabftract it from, fo 

 fmall a quantity of air as 120 cubic inches, which are only 

 equal to 40 grains in weight. The phenomena of the ther- 

 mometer feemed very well to accord with the fuppofition of 

 great heat or cold acting upon it for a few feconds only. 



Exp. II. Pnrfuing this idea, I imagined that if two ther- 

 mometers whofe bulbs were very unequal in magnitude were 

 inclofed together, the fmaller bulb ought to give the greater 

 variation: accordingly I inclofed two, the diameters of their 

 bulbs being .35 and .65 of an inch refpe&ively ; and having 

 exhaufted the air and reftored it again repeatedly in fuccef- 

 fion, and found a mean of the variations, that of the fmall 

 bulb was 2 ? .8, and that of the large, 2 .2. 



Exp. III. Repeated the exhauftion with the fmall ther- 

 mometer inclofed in three different circumftances fuccef- 



fively ; 



