268 Mr. Kir WAN's Remarks on fpecific Gravities 



point to oiij cubic foot and 66,46 cubic inches. And if the 

 dilatations are proportional to the degrees of heat throughout 

 the fcale, there mufl be an augmentation of 3,136 cubic inches 

 per cubic foot, produced by every 10 degrees of heat. Both 

 thefe points remain, therefore, to be determined ; firft, at what 

 temperature a cubic foot of water weighs exactly 1000 ounces 

 avoirdupois ; and,2dly, whether the dilatations produced by fuc- 

 ceffive degrees of heat are proportional to the degrees that pro- 

 duce them. This laft point has indeed been handled by others, 

 but with different views ; and their determinations are not eafily 

 applicable to the prefent queftion. 



To examine this matter experimentally, I ordered a hollow 

 tinned iron cone to be made, of four inches diameter in the 

 bafe, one-tenth of an inch diameter in the lummit infide, and 

 10 inches perpendicular height, Whofe folid contents fliould be 

 42,961 cubic inches, but by a flight diminution of the dia- 

 meter, and a protuberance ariiing from the foldering, I found 

 it to contain, in the temperature of 62°, but 42,731 cubic 

 inches, according to the eftimation of 1000 ounces to the 

 cubic foot ; and having filled it by immerfion in boihng water, 

 and taken it up at different degrees of heat, and weighed it 

 when cold, I found its contents as exprefled in the following 

 table ; the firft column of which fliews the degrees of heat at 

 which it was taken up ; the fecond, the weight of the water 

 contained in it ; the third, the diminution of weight occafioned 

 by thofe degrees of heat j the fourth, the fum of the diminu- 

 tions of weight in the cubic foot, by the preceding degrees of 

 heat ; the fifth fhews the weight of a cubic inch of water in 

 each of thofe degrees of heat; and the fixth, the augmentation 

 of bulk in the cubic foot by every 20° of heat. The horizontal 

 lines, marked thus ^j I have added from the experiments of Mr, 



z • BladHj 



