a Standard of Weight and Mcafure. 15 » 



The diverfity in the rcfult of thefe two experiments is deferving of notice, and muft be 

 explained. It may proceed from two caufes, which we will now inquire into. But firft 

 it may be obferved, that the accuracy in meafuring the dimenfions of thefe two bodies, as 

 well as the precifion in weighing them, has, I chink, been fuch as to put out of all doubt this 

 part of the experiment. From whence then does this difference arife ? Either of two caufes 

 may be fufpeited ; viz.. the preflure of the water againft the fides of thefe two bodies altering 

 their volumes, which, it may be prefumed, would have a greater efFedl on the cube, from its 

 figure, than on the cylinder, and in a direftion agreeable to this difference ; that is, it would 

 diminifh the capacity of the cube more than that of the cylinder, and thus make the apparent 

 weight of a cubic inch lefs in the experiment of the cube. But alfb we fee, that the cylinder 

 was weighed at a greater depth, by 1,2 inch, than the cube, below the fiirface of the water. 

 Now, if it be true that water is comprefEble *, it will become denfer, from its weight, at 

 different depths, and this circumflance would ad in the fame way with that juft mentioned ; 

 vix. would make the apparent weight of a cubic inch lefs from the experiment of the cube 

 than the cylinder, which we fee is the fa6l. 



(§. 25.) In order to diffipate thefe doubts, I caufed a very accurate hclPow brafs fphere to 

 be made, of about fix inches diameter, and of fuch thicknefs of metal, viz. 0,13 inch,, as to 

 be very little heavier than water, and yet of fuch flrength as, together with, its form,, to refift 

 any probable change of bulk by the preffure of water. 



This fphere, which has already been mentioned (§. 10.), was examined in the following 



manner. The fix-inch moveable bar r (Plate V. fig. 3.) of the guage, was compared with 



the divided fcale of iiKhes, fig. i. The microfcopes being adjufted to exa£tly fix inches, or 



the interval between 26 inches and 32 inches, and the bar placed under them, the excefs. 



above 6 inches was found to be as follows, by the micrometer, n o.. 



I ft trial. 2d trial, after r£-ailjuftment» 



inches. inches. 



6 + ,oo5S 1 6+,3o55 



,0056 \ 



,0054 -^ '- ,0054. 



'°°53 I thermom. ^°^l 

 '°°56 ^6^0,0 ,005s 



,0057 J ,0052 ^ 



therm. 

 '64",0 



Mean of the 1 ft trial = 6 ,00550 6 ,00536, 



Mean of the 2d trial = 6 >00536 



Mean of both, or 7 ;c 7 • ..t. ^ c ^ o 



length of bar 3 = ° P0543 ^ in the temperature of 64°. 



(§. 26.) The bar was then placed in the rectangular gauge k I m n, fig. 2. in the direflioa 

 ft ; and the end of the micrometer ferew brought to bear againft it repeatedly, fo as to touch 

 without force, or confiderable preffure : and the divifions f cut by die index, on the micron- 

 meter plate of the gauge, were as follow : * 



* See Mr. Canton's experinient in the Phil. Tranf. vol. LII. 



t Each thread of this fcrew is = ^J,^ inch, and each revolution of the fcrew is divided into 100 ; fo that 

 eifery divifioa on the micrometer plate is = ^^\^ inch. 



X 2 Trial 



