to ascertain a Standard of Weight and Measure. 157 
weighed at a greater depth, by 1,2 inch, than the cube, below 
the surface of the water. Now, if it be true that water is com- 
pressible,* it will become denser, from its weight, at different 
depths, and this circumstance would act in the same way with 
that just mentioned; viz. would make the apparent weight of a 
cubic inch less from the experiment of the cube than the cylin- 
der, which we see is the fact. 
(§. 25.) In order to dissipate these doubts, I caused a very accu- 
rate hollow brass sphere to be made, of about six inches diameter, 
and of such thickness of metal, viz. 0,13 inch, as to be very little 
heavier than water, and yet of such strength as, together with 
its form, to resist any probable change of bulk by the pressure 
of water. 
This sphere, which has already been mentioned, (§. 10.) 
was examined in the following manner. The six-inch move- 
able bar r, (Tab. V. fig. 3.) of the gauge, was compared with 
the divided scale of inches, fig. 1 . The microscopes being ad- 
justed to exactly six inches, or the interval between 2 6 inches 
and 32 inches, and the bar placed under them, the excess above 
6 inches was found to be as follows, by the micrometer, n 0. 
1st trial, 
inches. 
6+> 00 55 
>°°53 
,003b 
> oc 57 
> 
thermom. 
64°, o 
2d. trial, after re-adjustment, 
inches. 
6+ >0055 " 
,0032 
,0033 > 
,0034 
,0032 
therm.. 
64°, o 
Mean of the 1st trial = 6 ,00330 6 ,0033b 
Mean of the 2d trial = 6 ,00336 
Mean of both, or"! n 1 f ^ o 
length of bar |= 6 ’ 00 543 j m the temperature of 64 . 
* See Mr. Canton’s Experiment in the Phil. Trans. Vol. LII. 
