286 
ME. W. CROOKES ON THE ATOMIC WEIGHT OF THALLIUM. 
grain of air at 740 millims. and 0-1385 grain at 715 millims. ; the platinum weight 
will displace 0-56737 grain of air at 740 millims., and 0-054666 grain at 715 millims. 
The assistant weighing with brass weights would give then 
Under 760 millims. an error of 1-8276 grain") 
„ 740 ,, „ 1*7792 ,, Ion the 1000. 
„ 715 „ „ 1-7505 „ J 
The assistant weighing with platinum weights would give 
Under 760 millims. an error of 1-915329 grain') 
„ 740 „ „ 1-864863 „ on the 1000. 
„ 715 „ „ 1-834334 „ J 
Proceed on another supposition. Let the apparatus be weighed on two days, the 
barometer readings being respectively 740 millims. and 715 millims. Weighed on, say,, 
the first day at 715 millims., on the second day there would be an apparent increase of 
nearly 0-029 grain; and if weighed again on a third day at 760 millims. an increase of 
0-0771 grain, or against platinum weights nearly 0 - 081 grain. 
Had a specifically lighter fluid than water been taken, or exceptional cases exem- 
plified, it would have been possible to have largely increased these numbers ; but they 
are a fair average of the errors unrecognized in ordinary manipulation. 
To select another instance. In the determination of the amount of carbonic acid 
yielded by an organic body under combustion, the weight of the potash and bulbs is 
always sufficiently great to introduce an error arising from neglect of barometric varia- 
tion. Thus, taking the potash and bulbs to weigh 600 grains, there would be displaced 
0-336 grain of air at 760 millims. pressure, 0-327 grain at 740 millims., and 0-316 grain 
at 715 millims. Between 715 and 760 millims. there is an increase of weight of 0 - 02 gr., 
or, supposing 3-5 grs. of the substance under analysis had been taken, an increase of 
nearly 0-6 per cent. Again, a chloride-of-calcium tube, weighing, with its contents, 
350 grs., displaces 0-2135 gr. of air at 760 millims. pressure. Between 715 and 760 mil- 
lims. there will be an increase of 0-0127 gr., or of 0-36 per cent. Thus, in the estimation 
of the carbonic acid and of the water yielded by the organic body, the total error intro- 
duced by barometric variation is nearly 1 per cent. I need not quote illustrations of 
the effect of such an error upon the formula deduced ; it will perhaps account for the 
difference from theory often obtained in the results of carefully conducted analyses. 
The preceding calculations show that a simple formula may be stated which shall 
include the corrections on a certain volume of air for pressure and for temperature. It is 
W = 1 • 2 9 o 6 5 1 grm, V 1+0 -oo367T * 760 ’ 
where W is the weight in grammes of a volume (V) of air at a pressure P and tempe- 
rature T, 0'00367 being the coefficient of the expansion of air. For accurate experi- 
