[ 245 ] 
that particular thick?iefs or depth , ' of the imaginary 
fluid, which is, at any time , necefl'ary to render an 
entire column of it a count erpoife for an entire co- 
lumn of the natural air, is, at that time, the length 
of the fubtangent of the atmofpherical logarithmic . 
In fhort, the fubtangent of the atmofpherical logiflic, 
is the length of a column of fuch a fluid as I have 
fuppofed, which would be fuflained in the Torn- 
cell, an tube, by the preffure of the air, at the level 
of the iea, if we could fuppofe a tube of a fuffi- 
cient length. 
This is demonflrated by Mr. cotes, Harmon. 
Mei h p. i 8. and by no one elfe, that I know of, 
with equal fimplicity. 
It is a manifefc confequence from this, that the fub- 
taogent muft always be as the preffure of the whole 
cylindrical column (upon a given part of the earth’s 
furface) direfliy, and the denfity. of the air, at the 
furface, inveriely ; and it may therefore appear to be: 
repugnant to the theory already efiablifhed, to fup- 
pofe it iubjebt to variation. For if the preffure 
be always as the denfity, upon which, hypothecs 
the whole theory is founded, that which is 
always as the preffure diredlly, and the denfity in- 
verfely, can be no other than a conflant quantity.. 
But M. de luc’s experiments prove, beyond a 
doubt, that the atmofpherical fubtangent is variable? 
therefore the denfity of the atmofphere, at the fur- 
face of the earth, at different times, is not propor- 
tional to the whole preffure, at fuch times,, re- 
fpe&ively. And in this there is nothing inconfiftent 
with the foregoing theory, rightly underflood. When 
the denfity of the air is laid to be as the com- 
preffing force, this is to be underilood of air in the 
2 - fame 
