MECH 
giafs with fmooth and clean furfaces, having their Tides 
EF joined together with wax, and their Tides AB, CD, 
kept a little apart by another piece of wax W, To that 
their interior Turfaces, whofe common interleCtion is the 
line E F, may form a Tmall angle. When this apparatus 
is immerfed in a veffel MN full of water, the fluid will rife 
in fuch a manner between the glafs planes as to form the 
curve D qom E, which reprefents the turfaceof the elevated 
water. By meafuring the ordinates m n, op, See. of this 
curve, and alfo its abfciffae F«, F p, Sec. Mr. Haukfbee 
found it to be the common Apollonian hyperbola, having 
for its aflymptotes the furface DF of the fluid, and EF the 
common interfeftion of the two planes. To the very fame 
conclufion we are led hy the principles already laid down ; 
for, as the diftance between the plates diminifhesat every 
point of the curve DqomF, from D towards E, the water 
ought to rife higher at o than at q, ftill higher at m, and 
liigheft of all at E, where the diftance between the plates 
is a minimum. To illuftrate this more clearly, let ABEF 
and CDEF, fig. 29, be the fame plates of glafs, (inclined 
at a greater angle for the fake of diftinCtnefs;) and let 
E inq D and Eos B be the curves which bound the Turfaceof 
the elevated fluid. Then, fince the altitudes of the water 
in capillary tubes are inverfely as their diameters or the 
difiances of their oppofite fides, the altitudes of the water 
between two glafs plates fhould at any given point be in¬ 
verfely as the diftances of the plates at that point. Now, 
the diftance of the plates at the point m is obvioufly mo, 
or its equal np, and the diftance at q is qs or r 1; and, 
fince mn is the altitude of the water at m, and qr its alti¬ 
tude at q,we have mn : qr=np : rt ; but Fn : Fr=znp : rt\ 
therefore, mn : qr=Fn 1 Fr; that is, the altitudes of the 
fluid at the points m , q, which are equal to the abfeiflae 
F n, Fr, (fig. 28.) are proportional to the ordinates qr, mn, 
equal to F«, Fr. But in the Apollonian hyperbola the 
ordinates are inverfely proportional to their refpedtive ab- 
feiflse, therefore the curve FiqomF is the common hyperbola. 
As the plates are infinitely near each other at the apex E, 
the water will evidently rife to that point, whatever be the 
height of the plates. 
The phenomena which we have been endeavouring to 
explain are all referable to one fimple faff—that the par¬ 
ticles of glafs have a ftronger attraction for the particles 
of water than the particles of water have for-each other. 
This is the cafe with almoft all other fluids except mer¬ 
cury, the particles of which have a ftronger attraction for 
each other than for glafs. When capillary tubes therefore 
are plunged in this fluid, a new feries of phenomena pre- 
fent themfelves to our confideration. Let MN (fig. 30.) 
be a veffel full of mercury. Plunge into the fluid the ca¬ 
pillary tube CD, and the mercury, inltead of rifing in the 
tube, will remain ltationary at E, its depreflioti below the 
level furface AB being inverfely proportional to the dia¬ 
meter of the bore. This was formerly aferibed to a repul- 
five force fuppofed to exift between mercury and glafs ; 
but we fliall prefently fee that it is owing to a very dif¬ 
ferent c-aufe. ' 
That the particles of mercury have a very 'ftrong at¬ 
traction for each other, appears from the globular form 
which a fmall portion of that fluid alfumes, and from the 
refiftance which it oppofes to any feparation of its parts. 
If a quantity of mercury is feparated into a number of 
minute parts, all thefe parts will be fpherical; and, if two 
of thefe ipheres be brought into contaft, they will inftantly 
ruth together, and form a Tingle drop of the fame form. 
There is alfo a very fmall degree of attraction exifting be¬ 
tween glafs and mercury ; for a globule of the latter very 
readily adheres to the lower furface of a plate of glafs. 
Now, fuppofeadrop of water laid upon a furface anointed 
with greafe, to prevent the attraction of cohefion from re-' 
ducing it to a film of fluid : this drop, if very fmall, will 
be fpherical. If its lize be confiderable, the gravity of its 
parts will make it Ipheroidal; and, as the drop increafes 
in magnitude, it will become more and more flattened at 
its poles, like AB in fig. 31. The drop however, will ftill 
Vol. XIV. No. 2008. 
A N I C S. 721 
retain its convexity at the circumference, however oblate 
be the fpheroid into which it is moulded by the force of 
gravity. Let two pieces of glafs, oAm, pB n, be now 
brought in contafl with the circumference of .the drop : 
the mutual attraflion between the particles of water, 
which enabled it to preferve the convexity of its circum¬ 
ference, will yield to their fuperior attraction for glafs; 
the fpaces m, n, 0, p, will be immediately filled ; and the 
water will rife on the fides of the glafs, and the drop will 
have the appearance of AB in fig. 32. If the drop AB, 
fig. 31, be now fuppofed mercury inltead of water, it will 
alfo, by the gravity of its parts, alfume the form of an ob¬ 
late fpheroid ; but, when the pieces of glafs 0 A m, p B n, 
are brought clofe to its periphery, their attraftive force 
upon the mercurial particles is not fufficient to counteraCl 
the mutual attraction of thefe particles ; the mercury there¬ 
fore retains its con vexity at the circumference, and alfumes 
the form of AB in fig. 32. the fmall fpaces 0, p, being 
filled by the prefl'ure of the fuperincumbent fluid, while 
the fpaces m, n, ftill remain between the glafs and the 
mercury. Now, if the two plates of glafs A, B, be made' 
to approach each other, the depreflions m, n, will ftill con¬ 
tinue; and, when the diftance of the plates is fo fmall 
that thefe depreflions or indentations meet, the mercury 
will fink between the plates, and its defeent will continue 
as the pieces of glafs approach. Hence the depreftion of 
the mercury in capillary tubes becomes very intelligible. 
If two glafs planes forming a fmall angle, as in fig. 28. be 
immerfed in a veffel of mercury, the fluid will fink below 
the furface of the mercury in the veffel, and form an Apol¬ 
lonian hyperbola like DoE, having for its aflymptotes the 
common interfe&ion of the planes and the furface of mer¬ 
cury in the veffel. 
The depreftion of mercury in capillary tubes is evidently 
owing to the greater attraction that fubfifts between the 
particles of mercury than between the particles of mercury 
and thofe of glafs. The difference between thefe two at¬ 
tractions, however, arifes from an imperfeCt contact be¬ 
tween the mercury and the capillary tube, occafioned by 
the interpofition of a thin coating of water which gene¬ 
rally lines the interior furface of the tube, and weakens 
the mutual aCtion of the glafs and mercury; for this ac¬ 
tion always increafes as. the thicknefs of the interpofed 
film is diminifhed by boiling. In the experiments which 
were made by Laplace and Lavoilier on barometers, by 
boiling the mercury in them for a long time, the con¬ 
vexity of the interior furface of the mercury was often 
made to difappear. They even fucceeded in rendering it 
concave, but could always relfore the convexity by intro¬ 
ducing a drop of water into the tube. When the ebulli¬ 
tion of the mercury is fufficiently ftrong to expel all fo¬ 
reign particles, it often rifes to the level of the furround¬ 
ing fluid, and the deprefiion is even converted into an 
elevation. 
Newton, Clairaut, and other geometers, have main¬ 
tained, that the action of the capillary tube is fenfible at 
a fmall diftance, and that it is extended to the particles of 
fluid in the axis of the tube. Laplace and other philofo- 
phers, who have lately attended to this fubjeCt, fuppofe 
capillary attraction to be like the refractive force, and all 
the chemical affinities, which are not fenlible except at 
imperceptible diftances ; and it muff be allowed that this 
opinion is confident with many of the phenomena. It has 
been often obferved that water rifes to the fame height in 
glafs tubes of the fame bore, whether they be very thin or 
very thick. The zones of the glafs tube therefore, which 
are at a fmall diftance from the interior furface, do not 
contribute to the afeent of the water, though in each of 
thefe zones, taken feparately, the water would rife above 
its level. When the interior furface of a capillary tube is 
lined with a very thin coating of an unttuous fubftance, 
the water will no longer afeend. Now, if the attraction 
of the glafs tube were funilar to the attraction of gravity, 
of eleCtricity, or of magnetifm, it ought to aCt through 
bodies of ^11 kinds; and, notwithftanding the thin coat- 
2 X in** 
