Encyclopaedia Britannica, or, a Dictionary of arts, sciences, and miscellaneous literature : enlarged and improved. Illustrated with nearly six hundred engravings > Volume 8, ELE-FOR

E L E [
Ek drome- the diiTipation of fluid from a large fphere may in fa cl
t tC1' be greater tiian that from a fmall one in the lame given
^ time.
We have remarked above, that thefe experiments
were made in a particular date of the air; and the law
of diiTipation aicertaihed by them is of oourfe adapted
only to that given date. In a different date of the air,
even if this diould be impregnated with the fame pro¬
portion of moidure, the law of diidpation may be dif¬
ferent. The inference which M. Coulomb expected to
draw from his experiments was, that the ratio of diiTi¬
pation would prove to be Ids than the cube of lire
quantity of water held in folulion, except when that
quantity of water was what the air was capable of
holding in folution at the given temperature.
This is agreeable to obfervation 5 for we know that
air which is confldered as dry, that is, when it is not
nearly faturated with moidure, is the mod favourable to
eleflrieal phenomena.
Such is the general refult of Coulomb’s experiments
on the diiTipation of eleftricity into the air.
A he method in which M. Coulomb examined the
diflipation along imperfect conductors, by means of this
indrument, was, by completely inddating the ball /,
and then after obferving the lofs fudained by a body in
contact with it from the air, hiding a metallic rod
down the infulating dalk, till the diflipation began to
exceed what took place only by the air.
From his experiments refpefting the diflipation along
imperfect conductors, he found that this took place in a
different manner from that in which eledricily efcaped
by communication with the contiguous air. The elec¬
tricity feems to be diffufed chiefly along the furface of
the infulator, and appears principally to be produced
by the moiflurc that is more or lefs attached to it; M.
Coulomb illuftrates this in the following manner.
Water is found to adhere to the furface of all bodies
from which it is prevented by adheflon from efcaping
when the bodies are eleftrified, and is thus rendered
capable of receiving a greater degree of ele&ric power.
Let us fuppofe that the particles of moifture are difpofed
uniformly over the furface, with intervals between them;
the electricity that is communicated to one particle,
muft: acquire a certain degree of denfity, before it can
fly from this particle to the next, acrofs the intervening
infulating fpace. When an imperfeCt conductor of this
kind is eleCtrifled at one extremity, the communicated
eleCtricity, in pafling to the other extremity, mult be
weakened every ftep in pafling from particle to particle.
Suppofe we have three adjacent particles, which we
may call a, b, and c ; we infer from N° 374. of the ar¬
ticle Elixtricty, that the motion of b is fenflbly effect¬
ed, only by the difference of a and c; and therefore
the paffage of eleCtric fluid from b to c, requires that
this difference be fuperior, or at leaft equal to the force
neceffary for clearing this coercive interval. Let a par¬
ticle pafs over. The denfity of fluid of the particle b is
diminifhed, while the denfity of the particle on the
other fide of a remains as before. Therefore feme
fluid will pafs from a to />, and from the particle pre¬
ceding a to n ; and fo on, till we come to the electri¬
fied end of this infulator. It is plain, from this con-
fideration, that we muff at lafl: arrive at a particle be¬
yond e, where the Avhole repulfion of the precedinp-
Vol. VIII. Part I. . 6
9 1 E I. E
particle is juft Tifhcient to clear the coercive interval. Elsft
borne fluid will come over 5 and the repulfion of this,
acting now in the oppolite direction, will prevent any
fluid from coming to fupply its place in the particle,
which it has juft quitted •, the transference of fluid will
therefore flop here, and beyond this point the iniukition
will be complete. Hence we perceive that there is a
mathematical relation between the infulating power, and
the length of the canal 5 and this may be aicertained by
tiie theory which we adopted in the article Electri¬
city. We ihall here give an inltance of this invefli-
gation 5 and, for the fake of fimplicity, we Ihall take a
very probable cafe, viz. where the infulating interval,
or, as we may more properly call it, the coercive inter¬
val, is equal in every part of the canal.
Let R reprefent the coercive power of the infulator,
or the degree of force required to clear the coercive
interval between two particles. Suppofe a ball C, fig.
16. fufpended by a filken thread AB ; and let us de-
note the quantity of redundant fluid in the ball by C,
and let the denfities at the different points of the canal
be denoted by AH, P &c. ordinates to lome curve
1) d B, cutting the axis in B, the point where the thread
A B begins to infulate completely. Let Pp be an element
of the axis; draw the ordinate p f a tangent to the
curve dfY, the normal r/E, and draw f e perpendicu¬
lar to Pi/. Suppofe AC—r, AP—v, and P</—y.
Then we fiiall have Pprr.v, and dc~ —y It was
fliewn in N°(374. of the article Electricity, that the
only fen Able action of the fluid on a particle at P is
x'
when the action of the redundant fluid in the globe on
the particle at P, having the denfity y, is denoted by
7 | TI* T herefore — is —R, the coercive power of
(y-j-xj x r
the thread, which is fuppofed to be conftant, ———5
Pp
is therefore equal to feme conftant line R. But Pp
(or f e) : d e ■= Y d : PE. The fubnormal PE, is
therefore a conftant line. But as this is the property
of a parabola, the curve of denfity H d B muft be a
parabola, of which 2PE = 2R, is the parameter.
Cor. 1.—The denfities at different points of an im-
perfeeft infulator are in the fubduplicate ratio of their
diftances from the point of complete infulation: for
P d* : ADHrBP : BA.
Cor. 2.—The lengths of canal requifite for infulating
different denfities of the electric fluid are in the dupli-
yV J)* , *
cate ratio of their denfities: for AB=: —t--. ■■■, and PE
2PE’
Cor. 3.—The length of canal requifite for infulation
is inverfely as its coercive power, and may be repre-
- ., H* T, ^ DA* T>1
fentedby —. For AB
2PE'
B*
2R*
If we refleft on this theory, we fliall perceive, that
our formulae determine the diftribution of fluid along
the furface of an imperfect conductor, only in a cei>
tain manner, fuppofmg that the ball C has received a
certain determinate portion of fluid, for this portion dif-
fufing itfelf, particle by particle, through the conduc¬
ing matter, will extend to b in fuch a manner, as that
21 the