WITHDRAWN FROM THE ACTION OF GRAVITY. 347 
gradually absorb the excess by means of the small syringe, the operation pro- 
ceeds without difficulty so long as the elements of the meridian line which ter- 
minate at the edges of the disks deviate considerably from parallelism with 
the axis; but when they approximate to this parallelism, or, in other words, 
when we approach the portion of the unduloid which we wish to obtain, it is 
necessary to operate with greater precaution, as the figure might otherwise 
change spontaneously and disunite. By conducting the operation with care, 
and, towards the end, removing"the oil only in very small quantities, we arrive, 
as far as the eye can judge, at the desired portion of the undulaid, (Fig. 7,) a 
portion which varies in form by approaching or withdrawing from the cylinder, 
according as the diameter of the disks is greatgr or smaller relatively to their 
distance; but then the slightest cause, such as a minute movement communi- 
cated to the mass by the point of the syringe, is sufficient to produce the 
gradual alteration and destruction of the figure, which is seen to grow progres- 
sively thinner near one of the disks, the oil being transferred in greater quantity 
to the side of the other disk, and the mass finally separates into two parts. 
From the fact that, in the figure obtained as above, an alteration, occasioned by 
the most trifling cause, proceeds afterwards spontaneously, it would seem that 
the portion of unduloid comprised between the middle of one constricted por- 
tion and that of the next is at the limit of stability. 
We see, from what has just been said, why, in § 4, the adoption of a eylin- 
der as a solid system was recommended. With disks, there is need of the 
greatest circumspection and care to arrive at the point where the last elements 
of the meridian line are or appear parallel to the axis; while with the cylinder 
the figure is perfectly stable, and the required parallelism is established of 
itself. But it remains to be explained how the stability of thé figure can de- 
pend on the two circumferences along which the superficial stratum of the mass 
touches the cylinder, (§ 5.) This is easily done: in the case of the disks, 
when it happens, as has been said, that the figure grows thin spontaneously on 
one side, the elements of the superficial stratum which terminate at the edge 
of the disk near which this effect takes place are inclined towards the axis, 
(Fig. 11;) but, in the case of the cylinder, the last elements of the superficial 
stratum cannot be thus inclined, since they lie on the surface of the solid. 
Fig 11 Fig. 13 
This explanation naturally suggests the idea of substituting for thin disks 
thick ones, or, rather, portions of a cylinder ; for, by giving to the mass, at first, 
a sufficient volume for the oil to reach the edges of the faces of these thick 
disks opposite to those which front one another, and then removing so much of 
the liquid that the circumferences of contact shall fall on the thickness of the 
disks, the cause of stability, above indicated, will evidently exist just as well 
as with a continuous ¢ylinder. Now, this is fully confirmed by experiment; 
the disks which I used had each a diameter of 15 millimetres and a thickness 
of 8, and were fixed at a distance of 90 millimetres apart; the entire system is 
represented at Fig. 12. By causing to adhere to the whole a mass of oil, at 
first too great, then removing the excess, and lightly pressing the mass to right 
or left with the point of the syringe, so that the points from which the meridian 
line appeared to depart were nearly at an equal distance from the two bases of 
each disk, the figure produced evinced a perfect stability ; it is practicable, by 
