WITHDRAWN FROM THE ACTION OF GRAVITY. 329 
diately to approach that limit, so that the progress of the transformation after 
the anomalous mode originally impressed becomes more facile. 
§ 24. Thus the theory accounts for all the phenomena resulting from the 
action of vibrations on veins ejected in a descending vertical, for all those at 
least which Savart describes in a precise manner. We pass to veins ejected in 
other directions. And first, since, in these veins, there is equally a transform- 
ation into isolated masses, sounds must necessarily exert on them an influence - 
analogous to that which they exert on veins ejected vertically from above down- 
wards; No. 15 of § 3 has therefore no need of explanation, 
§ 25. But this is not the case with No. 16. If all the divisions, on attaining 
one after the other the extremity of the continuous part, became isolated in 
identically the same manner, and if all the masses parted from thence with the 
velocity precisely corresponding to the movement of translation of the liquid 
at that point, these latter would all describe exactly the same trajectory, and 
then the discontinuous part of the vein could present no dispersion or sheaf- 
like jet ; there are irregularities, then, as Savart remarks, in the emission of 
the isolated masses of the extremity of the continuous part; yet these irregu- 
larities must be very small, as the sheaf has no great extent. I had- thought 
at first that they proceeded from the same causes with those which were con- 
sidered in § 10. But if that were so, the suppression of the extraneous aetion 
would cause the sheaf to disappear and thus reduce the whole vein to a single 
jet; but this is what experiment has not confirmed: by employing, in regard 
to such a vein, the means used by Savart in the case of descending vertical 
veins—that is to say, by receiving the discontinuous part on a thick board, suita- 
bly inclined, and by placing soft bodies under the vessel from which the vein 
issues, under that in which it is received, and under the supports, I have not 
succeeded in producing any considerable diminution of the sheaf. We must 
infer from this that the irregularities are not owing to the vibratory move- 
ments, and that, consequently, they affect the action itself of the configurative 
forces. We perceive, in effect, that, considering the nature of the phenomenon 
of transformation, even slight disturbing causes must have an influence on the 
perfect identity of all the divisions which arise one after the other at the con- 
tracted section; we have seen, for example, in the experiments of §§ 50 to 55 
of the 2d series, an extraneous cause alters the equality of length of the divi- 
sions of a cylinder. This premised, we proceed to show that small differences 
of this nature in the incipient divisions of a vein, ejected under a suitable ob- 
liquity, must necessarily give rise to a certain dispersion of the discontinuous 
art. Us 
r Let us consider particularly two of the constrictions with the dilatation 
which they comprise between them. As we have seen, each of these two con- 
strictions, at first very feebly indicated on quitting the contracted section, after- 
wards deepens gradually in-the transit of the continuous part, by transferring 
half of its liquid to the dilatation; this then receives, by its anterior extremity, 
a portion of the liquid which is driven in a direction contrary to the movement 
of translation, and, by its posterior extremity, a portion which is driven in the 
same direction with that movement, so that its velocity of translation tends to 
be diminished by the first and increased by the second of these accessions. 
Now, although these two opposite actions are in general unequal, because the 
anterior constriction is, at each instant, in a little more advanced phase of 
transformation than the posterior, yet if the two constrictions were perfectly | 
identical at their respective inceptions, and if, in the sequel, they have under- 
. gone identically, though not precisely at the same instant, the same modifica- 
tions until their respective ruptures, it is evident that after these two ruptures, 
that is to say, at the moment when the dilatation exists in the state of an iso- 
lated mass, the sum of the quantities of movement supplied to this mass b 
the anterior constriction will have been absolutely compensated by that of the 
