t the
orbits of the asteroids discovered seemed to answer to these conditions,
and Olbers was even able to use his theory as a means of predicting the
position of yet undetected asteroids. Only Ceres and Pallas had been
discovered when he put forth his theory, but when Juno and Vesta were
found they fell in with his predictions so well that the theory was
generally regarded as being virtually established; while the
fluctuations in the light of Vesta, as we have before remarked, led
Olbers to assert that that body was of a fragmental shape, thus strongly
supporting his explosion hypothesis.

Afterward, when the orbits of many asteroids had been investigated, the
soundness of Olbers's theory began to be questioned. The fact that the
orbits did not all intersect at a common point could easily be disposed
of, as Professor Newcomb has pointed out, by simply placing the date of
the explosion sufficiently far back, say millions of years ago, for the
secular changes produced by the attraction of the larger planets would
effectively mix up the orbits. But when the actual effects of these
secular changes were calculated for particular asteroids the result
seemed to show that "the orbits could never have intersected unless some
of them have in the meantime been altered by the attraction of the
small planets on each other. Such an action is not impossible, but it is
impossible to determine it, owing to the great number of these bodies
and our ignorance of their masses."[7]

[Footnote 7: Popular Astronomy, by Simon Newcomb, p. 335.]

Yet the theory has never been entirely thrown out, and now that the
discovery of the light fluctuations of Eros lends support to Olbers's
assertion of the irregular shape of some of the asteroids, it is very
interesting to recall what so high an authority as Professor Young said
on the subject before the discovery of Eros:

"It is true, as has often been urged, that this theory in its original
form, as presented by Olbers, can not be correct. No _single_ explosion
of a planet could give rise to the present assemblage of orbits, nor is
it possible that even the perturbations of Jupiter could have converted
a set of orbits originally all crossing at one point (the point of
explosion) into the present tangle. The smaller orbits are so small
that, however turned about, they lie wholly inside the larger and can
not be made to intersect them. If, however, we admit a _series_ of
explosions, this difficulty