The calculations ending with these figures were made while the Sun's distance was still estimated at 95 millions of miles. Of course the reduction afterwards established in the estimated distance, entailing, as it did, changes in the factors which entered into the calculations, affected the results; and, though it was unlikely that the relations stated would be materially changed, it was needful to have the calculations made afresh. Mr. Lynn has been good enough to undertake this task, and the figures given in the text are his. In the case of Mars a large error in my calculation had arisen from accepting Arago's statement of his density (0·95), which proves to be something like double what it should be. Here a curious incident may be named. When, in 1877, it was discovered that Mars has two satellites, though, according to my hypothesis, it seemed that he should have none, my faith in it received a shock; and since that time I have occasionally considered whether the fact is in any way reconcilable with the hypothesis. But now the proof afforded by Mr. Lynn that my calculation contained a wrong factor, disposes of the difficulty--nay, changes the objection to a verification. It turns out that, according to the hypothesis, Mars _ought_ to have satellites; and, further, that he ought to have a number intermediate between 1 and 4.] [Footnote 19: Since this paragraph was first published, the discovery that Mars has two satellites revolving round him in periods shorter than that of his rotation, has shown that the implication on which Laplace here insists is general only, and not absolute. Were it a necessary assumption that all parts of a concentrating nebulous spheroid revolve with the same angular velocities, the exception would appear an inexplicable one; but if, as suggested in a preceding section, it is inferable from the process of formation of a nebulous spheroid, that its outer strata will move round the general axis with higher angular velocities than the inner ones, there follows a possible interpretation. Though, during the earlier stages of concentration, while the nebulous matter, and especially its peripheral portions, are very rare, the effects of fluid-friction will be too small to change greatly such differences of angular velocities as exist; yet, when concentration has reached its last stages, and the matter is passing from the gaseous into the liquid and solid states, and when also the convection-currents have become common to the whole mass (which they probably at first are not), the angular velocity of the peripheral portion will gradually be assimilated to that of the interior; and it becomes comprehensible that in the case of Mars the peripheral portion, more and more dragged back by the internal mass, lost part of its velocity during the interval between the formation of the innermost satellite and the arrival at the final form.]