the irregular nature of the surface of the Moon, the regions where the shadow was cast in the different observations being themselves at different altitudes. Even the maria and those crater floors which appear to us fiat are in reality decidedly convex as compared with the general spherical surface of the Moon. The results so far obtained, therefore, are not only uncertain, but from the fact that the altitude of the lower station is unknown are rather unsatisfactory. It is just as if we should state the height of our terrestrial mountains above the valleys near which they are situated, instead of giving their height above the sea-level.
While there is no sea-level on the Moon, yet we can adopt a spherical surface as a reference plane whose distance from the centre of the Moon shall be equal to the mean semidiameter of the Moon in a plane perpendicular to the line of sight. By measuring the displacement of a crater under different librations of the Moon, it is easy to compute its real height above such a reference plane. Such computations have been made by Franz ("Königsberg Beobachtungen," XXXVIII., Part 5) and by the author ("Annals of Harvard College Observatory," LI.), but they are rather too technical to be dealt with in this place.
These computations also seem to show that the Moon is slightly egg-shaped, the longest axis pointing toward the Earth. This, as we have seen in Chapter II., had been already predicted from theory. The amount of the projection toward the Earth is very slight—probably not more than a mile. For a more complete account of this subject, and also for a discussion of the effect of elevation on the computed longitudes and latitudes of the various lunar craters, the reader is referred to the voltune of the Harvard Annals above mentioned.
In closing this chapter it may be well to point out some of the respects in which this atlas differs from the lunar photographic atlases that have preceded it.
(a) It is the only complete photographic atlas of the Moon in existence. Not only so, but it covers the whole visible surface of the Moon five times.
(b) The plates are all arranged systematically, and in a form easily remembered, even without reference to the key given on page 91.
(c) The plates are of such a shape that all the objects shown are similarly illuminated. Had fewer, broader plates been used, many objects could not have been shown near the terminator. The similar illumination also permits the most suitable plate and exposure to be used for the whole of the particular region under consideration.
