In other ways as well the southern cap is the more self-contained. The rifts, indeed, break it up into separate portions and these in part remain as outlying detachments of the main body, as was notably the case in 1877 and in 1894, but they hardly have the permanency and importance of those similarly formed about the arctic pole. Nothing antarctic for instance compares with the subsidiary patch of the north polar regions lying in longitude 206°, which both in Schiaparelli’s time, and during the late oppositions as well was almost as fixed a feature of the arctic zone as the cap proper. Not quite so constant, however, and not so solid-looking a landmark is this patch for all its extent, which nearly equals the area of the more legitimate portion. It bears on its face a more pallid complexion as if it were thinner, and this is borne out by the fact that it occasionally disappears, an event which so far at least has never befallen the northern cap itself. Less constant the southern one is to its own minimum than the northern. In some seasons, in most in fact, it reaches like the other a more or less definite limit of diminution which it does not pass. But this is not always the case. In 1894 it disappeared entirely at the height of its midsummer. The season was probably unusually hot then in the southern hemisphere of Mars. In position the caps have something to say about physiographic conditions. Both caps at their minima are then irregular and the centre of the south one is markedly eccentric to the areographic pole. It lies some six degrees north along the thirtieth meridian. The northern one is also probably eccentric, but much less so, with a divergence not much exceeding a degree and of doubtful orientation. Not only are both caps not upon their respective poles but they are not opposite each other, the one lying in longitude 30°, the other in 290°. This speaks, of course, for local action. In some wise this must depend on the configuration of the surface, yet so far as markings go there is nothing to show what the dependence is.