We have however one quite natural fact on our earth which serves to illustrate one of these two features, the direction of the downward fissure. This is, the comparatively common phenomenon of basaltic columns and 'Giant's Causeways.' The wonderful regularity of these, and especially the not unfrequent upright pillars in serried ranks, as in the palisades of the Hudson river, must have always impressed observers with their appearance of artificiality. Yet they are undoubtedly the result of the very slow cooling and contraction of melted rocks under compression by strata _below and above them_, so that, when once solidified, the mass was held in position and the tension produced by contraction could only be relieved by numerous very small cracks at short distances from each other in every direction, resulting in five, six, or seven-sided polygons, with sides only a few inches long. This contraction began of course at the coolest surface, generally the upper one; and observation of these columns in various positions has established the rule that their direction lengthways _is always at right angles to the cooling surface_, and thus, whenever this surface was horizontal, the columns became almost exactly vertical. _How this applies to Mars._ One of the features of the surface of Mars that Mr. Lowell describes with much confidence is, that it is wonderfully uniform and level, which of course it would be if it had once been in a liquid or plastic state, and not much disturbed since by volcanic or other internal movements. The result would be that cracks formed by contraction of the hardened outer crust would be vertical; and, in a generally uniform material at a very uniform temperature, these cracks would continue almost indefinitely in straight lines. The hardened and contracting surface being free to move laterally on account of there being a more heated and plastic layer below it, the cracks once initiated above would continually widen at the surface as they penetrated deeper and deeper into the slightly heated substratum. Now, as basalt begins to soften at about 1400° F. and the surface of Mars has cooled to at least the freezing-point--perhaps very much below it--the contraction would be so great that if the fissures produced were 500 miles apart they might be three miles wide at the surface, and, if only 100 miles apart, then about two-thirds of a mile wide.[15] But as the production of the fissures might have occupied perhaps millions of years, a considerable amount of atmospheric denudation would result, however slowly it acted. Expansion and contraction would wear away the edges and sides of the fissures, fill up many of them with the debris, and widen them at the surfaces to perhaps double their original size.[16] [Footnote 15: The coefficient of contraction of basalt is 0.000006 for 1° F., which would lead to the results given here.]