Its behavior subsequent to recognition bore out the inference from its aspect when it first appeared. While for many days prior to its coming unmistakably into view it was impossible to say whether what was seen of the southern cap in 1903 and 1905 was cloud or snow; so even after it had definitely disclosed itself it continued to play at odds with the observer. Showing sharp at the edges one day it would appear but hazily defined the next, thus clearly demonstrating itself to be at the then unstable acme of its spread. Such a state of things we are only too familiar with in our own March weather when after days of sunshine that have melted off the winter’s white and fringed it with rivulets and awakening grass, a snow-storm falling upon it powders the ground again that was beginning to be bare and at one stroke extends the domain of the snow while mystifying the actual limits it may be said to occupy. The same condition of things, then, is not unknown on Mars, and to fix the precise date of so wavering a phenomenon is not so much matter of difficult observation as of physical impossibility. [Illustration: Hellas in winter.] Nor is the southern cap, at this the height of its winter expansion, confined strictly to its own proper limits. Faint extensions, now so connected with its main body as to form part and parcel of it, now so detached and dull of tint as to make the observer doubtful of the exact relationship, are generally to be seen attendant on it. Hellas in winter is much given to such questionable garb, and has in consequence been mistaken by more than one observer for the cap itself, appearing as it does well upon the southern limb and being often the only region to show white. Indeed, frost-bound as it then is, to consider it the polar cap, though possibly geographically incorrect, may climatologically be sustainable. Its northern extremity extends down to latitude 30°, a pretty low latitude for frost. Still such equatorward extension is not without corroborating parallel. In 1903, at what was in Martian dates April 26, the whole of the region south of the Solis Lacus and the Nectar showed white, with a whiteness which may as well have been hoarfrost as cloud. Now, the Nectar runs east and west in latitude 28°. So that in this instance, too, it is possible that arctic conditions knocked at the very doors of the tropics. Encroachment of the sort is equivalent to snow in Cairo and permanent snow at that; not an occasional snow flurry, but something to linger on the ground and stay visible sixty millions of miles away. [Illustration: White south of Nectar and Solis Lacus.]