Rightly viewed, no more subtle tribute could be paid to the remarkable character of the phenomenon of gemination than the scepticism with which it was immediately received and which it still continues to elicit. That the sight should be regarded as illusory speaks for its surpassing strangeness; and so far as oddity goes the encomium is certainly deserved. Of the bizarre features of this curiously marked disk, the double canals were at the time of their discovery the culmination, and though things stranger still if possible have since been seen there, it is not wonderful that doubt should still incredulously stare. If the mere account of them reads like romance, to see them is an experience. Nothing astronomical that I have ever seen has been so startlingly impressive as my first view of a double canal. Even in narration the thing justifies its effect. For a double canal consists of a pair of twin dark filaments, perfectly parallel throughout their course and inclosing between them ground of the same ochreish cast as that which lies without. Only on occasion is this tint of their midway departed from, and then only toward a darkening, never toward a lightening of it. Except for appearing paired, the lines resemble precisely the usual single canals. In length they vary from a few hundred to a few thousand miles, while in width each component, for narrowness, hardly permits of definite ascription. Compared for strength with the usual canal the lines of a double seem to hold on the average an intermediate position between the larger and the smaller of the single canals so far detected. Owing, however, to the massed effect of the pair by reason of their closeness, they have an advantage in showing over the singles of two to one. And this renders them among the most conspicuous and important meshes of the canal network. Like the single canals, they vary in strength with the Martian time of year; at certain seasons developing into heavy pencil lines and at others fading away to the merest gossamers, only just discernible like cobwebs stretched across the face of the planet. Although the individual constituent lines vary in aspect and never rise at their most to cognizable breadth, the distance parting their centres, or the width of the double, is quite measurable. The only difficulty in the way of its determination lies in the absence of a procurable unit small enough to mete it. The usual spider-threads of the micrometer are colossal in comparison with these filaments and present a standard only analogic at best. Nevertheless, by means of the finest threads that could be got, estimates of the distance between the pairs were made at Flagstaff in 1905, and the results agree as closely as the means permit with those got by measurement of the doubles as depicted in the drawings. [Illustration: Martian doubles.]