cal shape seems to be circular; for the better the atmosphere, the rounder they look. Under poor seeing they show as irregular patches smooching the disk, much as the canals themselves show as streaks; the spots differing from the canals in being thicker and not so long. As the seeing improves, the patches differentiate themselves into round dots and connecting lines. Such is the shape of the spots associated with single canals; that is, canals not double. In the case of the double canals, the spots look like rectangles with the corners rounded off. One of the most striking of all of them is the Trivium Charontis, which is nearly square.
Now it will be noticed that these shapes are as unnatural as they are definite, and that they all agree in one peculiarity: they are all convex, not concave, to the entering canals. They are not, therefore, mere enlargements of the canals, due to natural causes; for, were the spots enlargements of the canals, at their crossing-points they should be more or less star-shaped, or concave to the canals, whereas they are round, or roundish rectangles,—that is, convex to the same. Such convexity negatives, at the outset, their being purely natural outgrowths of the canals.
The majority of the spots are from 120 to 150 miles in diameter; thus presenting a certain uniformity in size as well as in shape.