ands of miles long, and nearly all forming
arcs of great circles.

"They cannot be cracks, for they are of uniform width throughout their
length, and always run direct from one definite point to another, no
matter how distant apart they may be.

"Cracks, such as we see on the moon, though sometimes straight, are
usually wide near the centre of disturbance which caused them, and
narrow off to a fine point, and often end anywhere out in open space;
moreover, they are usually very irregular in width, and take a zig-zag
course instead of a straight one. This, as I have said, is not the case
with a single canal line on Mars. If they were cracks, some at least
would be irregular and end in open space. The same remarks apply in the
case of foldings or ridges.

"The oases, once declared to be illusions, were then said to be large
openings in the soil at the junctions of the cracks; or they might be
craters, and so on. But this does not account for the appearance of the
oases between twin canals, or the systematic manner in which the canals
effect a junction with the oases. Again, therefore, the theory fails to
fit the known circumstances of the case.

"Dr. A.R. Wallace rather favours the idea of natural cracks or faults in
the surface of the planet; and suggests that the outer crust of Mars may
be a crystalline or similar formation which would lend itself to the
production of numerous cracks in the surface. He points to a few cracks
and faults in the earth's surface, all of small size, as confirming this
idea; but the cases he adduces only seem to prove that there is on our
earth absolutely no natural formation which can in any way properly be
compared with the lines seen on Mars. Moreover, there seems to me no
ground whatever, beyond the needs of the theory, for supposing that the
crust of Mars is of a crystalline nature, or such as would predispose to
the formation of cracks. On the contrary, all the evidence is against
it--the existence of vegetation in some parts, the general appearance of
the red portion, and the large clouds of sand which have been observed,
all being indicative of a sandy formation, in the red portion at least.

"The theory also fails to take into consideration the most important
point of all, viz. that every canal runs direct from one definite point
to another, perhaps over two thousand miles distant. In very many cases
numerous lines connect with one small area, or even with one point. The
Lucus A