ouble canals, or they may have been started from two or more
craterlets not far apart, the direction being at first decided by some
local peculiarity of structure; and where begun continuing in straight
lines owing to homogeneity or uniform density of material. This is very
vague, but the phenomena are so remarkable, and so very imperfectly
known at present, that nothing but suggestion can be attempted.

_Concluding Remarks on the 'Canals.'_

In this somewhat detailed exposition of a possible, and, I hope, a
probable explanation of the surface-features of Mars, I have
endeavoured to be guided by known facts or accepted theories both
astronomical and geological. I think I may claim to have shown that
there are some analogous features of terrestrial rock-structure to
serve as guides towards a natural and intelligible explanation of the
strange geometric markings discovered during the last thirty years, and
which have raised this planet from comparative obscurity into a position
of the very first rank both in astronomical and popular interest.

This wide-spread interest is very largely due to Mr. Lowell's devotion
to its study, both in seeking out so admirable a position as regards
altitude and climate, and in establishing there a first-class
observatory; and also in bringing his discoveries before the public in
connection with a theory so startling as to compel attention. I venture
to think that his merit as one of our first astronomical observers will
in no way be diminished by the rejection of his theory, and the
substitution of one more in accordance with the actually observed facts.

APPENDIX.

_A Suggested Experiment to Illustrate the 'Canals' of Mars._

If my explanation of the 'canals' should be substantially correct--that
is, if they were produced by the contraction of a heated outward crust
upon a cold, and therefore non-contracting interior, the result of such
a condition might be shown experimentally.

Several baked clay balls might be formed to serve as cores, say of 8 to
10 inches in diameter. These being fixed within moulds of say half an
inch to an inch greater diameter, the outer layer would be formed by
pouring in some suitable heated liquid material, and releasing it from
the mould as soon as consolidation occurs, so that it may cool rapidly
from the _outside._ Some kinds of impure glass, or the brittle metals
bismuth or antimony or alloys of these might be used, in order to see
what form the