ustful. Failing to see the
regular, uninterrupted road, he slants off to the right or left, he
escapes at a tangent. If I try to interfere and to bring him back to
the path of my choosing, he persists in his refusal, shrivels up, does
not budge, and soon the whole procession is in confusion. We will not
insist: the method is a poor one, very wasteful of effort for at best a
problematical success.

We ought to interfere as little as possible and obtain a natural closed
circuit. Can it be done? Yes. It lies in our power, without the least
meddling, to see a procession march along a perfect circular track. I
owe this result, which is eminently deserving of our attention, to pure
chance.

On the shelf with the layer of sand in which the nests are planted
stand some big palm-vases measuring nearly a yard and a half in
circumference at the top. The caterpillars often scale the sides and
climb up to the moulding which forms a cornice around the opening. This
place suits them for their processions, perhaps because of the absolute
firmness of the surface, where there is no fear of landslides, as on
the loose, sandy soil below; and also, perhaps, because of the
horizontal position, which is favourable to repose after the fatigue of
the ascent. It provides me with a circular track all ready-made. I have
nothing to do but wait for an occasion propitious to my plans. This
occasion is not long in coming.

On the 30th of January, 1896, a little before twelve o'clock in the
day, I discover a numerous troop making their way up and gradually
reaching the popular cornice. Slowly, in single file, the caterpillars
climb the great vase, mount the ledge and advance in regular
procession, while others are constantly arriving and continuing the
series. I wait for the string to close up, that is to say, for the
leader, who keeps following the circular moulding, to return to the
point from which he started. My object is achieved in a quarter of an
hour. The closed circuit is realized magnificently, in something very
nearly approaching a circle.

The next thing is to get rid of the rest of the ascending column, which
would disturb the fine order of the procession by an excess of
newcomers; it is also important that we should do away with all the
silken paths, both new and old, that can put the cornice into
communication with the ground. With a thick hair-pencil I sweep away
the surplus climbers; with a big brush, one that leaves no smell behi