and half as wide. It was not as heavy
as we had thought, which made it all the better for our purpose.

The remainder of our task, though tedious, was not unpleasant.

We first made the larger bones, which were to serve as the beams of our
raft, exactly the same length by filing off the ends of the longer ones
with rough bits of granite. I have said it was tedious. Then we filed
off each of the smaller bones projecting from the neural arch until
they were of equal length.

They extended on either side about ten inches, which, allowing four
inches for the width of the larger bone and one inch for the covering,
would make our raft slightly over a foot in depth.

To make the cylindrical column rigid, we bound each of the vertebrae to
the one in direct juxtaposition on either side firmly with strips of
hide, several hundred feet of which we had prepared.

This gave us four beams held straight and true, without any play in
either direction, with only a slight flexibility resulting from the
cartilages within the center cord.

With these four beams we formed a square, placing them on their edges,
end to end. At each corner of the square we lashed the ends together
firmly with strips of hide. It was both firm and flexible after we had
lashed the corners over and over with the strips, that there might be
no play under the strain of the current.

Over this framework we stretched the large piece of hide so that the
ends met on top, near the middle. The bottom was thus absolutely
watertight. We folded the corners in and caught them up with strips
over the top. Then, with longer strips, we fastened up the sides,
passing the strips back and forth across the top, from side to side,
having first similarly secured the two ends. As a final precaution, we
passed broader strips around both top and bottom, lashing them together
in the center of the top. And there was our raft, twelve feet square,
over a foot deep, water-tight as a town drunkard, and weighing not more
than a hundred pounds. It has taken me two minutes to tell it; it took
us two weeks to do it.

But we discovered immediately that the four beams on the sides and ends
were not enough, for Desiree's weight alone caused the skin to sag
clear through in the center, though we had stretched it as tightly as
possible. We were forced to unlash all the strips running from side to
side and insert supports, made of smaller bones, across the middle each
way. These we