t, and the fractured surfaces cemented together again. Parts of
bones, especially the interior, are often rotted into dust while the
harder outer surface is still preserved. The dust must be scraped out,
the interior filled with a plaster cement, and the surface pieces
re-set in position. Very often a steel rod is set into the plaster
filling the interior of a bone, to secure additional strength.

After this preparation is completed, each part being soaked repeatedly
with shellac until it will absorb no more, the bones can be handled
and laid out for study or exhibition. Then, if they are to be mounted
for a fossil skeleton, comes the work of restoring the missing parts.
For this a plaster composition is used.

Where only parts of one side are missing the corresponding parts of
the other side are used for model; where both sides are missing, other
individuals or nearly related species may serve as a guide. But it is
seldom wise to attempt restoration of a skeleton unless at least
two-thirds of it is present; composite skeletons made up of the
remains of several or many individuals, have been attempted, but they
are dangerous experiments in animals so imperfectly known as are most
of the dinosaurs. There is too much risk of including bones that
pertain to other species or genera, and of introducing thereby into
the restoration a more or less erroneous concept of the animal which
it represents. The same criticism applies to an overly large amount of
plaster restoration.

[Illustration: Fig. 42.--Bone-Cabin Draw on Little Medicine River
north of Medicine Bow, Wyoming. The location of the quarry is
indicated by the stack of crated specimens on the left, and close
to it the low sod-covered shack where the collecting party lived.
Beyond the draw lies the flat rolling surface of the Laramie
Plains and on the southern horizon the Medicine Bow Range with Elk
Mountain at the center.]

In some instances the missing parts of a skeleton are not restored,
because, even though but a small part be gone, we have no good
evidence to guide in its reconstruction. This gives an imperfect and
sometimes misleading concept of what the whole skeleton was like, but
it is better than restoring it erroneously. Usually with the more
imperfect skeletons, a skull, a limb or some other characteristic
parts may be placed on exhibition but the remainder of the specimen is
stored in the study collections.

[Illustration: Fig. 43.--A