So much for the probabilities as to the covering of animals known to us only by their fossil remains; but it is often possible to go beyond this, and to state certainly how they were clad. For while the chances are small that any trace of the covering of an extinct animal, other than bony plates, will be preserved, Nature does now and then seem to have relented, and occasionally some animal settled to rest where it was so quickly and quietly covered with fine mud that the impression of small scales, feathers, or even smooth skin, was preserved; curiously enough, there seems to be scarcely any record of the imprint of hair. Then, too, it is to be remembered that while the chances were very much against such preservation, in the thousands or millions of times creatures died the millionth chance might come uppermost.

Silhouettes of those marine reptiles, the Ichthyosaurs, have been found, probably made by the slow carbonization of animal matter, showing not only the form of the body and tail, but revealing the existence of an unsuspected back fin. And yet these animals were apparently clad in a skin as thin and smooth as that of a whale. Impressions of feathers were known long before the discovery of Archæopteryx; a few have been found in the Green River and Florissant shales of Wyoming, and a Hesperornis in the collection of the State University of Kansas shows traces of the existence of long, soft feathers on the legs and very clear imprints of the scales and reticulated skin that covered the tarsus. From the Chalk of Kansas, too, came the example of Tylosaur, showing that the back of this animal was decorated with the crest shown in Mr. Knight's restoration, one not unlike that of the modern iguana. From the Laramie sandstone of Montana Mr. Hatcher and Mr. Butler have obtained the impressions of portions of the skin of the great Dinosaur, Thespesius, which show that the covering of this animal consisted largely, if not entirely, of small, irregularly hexagonal horny scutes, slightly thickened in the centre. The quarries of lithographic stone at Solenhofen have yielded a few specimens of flying reptiles, pterodactyls, which not only verify the correctness of the inference that these creatures possessed membranous wings, like the bats, but show the exact shape, and it was sometimes very curious, of this membrane. And each and all of these wonderfully preserved specimens serve both to check and guide the restorer in his task of clothing the animal as it was in life.

And all this help is needed, for it is an easy matter to make a wide-sweeping deduction, apparently resting on a good basis of fact, and yet erroneous. Remains of the Mammoth and Woolly Rhinoceros, found in Siberia and Northern Europe, were thought to indicate that at the period when these animals lived the climate was mild, a very natural inference, since the elephants and rhinoceroses we now know are all inhabitants of tropical climes. But the discovery of more or less complete specimens makes it evident that the climate was not particularly mild; the animals were simply adapted to it; instead of being naked like their modern relatives, they were dressed for the climate in a woolly covering. We think of the tiger as prowling through the jungles of India, but he ranges so far north that in some localities this beast preys upon reindeer, which are among the most northern of large mammals, and there the tiger is clad in fairly thick fur.

When we come to coloring a reconstructed animal we have absolutely no guide, unless we assume that the larger a creature the more soberly will it be colored. The great land animals of to-day, the elephant and rhinoceros, to say nothing of the aquatic hippopotamus, are very dully colored, and while this sombre coloration is to-day a protection, rendering these animals less easily seen by man than they otherwise would be, yet at the time this color was developing man was not nor were there enemies sufficiently formidable to menace the race of elephantine creatures.

For where mere size furnishes sufficient protection one would hardly expect to find protective coloration as well, unless indeed a creature preyed upon others, when it might be advantageous to enable a predatory animal to steal upon its prey.

Color often exists (or is supposed to) as a sexual characteristic, to render the male of a species attractive to, or readily recognizable by, the female, but in the case of large animals mere size is quite enough to render them conspicuous, and possibly this may be one of the factors in the dull coloration of large animals.

So while a green and yellow Triceratops would undoubtedly have been a conspicuous feature in the Cretaceous landscape, from what we know of existing animals it seems best to curb our fancy and, so far as large Dinosaurs are concerned, employ the colors of a Rembrandt rather than those of a sign painter.

Aids, or at least hints, to the coloration of extinct animals are to be found in the coloration of the young of various living species, for as the changes undergone by the embryo are in a measure an epitome of the changes undergone by a species during its evolution, so the brief color phases or markings of the young are considered to represent the ordinary coloring of distant ancestors. Young thrushes are spotted, young ostriches and grebes are irregularly striped, young lions are spotted, and in restoring the early horse, or Hyracothere, Professor Osborn had the animal represented as faintly striped, for the reason that zebras, the wild horses of to-day, are striped, and because the ass, which is a primitive type of horse, is striped over the shoulders, these being hints that the earlier horse-like forms were also striped.

Thus just as the skeleton of a Dinosaur may be a composite structure, made up of the bones of a dozen individuals, and these in turn mosaics of many fragments, so may the semblance of the living animal be based on a fact, pieced out with a probability and completed by a bit of theory.