dreds of species. Sometimes very striking structural characters may be observed; but when assistance is sought from the literature on living forms, the student finds that characters of this class have been ignored, and it is necessary to make a fresh study of the modern genera before proceeding with the fossils. In order to do this, however, large collections are needed, and it is no easy matter to secure sufficient specimens. Thus, in one way and another, the opportunities for error in the study of fossil insects and plants are very many; so many, that it is easy to become discouraged, and yield to the temptation to confine oneself to the comparatively easy problems presented by modern types. In such times of discouragement, however, the student may be cheered by the discovery of some splendid thing, telling a tale beyond dispute; and so he returns to his labors, determined to unravel the secrets of the past, and to accept as philosophically as may be the inevitable results of his inability to avoid a certain percentage of error.
In an effort to reconstruct the landscape of the Miocene period in Colorado, we may well begin with the plants. The number of fossil plants described from Florissant is not nearly so great as that from (Eningen; but the deposits of the latter locality are, apparently, not so nearly contemporaneous; while, on the other hand, not nearly all the Florissant species that have been found have yet been published. The collections of the 1907 expedition, rich in new materials, have only