present the least variable features in those factors which determine relations to mechanical stress and physiological needs.
External organs are notoriously subject to variation, even under slight alterations of surrounding conditions, within the limits of the species or even within various stages of development of the same individual. From this it is clear that organs such as leaves must be very unreliable for phylogenetic purposes. It is, unfortunately, true that much of the paleobotanical work based upon a study of such parts must be of inferior value, and the conclusions drawn will require extensive revision when the more rigid tests to be applied through a knowledge of the stem structure are brought to bear.
The value of paleobotanical evidence consists in its ultimate correlation with known types of plants, and it is obvious that all such studies should be prosecuted with direct reference to the broader requirements of plant biology. This involves a comprehensive knowledge of the history of plant life from its earliest development; that the data derived from a study of living species should be correlated with the evidence obtained from fossilized remains. Existing vegetation shows a very incomplete record of plant life as a whole. Its history as known until very recent times, and even now to a very large extent, is displayed only through the medium of detached groups, and relates chiefly to the most highly organized types. Through the perspective afforded by paleobotany, it becomes possible to not only supply missing facts, but to establish what theory has for so long a time required a satisfactory demonstration of—a more or less continuous series of phenomena from the rudimentary forms to the most advanced organisms.
Until a very recent date the Linnæan division of plant life into two great phyla, the cryptogams and the phanerogams, was the prevailing conception of the constitution of the plant kingdom. This division recognized no connection between the two great groups, but regarded them as wholly distinct in origin as in character. But the rapid advances in a knowledge of plant anatomy, developed toward the middle of the last century, and especially the remarkable and epoch-making observations of Hofmeister respecting the process of reproduction, enabled him to break down the old barriers erected by the doctrine of the constancy of.species, and prove a genetic connection between the primary divisions of Linnæus. "With this starting-point, the cryptogams and the phanerogams were subjected to a severe scrutiny from an entirely new point of view, with the result that each underwent a revision which led to such a rearrangement of subdivisions as to present an entirely fresh conception of their relations to one another. The logical result was finally expressed in the subdivision of the plant world into four great phyla, which, in their evolutional sequence, came to be known as I., Thallophyta; II., Bryophyta; III., Pteridophyta; IV., Spermatophyta.
