As in the case of the Angiosperms, the fossils we have of Cretaceous and Tertiary Gymnosperms are nearly all impressions and casts, though some more or less isolated stems have their structure preserved. Hence our knowledge of these later Gymnosperms is far from complete. From the older rocks, however, we have both impressions and microscopically preserved material, and are more fully acquainted with them than with those which lived nearer our own time. Hard, resistant leaves, which are so characteristic of most of the living genera of Gymnosperms, seem to have been also developed in the past members of the group, and these tend to leave clear impressions in the rocks, so that we have reliable data for reconstructing the external appearance of the fossil forms from the Palæozoic period. The resinous character of Gymnosperm wood probably greatly assisted its preservation, and fragments of it are very common in rocks of all ages, generally preserved in silica so as to show microscopic structure. The isolated wood of Gymnosperms, however, is not very instructive, for from the wood alone (and usually it is just fragments of the secondary wood which are preserved) but little of either physiological or evolutional value can be learned. When twigs with primary tissues and bark and leaves attached are preserved, then the specimens are of importance, for their true character can be recognized. Fortunately among the coal balls there are many such fragments, some of which are accompanied by fruits and male cones, so that we know much of the Palæozoic Gymnosperms, and find that in some respects they differ widely from those now living. There is, therefore, much more to be said about the fossil Gymnosperms than about the Angiosperms, both because of the better quality of their preservation and because their history dates back to a very much earlier period than does the Angiospermic record. Indeed, we do not know when the Gymnosperms began; the well-developed and ancient group of _Cordaiteæ_ was flourishing before the Carboniferous period, and must therefore date back to the rocks of which we have no reliable information from this point of view, and the origin of the Gymnosperms must lie in the pre-Carboniferous period. The group of Gymnosperms includes a number of genera of different types, most of which may be arranged under seven principal families. In a sketch of this nature it is, of course, quite impossible to deal with all the less-important families and genera. Those that will be considered here are the following:—