Through all the Tertiary period (see [p. 34]) there were numerous flowering plants, and there is evidence that many families of both Monocotyledons and Dicotyledons existed in the Upper Cretaceous times. Further back than this we have little reliable testimony, for the few specimens of so-called flowering plants from the Lower Mesozoic are for the most part of a doubtful nature.
The flowering plants seem to stand much isolated from the rest of the plant world; there is no direct evidence of connection between their oldest representatives and any of the more primitive families. So far as our actual knowledge goes, they might have sprung into being at the middle of the Mesozoic period quite independently of the other plants then living; though there are not wanting elaborate and almost convincing theories of their connection with more than one group of their predecessors (see [p. 108]).
It is a peculiarly unfortunate fact that although the rocks of the Cretaceous and Tertiary are so much less ancient than those of the Coal Measures, they have preserved for us far less well the plants which were living when they were formed. Hitherto no one has found in Mesozoic strata masses of exquisitely mineralized Angiosperm fragments[8] like those found in the Coal Measures, which tell us so much about the more ancient plants. Cases are known of more or less isolated fragments with their microscopical tissues mineralized. For example, there are some palms and ferns from South America which show their anatomical structure very clearly preserved in silica, and which seem to resemble closely the living species of their genera. The bulk of the plants preserved from these periods are found in the form of casts or impressions (see [p. 10]), which, as has been pointed out already, are much less satisfactory to deal with, and give much less reliable results than specimens which have also their internal structure petrified. The quantity of material, however, is great, and impressions of single leaves innumerable, and of specimens of leaves attached to stems, and even of flowers and fruits, are to be found in the later beds of rock. These are generally clearly recognizable as belonging to one or other of the living families of flowering plants. Leaf impressions are by far the most frequent, and our knowledge of the Tertiary flora is principally derived from a study of them. Their outline and their veins are generally preserved, often also their petioles and some indication of the thickness and character of the fleshy part of the leaf. From the outline and veins alone an expert is generally able to determine the species to which the plant belongs, though it is not always quite safe to trust to these determinations or to draw wide-reaching conclusions from them.
In [fig. 59] is shown a photograph of the impression of a Tertiary leaf, which illustrates the condition of an average good specimen from rocks of the period. Its shape and the character of the veins are sufficient to mark it out immediately as belonging to the Dicotyledonous group of the flowering plants.
Seeds and fruits are also to be found; and in some very finely preserved specimens from Japan stamens from a flower and delicate seeds are seen clearly impressed on the light stone. In [fig. 60] is illustrated a couple of such seeds, which show not only their wings but also the small antennæ-like stigmas. Specimens so perfectly preserved are practically as good as herbarium material of recent plants, and in this way the externals of the Tertiary plants are pretty well known to us.
Fig. 59.—Dicotyledonous Leaf Impression from Tertiary Rocks
Fig. 60.—Seeds from Japanese Tertiary Rocks; at a are seen the two stigmas still preserved
A problem which has long been discussed, and which has aroused much interest, is the relative antiquity of the Monocotyledonous and the Dicotyledonous branches of the flowering plants. A peculiar fascination seems to hang over this still unsolved riddle, and a battle of flowers may be said to rage between the lily and the rose for priority. Recent work has thrown no decisive light on the question, but it has undoubtedly demolished the old view which supposed that the Monocotyledons (the lily group) appeared at a far earlier date upon this earth than the Dicotyledons. The old writers based their contention on incorrectly determined fossils. For instance, seeds from the Palæozoic rocks were described as Monocotyledons because of the three or six ribs which were so characteristic of their shell; we know now that these seeds (Trigonocarpus) belong to a family already mentioned in another connection ([p. 72]), the Medulloseæ (see [p. 122]), the affinity of which lies between the cycads and the ferns. Leaves of Cordaites, again, which are broad and long with well-marked parallel veins, were described as those of a Monocotyledonous plant like the Yucca of to-day; but we now know them to belong to a family of true Gymnosperms possibly distantly related to Taxus (the Yew tree).