166 J. Siarkie Gardner — Development of Dicotyledons. 



Before quitting the subject, it is well to revert to the danger of 

 determining the ages of fossil floras in remote parts of the world by 

 comparing and estimating the percentages common to those of 

 Europe. Not only have we to keep in mind the similarity that 

 dicotyledonous leaves belonging to different genera bear to each 

 other, a likeness increased by the process of fossilization where the 

 matrices are similar, but the fragmentary condition of the specimens 

 usually brought from distant countries. In the first five volumes of 

 the Arctic flora hardly a perfect dicotyledonous leaf is illustrated, 

 and it is most rare to find them collected with so important a 

 character as the petiole complete. In addition to this, were we to 

 take an armful of fallen leaves at random from each country, such as 

 Siberia, Japan, Sumatra, Australia, New Zealand, Madeira, Scotland, 

 France, Greece and the United States, and compare them together 

 after the manner of paleeontologists, is it likely that we should find 

 grounds for supposing that they all belonged to floras growing 

 synchronously ? The fallacy of such reasoning is even more appa- 

 rent when we realize that a considerable percentage of the species of 

 even Eocene floras, including such unmistakable forms as Osmunda 

 Javanica, are now living on the East Coast of Asia, for a comparison 

 by percentages would make the existing and the Eocene floras 

 contemporaneous. In like manner we have living Asiatic species at 

 Mull and Antrim, living Australian species at Bournemouth and 

 Sheppey. These must have migrated to or from their present 

 habitats, and might be found imbedded anywhere along the routes 

 they have passed over, and if so found might be of any age between 

 the Eocene and the present day. 



We are too accustomed to treat these floras as complete exponents 

 of the vegetation of the past. They are, however, but mere frag- 

 mentai-y indications of what existed. The flora of Gelinden 

 consists of, for instance, some 60 species, but only contains three 

 Ferns and one Conifer. Can we suppose that Ferns and ConiferEe 

 were not at least as richly represented then as now ? And if we 

 were well acquainted with the bulk of them, how entirely erroneous 

 our inductions from the few we do know might appear. Bourne- 

 mouth is a case in point, for on the right of the old river delta, 

 of which the present cliffs are a section, the flora is completely 

 different in character to that met with on the left. The latter 

 is a flora of leafy trees, and had it alone been accessible, the climate 

 might have been inferred to have been too temperate for the growth 

 of Palms, for none are found in it. Thus an utterly different view 

 of the character of the plant world of the Eocene period during the 

 Middle Bagshot would be deduced by a student of the flora of the 

 right bank, from that fairly deducible by a student of the left bank. 

 But, extraordinary to tell, within a few hundred yards of the latter, 

 still another completely different flora is found, in which occur Fan 

 Palms in place of Feather Palms, Araucaria Cunninghami, Eucalypti, 

 Avoids, and a whole host of tropical Ferns. Even now, after years 

 of collecting, new and astonishing plants are continually coming to 

 light, such as Hewardia regia, a South American type of Fern, and 



