J. Starkie Gardner — Fossil Flowering Plants. 501 



endemic to the Seychelles, from Sabal Blackbumiana of the Ber- 

 mudas, from a Desmoncus, an Areca, a Monoclora, and probably 

 of many, certainly of some other palms. When we consider that 

 probably many of the kinds of palm fruits would sink at once, we 

 realize how great an assemblage of this magnificent family is indi- 

 cated by the Sheppey drift. 



The difficulties we fear of determining anything but a fraction 

 of the Sheppey fruits must prove insurmountable. Their outer coats 

 are for the most part destroyed, and some part of their inner structure, 

 nearly always quite different in form from that which is external, is 

 revealed. Botanists have been able to determine but few of the 

 drifted fruits brought home by the Challenger, though these are more 

 perfect and of living species belonging to definite and known floras. 



The Bournemouth cliffs continue to furnish fresh forms, though 

 the leaf-beds are becoming more and more difficult of access. We 

 have especially enriched the series of Smilacece, and a complete 

 account of them has been presented to the Linnean Society. The 

 series now obtained falls little short of a hundred specimens, and is 

 by far the richest of fossil Smilaceas, perhaps of any family, ever 

 brought together. Such a material has enabled us to reduce the 

 number of distinct species to no more than five, most of which are 

 represented by foliage in all stages of development, from the largest 

 leaves measuring several inches, down to quite minute leaves from 

 near the extreme growing points. The necessity for such extensive 

 series when dealing with fossil leaves may not at once be apparent, 

 but the President of the Linnean Society expressed the opinion at 

 the meeting, that out of less material, not five, but five-and-twenty 

 species might have been made. 



The leaves of Smilaceee are highly characteristic, and can be 

 determined with a large degree of certainty ; but it is quite im- 

 probable that such will be the case with very many of the families 

 of Dicotyledons. There is, indeed, little hope that more than a very 

 few can be determined with anything like the precision required for 

 botanical purposes, unless we can call in aid the fruits or some other 

 organs. Thus if we may base a conclusion upon the large number 

 of the characteristic bracts, which envelope the seed in a section of 

 Flemingia that are met with in the Bournemouth flora, the leaves of 

 that genus should be far from uncommon, and they should also be 

 found in the Swiss Oligocene, yet no species of Flemingia has ever 

 been recorded from the Tertiaries. The leaves, however, may be 

 sought for among the supposed species of Populus and Carpinus. 



Fortunately fruits and even flowers are comparatively abundant 

 at Bournemouth, and we consequently anticipate little difficulty in 

 determining leaves belonging to such easily distinguishable fruits as 

 Alnus, Tilia, Acer, Carpinus. the Leguminosoe, and many others, but the 

 residuum with indeterminable fruits, or fruits that will not float, may 

 be very large. We are thus brought to the question, whether any 

 value beyond that of mere landmarks, or aids to the correlation of 

 rocks, can be attached to the determinations of fossil dicotyledonous 

 leaves arrived at when fruits are absent. Nearly every Tertiary and 



