Downloaded from http://jgslegacy.lyellcollection.org/ at University •~0 PROCEEDII~IGSCollege OF London THE GEOLOGICAL on July 2, 2016 SOCIETY. EApr. 28, APl~In 28, 1858. James Powrie, Esq., Reswallie, near Forfar, Marcus ttuish, Esq., Castle Donnington, Derby, l~enry D. M. Spence, Esq., Hyde Park Square, and Parkin Jeffcock, Esq., Derby, were elected Fellows. The following communications were read :-- 1. On some ¥EGETABIm RP.~AI~S from MADEIRA. By C~rA~nF.S J. F. Bvs~cav, Esq., F.R.S., F.G.S. TEE vegetable remains procured by Sir Charles Lyell and M. Hartung from the leaf-bed which they discovered* in the ravine of S. Jorge, in the Island of Madeira, were entrnsted to me for examination ; and I propose now to lay before the Geological Society the observations I have been able to make upon them. I have examined 140 speci- mens, by far the greatest part of them in a very imperfect state, mostly small fragments, often quite lmdeterminable, and, even when most perfect, no more than single detached leaves, of which, how- ever, the veins and margin are often very well displayed. Dico- tyledonous leaves predominate, but arc intermixed with numerous remains of Ferns, always, however, in small fragments, insufficient to give any idea of the general form of the frond. The very fragmentary and incomplete condition of the remains of Ferns in this bed is indeed striking. Something of it is attributable to the natnre of the stone, which is by no means fissile, but breaks quite irregularly; but something also would seem to be owing to the original conditions of the deposit. The ciremnstanee is unfortunate, because Ferns are so variable, and show such differences even in different parts of the same frond, that great uncertainty attends the determination of them from such small fragments. Of the Dicotyledonous leaves, a large proportion, as I have said, are too imperfect to be even described. Of those that are best pre- served, a few kinds may, with a certain degree of confidence, be referred to species now existing in Madeira ; others are manifestly different from any known to exist there; the rest have, in my opinion, no character sufficiently marked to determine their affinities. On this point, however, I am aware there may be different opinions. Botanists, indeed, are far from being agreed on the question, whether the affinities of Dicotyledonous plants can, as a general rule, be determined from the leaves alone; that is, whether the leaves by themselves afford characters sufficientlydefinite, and sufficiently in accordance with those afforded by the fructification, to be safe guides to the determination of genera or families. Some eminent pal~eobotanists of Germany and Switzerland, attaching great import- ance to the characters of leaves, have assigned generic names to the fossil plants of the tertiary age with a degree of confidence which, in not a few instances, appears rather surprising. This question cannot be completely settled without a very exten- See Lyell's Manual of Elementary Gcology, 5th ed. (1855), p. 518. Downloaded from http://jgslegacy.lyellcollection.org/ at University 1858.] ~vN~tmY--~OSSlI~College London L~AWS,on July 2, XA_I)~IaA. 2016 51 sive and careful survey of existing families of plants; and to go fully into it, even if I were at present completely prepared to do so, would be too great a digression from the immediate purpose of this memoir. I may observe, however, that, so far as I have hitherto been able to examine this subject, my conclusions are not favourable to the views of the naturalists above referred to. There are but few cases, as it appears to me, in which particular modifications of leaves are distinctly characteristic of natural orders or of natural genera. Certainly, in very many cases, wide variations in the form, venation, texture, and other characters of leaves are observable in the best- defined and most natural groups. For instance, in the Oaks, there are at least three distinct andwell-marked types of venation :--lst, the Chestnut type, in which the feather-veined ~ character is shown in the highest degree, and which is exemphfied in several of the Himalayan Oaks, such as Quereus serrata and Q. lineata, and less perfectly in some of the American kinds ; 2ndly, the Oa]c type proper, seen in the deciduous-leaved oaks of Europe; and 3rdiy, the Laurel type, in which the principal lateral veins combine into arches within the margin; this last form prevails in several of the evergreen oaks both of India and of America. Again, in the small and very natural genus Alnus, there are two quite distinct types of venation :--the feather-veined, in the common Alder, and in the Alnus.ineana and A. viridis ; the arch-veined, in the Alnus cordifolia and Alnus Nipalensis. So also in the genus Fagus : our common Beech and those of North America, and one of the antarctic species (F. Antarctica), agree in a well-marked type of leaf; but the Fagus Solandri of New Zealand and F. betuloides of Fuegia have leaves of so different a character, that their affinity to the first-mentioned species could never be inferred from those organs. We sometimes find certain fossil leaves spoken of as having the characters of the Proteaceous family. Yet, among the Proteacea" that I have examined, the leaves are so various in all respects, that I do not know anything approaching to a common character, unless it be the rigid eoriaceous texture; and this, though certainly very general in the order, is scarcely apparent in some, such as Lomatia dentata. ~Vhat is there in common between the leaves of Leuca- dendron argenteum and of Guevina Avellana ?--or of Knightia and any of the Serrurias ? Even in the same genus,--how could any one, from the leaves alone, infer the generic identity of Conospermum iongifolium and C. ericifolium ? The family of Rhamnece, another to which many fossil plants have been referred, includes several very distinct types of venation of leaves; and, even in the genus Rhamnus itself, the leaves of Rhamnus catharticus are materially different in their veining from those of Rh. Frangula. I do not dispute that there are some families and genera of plants which may be easily recognized by characters connected with the * See Lindley's Introduction to Botany, ed. 1., p. 93. I adopt the terms pro- posed by this author for describing the venation of lea~es. ~2 Downloaded from http://jgslegacy.lyellcollection.org/ at University 52 PROCEEDINGSCollegeOF London THE GEOLOGICAL on July 2, 2016 SOCIETY. [Apr. 28, leaves and their appendages. Such are the Myrtacev~~, the Mela- stomaeece, the Coniferw, the Rubiace~ (Cinchonacece of Lindley), the genera .STepenthes, Sarracenia, Bauhinia, Begonia, Cinnamomum, and to a certain extent the genus Fieust. But every botanist who has examined large unarranged collections of dried plants from foreign countries must be aware how difflenlt it is, in general, to make out the affinities of specimens without fruit or flower, and how often very similar leaves belong to plants of widely different families. This part of the subject has, in truth, been so ably treated by Dr. Hooker in the 10th volume of this Society's Journal, p. 163-165, that it would hardly have been necessary to touch upon it again, were not the confidence with which some eminent foreign geologists assign generic names to mere detached leaves calculated, in my opinion, to mislead the inexperienced ; wherefore I have thought it advisable to enter a fresh protest, and to remind geologists of the lmcertainty of the evidence we possess. The imperfect state of our materials must always be borne in mind. When we have before us only detached leaves, as most commonly happens in fossiliferous deposits, and as is the case, in particular, with all the specimens I have seen from Madeira, we lose the benefit of several characters which assist materially in the determination of recent plants: in particular, the character of insertion (alternate, opposite, or whorled); and the more important one of the presence or absence of stipule, and their nature when present. As an example of the importance of these organs, I may notiee, that the Magnolia family may be easily known, in general, by the pee~fiiarities of the stipules, taken together with those of the leaves; but the leaves separately would afford no sufficient indica- tion of the order. The same may be said, still more decidedly, of the Cinchonacece of Lindley. Another character, of great use in the detel,mination of recent plants, but almost necessarily wanting in the fossil state, is the nature of the pubesvence, the hair, down, scales, or other clothing of the surface. In those rare cases where the fossil leaves are so well preserved that the cellular structure of the epidermis, and its pores or stomata, can be satisfactorily examined, these may probably afford valuable aid towards the determination of affinities. Yet it is a point by no means determined, and one whieh ~l~serves careful study in recent plants, how far the characters ufforded by the epidermis and its pores are in accordance with those of other organs. It is certain, to mention one instance only, that the structare of the epidermis and position of the stomata in Salisburia are exceptional in the family of Conifers. In all that I have said, I have been speaking of the difficulty of determining plants by their leaves only. Where fruits of a marked * The peculiar intramarginal vein of Myrtace~, however, occurs in several species of other families. t I purposely omit to mention those genera which consist of only one species, such as I,iriodendron.