NATURAL SCIENCES OF PHILADELPHIA L15 



and thej r diverge and curve downwards just as these stipules do. Above tl 

 strong veins of the auricle are very weak veins, necessitating a \ i - ii.htdw 

 blade portion there, until another set of strong veins push out ami make i In- 

 main part of the lamina. 



If we press these auricles back against the petiole, and imagine an union 

 with it, then a separation from the main leaf blade, and an union of the ed 

 of the separated auricle, both above and below, we have a sheathed stipule 

 exactly as we find them, and we see how easily Magnolia Frazeri might be a 

 pinnate leaf of -five leaflets on the supposition that the stipular portions really 

 have taken the course we suppose these auricles might take. 



I suppose no one of experience in living plants doubts the possibility of the 

 adhesion of some parts and the separation of others, so as to make new parts 

 or organs. If such is desired, I would refer to the adhesion of the carpellary 

 leaves by their backs in the capsules of Staphi/lea trifolia ; and for separation 

 to the pinnate leaf often formed out of an entire blade in Fraxinus excelsior, 

 heterophylla, and many other plants with entire leaves which often have pin- 

 nate ones amongst them. 



It is scarcely possible, with these facts before us, to avoid the suspicion that 

 the stipules of Magnolia are not formed like the stipules of most plants, which 

 are perhaps leaf portions which have never beeu well developed, but rather are 

 the tolerably well developed side pinnules of a trifoliate or deeply auricled 

 leaf, which in an early stage had adnated with the petiole, and by their edges, 

 and thus formed the stipular sheath we see. The suppositious case 1 have 

 drawn from the auricles of M. Frazeri is still better illustrated by leaves of 

 some Ranunculaceous plants. For instance, Anemone Pennsylvanica. Lay the 

 lower lobes flat against the petiole, imagine the adnation by their backs, and 

 cohesion of the edges, and we have the idea clearly. 



It is difficult to conceive that these stipular sheathes could have been formed 

 in harmony with all the appearances we have detailed, in any other way ; but 

 ideas and possibilities are not as good as direct facts. These are furnished 

 in good part in other ways. 



In the East Indian species, M.fuscata, the flowers are axillary, not terminal 

 as in most other species. Three of the leaf axils on the growth of last year pro- 

 duce flowers. The lowest flower is the weakest, the upper the strongest. The 

 bracts which enfold the flower buds are of course transformed leaves ; and 

 here, in these weak flowers, where the tendency of the vital course is almosl 

 as near to foliar organs as to floral parts, we find these leafy-looking bracts 

 are trifoliate. The central lobe is composed of a short petiole, and a small 

 oval leaf blade. Sometimes this attempt of the lower axil to produce a flower 

 proves abortive. The already formed petals die away. In such cases the two 

 lateral leaflets die away also, and the little miniature central leaf goes on and 

 developes into one as large as the average on any part of the plant. But in 

 the stronger flowers we find, just in proportion to their strength, the two lat- 

 eral leaflets enlarge, and the central one diminish until at length it disappi 

 petiole and all. The laterals then adhere by their edges, become fleshy, and 

 end in bein- petals. These are clearly seen to be formed out of the adnated 

 lateral leaflets, which form the stipular sheaths in other cases, with the cen- 

 tral of the trifoliate leaf type absorbed. This observation, in addition to the 

 use I wish to make of it, confirms the views of some botanists, as 1 have 

 learned from Professor Asa Gray, that it is by metamorphosis of the petiolar 

 and stipular parts, rather than by modifications of the leaf blade, that petals 



are formed. ., , . , , 



From these facts we gather the certainty of a trilobate type of leaf, and see 

 the adnation of the edges; and only the dorsal adhesion to the petiole 

 which I have shown so probable as almost to amount to a certainty, is Left to 

 be established by actual fact. . 



This ternate division of the leaf is a marked character ,n Ranuncula 

 and with this exposition of a ternate type in Magnohaceae, its chum to a p 



1870. J 



