28 LEAVES OF LIRIODENDRON HOLM. 



ilar form, though with the exception that the bUide tapers distinctly into 

 the petiole in the leaf figured by Massalougo. And the next variety, 

 obtusiloha, figured on Plate xxxix, Figs. 3 and 5 I. c, shows again the 

 same kind of variation which occurs so often in the foliage of L. Tulipi- 

 fera, with the lobes more or less obtuse, an approach to the variety oh- 

 tusiloba described by Michaux. The two subvarieties siibattenuata and 

 rotundata (Plate xxxix, Figs. 3 and 5 /. c.) are distinguished by tlieir 

 blades tapering into the petiole or not. The last variety, incisa (Plate 

 xxxix, Figs. 4 and 6 I. c), shows a rather deep notch, but that is the 

 only diiference, and this seems to correspond to the leaves I have figured 

 on Plates vi. Fig. 26, and ix, Fig. 40. Furthermore, Ettingshausen* 

 has described a species Liriodendron Hauerii, from a leaf of which, 

 however, only the lower part is preserved and shows that the blade 

 tapers into the petiole, and that there are two pairs of relatively strong 

 lateral ribs, but without any lobes, since the margin of the leaf is en- 

 tirely wanting. The author sees, nevertheless, in this poorly preserved 

 fossil a difference not only from L. Tulqnfera, but also from L. Procac- 

 cinii. Another species is L. GardueH Sap., which has been mentioned 

 by Saporta,t and this leaf does not seem to differ in any degree from 

 the recent, at least not from the leaves figured on Plate vii, Figs. 26, 

 27, 29, 30. Saporta compares this form with the above-mentioned 

 variety chinensis, recently discovered in China, It is very interesting 

 to see this leaf of L. Gardneri figured together with three other leaves, 

 but representing L. Frocaccinii from Iceland, Eriz, and Meximieux, since 

 these four leaves illustrate, although in a small degree, the variation in 

 the foliage of our living Tulip-tree. The same author has also, together 

 with Marion, in their "Recherches sur les vegetaux fossiles de Mexi- 

 mieux,"! described leaves of L. Procaccimi, some of which represent 

 quite large leaves, especially Figs. 1 and 2 {I. c), but unfortunately very 

 defective, so that the lobes are not very distinct, but seem, however, to 

 have been somewhat obtuse. The two leaves. Fig. 3 and 5 (L c), are, on 

 the contrary, preserved very well, and show two very obtusely lobed 

 leaves, and here is to be observed the rudimentary prolongation of the 

 midrib. The last fossil European species of Liriodendron, L. Celalcovslcii, 

 has been described by Velenovsky in his "Flora der Boehmischen 

 Kreideformation,"§ though with little success, since, according to the 

 figure, it may never have belonged to any species of Liriodendron. It 

 seems, therefore, quite curious to see the following remark of Velenovsky, 

 concerning this leaf: "Von dem lebenden Amerikanischen L. Tidipi- 

 fera L. unterscheidet sich L. Celalcovsldi durch die form, obwohl der 

 Habitus und die Nervation in hochsten (Tradeuebereinstimmt," because 



* Constantiu von Ettingshausen : Die fossile Flora des Tertiiir-Beckens von Biliu, 

 III, 1869, Plate xia, Figs. 10 and 10 h. 



t G. de Saporta : Origine Pal6ontologiquc des aibrcs cultiv6s on utilises par I'liomuie, 

 1888, p. 266, Fig. 1. 



t Archives dn Museum d'bistoire uatnrelle de Lyon, I, 1872, PI. xxxiii, p. 271. 



§ J. Velenovsky : Flora der Boebm. Kreideform. in Beitriige zur Paleontologie des 

 Oesterreicb-Uiigarns und des Orient, iii, 1883. 



