230 THE FLORA OF THE DAKOTA GROUP. 



sentatives. This species is figured in PI. XXV, Fig. 1 ; PI. XXVI, Fig. 5, and 

 PI. XXVIl, Fig. 1, and its variety cruciformis in PI. XXVIII, Figs. 1,2. From 

 it is derived L. intermedium, a species with large leaves, narrowly bilobate on 

 each side, deeply emarginate at apex distinguished by the oblique direction 

 of the upper lobes, which are far distant from the lower, as shown in PI. 

 XXV, Figs. 2^ 3, and by the forking or the division of the lower lobes, 

 the variety hilohata (PI. XXVIII, Fig. 4). Then by the upper lateral lobes, 

 which are simple, linear, obtuse, and ascending in an acute angle of diA'er- 

 gence, while the lower ones, curving upward, reach nearly as high as the top 

 of the upper, the leaves take a peculiar appearance, like that of some species 

 of Aralia as seen ini. Wellinf/tonii (PI. XXVIII, Fig. 3). And now another 

 peculiar mode of division is observable in the leaves of the genus. They 

 become alternately cut into a number of lateral lobes, either short, round, 

 obtuse, or truncate at the sides, as in PI. XXVII, Figs. 4, 5, wliich is L. piu- 

 naU/idnm ; or they have the lobes cut not merely to the middle of the lamina, 

 as in the last species, but throughout to the median nerve, and are narroAver, 

 being long, linear or rather slightly enlarged from the base to the rounded 

 or obtusely lobateapex, parallel and eqiaidistant at base, appearing like leaflets 

 of a compound leaf, as seen in L. Snowii (PI. XXIX, Figs. 1 , 2).^ Add to these 

 remarkable specific forms those described by Dr. Newbeny in Ijulletin of the 

 Torrey Botanical Club, January, 1887, as Liriodei/dron ohhuifiifoHum and L. 

 (piercifoHHm, leaves which are oval, bordered with long, sharp teeth, and 

 resembling oak leaves ; then the leaves described as L. ohcordaftim, this 

 volume (PI. XXVIII, Fig. 7), and again, those of L. Beckwithii and L. pop- 

 idoides Lesq. (Cret. and Tert. Fl., PI. x. Fig. 1 ; PI. xi. Figs. 1, 2), and from 

 the whole series of forms there are evidently represented twelve well defined 

 species Avith marked varieties, and from this we may get an idea of the 

 extraordinary variability of the leaves of Liriodendron, at or near the origin 

 of that genus, or nearly, as far as we know, of the first appearance of 

 dicotyledonous plants. 



In the leaves legitimately referred to Sassafras by their peculiar char- 

 acters, and by the presence of fruits found vt'ith tliem in the same strata, 

 we have those of S. Mudgei and *S'. aeutilohum Lesq. (Cret. FL, PI. xix), all 

 trilobate with entire lobes. In S. cretaceum (loc. cit., PI. xi. Fig. 1), the 

 leaves, which are also trilobate, have the borders of. the lateral lobes slightly 



' The figures of this species of Liriodendron are mucli like that of Artocarpus, sp. uov., iu Zittel's 

 liaiidbnch d. Paleontologie, vol. "2, p. 478, tig. 5, which represents a leaf discovered iu the Cretaceous 

 of Greenland by Prof. Natborst. Generic difference isshowu by the nervation and the emarginate apex. 



