I902J NOTES ON THE PHYLOGENY OF LIRIODENDRON 59 



r 



f 



f 



were very closely related, and probably had a common ancestor. 



Among the leaves of L. T2ilipifcra we can find a perfect series 



from L, oblongifolitim \,o L. giganteiim and L, practiilipifenim on the 



one hand, to L. qiiercifolitim on the other. The latter was 



derived from Z. oblongifoHum through leaves with increased 

 lobation. 



The next fossil species is L. piniiatifidiun Lesq., and we note 

 that Lesquereux's two figures of this species differ considerably 

 from each other in shape. Both are fragments, and their nerva- 

 tion is dissimilar and remote from that which obtains in all the 

 other known forms of Liriodendron. While one may be a 

 Liriodendron, we are forced to consider the other as an alto- 



f gether different 



Quercus, at any rate 



I not related to Liriodendron. We note in passing that Les- 



quereux's Cissites obtiisilobus, also from the Dakota group (FI. 

 Dak. i6i. pL Jj.fig, 5), somewhat resembles the species under 

 discussion. Lesquereux'^ wrote later of Cissites as follows: 

 *'Velenovsky, in his Flora Bohm. Kreideformation, pt. 2, pi. 6. 

 fig^ 2, has a figure like this, and has named it Liriodendron Cela- 

 kovskii. It essentially differs in the lateral primaries being 

 basilar." We are inclined to think that this observation is a hint 

 at the true affinity of the specimen, for it has the indefinable look 

 of a Liriodendron leaf, and we have in our collection of L. 

 Tiilipifera leaves several that approximate Cissites obtusilobiis . 

 CissiUs alatus Lesq. (FI. Dak. group, i6o. pi. 23. fig. 6. 1891) 

 in all probability also belongs here. Lesquereux was inclined to 

 '■efer it to Liriodendron, and afterward compared it with L. 

 Gardneri Saporta, which it greatly resembles. We have also 

 L. Tulipifera leaves that resemble it, but somewhat remotely. 



The American Cretaceous species of Liriodendron at this time 

 diverge somewhat widely from the line of descent leading to 

 the modern form, and develop into some curious lobate forms 

 that will be considered later. 



We are obliged to go to Europe to find the stages subsequent 

 to L,j)blongifolitmi leading to L. Tulipifera. We find them nearly 



' ^lora Dakota group, p. i6i, footnote. 



I 



/ 



