54 BOTANICAL GAZETTE [JULY 



Somewhat aside from the preceding are those forms included 

 in Liriophyllum, a genus established by Lesquereux in i^']^^ 

 to contain certain leaves from the Dakota formation evidently 

 allied to Liriodendron. They are somewhat coriaceous in 

 texture, with obscure venation, and differ widely in outline from 

 any known forms of Liriodejidron, being (except L. obcordatnm^ 

 which is probably not a Liriophyllum at all) square or broadly 

 rhomboidal in general outline, and split at the apex along the 

 line of the midrib about half way to the base, the two lobes 

 being sublobate or bilobate, and separated by an obtuse sinus. 

 Just what is the relation between these leaves and Liriodendron 

 is hard to say. We know of no forms of Liriodendron, either 

 ancient or modern, that approach very closely the peculiar 

 shapes oi Liriophyllum populoides, d^nd L. Beckwithii. Fig. i shows 

 the nearest form to Liriophyllum I have been able to find in 

 some ten years collecting. In no case, however, is the resem- 

 blance very close, the nearest being those which would be 

 identical with Liriophyllum populoides were the base somewhat 

 wider. Bilobate leaves with a deep, wide sinus at the apex are 

 common enough, the leaves tending to assume that form in the 

 vicinity of flowers, or where the nourishment is defective ; but 

 none of them show any tendency to widen at the base, or con- 

 tract the width of the sinus as in the leaves of Liriophyllum 

 i^fig^ 2). There are two forms of Liriodendron leaves, either of 

 which I conceive could have been ancestral to Liriophyllum. 

 One is the common form of young leaves on modern shoots, and 

 needs but to become parted farther down along the midrib, and 

 to more largely develop the basal and apical lobes, to be a true 

 Liriophyllum. The other is some form similar to Lirioden- 

 dropsis angiisti folia, which, by a shortening and widening of the 

 blade, through some such form ?i% Liriodendronalatiim, might lead 

 to Liriophyllum. Of the two, the latter seems to be the more 



i 





tenable, for the orbicular notched leaves, first mentioned above, 



have never been found in the fossil state, unless PhylliUs orbicti- \ 



laris may be so considered. It requires but a slight enlargement 



'*Haydeii's U. S. Geol. and Geog. Survey, p. 482. 



■J 



\ 



