70 FLORA OF SOUTHEEN NEW YORK AND NEW ENGLAND. 



of Liriodendron J ' a in which he criticises their reference even to the Magnoliacese, 

 and says (loc. cit., p. 33) : "* ■* * there is good reason for considering some of 

 the obcordate leaves as belonging to plants of a quite different family, namely, if 

 we compare them with leaflets of the Leguminosae." 



The same author subsequently continued the discussion in a paper " On the 

 Validity of Some Fossil Species of Liriodendron ," & in which he calls attention to a 

 specimen figured by me, showing three leaves in close juxtaposition, and pertinently 

 remarks (loc. cit., p. 314): a * * * might we not then assume that they have 

 been situated close together, as they were found in the rock? They seem, indeed, 

 to have formed a trifoliate leaf, not unlike Desmodium, PJiaseolus, and others. Their 

 venation is much more like that of the Leguminosae than of any known Liriodendron. 

 Moreover we must not forget that notched leaves are not only common among the 

 Leguminosae, but exist in many genera of various families, e. g., Zygophyllum, Pas- 

 si flora, Akebia, etc. — which might also be taken into consideration." 



Leaves which are superficially indistinguishable from some of ours are described 

 and figured by Bayer from the Cretaceous of Bohemia under the name Bignonia 

 pulcherrima, d and it is interesting to note that in his fig. 126a he shows three leaflets 

 joined to a common petiole, thus forming a compound leaf. These figures are repro- 

 duced for comparison in our figures 2 and 3 on PI. XXV, together with Newberry's 

 type figure of Liriodendropsis simplex in fig. 1. 



A number of other fossil leaves, which have been described from time to time 

 under different genera, are impossible to separate from the general type represented 

 in Liriodendropsis. As examples in this connection may be mentioned Sapotacites 

 retusus Heer, e and Myrsinophyllum varians Vel./ a figure of which is reproduced 

 for comparison on PL XXV, fig. 6. 



Finally, attention may be called to the interesting comparison made by Ward 

 between certain forms of Liriodendropsis simplex and Chondrophyton laceratum Sap., 

 from the Cretaceous of Portugal,^ which latter he does not hesitate to rename Lirio- 

 dendropsis lacerata. 



In view, therefore, of the wide differences of opinion which have been expressed 

 in regard to the probable botanical affinities of these leaf forms and the impossibility 

 of separating one from another, except in the case of extreme forms, I have thought 

 it advisable to include all of the specimens from our vicinity under the generic name 

 Liriodendropsis, leaving it in the systematic position in which it was placed by the 

 author and separating it into as few species as possible, although doubtless some 

 authorities may be inclined to recognize additional species or varieties among the 

 many forms figured. 



oProc. U. S. Nat. Mus., vol. 13, 1890, pp. 15-35, pis. 4-9. 



b Bot. Gaz., vol. 20, 1895, pp. 312-316, pi. 23. 



c Liriodendron simplex Newb., Glen Cove, Long Island, N. Y. Trans. N. Y. Acad. Sci., vol. 12, 1893, pi. 5, fig. 2. See this 

 monograph, PI. XXIII, fig. 5. 



dStudien Gebiete Bohm. Kreideform. (Perucer Schichten), 1900 (1901), p. 156, figs. 126a, 126b. (Fig. 126a reduced in 

 size; fig. 126b nat. size.) 



e Fl. Foss. Arct., vol. 7, p. 32, pi. 61, fig. 10. 



/Kvet. Cesk. Cenomanu, p. 25, pi. 4, figs. 8, 9; pi. 5, fig. 12; pi. 6, figs. 10, 11. 



Sixteenth Ann. Kept. U. S. Geol. Survey, 1894-5 (1896), pt. 1, p. 540, pi. 107, figs. 6-8. 



