166 



SHORTER CONTRIBXTTIONS TO GENERAL GEOLOGY, 1921. 



Genus LIKIODENDRON Linne. 

 Liriodendron quercifolium Newberry. 



Plate XXXVI, figure:!. 



Liriodendron quercifoJium Newberry, Torrey Bot. Club 



liull., voL 14, p. (i, pi. (32, fig! 1, 1887; U. S. Geoi. 



Survey Mon. 2fi, p. 81, pi. 51, figs. 1-fi, 189(5. 

 lierry. New Jersey (Jeol. Survey Bull. 3, p. 138, pi. 



17, fig. 1, 1911; Torrey Bot. f/lub Bull., vol. 39, p. 



395, 1912; vol. 44, p. 182, 1917. 

 Lirio(lendro7i pinnatijidwm? Knowlton (not Lesquereux), 



U. S. Geol. Survey Twenty-first .'\.nn. Rept. pt. 7, 



p. 317, 1901. 

 Liriodcvdron snmvii Knowlton (not I/esquereux), idem. 



Leaves oblong, large, pinnately divided by 

 narrow sinuses into two to four lateral lobes. 

 Ape.x emarginatc. Base truncate to somewhat 

 cordate. Length along the midrib 7 to 9 

 centimeters and probably consideral)ly more 

 in some specimens, as one fragment measures 

 12 centimeters in width. Width in perfect 

 specimens about 9 centimeters. Lateral lobes 

 ovate, with very acute tips, some narrowed 

 proximad, giving them an almost obovate 

 outline; intervening lateral sinuses narrow and 

 deeply cut, some reaching nearly to the midrib, 

 rounded. Some specimens have only two main 

 lobes developed on each side and are then very 

 similar to the typical modern leaf of Lirioden- 

 drnii tulipifera. In these specimens, however, 

 the upper lobes are divided by a shallow sinus 

 into two acute lobules. Other specimens show 

 tliree lobes of equal magnitude on each side, 

 and one of the best specimens from the Wood- 

 bine sand has four nearly equal lobes on each 

 side, the basal and apical pairs being somewhat 

 shorter than the medial pairs. This form of 

 leaf is very suggestive of some species of 

 Quercus, but its variations, as well as its vena- 

 tion, show that it is related to Liriodendron. 

 The petiole is preserved for a considerable 

 length and is very stout, as is the midrib. 

 There is one main secondary traversing each 

 lol)e and running directly (o its apical point. 

 In addition there are one or more camptodrome 

 secondaries in each lobe which anastomose 

 with branches from the main secondary, their 

 number being dependent upon the relative 

 width of the lobe; they branch from the midrib 

 at angles of abotit 60°. 



At first sight this species appears to differ 

 considerably from Liriodendron ohlongifoUum 

 and from the modern form, but this difference 

 is 7iot nearly as great as it seems, and it is 

 proljable that Liriodendron qin rcifolium is 

 simply a variation from the conunon ancestor 



of the two species in the direction of Lirioden- 

 dron pinnatijidum Lescjuereux. Numerous 

 leaves of the modern tree can be found with 

 an incipient lobation suggesting Liriodendron 

 quercifolium. In these leaves, however, the 

 sinus is comparatively shallow and rounded, 

 so that the general appearance of the two is 

 not markedly similar. 



Knowlton recorded Liriodendron pinnalifi- 

 dum and Liriodendron snowii from the Wood- 

 V)ine sand, btit both of these prove to be frag- 

 ments of this species. 



Family TROCHODENDRACEAE. 

 Genus TROCHODENDKOIDES Berry, n. gen. 



This genus is proposed as a form gentis for 

 fossil leaves that appear to be referable to the 

 family Trochodendraceae. It is perhaps best, 

 for the present, not to attempt a definition. 

 Attention is called in a recent publication-' to 

 the possibility that certain Mesozoic forms of 

 dicotyledons commonly referred to Cehi.'^tro- 

 phi/lhnn. Populus, and Populophyllum repre- 

 sented ancestral forms of Tetracentron, Trocho- 

 dendron, and Cercidophi/lhim. A great many 

 CVetaceous j)lant species have been referred to 

 the existing genus Populus, and the evidence 

 for such a relationship is very slight in a iium- 

 ber of fonns, particularly among the older ones. 

 The plant from the Dakota sandstone described 

 by Les((uereux as Phj/JHtes rliomhoideu.'i, which 

 is present in the Woodbine sand, is here con- 

 sidereil the type of the genus and is for the 

 present the only species definitely assigned to 

 it. A critical survey of the late Lower Cre- 

 taceous and early Upper Cretaceous dicotyle- 

 dons would result in transferring a number of 

 forms to Trochod< ndroidef^, which may ser\-e 

 for the reception of aity fossil species of the 

 family. 



Trochodendroides rhomboideus (Lesquereux) Berry. 



I'late XXX\T, figure •",. 

 Ficuxf rhomlinith'UK Lesquereux, .\m. .lour. Sri., 2d ser., 



vol. 4(;, p. 9(1, KS(i8. 

 rhijJJilcx rhomboideus Lesquereux, Cretareous flora, p. 112, 

 1)1. 0, fig. 7, 1874. 

 Knowlton, 1^. S. Geol. Siu'vey Twenty-first Ann. 

 Uept., ]it. 7, p. 317. 1901. 



This species was described by Lestpiereux 

 from material found in the Dakota sandstone 

 at Decatuf, Nebr. lie at first referred it with 

 a query to Ficus, which it obviously does not 

 represent. Sidtsequently he transferred it to 



2' Borry, E. W., \m. Jour. Sci., 4lh ser., vol. 50, p. 49, 1H20. 



