SCITAMIXAI.ES. 



181 



sound, but tho pro'sont speric^s from Mississip|ii 

 can not ])o retained in tlie gtnius Calanioijsis, 

 even if that is a valid ^enns. 



Tlio present species ap|)ears to he closely re- 

 lated to tlie existing <:jenus Channvdorea of Will- 

 denow, a cjcmuis of small ])ainis with reedlike 

 stems, \s-iiicli conunonly spreads hy runncu's. 

 It has, in the existini; flora, ahout, 60 species, 

 ranging from central Mcvxico to Bolivia and 

 Peru in the Andes and in western Brazil. It is 

 richest in species in the hnmid mountainous 

 region of Central America, and several of these 

 Central American forms are practically indis- 

 tinguishable from this Wilcox species. 



Occurrence. — Ackerman formation, Coleraans 

 Mill, Choctaw County, Miss, (collected by E. W. 

 Hilgard). Wilcox group, Benton ( ?), Saline 

 County, Ark. (collected by R. E. CaU). 



Collections. — U. S. National Museum. 



Order SCITAMINALES. 



Family CANNACEiE. 



Genus C.4NNA Linne. 



C.\NNA EOCEXicA Borry, n. sp. 



Plate XV, figures 7 and 8. 



Description. — Leaves elongate-lanceolate, at 

 least 25 or 30 centimeters in length and proba- 

 bly longer. Maximum width not preserved. 

 Fragments show a width of 6 centimeters on 

 one side of the midrib without reaching the 

 margin. Margins entire, gradually narrowing 

 distad. Midrib stout below, at least 1 centi- 

 meter in width, becoming obsolete in tlie tip. 

 Secondaries thin, equally spaced, suhparallel, 

 and numerous, diverging from the midrib at 

 acute angles, 60° in large, presumably proximal 

 fragments, at intervals of about 2 millimeters, 

 becoming increasingly more ascending distad 

 until in the upper part of the leaf they are 

 approxinuitely parallel with each other and 

 with the long axis of the leaf. Tertiaries nu- 

 merous, thin, several between the secondaries, 

 with whiclr they are parallel. A specimen from 

 Old Port Caddo Landing measures 9 by 2() 

 centimet(>rs and indicates a leaf 12 by .56 centi- 

 meters. 



This species is not especially well presei'ved, 

 although it is represented by runnerous frag- 

 mc^nts in tlie clays of the Holly Springs sand at 

 Oxford, Miss. Larger fragments from the 

 Grenada foi-mation at Grenada, Miss., arc 



better preserv(vd than tlu^ specimcMis figured. 

 It may be compared with numerous (>Nisting 

 species of Cannacea? and Mai'antacea' and is 

 referred to Caiuia becavisc' it shows no generic 

 didereiices and also ])ecausc more coniplet(> 

 forms from (he Chiihorne group, (h'scrihed in 

 niainiscripl , ar(^ chvirly referal)l(^ to Canna. 



T]u^ g(>nus Caiuia is exclusively American and 

 contains from 25 to 50 species, mostly liygro]>hi- 

 lous in lial)it and confined to the Tropics and 

 suiitropics. One species. Canna flaccida Kos- 

 coe, penetrates northward as far as South 

 Carolina in the swamps that skirt the coast. 

 Among fossil forms that are referred to the 

 Scitaminales is the genus Scitaminophyton de- 

 scribed by Massalongo from the Italian Tertiary 

 and the genus Cannojihyllites of Brongniart, 

 with several species ranging from the Upper 

 Cretaceous to the Pliocene in the European 

 area. Fritel has recent ly shown that t he French 

 Ypresian species of the Paris Basin dcscril)ed 

 originally by Watelet as CannophyUif.es iingrri 

 is in reality based on fragments of an inideter- 

 minable palm.' Allied forms supposed to repre- 

 sent the family Zingiberacea' have been referred 

 to the form genus Zingiberites Ileer, which com- 

 prises several species, one Upper Cretaceous and 

 the otliers early Tertiary. One of these species, 

 Zingiherites Jubiiis Lesquereux,- is based on 

 very fragmentary material from the Denver 

 formation of Colorado not identical with the 

 present species. The Zingiberaceie is a large 

 family and is confined almost exclusively to the 

 Eastern Hemisphere. Though no competent 

 student would dispute its possible occiuTence in 

 the American Tertiary, the evidence should bo 

 more complete than that furnished by Les- 

 ([uereu.x to be at all convmcing. 



Another form genus for fossil leaves much 

 like the one under consiileration is Musoplvyl- 

 hi/n, fu'st described by Goppert for an un- 

 doubted Tertiary species of Musaoose from the 

 island of Java. Ten or a dozen species have 

 since been described froiu the Tertiary of Eu- 

 rope and one, Musophylhnn complicatum Les- 

 (pierenx, from the ejirly Eocene of the Rocky 

 Mountain region. Thougli it is beyond the 

 province of the present work to discuss at any 

 lengt-li the botanic; adinities of these sjXH-.ies of 

 Musoi)hyllum, I would at least point out that 



I Jour, botanique, vol. 22, pp. Un, Ul, flg. 1, IWIO. 

 -Lesquereux, Leo, The Tertiary llora, p. 95, pi. h',, fig. 1, 1S78. 



