I'LCRA OF THE CHEYENNE SANDSTONE OF KANSAS. 



217 



incurved and sometimes disappearing some dis- 

 tance above the point of insertion. To judge 

 by the lack of petiolules and the generally com- 

 plete character of the material it does not ap- 

 pear that the leaflets were normally shed, and 

 this is also indicated by the concrescence of 

 the terminal leaflets. The leaves are cori- 

 aceous and appear to have been stiff and strict in 

 habit. The rachis is stout and expanded prox- 

 imad. The midribs are excessively stout and 

 prominent on the lower surface. The seconda- 

 ries are relatively thin, straight, and subpar- 

 allel. They vary from camptodrorae to cras- 

 pedodrome. In the entire basal part of the 

 leaf and sometimes in the apex they are camp- 

 todrome. In many specimens one secondary 

 runs to each marginal tooth, although in other 

 specimens the camptodrome habit is retained 

 and a short branch enters the marginal tooth. 

 All these features are indicated in the accom- 

 panying figures. The tertiary venation is 

 usually obsolete, as the matrix is prevailingly 

 coarse. Occasionally percurrent nervilles are 

 seen. In specimens with broadly winged stripe 

 the venation of the leaf is continued in these 

 wings. 



This handsome species is represented by a 

 lara:e amount of material, which is fortunate, 

 as it would be almost impossible to correlate 

 fragmentary material. It is clearly a repre- 

 sentative of the genus Sapindop.'<is and would 

 well merit the specific name of rariabilis had 

 that not already been used for the type of the 

 genus, which came from the Patapsco forma- 

 tion of Maryland and Virginia. 



In the Patapsco formation the genus Sajnn- 

 dopsi.'i may be totally absent from a locality 

 or present in the greatest abundance, and this 

 is equally true of the Cheyenne sandstone of 

 Kansas, indicating possibly a gregarious habit. 



Various species of existing Sapindaceae 

 show similarities to the present species in form, 

 venation, and variation. All the previously 

 described species of Sapindopsis had entire 

 margins, and no trace of toothed margins has 

 been found in the material from the Atlantic 

 Coastal Plain. The existing genus Matayba 

 Aublet, with which I originally compared 

 Sapindopsis, has leaves with both entire and 

 dentate margins, and the general features 

 of Sapindopsis are siiared by other tropical 

 American genera of Sapindaceae. The genus 

 Matayba comprises about two score existing 

 32333°— 22 17 



species and is closely related to Cupania, also 

 exclusively American in the existing flora — 

 in fact, all the genera of the tribe Cupanieae 

 lomatorrhizae as segregated by Radlkofer are 

 confined to the warmer regions of the Western 

 Hemisphere. 



Occurrence: Localities 2221, 2224, 2229, 

 22.30, 7406, Medicine Lodge Creek, m draw 3 

 miles above Belvidere (Cheyenne sandstone 

 No. .3 of Hill) ; collected by Ward and Vaughan, 

 October IS, 1896 (unnumbered). 



Order MALVALES. 



FamUy STEECULIACEAE. 



Genus STERCULIA Linne. 



Sterculia towneri (Lesquereux) Berry. 



Plate LVII, figure 1 ; Plate LX; Plate LXI, figure 1. 



Aralia towneri Lesquereux, U. S. Geol. and Geog. Survey 



Terr. Bull., vol. 1, p. 394, 187-5 [1876]; Ann. Kept. 



for 1874, p. 349, pi. 4, fig. 1, 1876; CretareouB and 



Tertiary floras, p. 62, pi. 6, fig. 4, 1883; Flora of the 



Dakota group, p. 132, pi. 23, figs. 3, 4; pi. 31, fig. 1, 



1892. 

 Sterculia drakei f'ummings, Texas Geol. Survey Third 



Ann. Rept., p. 210, fig. 8, 1892. 

 Knowlton, in Hill, Am. Jour. Sci.. 4th ser., vol. 1, 



p. 213, 1895. 

 Slcrenlia siiovii Lesquereux, Flora of the Dakota group, 



p. 183, pi. 30, fig. 5; pi. 31, figs. 2, 3; pi. 32, figs. 1-4, 



1892. 

 HoUick, U. S. Geol. Survey Mon. 50, p. 94, pi. 34, 



fig. 20, 1907. 

 Aralia tovmeri HoUick, New York Acad. Sci. Trans., 



vol. 16, p. 132, pi. 14, figs. 11, 12, 1897. 

 Berry, New York Bot. Garden Bull., vol. 3, p. 92, 



1903. 



I have long thought that the Aralia towneri 

 and Sterculia snowii of Lesquei-eu.x represented 

 a single species but have never had a chance to 

 test this belief until I received the present 

 collections from the Cheyenne sandstone, 

 in which this is one of the most abundant 

 forms. It shows considerable variation in 

 size but obviously represents a single botanic 

 species. Unfortunately the name towneri ante- 

 dates snowiihy some 15 years, so that the latter, 

 which is much the better known of the two, 

 becomes a synonym. 



From the large amount of mateiial now 

 available the species may be described as 

 follows : 



Leaves of variable and often very large size, 

 palmately two to seven lobed. The lobes are 

 prevailingly conical and acuminate, occa- 



