220 THE FLORA OP THE DAKOTA GKOUP. 



It is composed of about twelve divisions or carpels placed around a central 

 axis. Of the numerous figures of Heer (loc. cit.), it essentially resembles 

 Figs. 2c and 10a, being only a little larger. But Heer^ acknowledges as 

 representing tlie same species, fragments of still larger seed than the one I 

 liave figured. He considered it first as a fruit of Diospyros. He compares 

 the fruits to those of Abeibopsis, described in V\. Tert. Helv., vol. 3, PI. 

 cxviii, and also to those of Cucumifes variabilis Bowerb., from the London 

 clay. The relation of these fruits to Nordenskioldia may receive a higher 

 degree of authority from the fact that fine leaves of Abeibopsis have been 

 found in the Dakota Group as well as in an upper stage of the Cretaceous, 

 as described below. 



The identity of the species with that of Heer is not positively ascer- 

 tained, though no appreciable difi^erence is to be remarked. 



Habitat: Kansas. Collection of Mr. R. D. Lacoe. 



Caepites cobdiformis, sp. nov. 

 PI. XXri, Fig. 9, 



Fruit deeply impressed into the stone, cordiform, separated in the 

 middle by a deep, linear furrow, as thougli composed of tA\'o narrowly 

 obovate ovules, which are straiglit and confluent in the middle, rounded 

 above, curved on tlie sides, pointed at the upper end, convex on the surface. 



The fruit is 12'"'" long and 10""" broad in the upjjer part, and is not 

 flattened, but each of the ovules is convex, as if connate in the middle along 

 the narrow line of separation. It seems thus conformed like the seeds of 

 Sapindus, comparable, for example, to S. falcifolius as figured by Heer in 

 Fl. Tert. Helv., vol. 3, PI. cxx, Fig. 8, which is, however, smaller and oval. 

 As Heer remarks, p. 61 (loc. cit), in some species of Sapindus, <S'. saponanus 

 L., for example, the seeds or ovules are united by twos along a thin, linear 

 clasp. Of the simple, detached seeds the author has also figured a number 

 (loc. cit., PI. cxxi, Fig. 2c), some of them rounded on one side, straight or 

 flat on the other, of such a shape that if two of them were connate along 

 the lineal side they would produce a fruit like that described above. 



As the leaves of Sapindus are abundantly found in the Dakota Group, 

 tlie reference of this fruit to that genus seems authorized. 



Habitat: Kansas. No. 4111 of the collection of Mr. R. D. Lacoe. 



'Fl. Fobs. Arct., vol. 1, Fl. xlvii, Fig. 5f; vol. 7, p. 125. 



