Half of a fruit, which has split in germination. In all the above characters, except 

 the shape, which is globose not cylindrical, the fossil agrees with Hippuris vulgaris 

 (figs. 22, 23). It also differs in having the mesocarp composed of a greater number of 

 layers of smaller cells, though the resulting thickness of the carpellary wall is the same in 

 both cases. The pericarp of H . vulgaris is composed of a thin layer of large cells with wavy 

 margins. These have been worn away in our fossil but their impressions can be seen. 



HIPPURIS Sp. 2? 

 PI. XIV, figs. 25a, b. 



A second fruit of Hippuris, globose like the last, but smaller, with thicker walls, 

 and with much smaller seed*cavity has been found at Reuver. It may belong to a second 

 species; but without more material we cannot decide whether these differences are 

 specific. 



ARALIACEAE. 



ARALIA RACEMOSA Linn.? 



PI. XIV, fig. 26. 



A complete fruit of Aralia has been found at Swalmen. There can be no doubt 

 as to its generic position, and except for size it is indistinguishable from the Chinese 

 A. racemosa (fig. 27). To this species we refer it with some hesitation, for the specific 

 characters in this genus are not well marked in the fruit. 



GENUS? 

 PI. XIV, figs. 28-30. 



In addition to the Aralia just recorded, we have single specimens of several 

 different fruits which may belong to the same family; but with this imperfect material we 

 can only figure them, in the hope that more specimens may be discovered (figs. 28—30). 



Another specimen which we at first thought to belong also to the Araliaceae 

 is shown in fig. 31. This has 6 bolster=shaped carpels, forming a syncarpous fruit, deeply 

 grooved longitudinally. We have nothing in our collection resembling this ; but among 

 some fruits of Zelkowa Keaki, received from the Yokohama Nursery Company, we 

 found, as an accidental impurity, a single withered fruit identical with our fossil. The 

 recent specimen shows that each carpel is constricted transversely, as if it contained 

 3 or 4 seeds; and the same constrictions can be found in our fossil. The occurrence of 

 more than one seed in each carpel would make the ascription to Araliaceae impossible ; 

 but we are unable to satisfy ourselves as to the internal structure of the badly ^preserved 

 recent specimen, and do not know to what family it can belong. 



124 



