of large loose cells of the pulp ; mesocarp lost except part of the inner layer, which 

 adheres to the endocarp ; endocarp black shining thin closely wrinkled concentrically, 

 wrinkles crossed longitudinally by the large irregular cells of the mesocarp. 



Length 3 mm. Swalmen. 



We have seen 3 specimens of this berry from Swalmen and a doubtful fragment 

 from Reuver. The berries are crushed so flat, and look so hopeless for determination, 

 that their botanical position was not recognised till it was too late to add an illustration 

 of the recent Ardisia. Though Avdisia is spoken of as having a stone*fruit, the endocarp 

 is often very thin, and our fruit seems to have the endocarp exceptionally thin; but 

 this is not a character of generic value. In the appearance of the style and general 

 character of the epicarp our fossil resembles A. japonica; though in the fossil the 

 epicarp is more thin*walled, coriaceous, and closelysswrinkled. These transverse wrinkles 

 are obviously original and not due to shrinkage of the fruit in drying, for the wrinkles 

 do not distort the large irregular meshes formed by the celkwalls of the mesocarp, 

 and these latter cells tend to be arranged radially, they are not elongate transversely. 



Ardisia is a large genus, now essentially tropical, but having its northern limits 

 in Japan. 



PRIMULACEAE. 



TRIENTALIS? Sp. 



PI. XV, figs. 27a, b. 



Seed irregularly ovate, facetted, ventral face angled at the elongate central 

 hilum ; sculpture of loose irregular hexagonal pits becoming elongate towards the hilum. 



Length 1.7 mm., breadth 1.6 mm. Reuver. 



A single seed, which in size, shape, surface sculpture, and form of the hilum is 

 exceedingly like Trientalis europaea (figs. 28a, b). We have not been able to compare 

 this seed with many species beyond those in our own collection, for in the herbaria 

 most Primulaceae are represented by flowers only. It evidently belongs to one of the 

 larger*seeded Primulaceae. The dorsal facetting shown in the fossil is probably due to 

 shrinkage and crushing, and is not original. 



SAPOTACEAE. 



MIMUSOPS SEPTENTRIONALIS Sp. nov. 

 PI. XV, figs. 29a, b 



Semina late ovata quadratave, basi truncata, lateraliter compressa, plano* 

 convexa, hilo maximo oblongo immerso fissura minuta obliqua angulo exteriore 

 apud raphem fissuraque simili angulo interiore ad extremum contrarium instructo. 



Seed large hard polished, broadly ovate or quadrate truncate below, laterally 

 compressed and plano-convex; testa minutely irregularly granulate; hilum very large 



129 



