part 2] GEOLOGICAL HISTOEY OF THE GENUS STEATIOTES. 123 



VI. A Subsidiary Etolutionaey Seeies. 



The species just discussed perhaps illustrate the main trend o£ 

 evolution in the genus Stratiotes ; but, in addition, a branch-line, 

 which has left no descendants, is represented by two species in the 

 Oligocene : S. websteri Zinndorf (Middle Oligocene) and S. thalic- 

 troides Brongniart (Upper Oligocene). S. tvebsteri is peculiar in 

 having a hilum which varies in position from the middle of the 

 dorsal margin to the apex : when apical it is almost always 

 associated with a beak, and this gives rise to a sigmoidal seed ; for 

 other positions of the hilum the beak is absent or feebly developed, 

 its absence giving rise to the common hooked form of seed. The 

 canal of the raphe is also remarkable : when the hilum is mid- 

 dorsal, it is marginal to the apex and then transverse ; for other 

 positions of the hilum it is transverse, either obliquely or directly. 

 It occurs in widespread localities — in the Bovey Tracey lignite of 

 Devon, the Hamstead Beds of the Isle of Wight, and the Middle 

 Oligocene Cyrena Marls of the Mayence Basin (PL V, figs. 7— 

 12, 31 & PL VI, figs. 1-5). 



S. thalictroides in the Upper Oligocene of the Paris Basin is 

 known only from casts, but it shares the same peculiarities of 

 form and structure. It is more elongate than S. websteri, the 

 average relation of length to breadth in S. websteri being only 

 2315 and in S. tlictlictroides 35, while in individuals of this 

 species it rises to 4 (PL V, figs. 13-14 & PL VI, figs. 6-8). 



The contrast between the short transverse raphe and dorsal 

 hilum of S. websteri, and the long raphe and basal hilum of 

 S. aloides, etc. led some investigators to separate the two types 

 generically. On these grounds, Zinndorf gave to S. websteri 

 the new generic name Stratiotites (60). As all other characters 

 of the seed agree with Stratiotes, such a separation seems un- 

 necessary, especially as a simple explanation of the short raphe can 

 be suggested. This depends on the fact that in fossilization the 

 outer layer of the testa — a conspicuous feature of S. aloides — is 

 nearly always destroyed (see p. 118). In S. headonensis, the 

 earliest-known form, the raphe lies on the extreme margin of the 

 woody layer of the testa, but sufficiently within to be preserved. 

 In the succession of species which lead from S. headonensis to 

 S. aloides, the raphe becomes more and more deeply sunk within 

 the woody layer (see pp. 121 & 122). If, now, it had happened 

 that in the offshoot represented by S. websteri the raphe, instead 

 of passing towards the interior, had moved even slightly to the 

 exterior, so as to lie between the median hard and the outer spongy 

 layers of the testa, then, in the process of fossilization, the spongy 

 layer and accompanying raphe would be lost, and the apical part 

 of the raphe (which in its passage to the seed-cavity traverses the 

 median layer) would alone be preserved, resulting in the form 

 seen in S. websteri. 



