186 C. H. Ostenfeld: 



Posidonia and Cymodocen have migrated through the Straits of 

 Gibraltar out along the nearest Atlantic coast, Posidonia going 

 northwards along the Iberian peninsula as far as the head of the 

 Bay of Biscay, and Cymodocea northwards to Cadiz and south- 

 wards alongl the African coast to Senegambia and the Canaries. 

 These interesting extensions of distribution are slight, and the two 

 species are essentially Mediterranean in type. 



As explained above, Posidonia has its only allied species in Aus- 

 tralian Avaters. C ymodocea nodosa is nearest allied to C. rotun- 

 data. one of the Indopacific species. Therefore it seems justifiable 

 to consider the Mediterranean species as derived from the Indo- 

 pacific one; or, perhaps more correctly, to derive both from an 

 Indopacific ancestor. The main point, however, is that both Poii- 

 donia and C ymodocea. must be supposed to have come to the Medi- 

 terranean from the Indopacific region, and that this happened at 

 a comparative early time, since the two Mediterranean species have 

 developed so far along their own line of evolution,- and are now 

 specifically well characterised. 



On the other hand, Zostera marina and Z. nana came into the 

 Mediterranean from the north and west, through the Straits of 

 Gibraltar. 



VII. The North-Pacific group consists of the genus Phyllo- 

 spadix. with its two very closely allied species — viz. : Phyllospadiz; 

 Sroulcri and P. Torreyi. 



Tlie genus Pliylloxpadi.r lias arisen from Zostera. It differs in 

 being dioecious and having a short, contracted rhizome (not the 

 long, straggling rhizome of Zostera). The two species are so close 

 togt'thcr. that an American lx)tanist, W. K. Dudley, who has 

 studied tliem in situ. doul>ts if they are really distinct. The one {P. 

 Scouleri) is very varial)le. while the other (P. Torreyi) is not. and 

 tlie amplitude of the variatitjn of the first species includes that of 

 tlie second one. Botli species inhabit the same part of the west 

 coast of Xo)-th America, from southern California to British 

 Colunihia ; and tlie variable (probably older) species occurs in 

 Japan, also making it probable that it, too, may be found in the 

 intermediate i-egion, along the shores of the Aleutian islands and 

 the south coast of Alaska. 



The genus l)eing resti-icted to the northern Pacific, and derived 

 from Zostera, lias })robably originated not in the tropics, but in 

 its present liome, and at a comparatively late time. 



Phyllospa/7ir serrulatu.s Rupr., of doubtful standing, came from 

 Alaska. 



