154 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



S. IXAEQTJALIFOLIA, Spr. 



S. TVallichii, Spr. 

 S. Wildenowii, Bak. 

 S. canaliculata, Bak. 

 S. Mettenii, A.Br. 

 S. Lobbii, Moore. 

 S. gracilis, Moore. 

 S. viridangula, Spr. 

 S. Cliilensis, Spr. 

 S. Victorias, Moore. 



ii. Geniculate. 



III. HoMOSTACHTS. 



IV. HETEROSTACHYt 



i. Bisnlcatas. 

 ii. Proniflorse. 



i. Flabellatge. S. caulescens, Spr. 

 S. Bra.unii, Bak. 

 S. Griffithii, Spr. 

 S. grandis, Moore. 

 S. flabellata, Spr. 

 S. Vogelii, Spr. 

 S. laevigata, Bak., var. 

 Lyallii, Spr. 

 S. viticulosa, Klotz. 

 S. erythropus, Spr. 

 £. haematodes, Spr. 

 (none examined). 

 (2 species, neither exam- 

 ;. ined). 



£'. bisulcata, Spr. 

 (none examined), 

 iii. Brachystackyte (none examined), 

 iv. Suberoste. S. suberosa, Spr. 



S. mollicejjs, Spr. 



S. Karsteniana, A.Br. 



S. stenophylla, A.Br. 



The above columns will show at a glance how difficult 

 it is to bring into accord a grouping based on external 

 niorphology and one based on anatomical structure. Of 

 course I do not aim at present at doing more than calling 

 attention to the fact that taking into account the anatomy 

 of the stem only we are led to group together species 

 widely separated by systematists and conversely to separate 

 species which agree strongly in external features. Even 

 in such a group as the Sarmentosce section of the subgenus 

 Stachygynandrum which so nearly agrees with the 

 anatomical section of which S. incequalifolia is the 

 type, we find S. Mettenii widely separated from that type, 

 and S. Wildenowii refused a place. A more prominent 

 case is however that of S. Icevigata, var. Lyallii, with, as 

 I have shown a perfectly distinct anatomical structure, 

 which stands side by side with species like S. Vogelii and 

 grandis, forms not to be separated from S. Martensii. 



