20 



DIAGNOSIS AND CLASSIFICATION. 



OKTnoRRnYNCELE : corolla tube much shorter ; filaments 



ted at or below 



mi 



of tube; beak in bud straight or flexuous, with ventral 



ddle 

 pposed to median line of lip and apex 



pect 



directed to edge of midlobe 



Similarly the ADUNC-E consist of two natural sections : 



Rhyncholoph.e ; corolla lip sessile or stipitate with usually rudimentary crests- 



filaments inserted at or below middle of tube; beak in bud falcate, with 

 apex apposed to median line of lip; hood sometimes beakless. 



§§§§ Bidentat^e; corolla lip stipitate, base erect, 2-cristate above; filaments inserted 



opposite top of ovary; hood beakless or shortly wide-beaked, always 

 2-dentate immediately below the apex. 



The EROSTRES constitute a single large section containing the whole of the forms 

 with corollas showing what appears to be the most archaic type of structure: 



Anodont^e; filaments inserted opposite top of ovary; hood erect, beakless or 



with very short rudimentary beak, edentulous or only toothed at the lower 



gle and with ap 



alway 



roun 



ded 



S 



The natural disposition of these 

 onantble and Orthorryncele are 



sections is not one of sequence 



S 



far as 



ed, the evidence of this is direct and 



abundant. Conditions intermediate between the types of beak aestivation in these two 



not only do not exist, but are hardly 



eivable, and they cannot have well descended 



the one from the other. On the other hand, the type of beak aestivation in Rhyncholoph^e 

 is precisely that which is capable of being modified in either of the directions assumed 

 in the other two types, and as a matter of fact the Rhyncholopioe include groups of 



species that 

 to S 



dicat 



of such transition 



Thus the OUganthce are clearly tr 



the Myriophyllce to Orthorrhynchje 



I 



the 



same 



way 



wh 



there 



transition to Bidentat 



d the 



is abundant evidence among the Anodont^e of 



as examples 

 verwhelming 



P. Langsdorfii, P. collata, P. fragilis may be quoted 



dence of transition from Anondontje to Rhyncholoph^: 



witness the groups Carnosce, Flagellares, Tristes. Recutitoe, and others 



th 



is no 



dence of transition from Bidentat^e to Rhyncholopbl 



i 







good 



the other hand, there 



is 



not the positive structural evidence which is to be obtained regarding Siphonanth.e 

 and Okthorrhynchjs that such transition is impossible. In the light, however, of the 

 great body of direct evidence of transition from Anodontjs both to Rhyncholoph^; 



d to I) 



and of the abse 



of 



dence of any direct connection bet 



Rhyncholoph,e and Bidentat^, we seem justified in assuming that both Rhyncholoph.e and 

 BidentatjE are derived immediately from Anodont^e. 



In 



accord to the Bidentat^: 



any 



we 



not entitled to 



position intermediate between Anodont^e and Rhyncholoph.e 



The chief evidence that has been advanced is the connection between P. cabulica Benth., (placed next to P. uncinata) 



and I , 



condition of P. dolichorrhiza. 



P 



i 



below 



Ibe conclusion to which the consideration of evidence from distribution leads is that the Bidentat* are of more 

 ,ngin ban the Rhyncholoph* are. And if this conclusion be justiEed (the evidence on which it rests is to be found 

 p. 56 ), it furnishes the direct proof necessary for the establishment of the hypothesis here advanced. 



