16 



DIAGNOSIS AND CLASSIFICATION 



the term is to be applied in the most provisional sense. At the same time it is very 

 evident from the morphological account of the genus how few m number the reliable 



diagnostic characters really are. 



In the present paper use has only been made of such characters as appear to 

 be absolutely constant within the limits of the genus. In using calycine characters for 

 example a calyx of which the limbus is cleft has been contrasted with one of which the 

 limbus is not cleft at all; no use is made of the extent of the cleavage. And in using 

 corolline characters a tube dilated towards the apex is contrasted with a tube not there 

 dilated, the amount of dilatation being considered of altogether subsidiary value. Then 



J elongation of the tube, as indicated by the relative length of calyx 



rnilcorollaTTrnot used in diagnosis, the portion of tube which elongates, as indi- 

 cated by the position of the staminal insertion, is employed. Only a subsidiary value is 



ned to the relative length of the lip when compared with the hood, while characters 

 derived from its outline (whether with lobes truncate or entire) and from its margin 

 (whether ciliatc or glabrous) have been used as diagnostic. Similarly a high diagnostic 

 value is assigned to the fact that a beak is present or is absent; also when the beak is 



the extent 



assiir 



present to the fact that it is lodged in the bud with its apex towards the edge of 

 the lip or towards the throat of corolla, and to the fact that its apex is entire or not 



entire while only a subsidiary value is assigned to the mere length of beak or to the 



extent of its bifurcation when it is bifid. Finally, great value is attached to the condi 

 tion of the filaments whether hirsute or glabrous (though when they are hirsute little 



ht is attached to the mere degree of pubescence), as well as to the surface inark- 

 f the seeds, which are, as Lange has pointed out, absolutely constant. 1 The 



wei & 

 ings 



fact of constancy, then, is tbe key to the value of the characters selected ; they are 

 such as appear to be constant not merely for individual species, but for the whole genus. 



One consequence of rigid adherence to the principle now explained is that certain 



forms usually accepted as specifically distinct are here accorded only varietal rank or 



are altogether reduced. Thus P. cabulica Bth. is joined to P. dolichorrhisa Schrenk; 



P, Semenowi Kegel to P. pycnantha Boiss. ; P. ochroleuca Duthie to P. maorantha 



Klotzsch: P. hbellata Jacquem. to P. rhinanthoidcs Schrenk. It does not follow because 



this has been done that the characters whereby these forms are separable are not 



constant so far as the two particular forms are concerned. The reduction indicates 



only that the characters on which it is necessary to rely in order to effect the diagnosis 



in these special cases are characters that are clearly not constant in certain other 



species of the genus. On the other hand, specific rank has in a few cases been accorded 



to forms already recognised as distinct, but hitherto only treated as varieties. Thus 



P. asplcniifolia var. albiflwa Hook. f. is here treated as a species apart from 



P. aspleni/olia Floerke, and P. furfuracea var. integrifolia Hook. f. as a species apart 



from P. furfuracea Wall., because in these instances the characters by which the forms 



are separable are characters which are constant in the genus as a whole. 



To this treatment of forms of the latter class exception will probably not be taken 

 since it is already certain that the forms are truly distinct. With forms of the first class 

 .it is altogether different, since the treatment here proposed for them involves a 

 fundamental difference of opinion as regards their rank. The reductions involved by the 



application of the canon have therefore only been effected in those species that actually 



within the area dealt with in this paper, because it seems improper to propose such 



" '"' " '■ ■ "■■ — I H I, ii,—— ■— — — ■ I — ■I ll !■ .M - ' ■ !■ W1 ■ | | ■ __ | , - | f^M !■■ . I ■ ^J M__MI_U_^M_I 



1 Lange, Kiaenkou, Kjobn. Bot. Tid$k\ iv, 24,7, et seq. 





