1 





r 





i 



t ^ 



i 



t 



1 



f 



^ ' 





t 



A revision of the genus Milella etc. 377 



In attempting to rehabilitate the old genus as Hooker, Gray and 

 Engler progressively adopted it, and in the hope of avoiding as far as 

 possible the common causes of error referred to above, the present work 

 is based in the first place upon the study of extensive material from the 

 whole geographical area of the genus and, in the second place, upon the 

 proposition that a single morphological character is ipsufficient ground 

 upon which to base generic or even sectional concepts. 



In seeking for a better basis upon which to lay down the lines of 

 genetic development in the genus, it was hoped that anatomical structures 

 of value in classification would be found. Nothing of importance, however, 

 has been revealed aside from bringing to light the close relationship existing 

 between the species M. nuda and M. caulesceiis. — It should, however, 

 be added that herbarium material proves unsatisfactory for close anatomical 

 work and that ample fresh material might give additional results. — In 

 a similar manner it has been shown that purely vegetative characters are 

 wholly inadequate and of value only in differentiating the species. 



It is therefore clear that the floral characters are the only ones that 

 can be relied upon. A comparative morphological study of the various 

 organs of the flowers of all the species reveals the fact that in the arti- 

 ficial and arbitrary division of the genus based upon position and number 

 of stamens, morphological structures of far greater importance for classi- 

 fication purposes have been ignored. Briefly stated, these are found to 

 be: (a) shape of the floral axis, (b) form and venation of the sepals, (c) 

 structure and division of the petals and (dj form of the anthers. 



An examination of figures i— 8, following the systematic part of this 

 work, will show at once a remarkable agreement in the ground-plan of 

 these structures which seems quite significant, and one that cannot be 

 ignored in dealing with affinities. It is on the basis of a combination of 

 these characters that we present the following new alignment of species 

 and reduction of the number of sections. Their merit will be argued in 

 the discussion on relationship in Part III. 



Mitella 



Miti 



Dminimojidia DC. Prod. \\. 



49. 1830. — Pectiantia Ilafin. Fl. Tell. II. 72. 1836. — Oxomelis Rafin. II. 

 73. 1836. — Mitdlopsis Meisner, PI. Vase. Gen. 100. 1838. — Mitelalsim 



(T et G) Howell, Fl. NW, Am. 201. 1898. 

 Axis cup-shaped, saucer-shaped, campanulate or turbinate, more or 

 •ess united with the ovary. Sepals 5, oblong, ovate or triangular, valvate 

 "^ the bud. Petals 5, pinnately cleft, pinnately divided, trifid or entire 



<) The name is derived from the Greek word [xtTpa, meaning a headband or a 

 turban, it was apphed by Tournefort on account of the resemblance of the dehiscing 

 capsule to the mitre or bishops cap. 



.- /- 



V 



V£ 



