212 short or lon^- tube, end before dividinf^ into lobes. 



(b) The retrls are free almost or quite to the bese, or 

 ere united belov into p short or lone: tube, end are 

 inserted in the anp-le vhere the cf^lyx unites with the 

 ovr-rv, or on +>ie sides ot- -t the apex of the celyx- 

 tube. 



(c) "^he steraens are erect, or ^re incurved over or bent 

 dovm into the cur or tube of the calyx. 



(d) A true non-sti£?matic style is present or it is absent, 

 end the stigmas (the so-called styles of books) are 

 stigma tic to the base. 



(e) The stiPTia is one only and sessile and circular or 

 elliptic in outline, or the stigmas (v/hether a style 

 is riresent or not) are 4-20 and vary in form. 



(f) The ovrr^'' is inferior, or it is partly or more than 

 half superior. 



(f^) The fruit is Indehiscent, or it opens by valves, 

 (h) The fruit is one-celled, or it is 4-20-celled. 

 (i) The valves of the fruit (capsule) are provided vdth ex- 

 panding keels, or those orrsns are modified into non- 

 exT?nding vdnps. 

 (j) The cells of the capsule are roofed v/ith cell-v/ings, 



or they are open and without cell-wings. 

 (k) The opening of the cells of the capsule is nearly 



closed by a tubercule, or there is no tubercule at the 

 openin--^. 

 (l) The seeds are vdngless, or are broadly winged. 

 (m) The seeds are regularly arranged in a single )A*iorl 



around the centrrl axis, or they ere irregularly placed 

 in the cells, 

 (n) There is also the very important general character and 

 nature of the leaves and i-hether they ere alternate or 

 OP'^osite, or dotted, or are v/ithout dots. 

 Of these differences only three or four have been even men- 

 tioned by authors end have apparently been over-looked. Yet in 

 no other order of relents do v'e find that such a multitude of 

 perfectly distinct characters are allowed to remain in combina- 

 tion in the same genus. I therefore rropose to utilise these 

 cherocters in vpricus combinations for the purpose of dividing 

 this hitherto unv:iedy heterogeneous group into smaller and more 

 uni'Porm e'enere , so thpit speces csn be more easily referred to 

 their r>roper position, ^or in the monographs as the:/ at present 

 exist, I find the arrangement so chaotic that totally different 

 t"','Tes of structure ere mixed together in the same group, ^nd 

 at the •^resent tine species are being described and referred to 

 this or that group or supposed affinity in a most hapharzerd 

 manner, or are described without any indication of affinity, ren- 

 dering it almo."t imrossible to identify the species v/ithout see- 

 ing the- type specimen. And ^ myself must plend tributed to the 

 ch.^os. This, however, has been unaboidable, as there is no 

 meens of discovering from, books the true affinity of many of the 

 species. Added to this there is the i"urther difficulty of knov/- 

 inp" to v'hpt p-enus some of the srecies described by Alton and 

 Hav.'orth pointed should be referred, ^or, as ^ have elsewhere 

 pointed out, the tv'o most recent monographs of i^-esembryanthemum — 

 thpt of Sender and of Berger — are compilstions ^^'ithout any orig- 

 • inal research work embodied in thegi, chiefly based upon the ex- 

 cellent figures and descriptions in Salm -^yck's work, which, 



