212 fit state for examine tion, and ^ would here like to acknowledge 

 niy indebtedness to and to thank those kind friend in South -i'i-frica 

 v'ho heve surrlied me vdfh material, vixJ — ^--r. ^^. J", -^ustin, 

 Professor H. Comnton, Mr,- T, N. Leslie, "^t. ^. --arloth, I-r. 

 ^\ 5. Fillans, Dr. J. B. Pole Evans, Mrs. S. Hood, Mrs. I. 

 V/iggett, and esr^eciallj'' to -r. 5". ?rith for generously placing 

 at my disriosel the flowers ^nd fruits of the fine collection of 

 sT^ecies he brought to exhibit in the South African section of 

 the '''embley "S-diibition in 19^:4, vithout vrhich some of the v/ork 

 to foT.lov' could not have been acconrlished. 



-^ Is certainly remarkable that' di.u?ing the one hundred and 

 seventy-tv.'o years that have elapsed since Linne established the 

 genus J-esembr7?"antheraura no monographer seems to have examined 

 the details of the floral structure of the vrrious species, 

 while the characters presented by the fru.its and seeds have been 

 entirely ignored. 



Up to the time of connencinr- the articles I have already 

 T^ublished, I v.-as under the impression that ^e Clnndolle, ^entham 

 and Hooker, bonder, and Berger, v/hen describing the characters 

 o:^ the genus and species, must have examined the structure of 

 the flov.-ers (especially the ovary) and fruit, and that there 

 was nothing more to be discovered concerning that structure. 

 I vres therefore much surr^rised to find hovr greatly various groups 

 of srecies differed in the placentstion and structure of the 

 ovary and fruit, apart from other differences in the structure 

 of the fla-'er, such differences being usually accomxpanied by a. 

 consT^icuous difference in ha^bit. I believe that this is the 

 only large genus that has endured for so long a time as Linne 

 left it. If it had belonp'ed any other author it is almost cer- 

 tain that it would have been divided into several genera long 

 ago, for there are eharacters for their separation in plenty when 

 looked for. Having realised this, I proceeded to examine the 

 flo-"'ers and fruit of all the species nlpGed under the genus so 

 far as material has been available. This has occu^-ied a very 

 srre-'t amount of time and much remains to be done, because I have 

 been unable -to obtain from South •'-frica the necessary flov/ers 

 and fruit of som.e of the sectional types for examination. So 

 thpt this must not be looked upon as a revision of the whole 

 grour, but merel3'- as a preliminary effort to divide it into nc.t- 

 ural genera, and to define them structural Ij'', so that the proper 

 r>osit?on of nev* or imperfectly'- known species may hereafter be 

 more easily and more accurately determined. 



The result of my examination demonstrates that ^-esembryan- 

 themum as at present understood consists of tv o well-defined end 

 easily recognised primary groups, viz.? — 



1, Species vdth the ovules and seds borne on the central 



axis of the ovrry end fruit, that is, with axile placen- 

 tation. 



2. Species with the ovules and seeds borne along a central 



line on the outer wall or floor of the cells of the 

 ovary and fruit, a rather peculiar form of parietal 

 placentation. 

 "^ach of these groups can be further sub-divided into genera by a 

 combination of some of the structural differences found in their 

 flo-t'^ers and fruits, which are as follow^ — 



(a) 'The ofl.-^n. is lobed dov^-n to its union with the ovrry, 



or is produced above that union into a shallow cud or a 



