211 fruits and seeds of these plants, ivith the result that I soon 

 perceived that diirin^ the one hundred and seventy-two ye-^rs that 

 has Passed since the genus I'-esemhrYanthemum v/as established by 

 Linne, no author (including myself J v.Titing about this group 



of plants appears to have properlj'' dissected the flowers and 

 exanined the relationship of their parts as compared Vidth those 

 o"^ othpr species, v/hile the structure of the fruits and nature 

 of the seeds has been utterly ignored. aH of us have follo\;ed 

 one another blindly b'^ accepting the generic characters as given 

 in boolcs T.'ithout investigating their applicability as a whole to 

 the vprious species. 



The discovery I allude to above vas that out of seven speci- 

 mens received from South --Africa, all named I-esembryanthemum pom- 

 eridienum, tv/o has fruits and seeds which I 



212 noticed were so utterl^r different from those of that well-known 

 r>l<^nt th? t thej" coul-^ not possibly belong to the same genus. 



1 then examined the flowers of the seven specimens and found them 

 to represent two distinct t^'-pes of structure. Certainly the 

 flowers and fruits and seeds of these seven specimens could not 

 have been examined or they v/ould not all have been deemed to be 

 the same species, although they somewhat resemble one another, 

 in general appearance. 



■^.is discovery caused me to malse a preliminary examination 

 o:^ several other species that were dist net from one another in 

 their type of habit, and mj'" investi'-- tion soon convinced me 

 that it is eas3'' to demonstrate that, in accordance with the manner 

 in which genera are made and characterised in other natural 

 orders, the genus J-esembrvrnthemum as at present understood in 

 books represents a sm8.11 ^^atural '^rder and unquestionably con- 

 sists of a num-ber of perfectly distinct and usually easily rec- 

 ornisable genera, which have all been assigned to the one genus 

 . merely because their flowers have a superficial resemblance to 

 one another. 



As I have already pointed out in these articles, Haworth 

 considered that some of the sections into v/hich he divided the 

 f^enus were really distinct genera, although he did not character- 

 ise them by structural details, and with fev/ exceptions he did 

 not propose names for them, and such names as he did propose have 

 not been taken up by any author before I did so in- these articles. 

 I ^ird, hov-ever, that Hav/orth v/as perfectly correct in his viev;s. 

 "^et, although his sections are accepted, the differences assigned 

 to them are based almost entirely upon habit and vegetative 

 characters, the important structural characters that accompany 

 many o-f* the vegetative characters being overlooked or ignored. 

 If, ho"-ever, we turn to ^rders \7here a ler^e number or most of 

 the genera composing them, have flowers that ere very much alike 

 in General appearance, such as are found in Gonvolvulaceae, ^el- 

 vaceae, G!-r.uG3 ferae, Umi^elliferae, eguminosae and Gompositae, 

 vre find that the genera of these orders are based upon structural 

 differences in the flov^er or fruit that are constant, even if 

 small and often not observable without dissection, but ere often 

 ecGompanied by a difference in habit or vegetative character. 

 l^To'- I find that Uesembryanthemum as hitherto constituted 

 is an exact parallel of such orders, for although the flovicers 

 and even the fruit in some cases of the various sections have much 

 superficial resemblance to one another, yet I'hen examined, are 

 in many instances found to have very different structural char- 

 acters. But - have had considerable di:"fieulty in obtaining flov;- 

 ers and fruit — especiallly fruit and seeds — of many kinds in a 



