LINNEAN SOCIETY OF LONDON. 6 1 



will vanish. The Ranaliau families to the older systeniatists were 

 somewhat of a stumbling block, as their flouers were difficult to 

 harmonise with the formal flower of alternating whorls. To 

 those who embrace the anthostrobilale theory of the flower, these 

 families instead of mystifying supply the key to the whole. 

 Indefiniteness in all parts of the flower is what we are on the 

 watch for and requires no explaining away. 



Engler's system really owed its origin to a praiseworthy ett'ort 

 initiated by Brongtiiart to abolish the Apetalse. The latter re- 

 cognised that most apetalous flowers were reductions from 

 polypetalous types, and should therefore be capable of being 

 intercalated among the Polypetalje. Instead of keeping strictly 

 to this progressive idea and at the same time retaining the Eanales 

 at the commencement of the sequence, Engler and his school 

 diverged on novel lines, postulating the prin)itiveness of the uni- 

 sexual flower without, or with merely a sepaloid, perianth. He 

 passes from such forms to families possessing a uniseriale petaloid 

 perianth, and then to ones with a definite calyx an;l corolla. 

 Superficially the system a])pears so far to run smoothly ; but now 

 comes the \\eak link in the chain — the Ranalian families have to 

 be inserted. They break the progressive nature of the sequence 

 with their indotinite perianth, Without them the gradual 

 perfecting of the biseriate perianth would have followed through 

 the polypetalous faniilies up to the Sympetala?. The Kanales 

 stand inconveniently in the way, just as they did with the old 

 formalists. A linear arrangement of families is, of course, 

 merely a makeshift, but at the same time an unavoidable one. As 

 far as possible it shoidd follow evolutionary lines, and after that 

 its pi-actical value should be consulted. Engler's system in the 

 writer's opinion fails to fulfil the first condition, and granting this 

 there is no reason in retaining it on the second account, for it is 

 no improvement on Bentham and Hooker's arrangement from the 

 practical point of view. It is a matter of regret that certain 

 recent English systematic publications have been arranged on 

 Englerian lines, thus departing from the long continued practice 

 of commencing British floras with the Eanunculaceaj. 



It is interesting to note in the history of taxonomy that each 

 big forward move has usually been accompanied by a step back- 

 wards. Engler's system will prove no exception. The forward 

 move consists in a better grouping of the families into cohorts 

 (orders), and the backward step in allowing the catkin-families to 

 usurp the place previously occupied by the lianales. 



The German school of taxonomy, associated with Tx'eub's 

 classical work on Casiiarina, has had, however, this merit. 

 Attention became focussed on the Amentiferous families with the 

 hope of establishing their primitiveness, ol elucidating the nature 

 of the Anglospermons embryo-sac and of connecting them 

 phyletically with the Gnetales, thus affording a real clue as to the 

 origin of tiowering plants. None of these expectations have been 



