20 SCIENTIFIC WOKK, 1860-1865 



But his dictum gave no clue to the principle on which this 

 grouping should be made. What was the true test of affinity ? 

 It was something more than the casual resemblances which in 

 early days led to the division of plants into trees, shrubs, and 

 herbs. 



As knowledge advanced, Linnaeus, with his gift of lucid 

 discrimination and concise terminology, was able to mark off 

 species clearly by their structure and group them. Here was 

 a firm step for future advance to a natural system ; but that 

 advance was stopped by the very success of his non-natural 

 identification system. 



He applied the strict principle of formal logic, whereby a 

 species is defined as possessing the attributes common to a 

 wider class (genus) together with the attributes peculiar to 

 itself. Hence his scheme of the two names, one generic, the 

 other specific, which labelled and ' placed ' a species, the 

 1 barbarous binomials ' of a later sneer, which ignored Lin- 

 naeus' logical mind and the orderly basis he laid for future 

 workers. 



De Jussieu revived and filled out a conception which had 

 already been partly applied. Nature had given plants as 

 they germinated, either one seed-leaf or two or none. Here, 

 then, were the three primary groups of De Jussieu' s system 

 (1789), monocotyledons and dicotyledons (together the flower- 

 ing plants) and acotyledons (the cryptogams, mainly). He com- 

 pleted his subsidiary grouping by dividing the flowering plants 

 into fifteen classes, somewhat artificially arranged, and these 

 again into 100 natural orders, each made up of a group of genera 

 with characters in common. 



This system De Candolle recast. De Jussieu's classes were 

 scarcely satisfactory ; the addition of whole new floras, such as 

 those of the Cape and Australia, meant much reorganisation. 

 The great virtue of De Candolle's system was that, in the main, 

 it was established on a morphological basis. True that he 

 employed physiological characters as well for some of his 

 definitions, but he recognised the comparatively small value of 

 these in classification, unlike Lindley, none of whose classifi- 

 catory schemes held good, for the most diverse plants may show 



