CHAP. .] EVIDENCE, INDUCTIVE 91 



leaves of a very delicate nature and presenting 

 a venation like that of a dicotyledon. If this be 

 a correct description, then our existing duck- 

 weeds or Lemncc are still further degradations. 



My contention is that monocotyledons were 

 descended from very early aquatic dicotyledons. 

 The many reasons for so thinking are based 

 entirely on induction, as it is in but few cases 

 possible to prove it experimentally ; but the 

 number of coincidences in point of structure 

 between the morphological and histological details 

 are so extremely numerous that they supply a 

 very large amount of " cumulative probabilities " 

 leading to a logical " moral conviction " that 

 such has been the line of descent. 



Of course the primary morphological distinc- 

 tion of monocotyledons is the single cotyledon, 

 from which follows naturally the ternary arrange- 

 ment of the floral whorls, through the phyllotaxis 

 (J and J), 1 and the parallel venation and forms 

 of the leaves (really phyllodes). 



The anatomical structure of the stems is the 

 next prominent feature, the woody bundles being 

 separate and scattered, and not in well defined 

 cylinders as in all our trees ; but there are many 

 other minute details, to be alluded to hereafter. 



A significant fact must here be stated. If 

 monocotyledons were descended from aquatic 

 dicotyledons it would be an a priori probability 



1 See " Origin of Floral Structures," p. 28 ; and Trans, Lin. Soc. 

 Sec. Ser. Bot., vol. vii. p. 153 (1907). 



