THE SUDDEN ORIGIN OF NEW TYPES 399 



Arber^) foreshadowing the hermaphrodite flowers of Angio- 

 sperms. It will be evident from the principle already enun- 

 ciated above that the very fact of a repetition of protecting 

 bracts (or primitive perianth) as well as a repetition of the male 

 and female elements (microsporophylls and megasporophylls) 

 must have induced a high degree of variability in all these 

 structures and in the resultant strobilus as a whole. In this 

 connection it is of importance to note that the American 

 Benncititccc, according to Wieland's researches, show a remark- 

 able variety both of form and structure. 



A most significant result of his observations is that the male 

 and female elements are not matured at the same time. This 

 protandry would appear, in Wieland's opinion, to indicate (as 

 it does in the Angiosperms of the present day) the dependence 

 of these primitive flowers on insect-visits for fertilisation. The 

 abundance of pollen or microspores produced by the pinnate 

 microsporophylls renders it likely, as Arber {op. cit.) suggests, 

 that pollen and not honey was the original attraction of these 

 early flowers to insects. Dr. Scott- has in this connection 

 remarked on the bright colours shown by living species of 

 Cycas and Encephalartos ; and it seems therefore not unreason- 

 able to infer that these primitive flowers also displayed colours 

 and laid themselves out for insect-visits. This entomophily 

 must have been a most potent factor in the action of natural 

 selection upon a plastic and variable group. The presence 

 of protandry in itself implies a long-established succession of 

 insect-visits to the ancestral Bcnucttitcce, for it is a well- 

 known law^ that the primary eff'ect of intercrossing is to 

 enhance the size of the corolla, to give a preponderance 

 to the andrcecium, and to cause protandry by checking the 

 growth of the gynoecium. Now since this is already the state 

 of things in the Bennettitece (excepting only that the numerous 

 floral leaves are long rather than broad), it would follow as 

 a matter of course that the agency of insect-visits must have 

 been at work upon the ancestral forms for a considerable time 

 previously. 



Perhaps it might be travelling too far into the region of 



> Op. cit. p. 58. 



- Jouni. Roy. Microsc. Soc. April 1907, p. 139. 



' Henslow, Rev. G., The Origin of Floral Structures., etc. p. 20. London, 

 1S88. 



