PHYLOGENY OF THE ANGIOSPERMS 469 



dence accumulating from the fossil record that the angiosperms had 

 their origin much further back in geological time than has been com- 

 monly assumed, perhaps to the Permian. 



In the search for possible ancestors of the angiosperms, morphologists 

 have been neglecting three aspects of stamen structure that seem of 

 major importance: (1) the laminar form of the microsporophyll; (2) 

 the wall-less, sunken position of the microsporangium; (3) the type of 

 pollination in families with primitive stamens. The laminar stamen is 

 surely primitive, as shown by Degeneria, Himantandra, the Magnolia- 

 ceae and Nymphaeaceae, and by the laminar form of the carpel. The 

 wall-less character of the sporangium — clearly evident in the laminar 

 sporophylls — is concealed, in typical anthers, by the interpretation of 

 the anther-sac walls as sporangium walls. The megasporangium also is 

 wall-less and sunken, if the nucellus is interpreted as an emergence of 

 the lamina, as it seems to be, from comparison of micro- and mega- 

 sporophylls. 



Important in the evolution of angiosperm flowers have been the 

 biological aspects, especially those aspects related to pollination. The 

 history of pollination in the angiosperms has surely played a prominent 

 part in major changes in the structure of the flower and the inflores- 

 cence. The story is one of gradual modification, accompanying the 

 change from pollination by beetles to pollination by higher insects, 

 birds, bats, wind, and to self-pollination. Note the major change from 

 the nectarless Eupomatia and Calijcanthus, with their supply of food 

 bodies for chewing insects and their heavy, penetrating odors — attrac- 

 tive to beetles only — to nectary-bearing, fragrant flowers — attractive to 

 bees — and to nectarless, odorless, wind-pollinated flowers. The develop- 

 ment of nectaries and other early changes parallel the evolution of the 

 insects. Beetles first appear in the fossil record in the Permian; other 

 insects in later periods. 



A line between the earliest angiosperms and the stock from which 

 they arose perhaps cannot be drawn. But from a general survey of what 

 appear to be the more conservative characters in living taxa, some 

 suggestions can be made as to the characters of a possible ancestral 

 stock. 



General body habit, especially anatomy, suggests that angiosperms 

 belong in the fern line of vascular plants. In stelar structure and vascu- 

 lar histology, they are fernlike, rather than cycadlike. Amphiphloic 

 steles are frequent, and the primitive vessel elements are closely like 

 those of Pteridium. Basic nodal structure is tvvo traces to one gap. 

 Laminar placentation suggests the soral distribution of ferns. (The dis- 

 tribution of microsporangia — four, in two pairs, median on the sporo- 

 phyll, not known in ferns — is perhaps a critical character, one that must 



