370 ANNUAL, REPORT SMITHSONIAN INSTITUTION, 1918. 



Cretaceous, so that history fails to shed any light on the problem. 

 Formerly, gyninospernious and other remains from the older rocks 

 were thought to represent monocotyledons, but none such are now 

 recognized. Modern opinion tends in the direction of regarding the 

 distinctions between these two groups as largely cumulative, and that 

 the number of the cotyledons and the correlative characters are of 

 less phylogenetic importance than was formerly believed. The view 

 is here advocated that the angiosperms are strictly monophyletic 

 that the monocotyledons are not primitive, nor did the primitive an- 

 giosperms exhibit combined monocotyledonous and dicotyledonous 

 characters, but that the dicotyledonous type is the more primitive, 

 and that it has given rise to the monocotyledonous stock, not, how- 

 ever, as a single or monophyletic line of evolution, but by one such 

 line combined with several reduction series derived from different 

 regions in the dicot3 T ledonous plexus. Systematists have long rec- 

 ognized the similarities between the monocotyledonous arums, pond 

 weeds, and screw pines on the one hand, and the dicotyledonous pep- 

 pers and willows on the other, as well as the convergence of certain 

 Ranalian groups and the water plantains. The case of the water-lily 

 family is a classic instance of a group of dicotyledonous origin which 

 has essentially reached the monocotyledonous category. 



Analogies between the amphisporangiate " flowers " of the aber- 

 rant Cycadeoideas of the Mesozoic and such dicotyledonous flowers 

 as those of the Magnolia have furnished a basis for a theory of angio- 

 sperm descent which, while fascinating, is believed to be illusory. 

 More definite evidence for a revival of the old view that the angio- 

 sperms are related to the gynmosperms through the anient if erous 

 dicotyledons and some ancient gnetalean type similar to Gnetum has 

 recently been accumulated. This evidence comprises the character 

 of the inflorescence, the floral morphology, the details of sporogene- 

 sis, fertilization and embryogeny, the organization of vessels in the 

 wood, the broad rays, companion cells in the bast, the habit and foli- 

 age, the dicotyledonous embryo, the elimination of archegonia and 

 the organization of eggs. The foregoing considerations are derived 

 entirely from a study of recent forms, since no certainly determined 

 fossil forms of Gnetales are known, and no known fossil angiosperms 

 throw any light upon these features. 



The monocotyledons and dicotyledons may now be briefly charac- 

 terized. In the former many of the forms are herbaceous, which 

 argues for modernity. Anatomically the bundles are closed (am- 

 phivasal), are without cambium, and are scattered through the paren- 

 chymatous ground mass of the central cylinder. Sometimes the 

 bundles indicate a primitively circular arrangement such as obtains 

 in the dicotyledons, and the theory that a tubular central cylinder 

 with foliar gaps was an ancestral condition accords with the view of 



