CONSTANCE: SYSTEMATICS OF THE ANGIOSPERMS 421 



1907, 1908; Hallier; Berridge, 1914; Engler; Fisher, 1928; Boothroyd, 1930; 

 Hallock, 1930; Abbe, 1935; Mez, 1936; Manning, 1938, 1940; Abbe and Earle, 

 1940; Gunderson; Langdon, 1947; Corner, 1949; Steb])ins, 1950). Both Parkin 

 (1914) and Engler recognized distinct "cymose" (determinate) and "racemose" 

 (indeterminate) types of flower arrangement. Engler thought both could be 

 derived from a primitive panicle, and Parkin suggested the racemose might have 

 been produced from the cymose by way of a panicle. In Gramineae, Ziegenspeck 

 (1938) advocated the derivation of both the spike and the superfinely branched 

 panicle from a basic panicle, itself perhaps a descendant of the Commelinaceous 

 cincinnus. 



In the best available review of inflorescences, Rickett takes a sharply critical 

 view of his subject, finding it impossible to draw clear and meaningful distinc- 

 tions between "cymose" and "racemose," "simple" and "compound," "axillary" 

 and "terminal" categories, and suggesting that "inflorescences are rarely as 

 simple as they seem" (1914, p. 211). He thinks the simple dichasium may afford 

 at least a starting point for the investigation of inflorescences, and believes that 

 the reduction of individual dichasia and the grouping and shortening of leafy 

 branches bearing them terminally, with the concomitant reduction of leaves to 

 bracts, may afford an explanation for various existing complex arrangements. 

 He concludes that it may be "idle even to speculate on the origin of inflorescences, 

 since we know so little of the relationships of the families of flowering plants, 

 and since by reduction in number of flowers and condensation of branches the 

 same patterns may be attained from different beginnings" (p. 211). This idea 

 of polyphylesis of inflorescences was applied to the "raceme" by Parkin (1914). 

 Woodson, also, has cautioned us (1935, p. 35) : 



It would appear a fruitless task to search for the earliest indication of the inflorescence 

 among the extant flowering plants: the origin of the inflorescence is at least as remote 

 as the origin of the flower, and a greater antiquity seems probable from the evidence of 

 paleobotany. 



On a more practical level, the separation of Amaryllidaceae from Liliaceae 

 by Hutchinson (1934, 1935) is a classic instance of the utilization of inflorescence 

 characters. Philipson conducted a series of investigations on the capitula or 

 similar structures of Compositae (1946, 1948a), Dipsacaceae (1947a), Valeriana- 

 ceae (1947b), and Campanulaceae (1948b). On this basis he suggested the 

 fundamental similarity of the Campanulaceous and Composite types, and their 

 basic differences from those of the other two families; that of Dipsacaceae, he 

 thinks, could have been derived from a Caprifoliaceous type. 



Floral Morphology and Anatomy 



Inasmuch as the classification of angiosperms has, for the past two hundred 

 years, emphasized floral structure, sometimes almost to the exclusion of other 

 features of the plant, it is not surprising that the literature on this subject is 

 appallingly voluminous. Attention will be given here only to those aspects of 

 flowers and flowering wliich appear to have a fairly direct bearing on problems 

 of classification and phylogeny. 



Floral constitution, as it has generally been employed in classification, has 

 been based on the classical morphological interpretation credited to Wolff, Lin- 



