420 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



nary studies of Magnoliales by Eao (1939) and Bondeson (1952) indicate the 

 existence of a diversity of stomatal types in woody Ranales, which may prove 

 of phyletic interest. However, as remarked by Bailey and Nast (1945b, p. 149) : 



Not until the stomata of a wide range of the Ranales and other orders have been care- 

 fully reinvestigated will it be possible to assess the phylogenetic significance of different 

 stomatal structures in discussions regarding the origin and the relationships of the 

 dicotyledons. 



The recent studies by Foster (1944, 1945, 1946) have focussed attention on 

 foliar sclereids as a source of systematic information. A form-classification of 

 these structures has been undertaken by T. A. Rao (1951), w^ho recognizes four 

 principal types based on ontogeny and subdivides them with respect to size and 

 shape. However, as remarked by Foster, "it is hazardous or indeed impossible in 

 many instances to generalize with respect to the major trends in morphological 

 specialization of foliar sclereids within systematic units" (1946, p. 253). Model 

 examples of the successful systematic employment of such foliar characters as 

 shape of midrib xylem, stomata and stomatal crypts, structure of hypodermis, 

 shape of terminal sclereids, and distribution of free sclereids, together with evi- 

 dence from floral anatomy, are provided by the studies of Morley (1953a, 1953b) 

 on Melastomaceae. 



Of a miscellaneous character are the attempts to connect Curcurbitaceae with 

 Passifloraceae by the parallelism of their tendril structures (Hagerup, 1930), 

 Cactaceae with Portulacaceae by similarity in emergences (Chorinsky, 1931), 

 and Julianiaceae to Juglandaceae by their mutual possession of resin canals 

 (Stern, 1952). Although the nature of trichomes is taxonomically useful within 

 a limited frame of reference and some family distinctions are to be noted, even 

 among the monocotyledons ( Staudermann, 1924), Heintzelman and Howard as- 

 sert that at least in Icacinaceae "no broad phylogenetic lines of specialization 

 can be drawn from the study of pubescence" (1948, p. 51). Chattaway (1937) 

 stated that in Sterculiaceae the kind and distribution of crystals in vegetative 

 parts "appear to have little phylogenetic significance" (1937, p. 363). 



Inflorescence 



The primitive condition of the angiosperm inflorescence has been variously 

 postulated as having been a solitary flower, usually terminal to a leafy shoot 

 (Arber and Parkin, Ilallier; Parkin, 1914, 1923; Hutchinson, Sprague, Gunder- 

 son), some kind of a panicle (Engler; Zimmerman, 1935; Rickett, 1944), or a 

 dichasial cyme (Woodson, 1936). Croizat (1943) emphatically doubts that 

 either the solitary terminal flower or the cymose raceme is to be regarded as the 

 genesis of all inflorescences. Similarly, both Wettstein and AVoodson suggested 

 the solitary-flowered inflorescences are usually derived from pluriflorous ones 

 through reduction, a view expressed by Bailey and his associates in regard to 

 Winteraceae (Nast, 1944; Bailey and Nast, 1945a) and Degeneriaceae (Bailey 

 and Smith, 1942) and by Melchior (1932) Avith respect to the genus Viola. Most 

 authors seem to have agreed that aments, capitula, spikes, umbels, and other 

 compressed or highly branched inflorescences with small flowers are specializa- 

 tions, often showing a combination of aggregation, reduction, and condensation 

 from either angiospermous or gymnospermous ancestors (Arber and Parkin, 



