424 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



ture and often of improving classification thereby, viz., Ranales (Smith, 1926; 

 Rassner, 1931; Schoffel, 1932; Bronland, 1935; Cliapman, 1936; Reece, 1939; Ka- 

 sapligil; Tepfer), Rhoedales (Eggers; Norris, 1941; Stoudt, 1941), Malvales 

 (C. V. Rao, 1952), Geraniales (Moore, 1936b), Caryophyllales (Laiibengayer, 

 1937; Mattfeld, 1938a, 1938b; Thomson, 1942), Ericales (Copeland, 1935b, 1937, 

 1938, 1939, 1941, 1943, 1947; Falser, 1951; Chou, 1952), Primulales (Dickson, 

 1936; Douglas, 1936), Gentianales (Lindsey, 1938, 1940; Woodson and Moore, 

 1938), Polemoniales (Copeland, 1935a; Dawson, 1936; Lawrence, 1937), Rosales 

 (Jackson, 1934; Moore, 1936a), Rubiales (Wilkinson, 1949), and Asterales 

 (Koch, 1930), and such distinctive families as Proteaceae (Kausik 1938a, 1938b) 

 and Thymelaeaceae (Leandri, 1930; Heinig, 1951). 



A. Perianth: A wide variety of explanations has been offered as to the origin 

 and primitive constitution of the perianth. That a sterile perianth, well differen- 

 tiated from both foliage leaves and sporophylls, was present in the first angio- 

 sperms was postulated by Arber and Parkin, Wernham (1911-1912), and Bes- 

 sey. It has been suggested that the dicot perianth was initially divisible into 

 calyx and corolla (Hutchinson; Stebbins, 1951), and that it was not (Hallier, 

 Bessey, Gunderson). Gliick (1919), following Prantl, supposed the whole peri- 

 anth to be derived from foliar bracts, while Worsdell (1903, 1907) metamor- 

 phosed the perianth in toto from the androecium. More popular, however, has 

 been the alternative of obtaining sepals through modification of bracteoles, and 

 petals by sterilization of stamens, wdth which they usually agree in possessing a 

 single trace (Rendle, 1903; Engler; Smith, 1926; Troll, 1939a; Eames, Wett- 

 stein). To Sprague, "the hypothesis of a single primitive type of perianth seems 

 superfluous" (1925, p. 113), a view in which he has been joined by Arber and 

 Parkin, Mattfeld, Ehrenbergh (1945) and, I suspect, many modern workers. 

 Much has been made of the fact that Alismatales possess a heterogeneous floral 

 envelope like that of most dicots, whereas that of most other monocots is homo- 

 geneous and hence allegedly derived wholly from androecium (Nicotra, 1909- 

 1910; Salisbury, Hutchinson; Eber, 1934; Markgraf, 1936). On the contrary 

 Gliick and Plantefol viewed the entire monocotyledonous perianth as foliar, the 

 latter on phyllotactic grounds; Puri (1951) regards it as composed wholly of 

 "tepals." Attention should perhaps be called to the anomalous genus Trimenia, 

 where bracteoles intergrade to tepals (Money, Bailey and Swamy, 1950), and to 

 Winteraceae, where the primitive calyx is synsepalous although the numerous 

 petals are free (Bailey and Nast, 1945a). In Caprifoliaceae, even within the 

 same genus, the vascular supply of a sepal may be reduced from three traces 

 to one (AVilkinson, 1949). Mattfield (1938a, 1938b) derived the petals in Caryo- 

 phyllaceae and other families from the adaxially fused stipules of the alterni- 

 sepalous stamens. Woodson and Moore (1938) interpreted both calycine and 

 coralline scales in Apocynaceae as stipular; Heinig (1951) thinks the petaloid 

 scales in Thymelaeaceae possibly represent sepalar stipules, whereas Leandri 

 (1930) apparently considered them abortive petals. 



Once a double perianth has been formed, however, there is a rather general 

 consensus that the principal trends of evolution are from spiral to cyclic ar- 

 rangement, hypogynous or perigynous to epigynous, pentacyclic to tetracyclic, 

 polymerous to oligomerous, choripetalous to sympetalous, and actinomorphic 

 to zygomorphic. These modifications are customarily visualized as owing to adap- 



