418 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



seedlings. The basic cliaracter of the trilacunar node appears to have been 

 ratlier generally accepted, as by Hunt (1937), Dormer (1945), and Sporne, but 

 Croizat (1940) and Ozenda (1949) both regard the multilacunar node as the 

 original type. Ozenda believes that the return to a multilacunar condition in 

 such dicot families with slieathing leaf bases as Polygonaceae and Umbelliferae 

 is due to a "surevolution" or reversion to a quasiprimitive state. Sinnott men- 

 tioned the plastic and variable condition of the node in woody Ranales and Ku- 

 mazawa (1930) reported the occurrence of both trilacunar and unilacunar situa- 

 tions within Ranunculus. In a resume of their work with woody Ranalians, 

 Bailey and his associates (Money, Bailey and Swamy, 1950) report that those 

 families with monocolpate (or derived) pollen and ethereal oil cells form two 

 presumably natural alliances of ten families each, the one showing unilacunar, 

 the other tri- or multilacunar nodes. Whitaker (1933) thought that he had 

 found some correlation among these families between the possession of trila- 

 cunar nodes and 19 pairs of chromosomes, and unilacunar nodes and 14 chromo- 

 some pairs. In a recent review of petiolar anatomy, Hare (1944) concluded 

 that, although the petiole provides a fresh set of supplementary characters for 

 purposes of classification, it should be regarded primarily in terms of mechani- 

 cal adaptation and as having "little phylogenetic significance" because of the 

 abundant parallelism in unrelated families, a view shared by Dehay (1941) and 

 Metcalfe and Chalk. The wide occurrence of both the unilacunar and the tri- 

 lacunar node in Ranalian groups suggests that these conditions may be equally 

 primitive. 



From the major and diverse uses which have been made of them from the 

 very dawn of plant classification, one may readily agree with Arber that "there 

 seems to be no room for doubt that phyletic indications, external and internal, 

 are carried by the leaves" (1925, p. 8). Attempts to place these evolutionary 

 trends in a generally agreed upon order have been conspicuously less successful. 

 In regard to leaf position, Schellenberg (1928) viewed the distichous arrange- 

 ment, Salisbury (1926) and Winkler (1936a) the tristichous, as basic; von Veh 

 (1930) believed that all monocotyledons could be derived from a distichous con- 

 dition, all dicotyledons from a tetrastichous one. Haccius (1939) suggested that 

 all dicot arrangements could be derived from two fundamental types, and that 

 a truly distichous condition can be attained by different routes. An evolutionary 

 sequence from spiral or alternate to decussate or verticillate is visualized by 

 Hallier, Hutchinson, Dormer (1945), and Sporne, and a trend in the reverse 

 direction by Bessey (1915) and Sprague. 



Concepts as to the primitive characters of angiosperm leaves lean heavily 

 upon the presumed antecedents of the group. Thus, Arber and Parkin, and 

 Wieland (1929, 1931, 1933) suggested that if leaves of early angiosperms were 

 compound, the existence of the group might have been long disguised in the 

 fossil record by their resemblance to those of seed ferns or cycadophytes. Hal- 

 lier (1912), reversing his earlier advocacy of simple Magnolian leaves as primi- 

 tive, read archaic qualities into the pinnately compound foliage of Berberida- 

 ceae. Several authors have indicated a belief that palmately or ternately divided, 

 lobcd, or at least veined leaves were either primitive in flowering plants (Sin- 

 nott, 1914; Coy, 1928; ^Yinkler, 1936a; Gunderson, 1943), or at least intermedi- 

 ate between ancestral compound leaves and more modern, simple ones (Wieland). 



