CONSTANCE: SYSTEMATICS OF THE ANGIOSPERMS 419 



On the other hand, apparently drawing a parallel with modern tropical forest 

 types, Bessey, Parkin (1923), and Bews conceived of ancestral angiosperm 

 leaves as being simple, entire, and evergreen. Winkler and Gunderson regarded 

 pinnate division, lobing, or venation as derived from palmate, while Corner 

 believes the palmate derived from the pinnate by shortening of the axis. The 

 fossil record, according to Croizat, indicates that penninerved leaves are fully 

 as primitive as either palmate or lobed ones, and questions "that the evolution 

 of the leaf of existing Angiosperms began with ancestors of a comparatively 

 uniform leaf -pattern" (1940, p. 56), a point emphasized also by Sprague. The 

 relative antiquity of stipules was stressed by Thomas ( 1932 ) and Sporne. Within 

 Leguminosae, Dormer (1945, 1946) supports a sequence from highly compound, 

 pulvinate, and stipulate leaves to less compound, estipulate, and epulvinate. 

 Arber considers the leaves of monocotyledons to be "replacement organs," which 

 are not strictly comparable with or homologous to the leaves of dicotyledons. 

 Leaf venation of distinctive pattern has customarily served taxonomists and 

 paleobotanists as an indispensable means of quick recognition of certain plant 

 families, and the presumably parallel venation of monocots as opposed to the 

 reticulate nervation of dicots has been overemphasized and oversimplified. The 

 recent studies of Foster (1950a, 1950b, 1951) on foliar venation in Quiinaceae 

 suggest that much of value may be expected from this field of investigation, but 

 that it is far too early for useful phylogenetic conclusions. The situation ap- 

 pears to have been ably summed up by Foster in the following words (1931, 

 p. 244) : 



It seems almost unnecessary to state in these days when all phylogenetic systems are 

 experiencing revision and reexamination in the light of new facts that we are very far 

 from an understanding of the details of foliar evolution. 



Although Odell (1932) questioned that even living species could be recog- 

 nized satisfactorily on the basis of leaf form, venation, and epidermal structure, 

 Edwards (1935) has endorsed the value of cuticular characters, especially in 

 the critical determination of fossil angiosperm leaves, provided a sufficiently 

 broad and detailed comparison is made with living types. An impressive utili- 

 zation of cuticular and epidermal features in attempting a more natural clas- 

 sification of Gramineae is represented by the studies of Prat (1932, 1936). The 

 so-called "spodogram" technique of Molisch, which consists in revealing the pat- 

 tern of mineral residues in the epidermis by ashing, has been applied systemati- 

 cally to the Urticales by Bigalke (1933) and to several other groups by Japanese 

 workers. 



That the stomata of angiosperms are of diverse types is generally recognized, 

 and Solereder designated four basic kinds in dicotyledons: Eanunculaceous, 

 Cruciferous, Caryophyllaceous, Kubiaceous. Metcalfe and Chalk point out that 

 these types are by no means confined to the families for which they are named, 

 and proposes to substitute new terms for these family designations. Hutchinson 

 believed that the existence of stomatal differences strengthened his fundamental 

 separation of woody and herbaceous lines of dicotyledons. An alignment of Com- 

 melinaceae with Gramineae, and of Juncaceae with Cyperaceae, according to 

 Ziegenspeck (1938), is confirmed by stomatal structure. The gratifying results 

 obtained with the phyletic indicator-value of stomata in gymnosperms suggests 

 that further investigation may extend their utility in angiosperms. The prelimi- 



