422 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



naeus, and Goethe and paraphrased by Eames: "The flower morphologically is a 

 determinate stem with appendages, and these appendages are homologous with 

 leaves" (1931, p. 147). As a corollary he adds, "This commonly accepted view 

 of the nature of the flower is sustained by its anatomical structure. Flowers, in 

 their vascular skeletons, differ in no essential way from leafy stems" (p. 147). 

 Granting this classical homology of vegetative shoot and flower, the primitive 

 state of the latter would logically be that most similar to the former. Thus, we 

 might anticipate as primitive a radially symmetrical flower with an elongated 

 axis bearing numerous, indefinite in number, separate, leaflike members ar- 

 ranged in regular spirals, or cycles, or pairs, depending upon the vegetative 

 phyllotaxy. Such an archetypic flower would possess essentially that central 

 floral plan from which de Candolle regarded all other kinds as derivative in 

 consequence of cohesion, adnation, abortion, or change of symmetry. It is also 

 essentially the eucmthium (euantJiostrohilus) of those numerous authors who 

 have thought the primitive flower to be a simple bisexual strobilus, perhaps best 

 represented among living forms by apocarpus, entomophilous, hypogynous Ra- 

 nales (Polycarpicae) of the dicotyledons and Alismatales (Helobiae) of the 

 monocotyledons. Even those who have regarded some other basic floral type as 

 more archaic, or who have accepted a polyi^hyletic origin of angiosperms, have 

 usually held the Ranalian flower to be the prototype of some or most other 

 flowering plants. Perhaps it should be emj^hasized at this juncture that Arber 

 and Parkin cautioned: "As we have pointed out, there is no reason to believe 

 that any Angiosperm with a complete assemblage of primitive floral characters 

 is to be found today, nor indeed that such a flower ever existed" (1907, p. 45). 

 Rather, the presumed ancestral traits are to be discovered dispersed among exist- 

 ing dicotyledons and monocotyledons, especially Ranales. 



The principal dissent from the acceptance of the "complete" hermaphrodite 

 flower as primitive initially came from those who, following Eichler, Engler, and 

 Wettstein, attempted to find homologies betv/een unisexual gymnospermous in- 

 florescences or strobili of vascular cryptogams and a wind-pollinated, unisexual 

 flower with little or no floral envelope {27seudanthium,) . According to Engler, 

 "it is especially not to be conceded that families with prevalently wind-polli- 

 nated plants without au}^ or only simple perianths could have developed from 

 insect-flowers with simple or double perianth" (p. xxiii). The difficulties in the 

 way of manufacturing a bisexual flower from unisexual inflorescences or strobili 

 (Karsten, Vuillemin, Neumayer, Wettstein, Emberger) or of an angiospermous 

 gynoecium from naked ovules or sporangia (Thomas, Hagerup, Hjelmqvist, Jan- 

 chen), has led to the spinning of ingenious but tortured "character-phylogenies" 

 which strain credulity. Calestani (1933) has raised the interesting suggestion 

 that the ancestral angiosperms must have been entomophilous and bisexual 

 in order to have had a selective advantage over the anemophilous, unisexual 

 gymnosperms. 



In recent years, there have been numerous attempts to explain the structure 

 of angiosperm flowers on the basis of the features of Devonian Psilophytes, 

 assuming that all the organs of vascular plants are referable to aggregations of 

 fertile and sterile "telomes" (Zimmermann; Thomas, 1934, 1936; Hunt; Chade- 

 faud, 1946, 1947; Bertrand, 1947a; Lam; van der Hammen, 1948; Emberger, 

 1950, 1951; Wilson; Suessenguth and Merxmiiller, 1952; Takhtajan, 1953). To 



