THE CARPEL 221 



pinnate. The ovule of the angiosperms has two integuments; that of 

 the Caytoniales, one. 



It is claimed that this theory of the origin of the carpel was not 

 originally considered to suggest the origin of the angiosperm carpel 

 itself from the Caytonialian megasporophyll, but as a possible method 

 of origin of the carpel from a pteridosperm sporophyll. But, in recent 

 years, it has been suggested many times that Mesozoic pteridosperms 

 are possible ancestors of the angiosperms, and the theory has been 

 considered to imply this origin. 



Homologies of the Carpel with Gymnosperm Structures. Comparison 

 of reproductive structures of some living gymnosperms with simple 

 angiosperm flowers has been believed to demonstrate homology between 

 the integument of the gymnosperm ovule and the angiosperm carpel. 

 In Ephedra, the pair of bractlets that enclose the naked, apparently 

 terminal, ovule are considered homologous with the carpels of a two- 

 carpellate flower of Pepcromki, demonstrating the integumentary nature 

 of the carpel. As a part of the search to find in the gymnosperms the 

 origin of the carpel, the cone of Junipenis has been claimed to give 

 important evidence. The ovules of this conifer are "basal" on the fertile 

 scale and interpreted as cauline ovules surrounded by sterile scales, the 

 homologues of angiosperm carpels. Ephedra and Junipenis are con- 

 sidered to show that the carpel is a sterile appendage which surrounds 

 cauline ovules. But the plants used to demonstrate these homologies are 

 highly specialized: Ephedra, a surviving remnant of an ancient stock; 

 Jiiniperus, a conifer with greatly reduced cones; the Piperaceae, angio- 

 sperms with greatly reduced flowers. The simplicity of great reduction 

 suggests a similarity in morphological structure that does not exist. 



Carpel Polymorphism. The theory that the gynoecium consists of 

 carpels of more than one basic type was proposed by E. R. Saunders 

 in 1923, and, for fifteen years following, many descriptive articles and, 

 ultimately, a two-volume book, "Carpel Polymorphism," were published. 

 This theory presented an entirely new interpretation of the gynoecium, 

 one wholly at variance with the classical view. Two basic types of 

 carpel, the valve and the solid, were described and, later, a third, the 

 semisolid, was added. Recognition of these types multiplied twice to 

 several times the number of carpels in the gynoecium of most families 

 and commonly doubled the number of whorls. For example, the gynoe- 

 cium of most legumes consists, according to this theory, of two carpels, 

 one valve and one solid, that of the Liliaceae of six, rather than three, 

 in two whorls, one valve, one solid. Large numbers of carpels were 

 described in gynoecia commonly considered to have one or two: 10 to 

 14 in the peanut, Arachis, and Scorpiurus; 16 in Brassica; 40 to 50 in 

 Raplsfrum, 



