CONSTANCE: SYSTEMATICS OF THE ANGIOSPERMS 437 



spermous and microsperinous plants, and adds that niicrospcrmous herbs could 

 not have given rise to megaspermous trees. Like all phyletic schemes based on 

 a single feature, this Durian theory is more entertaining than convincing. 



The standard textbook classifications of fruits and seeds are obviously arti- 

 ficial, teleological, and thoroughly unsatisfactory. Two modern students, Wink- 

 ler and Guguleac, have attejnpted to subordinate the fascinating biological 

 aspects of fruit- and seed-dispersal, and to relate fruit classification squarely 

 to the structure of the floral gynoecium. The fruit, it appears, is almost as dif- 

 ficult to define as is the flower. Knoll (1939), following Gaertner and Goebel, 

 designated as fruit all parts of the flower remaining at the time of seed ripening; 

 Winkler (1939, 1940) and Gusuleac (1938a, 1938b), on the example of Pax, 

 restricted the fruit to that structure which arises from the gynoecium as a con- 

 sequence of fertilization or parthenocarpy. Accepting the foliar interpretation 

 of the carpel, both Winkler and Gu§uleac regarded as primitive those fruits de- 

 rived from apocarpous gynoecia, those from syncarpous (paracarpous, coeno- 

 carpous) gynoecia as advanced. The classification proposed by Winkler (1939, 

 1940) has tw^o main divisions, the apocarpous Sammelfrucht and the syncarpous 

 Einheitsfnicht. These two major divisions are then each subdivided on the basis 

 of superior versus inferior ovary, and the resultant four subdivisions on the 

 criterion of dry versus soft-fleshy texture. Guguleac (1938a, 1938b) proposed 

 four principal categories: Apokarp, Eusynkarp, and Apokarpoid (syncarpous 

 but separating into carpellary units), each representing the entire gynoecium 

 of a single flower, and ZonantJiokarp, an artificial grouping of all "false fruits" 

 derived from the gynoceia of two or more flowers. The apocarpous group is bi- 

 sected according to w^hether the gynoecium consists of a single, or of two or 

 more carpels; the syncarpous group according to whether the gynoecium is pluri- 

 locular or unilocular; and the apocarpoid group according to whether the disin- 

 tegration of the mature fruit yields pieces equivalent to whole or to only partial 

 carpellary units. Each of these six subdivisions is then redivided into a capsule-, 

 a nut-, and berry-, and a drupe-series. The three prime categories of this system, 

 as thus aligned, are supposed to represent a phylogenetic series, but in a second 

 article the same year, the Eusynkarp and the Apokarpoid groups were trans- 

 posed. Winkler stressed the follicular carpel as the basic unit of all angio- 

 spermous gynoecia and attempted to show that the elements of a septicidally de- 

 hiscent capsule are each fully equivalent with such a free carpel (1936b, 1939, 

 1940, 1941; Juhnke and Winkler, 1938). To Winkler, Guguleac's fruit-series 

 should then be read in accordance with the advancing grade of syncarpy, viz. : 

 (1) carpels free — apocarpous, choricarpellous; (2) carpels weakly united and 

 separating in fruit — apocarpoid, dyssencarpellous; (3) carpels strongly united 

 and not separating into units corresponding with carpels — eusyncarpous, syn- 

 carpellous. Both authors, it is to be noted, accepted the view that analogous de- 

 hiscent or indehiscent, dry or fleshy, and one- or several-seeded fruit types can 

 arise anywhere along this sequence, presumably as a result of the selection pres- 

 sure of biological demands. The loculieidally opening capsule, also, is regarded 

 as a biological rather than a basic, phylogenetic type. The evolutionary validity 

 of such classifications as the two just described is attested by the fact that mem- 

 bers of the same taxonomic groups are found to be characterized by the same or 

 closely related fruit types. This situation is strikingly different from that which 



