FLORIN: SYSTEMATICS OF THE GYMNOSPERMS 353 



and that two different types of polyembryony had derived from it. Allen 

 (1946b) found the relation of simple to cleavage polyembryony still obscure, 

 and suggested that the former might represent a less specialized condition than 

 the latter. According to Thomson (1945), Buchholz's explanation is essentially 

 a defense of the primitiveness of the embryogeny of Finns and of the phylo- 

 genetic significance attached to this genus by Jeffrey (1917). The relative fre- 

 quencies of simple and cleavage polyembryony in lower and higher gymnosperm 

 groups indicate that the simple type is primitive and has been replaced by the 

 cleavage type. Sufficient proof that simple embryogeny in conifers differs in 

 origin and character from that in other plants has not been produced. Cleavage 

 polyembryony might, moreover, have originated independently in various fami- 

 lies, and TRa.y thus be of less phylogenetic value than has generally been supposed. 



Interest in palynology in general was strongly promoted at the beginning 

 of the modern period, chiefly by the appearance of a manual by Wodehouse 

 (1935). He discussed the principles involved in the study of pollen grains, fur- 

 nished a method of approach, described the grain forms of various families, 

 and discussed evolutionary tendencies and relationships within and between the 

 groups. His treatment of the morphology of the grains in lower gymnosperms 

 was incomplete and misleading, however. 



The microspores of pteridosperms were studied by Halle (1933), Scliopf 

 (1948), and especially by Florin (1937), who emphasized the occurrence of two 

 main types in this group, one of which — large, ellipsoidal grains, provided with 

 a monolete, often somewhat deflected mark near the middle, and furrowed dis- 

 tally — appears to be characteristic of the Medullosaceae. Florin (1936b, 1938- 

 1945, 1951) further investigated the microspores in paleozoic conifers. They are 

 of the same general type as those of the cordaites, and have an annulate air sac, 

 interrupted only at the distal pole. Upper permian conifers, on the other hand, 

 have pollen grains with two air sacs. The former type is therefore relatively 

 primitive in the conifers, while the grains of the living Pinaceae and Podocar- 

 paceae — with two or three smaller sacs, or with no air sac at all — are reduced 

 structures. R. Potonie (1952) then discussed the ontogeny, origin, and evolu- 

 tion of the air sacs, and Schopf, Wilson and Bentall (1944) classified micro- 

 spores occurring isolated in paleozoic deposits. Miiller-Stoll (1948) distinguished 

 pinoid, laricoid, and taxoid conifer pollen in relation to wall structure and be- 

 havior at germination, and regarded as most primitive the monocolpate pollen 

 of cycads. Ginkgo, bennettites, and cordaites. Palynological studies of living 

 conifers were in particular carried out with reference to the Pinaceae, Taxo- 

 diaceae, and Podocarpaceae (Campo-Duplan, 1950, 1951; Ueno, 1951; Cran- 

 well, 1941). 



Emberger (1944 and earlier) brought up for discussion the nature of the 

 lower gymnosperm seeds. The "seeds" of pteridosperms and cordaites, although 

 externally resembling true seeds, are really of the nature of fertilized ovules 

 at the shedding stage. These plants thus shed megasporangia, each enclosed in 

 an integument, instead of true seeds containing embryos. To Emberger this 

 meant that the lower gymnosperms were praephanerogams, with a position in- 

 termediate between vascular cryptogams and phanerogams. Favre-Duchartre 

 (1943; ef. T.-L. Li, 1934) asserted that the seeds of Ginkgo behaved in a similar 

 way. Chadefaud (1944) regarded moreover the cycads as representing the 



