354 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



"praephanerogamous stage," Mangenot (1952) supported the conception of Em- 

 berger in contradistinction to Ganssen (1944-1952) and Lam (1948). Ar- 

 nold (1948), Martens (1948), and Walton (1952) did not believe that there was 

 any genuine difference between the seed-like ovules of paleozoic plants and the 

 seeds of modern ones, but only a delay in embryo development until the termi- 

 nation of a rest period. Emberger (1949, 1952b) defended, however, the notion 

 of praephanerogams and their position as a distinct large natural group of 

 vascular plants. A critical analysis of the problem was then presented by Mar- 

 tens (1951). As regards external characters, the ovule and the seed — as well as 

 the praephanerogams and the phanerogams — are contradistinctive, but this does 

 generally not apply to other features. Embryo formation before shedding is a 

 character common to conifers and bennettites, but does not apply to chlamydo- 

 sperms. The contrasted feature — embryo formation after shedding — is extremely 

 variable. The praephanerogams shed either spores or prothallia that have just 

 been fertilized, or even embryos, while the phanerogams in some cases shed just 

 fertilized prothallia {Gnetum), but nearly always embryos. The criterion of a 

 true seed, based on the accumulation of food-reserves in the prothallium, is valid 

 for the conifers, but not for all chlamydosperms {Ephedra). The character of 

 the integument attrilnited by Emberger to the ovule of the "praephanerogams" 

 is valid, but the contrasted feature is invalid in numerous conifers as a criterion 

 of true seeds. Martens moreover pointed out the difficulties involved in classi- 

 fying the bennettites with the true phanerogams, in separating the cordaites and 

 ginkgoes from the conifers — and the cycads from the bennettites — and in putting 

 together such widely dift'erent groups as the cordaites and the pteridosperms. 

 Hagerup (1933) interpreted the integument of a conifer ovule as a mega- 

 sporophyll carrying the megosporangium on its ventral side. It differs distinctly 

 from the corresponding organs of cycads and ferns, and conifers and cycads 

 can therefore not be referred to the same higher group, the gymnosperms. 

 Hagerup 's theory was accepted by Emberger (1944, 1950), but in most other 

 quarters it now appears to be rejected. Halle (1937) and Walton (1952) re- 

 garded the integument of pteridosperm ovules as a syntelome surrounding a 

 fertile telome. In Forin's (1951) opinion, the integument of the ovule in cor- 

 daites and conifers is formed by collateral fusion of two uninerved branches 

 (sterile telomes) of the megasporophyll enclosing the single terminal megaspo- 

 rangium (fertile telome). The megasporophyll of the conifers constitutes a 

 telome system, producing as a rule one terminal ovule by dichotomy, overtop- 

 ping, and aggregation of telomes. In the taxads, however, the integument is 

 probably formed out of two or more sterile, aggregated telomes — or in certain 

 cases small telome systems — which are overtopped branches of the floral axis. 

 The position of each component corresponds to that of a megasporophyll (spo- 

 rangial truss) in the cordaites and conifers. According to Eames (1952), the 

 ovular integument of Ephedra is also made up of two components. 



The conifer pollen grains and ovules have moreover been studied from the 

 point of view of the evolution of pollination mechanisms. Doyle (1945) sug- 

 gested that the micropyle of the erect ovule of the paleozoic Lehachia, which pos- 

 sesses more or less erect female cones, exuded a pollination fluid in which the 

 wind-borne pollen grains were caught. The annulate air sac caused the grain 

 to float with its distal germinal zone directed towards the nucellus. Pollina- 



