FMBRYOGENESIS IN GYMNOSPERMS 191 



again divide, three very regularly arranged tiers being thus formed. A 

 division in the lowest tier completes this proembryo phase of develop- 

 ment. Fig. 37. The lowermost tier will give rise to the embryo proper, 

 the middle tier to the primary suspcnsor, the third tier to the rosette, 

 while the uppermost tier which is unwalled on its upper surface will 

 soon degenerate. The rosette is thus adjacent to a wall described as the 

 basal plate which separates the embryo from the unorganised remains 

 of the proembryo. Rosette cells are present in Pi/ius, Cedrus, Tsuga and 

 Pseudo/arix; but in Abies, Picea and Lari.x, in which they are initiallv 

 present, they soon degenerate. Both the distal embryonic cell tier and 

 the rosette tier may give rise to embryos, the development of the latter 

 taking place more slowly. We shall return to them later. 



Attention must now be directed to the lowermost tier, for it gives 

 rise to all the organs of the embryo proper, including the shoot, 

 cotyledons, leaves, roots and secondary suspensor cells, the upper three 

 tiers contributing nothing to these developments. These tiers, neverthe- 

 less have important functions. The uppermost tier, of which the distal 

 ends are in open contact with the egg cytoplasm, is apparently active in 

 transmitting nutrients to the growing embryo below; and this appears 

 to continue as long as any food reserves remain in the micropylar end 

 of the proembryo. The cells of the tier below this absorptive layer, 

 known as the 'rosette' because of its appearance in vertical view, are 

 actively meristematic and often develop into embryos. The cells of the 

 penultimate basal tier undergo a quite remarkable elongation and 

 constitute the primary suspensor. At this stage in the development, 

 where the embryonic region is about to be thrust down into the pro- 

 thallus, the proembryo phase may be considered, arbitrarily but con- 

 veniently, to have come to an end. 



Polyembryony is a very general phenomenon in the Coniferales. It 

 may be due to the development of several fertihsed eggs, this being said 

 to constitute simple polyembryony : or it may result from the longitudi- 

 nal subdivision and separation of several embryos from a single pro- 

 embryo, this being described as cleavage polyembryony. Both types are 

 of wide occurrence: simple polyembryony is general in Larix, Picea, 

 Pseudotsuga and Abies', while cleavage polyembryony is a constant 

 feature in Pinus, Cedrus, Pseudolarix and Tsuga. Buchholz (1933) has 

 made the distinction between determinate and indeterminate cleavage 

 polyembryony in conifers. In the latter there are no indications that 

 any one of the several embryos, derived from a single zygote, has a 

 distinct advantage during development. In the former, one embryo, 

 usually the terminal one, is more favourably situated than the others, 

 and tends to be the one which comes to maturity in the seed. The 

 position of the various cleavage embryos is thus important in determining 



