THE GYMNOSPERMAE : CONIFERALES AND TAXALES 703 



slowly grows and surrounds the whole ovule. This is commonly called the 

 aril, but it is a misnomer, for the structure is not the same as those outgrowths 

 of the seed coat which bear this name among the Angiosperms. It has been 

 more justly compared with the epimatium, secondary covering of the ovule 

 in the Podocarpineae. Some consider it to be an outer integument, but in 

 the absence of any decision as to its true nature we shall here refer to it as 

 the cupule. Although it starts as a ring, its upward growth is slightly two- 

 lipped and it receives two minute and almost abortive strands of scattered 

 tracheids, arising as branches from the base of the integumentary strands. 

 It may therefore be derived from two fused bract scales, just as the integument 

 itself may be similarly derived. One may recall in this connection the two- 

 lipped formation of the integument in Piniis. 



The integument later develops a very hard, stony layer outside the 

 vascular bundles, and this is covered externally by three layers of cells forming 

 a thin brown skin which is soon detached, leaving the stony layer as the 

 exterior coat of the ripe seed. 



During the lignification of the stony layer a tube-like sheath of cells, with 

 thick yellow cuticle, develops around the micropyle, the resistance of 

 which prevents the micropyle from being closed by the growth of the 

 integument. 



The vascular bundles of the integument penetrate this stony layer at the 

 base through two pores or foramina. These may be compared with the 

 lateral foramina in the seed of Torreya (see p. 696), which also admit vascular 

 bundles from outside, but lie near the apex of the seed. Sahni has suggested 

 that this condition in Torreya may have been derived from the Taxus con- 

 dition by the expansion of the seed base. This corroborates Oliver's theory 

 that the whole of the lower part of the seed in Torreya is a secondary develop- 

 ment, added, in the course of evolution, from seeds of the Cordaitalean type, 

 in which the bundles penetrate the stony integument at the base (Fig. 706). 



During the development of the seed in Taxus there is apparently a con- 

 siderable amount of intercalary growth at the base. While the whole of the 

 nucellus is free from the integument in the young state, the free portion is 

 reduced to a small apical zone in the ripened ovule, the lower portion between 

 the base of the seed and the limits of the free nucellus having developed 

 secondarily. It is interesting to note that the two integumental bundles are 

 bent into a knee-like flexure just above the point at which they penetrate the 

 stony integument and that they are thickened at this bend. Sahni considers 

 that this marks the place at which the vanished nucellar strands were formerly 

 given oft, such as existed in the Cordaitalean seeds. The nucellar traces in 

 Torreya are probably the only relic of this interior system now surviving in 

 Taxales or Coniferales, which may have been important only when fertiliza- 

 tion by free antherozoids was the rule, for abundant secretion of moisture 

 by the nucellus would then be essential. 



The archesporium may be recognisable in the young ovule before the 

 integument is mature, but the development of the female gametophyte is not 

 completed until the following spring. 



