728 TRANSACTIONS OF SECTION K. 



described some years ago in America, the archegonium on the prothallus of the 

 retained megaspore is fertilised by sperms liberated from microspores which 

 become caught in the lips of the open megasporangial wall. This analogy suggests 

 to us that the pollen-chamber cavity may be a relic or modification of the original 

 place of dehiscence. If this conjecture be true, we have here what was once an 

 exit-pore converted to the purposes of ingress, just as we find, in so many Thallo- 

 phytes, tubes and beaks, once, as it is supposed, the orifices of zoospore discharge 

 now serving for the reception of male gametes. 



A great feature in the early seed types was the complexity of the integument, 

 ■ and this still holds good in recent Cycads and some other Gymnosperms. 

 Protective envelopes are so commonly associated with reproductive organs, and the 

 nutritive conditions are so favourable to their production, that a naked nucellus 

 strikes one as anomalous. If future research confirm the supposition that the Ferns 

 which stand in possible relation to early seed-plants were ex-indusiate, like the 

 Marattiacese, recent and fossil, then no doubt the seed-coat is a new formation, 

 having no true homology with, but merely homoplastic resemblance to, ordinary 

 Fern-indusia. The only case of a naked nucellus that recalls itself is the rather 

 mysterious instance of Lepidocarpon in which Dr. Scott reports the not infrequent 

 occurrence of non-integumented megasporangia with the prothallus fully developed. 



The robust nature of the seed envelope, which was often drupaceous, is in com- 

 plete harmony with the whole character of the seed if you regard the habit at its 

 inception as a xerophllous adaptation. And such no doubt it was, an improved 

 method whereby the plant became independent of chance water at a very critical 

 stage in the life-history. Some of the peculiarities of fossil seed-coats, especially 

 the ribbing of the Lagenostomas and several other genera, have been attributed 

 to the multiple origin of this structure, at any rate in some cases. The remarkable 

 circlet of tentacles which surrounds the summit of Lagenostoma physoides (best 

 known by Williamson's earlier name Physostoma elegans) suggests that a number 

 of foliar lobes have been incorporated in the seed, whilst the presence of peri- 

 micropylar ridges and the septate canopy in allied forms may be taken as only a 

 less evident indication of the same thing. 



The relation between the integument and sporangial body of recent Gymno- 

 sperm seeds is found to be an inconstant character, and the same is true of the 

 fossils. In general character the relationship recalls that which obtains between 

 the ovary and receptacle of an Angiosperm. Whilst the Lagenostomas resemble 

 Cycas and Pinus in having the integument free at the apex only, Taaits, Phyllo- 

 cladus and Araucaria are in agreement with the Trigonocarpons and other seeds, 

 which are generally attributed to Medullosese, in having an integument which 

 rises freely from the chalaza. It is interesting to note that the fossil seeds of the 

 latter group show an additional complexity in the wall of the nucellus. For in 

 them a series of tracheal strands or even a mantle of tracheides is found running 

 up from the chalaza to the pollen-chamber. It is evident that nothing was spared 

 in these older seeds to ensure adequate access of water to the pollen-chamber 

 where the sperms must have been liberated. 



In due time the protective sheath, or testa, appropriated other functions supple- 

 mentary to that of protection. Of these the most important must have been the 

 reception of the pollen. A very striking feature in all the Lagenostomas is the 

 way in which the tip of the nucellus (where the orifice of the pollen-chamber 

 is situated) projects beyond the integument. In these seeds the microspores 

 had direct access to the pollen-chamber without first descending a micropylar 

 canal. 



In the MeduUosean seeds also the nucellus is distinguished by a long beak, as 

 Dr. Scott and Mr. Maslen have shown recently for Trigonocarpon, and, as we know, 

 in Stepha7iospermum, and many other cases. So far as we know, this beak does 

 not extend to the surface, though it engages with the micropylar canal, and is 

 continued some distance up. 



Though it can hardly be supposed that the long beak has been inherited from 

 the ancestral sporangium, its presence may be none the less significant of what 

 took place when the seed method was initiated. The direct pollination in 



