22 Transactions. 



and the nucleus retreats (at first, at any rate) to the inner end of the 

 cell. The ovum surrounds itself with a delicate membrane, which arches 

 up somewhat into the base of the neck-canal of the archegonium and at 

 that point thickens. It grows to a fairly large size before it segments, 

 somewhat, though not to the same extent, as Bruchmann has described 

 in the case of Lycopodium clavatum and L. annotinum (Bruchmann, 1898). 

 The first division wall to be formed is more or less transverse to the axis 

 of the archegonium, and seems to be approximately in the middle of the 

 cell (figs. 52, 53). This wall thus divides the embryo into what we may 

 speak of as the lower and upper regions. This first division may be 

 clearly traced afterwards in older embryos. The next division wall to 

 appear is in the lower half, and extends at an angle from the first 

 wall to the lower end of the embryo (figs. 52, 53). It also may be clearly 

 seen in older embryos. No embryos were found in transverse section, 

 so that this description of the earliest stages in segmentation can only 

 refer to the appearance of the embryo in longitudinal section. Still older 

 embryos are shown in figs. 54, 55, and 57. I find it difficult to describe 

 with any degree of certainty the sequence of segmentation which has taken 

 place either in the lower or in the upper parts of these embryos. 



In addition to the section of the embryo shown in fig. 55, already referred 

 to, a second section (fig. 56), obviously not so nearly median, shows a part 

 of the same embryo which I am inclined to think is the stem-rudiment. In 

 it there are two main walls intersecting at right angles, and in one of the 

 cells so formed another wall has appeared cutting out what might well be 

 an apical cell. This part of the embryo took the haematoxylin stain rather 

 more darkly than did the rest, and the nucleus of the " apical " cell was 

 conspicuously large, the suggestion being that this part was forming rapidly. 

 It will be evident from a comparison of the two sections of this embryo 

 that this portion which we are now considering belongs to the upper region 

 and has arisen laterally from it. If it proves to be correct that the shoot 

 originates from the upper half, this fact would distinguish the embryo of 

 Tmesipteris from that of the Lycopodinae, where the upper primary segment 

 constitutes a suspensor, but would, on the other hand, suggest the embryo of 

 Equisetum and the Ophioglossaceae. Of course, one main reason why the em- 

 bryo of Tmesipteris is likely to prove of special interest is the fact that the 

 adult plant has no root, consisting only of an underground branched rhizoid- 

 bearing rhizome and an aerial branched leaf-bearing portion. Anticipating 

 here what I shall be bringing forward in connection with the developing 

 plantlet, we may say that the young plant of Tmesipteris is " all shoot," 

 just as the embryo of certain members of the Ophioglossaceae has been 

 described as " all root." The question naturally arises whether there is 

 in the embryo of Tmesipteris anything which may be interpreted as the 

 undeveloped rudiment of a root. Only a much fuller study of the develop- 

 ment of the embryo than that given above can satisfactorily decide this 

 point. I hope to be able to gather more material for such a study. The 

 stages described above stop short at a most interesting point, and I have 

 found it difficult to interpret sonic of them. Keeping pace with the growth 

 of the embryo, the surrounding prothallial cells rapidly subdivide, so that 

 the embryo is enwrapped by a small-celled tissue which soon begins to 

 project as an eminence from the side of the prothallus (figs. 55, &c). 



Before passing on to the section of this paper which deals with the 

 developing plantlet there is still an important and interesting point to 

 be brought forward which concerns the question of the " foot " of the 



