Anatomy of the Genus Selaginella , Spy. 25 
sporangium has, as a rule, three lobes, instead of four. Occasionally there 
is evidence of the fourth abortive spore. 
3. Reduction to two equal megaspores. 
S. molliceps (2 cases), S. patula (1 case), S. flabellata (1 case). 
In vS. rupestris there are normally two spores. 
4. Reductions to one megaspore of correspondingly large size. 
S. sidcata , S. molliceps , 5 . rupestris. 
Cases of increase in the number of spores are very rare. I have only 
discovered two—in Vogelii (12 spores) and vS. involvens (8 spores). 
These are suggestive of the common ancestry of the micro- and mega¬ 
sporangia, whilst the reductions are indicative of a tendency towards the 
seed habit, specially exemplified in S. Bakeriana and vS. rupestris. It is 
difficult to say how these reductions have arisen. In the case of two equal 
spores, one is tempted to explain them by premising one division only 
of the spore mother-cell. This has been proved conclusively for S. rupes¬ 
tris (6), none of the spores developing if two divisions occur. However, 
in face of the frequent occurrence of three equal spores, and occasional small 
abortive ones, it seems more probable to suggest, as the general rule, that 
division into tetrads takes place, but that some develop excessively at 
the expense of the others. 
The reduction to one megaspore in certain species, and the confirmed 
tendency to reduction in others, become more significant when taken 
in conjunction with the fact that not only are archegonia developed, but that 
fertilization and development of the embryo with roots and cotyledons 
frequently take place whilst the spores are enclosed within the sporangium 
and still attached to the strobilus (6). With particular reference to S. rupes¬ 
tris Miss Lyon remarks : * The strobili may fade and separate from the 
plant before fertilization, but the spores do not fall from the sporangium. 
In no case is there evidence of fertilization in the spores that are shed ’ (6). 
6'. rupestris has, therefore, progressed a long way in the direction of the 
seed habit. Indeed if one could conceive a hypothetical case of a unispored 
megasporangium of Selaginella , whose spore had been fertilized and had 
developed an embryo whilst still surrounded by the protective sporangium 
wall, we should have what might be regarded as the morphological equiva¬ 
lent of a seed. With respect to this seed-bearing habit, the Selaginellaceae 
' are far ahead of their lower allies, the Lycopodiaceae, and suggest a parallel, 
even though a remote one, with the Gymnosperms. 
Detailed structure of the sporangium wall, and mechanism for casting of 
spores. 
Both micro- and megasporangia are stalked, attached to the stem just 
above the origin of the leaf. The megasporangium is definitely four-lobed, 
owing to the relatively large size of the four enclosed spores, whilst the 
