Leaf and Sporocarp in Marsilia qnadrifolia , L. 139 
these. Another longitudinal row of the same cells gives rise 
to each of the tannin-sacs. 
The pinnae or divisions of the lamina are formed by the 
continued activity of the marginal cells of certain segments, 
but their limits do not correspond exactly with those of the 
segments, the lower pair being nearly two segments in length 
and the upper pair about three. 
In its mode of origin, then, the leaf of Marsilia agrees 
closely with that of other leptosporangiate Ferns, as it does 
also in its further growth by the segmentation of a two-sided 
apical cell. But the position of the first division-walls in 
these segments, while very like that described for Asplenium 
serpentini by Sadebeck (’74), is apparently quite unlike that 
described for Ceratopteids by Kny (’75), for Onoclea by 
Campbell (’87), and that given by Campbell (’95, p. 325 ) 
for the Leptosporangiates in general. In the development 
of the lamina also Marsilia is unlike other described forms 
except Ceratopteris , since the pinnae are not co-extensive with 
the segments as in Onoclea and Asplenium , though all agree 
in having the pinnae formed by the activity of a series of 
marginal cells. There is however great need of more detailed 
work on the origin of the leaf and the differentiation of this 
into petiole and lamina. 
The sporocarp of M. qnadrifolia is developed from a trans- 
versely placed apical cell, arising in a marginal cell on the 
inner side of the young leaf. The second sporocarp when 
present (usually) arises in the same way from a marginal cell 
of the first. The two are thus respectively primary and 
secondary branches of the leaf. 
More rarely we may find two or more sporocarps inserted 
separately on the petiole, both on the same side. Then the 
suggestion is a tempting one, more especially in cases like 
M. polycarpa , where ten or more sporocarps may be borne 
in the same way, that the sporocarps represent pinnae homo- 
logous with those at the tip of the petiole, and the study 
of abnormal pinnae by Biisgen may perhaps seem to favour 
this. But before accepting this we have to account for the 
