or u von ; c:t.ui-u.i1 G'jj],TricntG o^ oucii uiau ol the cr.psule. 

 We liave seen above that sectioi' I in the coral scgricnts, 

 au well aG in tlie other Gej^r.icnts of this part of the spoi'O- 

 carp, developes much as in the seg-ments of tlio stalk or 

 of tlie petiole except for the peculiar specialization of 

 the hypodermis. All the other sections on the contrary 

 have a peculiar liistory differing from those of either 

 th.e stalk or the loaf. Sections III , IV and V , dor- 

 sal to the marginal coll, v/iden rapidly at their outer 

 ends, imd, as sections II and V do not v;iden in th.is 

 way, the ultimate marginal cell is pushed into a vsnti-al 

 position (Figs. 45, 46, 51 )• Of coiu-se the interpolation 

 of the e.itra sec' ion IV , dorsal to the marginal cell^ 

 would give the lettter a more ventral position than it has 

 in the petiole (Figs. 15, IC), and Ihis is evidently tlie 

 use of its formation out of the normal sequence, "but this 

 result is greatly enlianced by its very vigorous expansion 

 at the outer end. Finailly section VI grows out at tlie 

 ventral end, beyond the cells arising from the basiscopic 

 marginal coll and spreads over into contact with the cells 

 of section V at the ventral surface (Figs. 53, 54). Tlie 

 plerome of the basiscopic oi- soral half of section III 

 breaks up by one tinticline and then by periclines (Figs. 

 51, 53) to foiTri tlir dorsal part of the lateral bundle. 

 The more dorsal paru of section IV does likev/ise, v/hiie 



