TROPAOLUM -PEREGRINUM AND T. SPECIOSUM. 229 
In Plate XIV. fig. 8, the germ is further advanced; the extra-seminal process 
having now run some little distance in the substance of the carpel, keeping just 
outside the inner surface of the cavity of the germen. 
In the still later stage, represented in Plate XIV. fig. 9, the extra-seminal 
process is greatly elongated, continuing to run in the substance of the carpel 
just outside the inner surface of the cavity of the germen. The elongated base 
exhibits no change. The suspensor is somewhat longer and thicker, this being 
apparently due solely to enlargement of its cells. The young cotyledons are 
| now. distinctly developed, the “embryonal globule” stage being now passed. 
| The subsequent stages need not be minutely detailed here. They are cha- 
_ racterised chiefly by the further elongation of the extra-seminal root process, 
which ultimately runs the whole length of the carpel; and by the enormous 
_ development of the fleshily thickened cotyledons, between which a well-deve- 
_loped plumule lies. The radicle, as in 7. majus, is but slightly developed. 
In Plate XIV. fig. 10, is represented an exceedingly interesting abnormality, 
| which appears to illustrate, negatively, the function of the extra-seminal root. 
_ Here this process has attained a considerable length, but has failed to make its 
| way out of the seed—is zntra-seminal, in fact,—and, apparently in consequence 
| of this, the suspensor and embryo are misthriven and shrivelled as if they had 
suffered from want of that nourishment which it is doubtless the function of the 
| extra-seminal process to obtain from outside the seed. 
| One of the most striking peculiarities of the germ of this species, as compared 
| with those of 7. majus and T. peregrinum, is the total absence of the placental 
_root-process, not the slightest trace of which is to be observed at any period of 
) development. In looking at some of the more advanced stages in 7. speciosum 
| (eg., Plate XIV. fig. 9), one might at first sight imagine that the elongated, 
curved, and pointed base represented the placental root-process in the other 
two species; but in tracing it back through its developmental stages, it is demon- 
| strably nothing more than an elongation of the pointed base of the originally 
flask-shaped germ—retaining throughout, moreover, its original position, 
impacted in the pointed apex or micropylar extremity of the embryo-sac. 




Conclusion. 


The phenomena above described suggest many interesting considerations. It 
‘cannot be doubted that the lateral branch or branches from the base of the 
| suspensor in 7ropw@olum serve as organs of nutrition for the developing embryo 
'—as fetal roots, in fact. It seems probable, as has already been remarked by 
‘Scuacur and myself, that their presence is physiologically conditioned by the 
absence or, at most, the very slight development of even transitory endosperm. 
In treating of this subject, Scnacut has pointed out that those very different 
structures, the czecal dilatations of the embryo-sac in many Labiate, Scrophu- 
