TROPAOLUM PEREGRINUM AND T. SPECIOSUM. 225 
numerous cases already observed, both in plants and animals, of an abnormal 
development in one species, shadowing forth, so to speak, the normal condition 
in another. In this abnormal penetration of the carpellary tissue by the extra- 
_ seminal process in 7. majus, we have an imitation, though a feeble one, of the 
| normal penetration which occurs in 7. peregrinum and T. speciosum, to which 
_ Ihave now to direct attention. 
Tropeolum peregrinum, L. (Canary Creeper). 
| The investigation of the embryogeny in this species is much more trouble- 
some than that in either 7. majus or T. speciosum. The delicate and slender 
| character of the germ, together with the frequent difficulty in getting exactly 
| mesial sections of the fruit-lobes from slight assymmetry, if these, as is often the 
| case, are not all three equally developed, has forced me to be content with a 
comparatively imperfect result. 
In Plate XV. fig. 20, I have given an outline figure of a young germ, the 
parts of which, though somewhat different in form, may at once be identified - 
with the corresponding parts in the germ of 7. majus, a similar outline of which, 
at about the same stage of development, I have also given for comparison 
_ (Plate XV. fig. 19). In 7. peregrinum the basal projection (0) is very distinct, 
but is more knob-like and less pointed than in the common species, There is 
' also considerably more marked obliquity of the body of the germ in 7. pere- 
\grinum than in T. majus; the basal point (6) in 7. majus being nearly in 
| line with the suspensor, while this is far from being the case in 7. peregrinum. 
_ At this stage the extra-seminal and placental root-processes are distinctly indi- 
cated. The latter process (p/r) is still rounded in the germ of 7. majus figured, 
while it has become pointed, although not yet elongated, in that of 7. peregri- 
nun. 
| In Plate XVI. fig. 21 is represented a section of a somewhat advanced fruit- 
/ lobe, where I have succeeded in displaying all the parts of the germ in situ, and 
in tracing both root-processes to their extremities, so as to enable me to give a 
‘complete figure of the germ at that stage. Here the suspensor has become 
greatly elongated, and the embryo at its extremity exhibits rudiments of the 
'two cotyledons. The placental process is now somewhat elongated, and has 
/made its way to the placental vascular bundle, along the course of which it 
runs, all exactly as in 7. majus. The extra-seminal process bores through the 
' seed-coats, also exactly as in 7. majus, making its way into the cavity of the 
seed-vessel. Unlike that in 7. majus, however, this process, after running a 
| short distance in the ovarian cavity, penetrates the carpel (say a little above its 

| 
middle third),and pursues the remainder of its course embedded in the carpellary 
‘tissue. The direction of this process, after dipping into the carpel, is somewhat 
variable; sometimes it runs along just outside the inner surface of the carpel, 
