39§ 
Notes. 
appearance to those in the skeleton. It is evident that such per- 
forations must accelerate the general decay of the epicarp, after which, 
however, all traces of them must disappear, as nothing is left but the 
netlike skeleton and the endocarp. Long before the decay of the 
epicarp has become general, the endocarp has split from the base 
towards the marginal ring of the capsule which bears the calyx-lobes. 
This splitting takes place in the same way as in Z. hypoleuca and 
other species, but the tenderness of the outer tissues of their capsules 
causes them to tear as the valves expand, whereas the skeleton of 
Z. macrostachys is sufficiently strong to resist that pressure. Thus the 
seeds are freed and find their way out, through the pores of the 
skeleton in Z. macrostachys, through the ruptures in Z. hypoleuca . 
The next question is if the normal dehiscence by two valves on the 
top takes place in Z. macrostachys . So far as the material at hand 
and Hildebrand’s observations go, it is not the case, though the 
anatomical structure is not of a kind to prevent it or to make it 
improbable. On the contrary, the very fruit which Dr. Zahlbruckner 
kindly sent, split after boiling, with a very slight pressure, along a line 
running from the persistent base of the style to the periphery of the 
top along which the normal dehiscence was to be expected. 
The obliteration of the normal terminal dehiscence in combination 
with the peculiar development of the mesocarp and the mode of 
dissemination depending thereupon seems to deviate from what we 
find, for instance, in Z. nicotianaefolia and Z. hypoleuca . But I must 
now introduce another species which forms a connecting link in 
a striking way, Z. Gaudichaudii , also a native of the Hawaian Islands, 
the close affinity of which to Z. macrostachys cannot be doubted. 
Here we have a thin cartilaginous endocarp splitting loculicidally to 
the base, and a network of vascular bundles which though finer and 
less strong than in Z. macrostachys acts in a similar way. It is strong 
enough to resist the pressure of the expanding valves, for a long time 
at least. The parenchyma decays more or less in the end, but without 
leaving such distinctly circumscribed pores, and its decay is more like 
that caused by ordinary maceration. The top, which is produced into 
a beak, opens either in the normal way, or it remains altogether entire 
or at least so long that meanwhile the dissemination by way of the 
decayed pericarp has already begun. With regard to Z. Gaudichaudii 
we might say it is in a state of transition towards becoming inde* 
hiscent, a state which seems to have already been attained by 
