6i 
cal, with three longitudinal furrows or folds. Just occasionally 
two of these furrows may be seen as the grain lies on its side 
The surface of the grain is delicately reticulated. If pollen from 
an early flower of a well nourished plant be examined, numerous 
slender glistening viscin threads will be seen connecting the grains 
together. In pollen from late flowers or ill nourished plants 
threads are few or wanting, but the grains still stick together in 
groups. I take it that in these cases the viscin (or the cells provid- 
ing it) is absorbed more completely as nourishment by the grains, 
and only sufficient remains to render them just sticky. Late 
pollen is more dust-like than early pollen, and the anthers of late 
flowers empty more quickly than those of early ones. Thus it 
seems at first sight that early flowers have the best chance of 
pollination, but the numerousness of late flowers tends to counter- 
balance the dryness of their pollen. So far as I can judge any giv- 
en flower on the panicle has as good a chance of pollination as 
any other in the presence of suitable Insects. I have for the 
past three years kept a number of bushes under observation right 
through the flowering season, visiting them once or twice per 
week. When I have found flies plentiful on a shrub throughout 
the season, flowers on all parts of the panicle produce fruits with 
apparently equal freedom. I have not found time to make an 
exact census, but I feel satisfied that, if early flowers have any 
advantage, it is only a small one. Always those shrubs on which 
I have not seen pollinators produce few or no fruits. Sometimes 
mid-season and sometimes late flowers only produce fruits. The 
inference is that the shrubs were visited by flies only at these 
times. No specially watched shrubs have behaved so. 
Now what induces the flies to visit the Marianthus blossoms ? 
The flowers are not conspicuous from a little distance to human 
eyes, nevertheless I think appearance plays some part in attract- 
ing, because I have noticed that the most conspicuous of 
a group of shrubs is most freely visited by flies. Early 
in February, 1907, I placed some flowers in a vase of 
water in a cool cellar in order to keep them fresh. When I went 
to them next day I noticed a delightful though faint limonaceous 
odour, evidently emanating from the flowers. Up to this time I 
had regarded the blossom as odourless, for I had never been able 
to detect the least scent, though I had repeatedly smelled flowers 
carefully, before and after gathering them. I have several times 
since noticed the odour on perfectly calm days, when standing 
close to a shrub in the bush, without deliberately smelling the 
flowers. The perfume much resembles that of Pittosporum and 
is always developed when flowers are kept in a cool place free from 
draughts. The least breath of air disperses it. It seems to be 
particularly pleasing to house flies. If a branch of bloom with 
the scent well developed be taken into a room the flies fly to it 
from all quarters. I left a panicle dried in sand with the shape 
