e delicate and perfect things to be brought up by watch through day and
night, than her recumbent clusters, trickling, sometimes almost gushing
through the grass, and meeting in tiny pools of flawless blue.

10. I will not attempt at present to arrange the varieties of the
Giulietta, for I find that all the larger and presumably characteristic
forms belong to the Cape; and only since Mr. Froude came back from his
African explorings have I been able to get any clear idea of the brilliancy
and associated infinitude of the Cape flowers. If I could but write down
the substance of what he has told me, in the course of a chat or two, which
have been among the best privileges of my recent stay in London, (prolonged
as it has been by recurrence of illness,) it would be a better summary of
what should be generally known in the natural history of southern plants
than I could glean from fifty volumes of horticultural botany. In the
meantime, everything being again thrown out of gear by the aforesaid
illness, I must let this piece of 'Proserpina' break off, as most of my
work does--and as perhaps all of it may soon do--leaving only suggestion
for the happier research of the students who trust me thus far.

11. Some essential points respecting the flower I shall note, however,
before ending. There is one large and frequent species of it of which the
flowers are delicately yellow, touched with tawny red, forming one of the
chief elements of wild foreground vegetation in the healthy districts of
hard Alpine limestone.[26] This is, I believe, the only European type of
the large Cape varieties, in all of which, judging from such plates as have
been accessible to me, the crests or fringes of the lower petal are less
conspicuous than in the smaller species; and the flower almost takes the
aspect of a broom-blossom or pease-blossom. In the smaller European
varieties, the white fringes of the lower petal are the most important and
characteristic part of the flower, and they are, among European wild
flowers, absolutely without any likeness of associated structure. The
fringes or crests which, towards the origin of petals, so often give a
frosted or gemmed appearance to the centres of flowers, are here thrown to
the extremity of the petal, and suggest an almost coralline structure of
blossom, which in no other instance whatever has been imitated, still less
carried out into its conceivable varieties of form. How many such varieties
might have