454 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
must own that I have done in the past, as my own garden will 
testify, more bad work than good. 
First, let me say that in dealing with comparatively small 
stones, say anything under 2 or 3 cwt., it is, I think, a mistake 
to seek to get any large effect " out of them. Such stones are, 
perhaps, as usefal, " culturally," as rocks of as many tons ; but 
" effect " in any large sense can rarely be had from them ; and 
when this is not recognised, affectation and failure of attempt 
are the first suggestions conveyed, it may be unconsciously, to 
the onlooker. In such cases it is far better to let the stones be 
seen but little, even not at all. I hold it an error to make such 
stones (as is said) "go as far as they can," as by exposing the 
whole of them or of their largest sides. A method precisely the 
contrary should, I think, be pursued. Bury in the surface, and 
for from half to two-thirds of their bulk, any stones of compara- 
tively mean size. They will thus equally, or better, fulfil their 
general functions. Thus to leave but their corners exposed 
above ground is the best means of concealing their smallness, 
for the corner of a large rock buried deep looks but as big as that 
of a stone buried less. 
The stones should in general rest on their largest flat side, 
both for reasons of stability and of " naturalness." The " tilting 
on end " plan, which most of us have from time to time adopted 
" under stress," cannot, I think, be otherwise defended, and 
should be masked by suitable overgrowth where it has been 
unfortunately (or necessarily) adopted. The general " contour " 
of a rockery should, in my view, never be, avoidably, unnaturally 
regular. On the other hand, we should never, in our efforts 
to avoid that fault, be drawn into making it bizarre, or even 
unnaturally varied. Frost, weather, and water, are the tools 
with which nature shapes her rock-gardens, their lines and 
surfaces ; and if these tools will not suffer hardness of outline 
they will as insistently disallow freak and fancy in the work. 
And this leads mo to my last observation in this connection, 
as regards the shaping of the surface soil. At least in the cheap 
and useful" class of rockery which I am now considering, the 
soil surfaces left exposed should, in my view, be relatively large. 
Indeed, in many cases (as when little good stone is to be had) 
the tops of, say, the mounds may often best be furnished with 
plants only, to the exclusion of stone. I am, of course, speaking 
