INTRODUCTION. 
tivators will easily come to a rough general agreement on matters 
of beauty, the problem of size and admissibility will be settled 
differently by each gardener, according to the size of his garden. 
Speaking generally, my own rule has been to concentrate most 
of my attention on plants that do not exceed six or nine inches, 
and to deal more cursorily with larger subjects. To this vague rule 
various exceptions will be found; but I have tried to keep in the 
centre of my eye the needs of the rock-garden proper, rather than 
that of its less-exclusive fringes ; and this intention, for instance, 
has been my chief guide in selecting from among the beauties of 
such large and difficult families as Pentstemon, Salvia, Digitalis, 
and Verbascum, that usually tend more towards the herbaceous 
border than the rock-garden, so far as their stature is concerned. 
Various omissions and brevities may thus be accounted for; 
while yet others may be attributed to the fact that such plants 
are already dealt with adequately and copiously by catalogues. 
Against these there is no need to compete, and readers need not 
be troubled with long disquisitions on Squills, when these are 
fully treated by every bulb-list. As for such things as Snowdrops, 
catalogues here again offer the best forms; the rest are differen- 
tiated by such minute frecklings as to make their deciphering 
rather the delight of the expert, than any matter of moment to 
the general gardener. 
With regard to the matter of arrangement, I have usually 
followed a strictly alphabetical order. In certain races, however, 
certain very definite groups are formed, into which the species 
more conveniently fall. Here, accordingly, I have followed 
the distribution, arranging the species sometimes alphabetically 
in their classes (as in Draba), and sometimes botanically, as in 
the overcrowded race of Sempervivum; in which the members 
are far too subtly differentiated for each to be described at length, 
unless at a paralysing cost of space. Here, then, it is very much 
easier to have all the hairy-leaved species compendiously grouped, 
and then all the glabrous ones; a glance through the rapid list 
will at once deviate the inconvenience of the apparently dis- 
ordered arrangement. For larger and more specifically important 
xxii 
