PARK AND CEMETERY. 
405 
CYTISUS SCOPARIUS. 
VAR. ANDREANUS. 
Garden Plants, Their Geography, XIV. 
ROSALES (a.) 
THE GENISTA, ROSA AND DROSERA ALLIANCE. 
This group as 
it is constructed 
by botanists is 
one of the larg- 
est in the vege- 
table kingdom, 
containing no 
less than 47 
tribes, 666 ge- 
nera and 9,589 species. It is 
very fully represented too in 
nearly all parts of the world, 
except, it is said, that the small 
islands of St. Helena and 
Tristan d’ Acugna, are desti- 
tute of legumes. The species 
included, however homoga- 
mous they may be, are ex- 
tremely heterogeneous in as- 
pect, and so numerous that 
close selection becomes imperative for garden pur- 
poses. In arranging on the ground the mass of 
material is found so unwieldy that it is usually 
necessary to break the planting into semi-detached 
groups. None of the alliances are more fully 
represented in trees, shrubs and herbs, none are 
more beautiful in all sections, and but few are so 
full of plants useful to man. 
The determination of scientists has rested 
largely, I believe, upon the amygdaloid embryos, 
and these are often microscopical. It is a question 
if it would not often be better to seize upon more 
obvious characters. The papilionaceous flower, or 
leguminous fruit is pretty sure to obtain through 
nearly 9,000 of the species, a number sufflciently 
large and important to constitute an alliance, and 
upon the whole they seem sufficiently distinct from 
the bulk of rose flowers proper. I have never been 
able to see with the naked eye any great resem- 
blance twixt a strawberry and a bean, or between a 
Kentucky coffee tree and a sedum. Yet, they are 
forced into the same alliance, and it must often per- 
plex the tyro to know why. Microscopical investi- 
gation is impossible to the mass of mankind, and 
impracticable for most gardeners, who, though 
often willing enough to take the dicta of eminent 
investigators on trust, will often feel a shock to 
their understanding when they find the honey locust 
and the sundew grouped under the name of the 
rose. It does too much violence to the common 
understanding. 
I am led to these remarks by the very immensity 
of the Rosal Alliance, as well as by the diversity of 
the species. 
Nothing short of a woodland and an immense 
garden in departments could accommodate a fair 
representation of the tribes, and even for purposes 
of pure ornament and instruction, with the closest 
selection of species and varieties, their claims to 
recognition are so pressing, especially in the woody 
sections, that for garden purposes and the exigen- 
cies of embellishment it nearly always becomes nec- 
essary to break them into separate groups. The 
New York Botanical Gardens largely run to quarter 
million buildings, but the tender plants, their raw 
products and manufactures would fill them, and 
the 45 acre Arboretum would fail to ornamentally 
accomodate the hardy Rosales. 
The dictamen of the botanists has been spread over 
the earth, however, and it is not well to change it; 
topsy-turvy classification is especially confusing, 
otherwise I could wish that for garden purposes 
Rosales could be divided into (a.) Legumales; (b.) 
Rosales proper, and (c. ) Saxifragales. 
The alliance commences with two tribes, ConnarecB 
and Cnestidece, having about 170 species distributed 
over the tropical regions; none are nearer to the 
United States than Central America. 
Piptanthus Nepalensis is a monotypic genus 
from the Himalayas; it is known as the evergreen 
laburnum, and should be tried at the south. 
Therntopsis has thirteen species in Eastern and 
Northern Asia and in North America, mostly from 
