368 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
in this country alone, and that each of these types was in the posses- 
sion of more than one nurseryman at least. 
(4) That as there appeared to be considerable variations in size 
of leaf, tendency to " feathering " &c, even in individual types, a 
very strict and detailed method of classification was required from 
the outset. 
Thus the preliminary examination demonstrated very clearly 
the need for this work, and supplied ample justification for the very 
detailed methods which it was proposed to follow. One instance 
is sufficient to demonstrate the misapprehensions under which fruit- 
growers are suffering. ' The Broad-leaved English Paradise ' is with- 
out doubt the most popular amongst growers to-day. It is the most 
asked for. Whilst we nominally received it in twenty-one collections 
called ' Broad -leaved English,' we actually received it only three 
times as a true sample. In the majority of cases the ' Doucin ' 
was sent as the ' Broad-leaved.' What the respective merits of 
these two Paradise stocks may ultimately be proved to be, we are 
not yet in a position to say, but that the stocks are very easily dis- 
tinguishable and possess widely different characters will be shown. 
Moreover, it is clear that the respective root systems of these two 
stocks, after four years of normal growth, develop along strikingly 
different lines, and it seems probable that this must affect the 
maturing tree. Yet growers to-day are largely planting trees on stocks 
that are not guaranteed as true to name, and which are, moreover, 
not infrequently mixed with several other varieties of ' Paradise.' 
It will at once be realized how difficult it is to make any progress 
in the comparative study of relative growth, hardiness, and economic 
value generally until this initial matter is set in order. 
VI. Cultural Methods Adopted. 
Though all the collections of samples did not arrive the same year, 
the general treatment adopted was uniform. Each collection, after 
being numbered and entered in the record book, was planted out as 
before described, and allowed to stand for one year. The following 
spring each stock was cut back to within two or three buds of the 
ground level. This caused the stocks to send out numerous shoots, 
which were moulded over several times during the summer. This 
was a simple method of forming stools, and of encouraging the young 
shoots to send out adventitious roots, where they were layered 
or moulded up. (Figs. 66 and 67.) These rooted shoots or layers, 
when severed from the parent plant at the end of the growing season 
and after leaf-fall, have afforded the means of multiplying each col- 
lection for the purpose of obtaining sufficient quantities of each type. 
The process has formed a very useful means too, as will be seen later, 
of judging comparative rooting vigour. 
Stocks treated in this way form permanent stools, and continue 
to send up a new supply of shoots yearly. 
