392 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
demonstrate the very wide choice of a root system which is open to the 
fruit-grower even amongst Paradise types. There are the very vigor- 
ous types amply furnished with fibre, such as Types I. and VI.; there 
are Types III. and V., of more moderate vigour but with a deep-rooting 
tendency ; and there is Type IV., the main lateral roots of which were 
found generally not more than four or five inches below ground. Types 
IV., VIII., and IX. probably represent the most dwarfing systems, 
whilst Types II. and VII. would seem to afford quite a different choice. 
At the present moment the lack of any commonly recognized 
nomenclature, together with the knowledge of probable mixtures 
amongst Paradise stocks, makes it very difficult to generalize from 
existing bush plantations as to what the actual effects of these various 
types have been in the past. This information could, of course, be 
obtained through a very extensive investigation by means of taking 
root cuttings from existing trees. Certain facts are indeed patent 
from established plantations. There is ample evidence of inequalities 
of growth and maturity due to a mixture of stocks, a fact too not 
infrequently betrayed through an examination of root suckers. The 
suckers of Types II. and III. are often seen, though it is not always 
an easy matter to recognize types in immature suckers. There may 
be some slight evidence for surmises, such as that on a strong soil the 
Doucin seems to make a very large bush ill-furnished with fruit 
spurs, and that many of the bushes that " never do " are on French 
Paradise stocks which are to be found unsuspectedly mixed with 
other types. But for the confirmation of such surmises and the 
impressions of growers there are at present few scientific data. Now 
that the chief Paradise types are easily identifiable, all reliable evidence 
along these lines will be of value. It is in the same light that one can 
re-read with profit the impressions of the old "lovers of planting." 
The aim should be, if possible, to regulate our varied requirements 
through the very diverse root systems at our disposal. If trees are 
worked on Paradise for purposes of garden dwarfs, to act as temporary 
" fillers " in the commercial plantation or to be the permanent trees, 
it would seem worth while making a study of the root system, even 
apart from possible refinements in improvement of colour, flavour, 
and the like. 
XVI. Summary of the Nine Types, 
Considerable space has already been given to the minute details 
of each Paradise type, details which it is hoped may serve as an 
ultimate guide as to type. At the same time it has been thought 
advisable to collect the most salient features distinguishing each 
type so as to present some sort of general impression. These 
summaries should afford a preliminary means to identification, to be 
supplemented, when a doubt exists, by the more minute details given 
in the comparative tables. Undoubtedly summer is the most favour- 
able time for beginning identification work. In July or August the 
leaf characters have assumed their most typical aspect. For this 
