THE WOIU RX EXPERIMENTS: THEIR OIUECT AND METHOD. 35 
tied in accepting them. We Lave, however, compared them in some 
fifty cases with the values obtained by determining the amount of new 
wood formed, as measured both by the number of new shoots and by the 
length of new wood on the tree, also, in a more limited number of cases, 
with the thickness of the leaf and the total leaf area on the tree. For the 
details of this comparison we must refer to our first " Report " (1897*) ; it 
must be sufficient to say here that the results are highly satisfactory, and 
that the values obtained by different methods differ only in cases where 
such differences might reasonably be expected. We are probably well 
within the limits of the truth when we estimate the error of our mean 
results with leaf measurement in any single experiment on eighteen trees 
to be within 5 per cent, of the true values. 
In certain cases, of course, such a method will and must fail ; but it 
is all-important to have obtained some means of general, though, 
perhaps, not of universal, application, whereby the results of horticultural 
methods may be practically gauged. Without such a method no reliable 
record of results is possible ; personal mipressions and recollections are 
useless in these matters. 
It is impossible in a short paper of this character to give any general 
account of all the subjects which are under investigation on our farm, 
and I have thought it better to try to give a correct idea of the general 
character of the work by entering into the details of the methods adopted 
in particular cases. One class of experiment we have intentionally avoided 
as far as possible, namely, the testing of varieties. It appears to us that 
very little good can come out of such testings. There is no doubt that 
the man who could conclusively prove the respective merits of the 
enormous number of varieties of certain fruits now known, and could 
reduce the list of valuable ones to five or ten per cent, of its present 
length, would be a public benefactor ; but this could only be done 
efficiently by means of numerous testing stations throughout the country 
all under one organisation. For one man to attempt to determine the 
merits of different varieties by planting them together in one particular 
field is absurd, and he might just as well leave the question of good and 
bad varieties to solve itself, as it is bound to do in the course of lime. 
We have, of course, collections of considerable numbers of varieties 
of most of the hardy fruits, but in making these w^e have always aimed 
at doing something more than the mere " testing " of varieties. For 
instance, with apples, we have over 100 varieties of the reputedly best 
sorts in one plot of ground, but each variety is grow^n side by side on 
different stocks ; a portion of each lot on each of the stocks will be sub- 
jected to a different system of pruning, while the w'hole plot can be 
divided into two or four similar sections for comparative experiments with 
different insecticides. A collection of pears, all of the same age at 
planting, has been made on similar principles. 
With strawberries w^e also have a collection of about a hundred 
varieties, and a fresh lot of each variety has been planted every year, so 
that we now have plants of all of them in five different ages, side by side, 
and the results, we hope, will give us data as to the powers of lasting, as 
well as of those of cropping, of these varieties. 
Eyie & Spottiswooile. 
P 2 
