30 JOURNAL OF THE EOYAL HORTICULTURAL SOCIETY. 
they rest on the same broad basis — a basis of trial and error — but that 
with the practical man the results are reached by a longer process. One 
horticulturist finds that he obtains greater success when he adopts one 
method than when he adopts another, and he therefore concludes, per- 
haps too hastily, that the first method is the correct one ; his fellow 
worker, however, working under slightly different conditions, may come 
to an opposite conclusion, and it is not till the various experiences of 
many workers, extending often over a great length of time, are brought 
together and compared, that the proper method of procedure can be 
settled. Experience gained in this way must, no doubt, lead to correct 
conclusions in the end, but it is a slow process, as we must acknowledge 
when we consider how horticultural practice has varied from generation 
to generation, and how, even at the present day, eminent authorities hold 
different opinions on some of the most elementary points of their craft. 
The object of the scientific worker is simply to shorten this process, and, 
instead of merely drawing conclusions from what may happen to come 
under his notice, he carries out experiments specially designed to lead to 
definite conclusions. Once the facts are established, theory may certainly 
come in to help us to explain these facts and to suggest further investi- 
gations ; but such theories are not the empty phantasies which are 
sometimes characterised by such a name, nor can they ever, if true, 
be opposed to facts, being, indeed, merely the co-ordination of facts 
themselves. Nothing is so procreative of life and development in any 
branch of knowledge as a thoroughly good theory, for it knits together 
hitherto disconnected facts, and explains their meaning and their depen- 
dence on each other, and, perhaps, the next best thing to a thoroughly 
good theory is a thoroughly bad one, for it sets everyone to work to upset 
it, and the result is a healthy spirit of investigation and activity. 
With rigorous scientific experiments in agriculture we have long been 
familiar, thanks, in the first place, to devoted EngHsh investigators, and 
more recently to foreign and colonial Governments, who have far out- 
stripped us in their efforts to place agriculture on a firm basis of 
knowledge. In horticulture, however, such work is but little known in 
England, though it forms an important subject for investigation at most 
of the foreign experimental stations. The^e are, indeed, many difficulties 
in dealing with horticulture in the same way as agriculture. Foremost 
amongst these is the great space occupied by a tree as compared with that 
occupied by plants used in agriculture. Every tree or plant, just as in 
the case of an animal, possesses a certain individuality, and exhibits 
differences in its constitution, and differences in behaviour under the 
same circumstances ; it is, therefore, impossible to arrive at any reliable 
conclusions from experiments which are not made on a large number of 
individuals. In agriculture this is not difficult. A quarter-acre plot of 
wheat will contain some 300,000 separate plants, and the mean of the 
results obtained from such a number will render insensible the variations 
due to the individual personalities of the separate plants ; but a single 
experiment with the same number of trees at 10 feet apart would occupy 
700 acres, and to undertake a set of experiments on such a scale would 
be somewhat impracticable ; while, if a quarter of an acre is assigned to 
the experiment, we should get into it only 100 trees, and the result 
