Results of Water Culture Experiments. 435 
often the case, where one factor is predominatingly the limiting one, to show 
this is so, and it is also possible to show that a factor is not the limiting one 
when alterations in its magnitude have no effect on growth. 
The Limitations of the Water Culture Method. 
It must be acknowledged at once that experiments with water cultures 
have in the past yielded information of first importance. Thus it was 
by their use that the elements necessary for plant growth were demon- 
strated. More recently in Dr. Brenchley’s hands they have given much 
information as to the toxic properties of various substances. But it is 
necessary to bear in mind the limitations of the method. In the form 
in which it has hitherto been employed, the essential of the water culture 
method is to control the amount of substance added to the nutrient solution 
and to measure the growth corresponding to this. But growth is the result 
of such a complex of factors, scarcely any of which are completely under 
control and many of them not at all. An understanding of the physiology 
of the plant can only be brought about by careful analysis of factors 
carefully controlled, and by accurate quantitative measurement. 
Again, owing to the variability of individual plants, the method leaves 
much to be desired as regards accuracy, unless large numbers of plants are 
dealt with and the probable error of experiment calculated. This makes 
the method very laborious. 
Dr. Brenchley’s last published results are probably as accurate as can 
be obtained by the method, and she has very properly calculated the 
probable error. The mean of each set of ten plants growing under the 
same conditions had a probable error range which was often about 4 per 
cent, of the mean value, rarely below that, and which was frequently 
as much as 9, 10, or even 11 per cent, of the mean. Hence for a difference 
between two mean values to have a reasonably certain significance in these 
experiments it would be necessary for the means to differ by about 10 per 
cent, to 15 per cent, in favourable cases, and by as much as 30 per cent, or 
even more in unfavourable ones. 1 This cannot be regarded as very satis- 
factory in comparison with the decree of accuracy which can be obtained 
by other physiological methods which at the same time are much less 
laborious. Nevertheless for the solving of some problems the method 
may be the only one available. 
Summary. 
1. Growth, the resultant of a number of different processes, is dependent 
ultimately upon many separate factors, each one of which may limit 
the rate of growth. 
1 See, for example, Gray, F. W. : A Manual of Practical Physical Chemistry, London, 1914, 
chapter i. 
