The problem of immunity to disease in plants is now receiving a large amount of attention from plant-pathologists. It has long been known that different varieties of the same species of plant may differ in their susceptibility to the attacks of fungi and possibly other parasites, but even now practically nothing is known of the factors which determine whether a plant will prove immune to a disease or not. The following paper is a record of investigations which it was hoped would contribute something towards the solution of this problem, or which might at least suggest lines of further research on this subject.

During the past few years we have grown a large number of plants in partially sterilised soils alongside of others in untreated soils, in order to compare the total weights of dry matter produced. While the experiments were in progress certain qualitative differences in growth and habits of the plants revealed themselves; we propose in this paper to record these differences and to set out certain of the hypotheses that have been put forward to account for them. It was soon found that several distinct problems in plant physiology were concerned, and that any attempt to test the hypotheses experimentally would involve us in a number of side issues and lead us away from our main purpose. We therefore for the present confine ourselves to a statement of the facts observed, leaving their more complete elucidation for later work.

The present paper is an account of observations and experiments in connexion with Silver-leaf Disease which have been made since the publication of my paper on the same subject in the Journal of Agricultural Science for 1911. I am indebted to the Board of Agriculture for giving a grant to assist this work and to the Cambridge University School of Agriculture for placing at my disposal a plot of land on its farm for most of the fruit trees used in the experiments. I have also to thank Mr R. I. Lynch, Curator of the Botanic Gardens, Cambridge, for providing accommodation for other trees.

In a previous paper on the “Interpretation of Milk Records” the following points were dealt with:

I. Selection of a figure definitive of a cow's milking capability.

II. The influences affecting such a figure.

(a) The usual methods of describing a cow by her total yield per calendar year, per lactation, or per average week are inconvenient both for practical breeding on a large scale and for definite enquiry on the inheritance of milk yield or its possible correlation with other characters. The figures are inconvenient because a variety of circumstances which affect them must be stated for every individual case before such figures can be reliable. Chief of these circumstances are age of cow, length of lactation, number of weeks dry before calving, interval between calving and subsequent service, and time of year of calving.

The practical advantages accruing from the use of lime are now generally recognised, but on account of the complexity of the changes induced it is by no means easy to determine, in the case of a field soil, the exact extent to which one effect or set of effects has contributed to the general improvement of the soil in relation to its crop-producing power.

Results of our previous work were published in the Journ. of Agric. Sri., IV, 1911. There, also, a brief summary of the chief papers published on the subject of the fat globules was given, to which it is unnecessary to refer in detail here.

When the work was first commenced in 1909, the problem before us was the consideration of the variation in the size of the fat globules, with relation to churning, as regards the different breeds of cattle. The breed was considered because it was usually supposed that this was one of the chief factors which influenced churning. The most definite result of our work was that it was shown that consideration must be given to the character of the milk, irrespective of the breed. This conclusion is quite contrary to that of other workers, but an examination of their figures shows, undoubtedly, that this is actually the case: the results given by Woll (Digestion Expts., Seventh Annual Report, Agric. Expt. Stat, Wisconsin, 1890, 238; also Agric. Sci., 1892, vi, 445) emphasise this point particularly. It is also shown by this year's work, as may be seen from Tables V–XI.

The comparative size of the globules has been worked out very thoroughly by Gutzeit, who measured the mean volume of the globule. Other workers give their results as “relative sizes.” In our work we attempted to ascertain the distribution of the fat in the globules, and to this end the number of globules of each size was determined, and curves were drawn. The result of this, however, was negative. At that time we were considering the breeds of the cows from which the milks were obtained; had we considered them solely as milks of a certain mean size of globule, much more might have been achieved.

An apparatus was devised to give an absolute figure for the churnability of any cream, but, until the effect of the other factors has been determined, it is impossible to interpret the results. Some experiments to ascertain the optimum temperature were described also.

In investigating the influence of various factors on the character of milk, it is now recognised that it is insufficient (at least for many purposes) to make comparisons of the percentage of fat, but some regard must be paid to the distribution of the fat in the globules.