Crop scientists use statistically controlled field trials to make all sorts of comparative measurements such as the yields of cultivars, the best spacing between rows, and within rows, the optimum fertiliser applications, and so on. The statistics are mathematically quite complicated, but modern computer software has largely eliminated this difficulty. These statistically controlled trials can be very accurate. However, the statistics are not accurate and, indeed, are positively misleading, when it comes to measuring pests and diseases.

Vanderplank (1963) first recognised this problem, which he called the ‘cryptic error’ in field trials. The problem is caused by the fact that crop parasites can move from one plot to another within a field trial, and this can ruin statistical analyses. The phenomenon was renamed inter-plot interference and, finally, became parasite interference, when it was appreciated that it was also highly relevant to the screening of individual plants in recurrent mass selection. This is because parasites can also move from plant to plant, and a resistant plant surrounded by susceptible plants can have its resistance obscured by parasite interference. Parasite interference can increase the amount of damage by several hundred times (James, et al, 1973) and, if not recognised can be incredibly misleading .

Parasite interference is possibly at its most misleading in a wheat breeding technique known as ‘head to row’ screening, also known as family screening. In this technique, all the seeds from one head of wheat are sown in a single row, or family. The idea is to select the best families first, and then select the best individual in each family. Genetically, this makes a lot of sense but, in breeding for resistance, it is a disaster. Consider the following diagram.

Each of the boxes labelled A-E represents a single row, a single head of wheat. A red box has the maximum parasitism, while a green box has zero parasitism. A yellow box has a level of parasitism approximately halfway between the maximum and the minimum levels of parasitism. The red boxes A, C, & E all have vertical resistance gene 1 and their resistance has been matched. They have low levels of horizontal resistance and, consequently, they have the maximum parasitism. They are also interfering with Rows B and D, and are loading them with parasites in the process of parasite interference.

Row B also has vertical resistance gene 1 and the parasite interference is consequently matching allo-infection. But row B also has a high level of horizontal resistance. Unfortunately, this horizontal resistance cannot be observed because it has been swamped by matching parasite interference. This is the family that the breeder should have kept but which was invariably thrown out.

Row D, on the other hand, has vertical resistance gene 2 and the parasite interference is consequently non-matching allo-infection, and D looks perfect, except for damage due to hypersensitivity flecks (see below). But its very low level of horizontal resistance is completely obscured, and its vertical resistance is unstable. It is the very opposite of perfect. This is the family that the breeder should have thrown out but which was invariably kept.

When working with head-to-row plots, the parasite interference can be so intense that the hypersensitive flecks of non-matching allo-infection appear to be a serious disease. Small grain cereal breeders used to point out that this was not true disease, and that such damage would not occur in farmers’ fields, where the non-matching allo-infection was negligible. But, it seems, they never did consider the effects of a comparable interference on a high level of horizontal resistance, when the allo-infection was all matching infection.

These misleading effects of parasite interference are the main reason why crop scientists neglected horizontal resistance in favour of vertical resistance for most of the twentieth century.

Finally, when working with recurrent mass selection, it is imperative to use relative assessments of resistance. That is, the least parasitised individuals are selected, regardless of how severely parasitised they may be. Most of that parasitism will be the result of interference which will not occur in farmers’ fields.

References:

Van der Plank, J.E. (1963): Plant Diseases: Epidemics and Control. Academic Press, New York & London, 349pp.

James, W.C., Shih, C.S., Callbeck, L.C., & Hodson, W.A. (1973): Interplot interference infield experiments with late blight of potato (Phytophthora infestans), Phytopath., 63: 1269-1275.