An epidemiological analysis of the Phytophthora and Alternaria blight pathosystem in the Natal Midlands.

Abstract

The history of the development in Natal of a forecasting service to warn of outbreaks of late blight disease caused by Phytophthora infestans is presented. The late blight pathogen and Alternaria solani, the causal organism of early blight disease, interact on potatoes and tomatoes to form a blight disease complex. Evidence is presented to show that it is expedient to manage this blight complex as a whole rather than to direct control at only one of the components in ignorance of the consequential enhancement of the potential of the other. In a search for an improved blight complex management strategy, factors concerning the possible existence of an annual migration of Phytophthora infestans inoculum, first postulated in the 1960's, along an east-west route across Natal, are collected and collated. Corroboration of the existence of the Phytophthora-pathway is given, inasmuch as it represents a serial outbreak of late blight along a temporal gradient. The possibility that the pathway is a manifestation of disease resulting from the erruption of pre-existing inoculum along an environmental gradient, can not specifically be excluded. However, the peculiar pattern of anabatic and katabatic winds along a river-valley network, superimposed on a continuous cropping pattern and its concomitant opportunity for blight to be endemic in the province, supports the postulated Phytophthora-inoculum pathway A fungicide spray trial was conducted in order to investigate the possibility of us i ng the pathway phenomenon as the framework for an improved blight control strategy and to explore the nature and level of the competitive interaction between Phytophthora infestans and Alternaria solani. This trial revealed that the interaction between the components of the blight complex was differentially altered by weather patterns and fungicide combinations. Treatments in which metalaxyl (Ridomil) alone was used for the control of late blight, gave a yield similar to those with propineb (Antracol), which inhibits A. solani primarily but also hus some negative effect on P. infestans. The yields from both these treatments were siguificant ly (p < 0,05) better than the yields recorded in the unsprayed control plots. A treatment in which Ridomil and Antracol were combined such that each was applied according to its recommended concentration, gave yield increases of 32,3% over the unsprayed control, although the yield from the Ridomil/Antracol treatment was not significantly greater (p < 0,05) than the yields recorded where either Ridomil or Antracol were used. A computer simulator, named GAUSE, was developed to simUlate the consequences of the competition between various combinations of P. infestans and A. solani. Results simulated by GAUSE corroborated those obtained from the field trial and support the conclusion that diseases of complex etiology require more than simplistic, univariate analysis of single cause-and-effect pathways. The competition quotient CQ is developed as a new parameter of competitive interactions. It is calculated as the ratio of the amount of disease in the absence of competition, to the amount of disease when the causal pathogen is competing with another pathogen in the same niche. The CQ may be calculated from various standard epidemiology statistics and it is used to demonstrate that the competitive displacement phenomenon places constraints on the interpretation and application of Vanderplank's basic epidemiology equations. A new pathosystems management concept namely the pathotope (pathos = suffering; topos = place0 concept, is introduced, having developed from the notion that epidemics have spatial as well as temporal attributes. Accordingly, an area in which individual farms are at the same level of probability at risk to disease, delimits the pathotope. The concept can be described at many integration lsvels and is presented as an important quantitative unit of comparative epidemiology. The pathotope concept accomodates such notions as are contained in the postulated Phytopnthora-pathway and is especially suited to integration with disease forecasting methods. An example of the application of the pathotope approach is presented and a strategy is proposed by which fungicide spraying is initiated and applied synchronously as determined by the degree of communal risk to attack and epidemic increase of disease. Within a pathotope, several common factors collectively determine the vulnerability of the group to disease. If a coherent, uniform strategy is to be developed and implemented by pathotope members, it is necessary that all members have access to the relevant information and that it be collected and disseminated conveniently and rapidly. A computer-based disease monitoring and mapping system which achieves these objectives is presented.