Agriculture Reference
In-Depth Information
Figure 2.10. Yield levels and crop loss (Campbell and Madden, 1990b).
Experimental methods are mainly based on yield comparisons between infected
and healthy plants or between plants with different disease severities using field
plots, microplots (hill plots), single plants or tillers; between resistant and
susceptible varieties; between infected plants and plants treated with fungicides; or
between healthy plants and plants where disease damage has been simulated by the
removal of essential plant organs, such as the flag leaf on a cereal plant. In all of
these methods, it is important that experiments are properly designed so that results
can be analyzed statistically and, if possible, they should be repeated over a number
of seasons and in different areas. Gaunt (1995) pointed out that in the general area of
yield loss experimentation, two methodologies have gained less acceptance than
might have been anticipated: the use of disease gradients in large field plots and the
use of non-replicated designs. In both cases, difficulties have arisen with experi-
mental design and interpretation of results.
2.6.4 Empirical yield loss models
Each experimental method, such as those described above, will always have
advantages and disadvantages but conventional field experiments will probably
continue to provide most information for the mathematical modelling of disease-
yield loss relationships. Most models describe losses at the field level; the type and
complexity depend on the pathosystem and the disease descriptor employed as the
independent variable using least-squares regression analysis. Such models do not
directly use information on crop physiology and are, therefore, empirical or
descriptive rather than mechanistic in nature. However, they can be consistent with
