Agriculture Reference
In-Depth Information
because the rate of evolution is proportional to the additive genetic variance
(Hedrick, 2004). However, interpretation of effective doses other than the ED
50
is
not necessarily straightforward. The widely-used probit model assumes that the
logarithm of
tolerances
is normally distributed in the pathogen sample (Finney,
1971), an individual's tolerance being the dose of fungicide which is just sufficient
to kill it. If tolerance is not normally distributed - if, for example, there are several
discrete levels of resistance in a fungal population, as is the case for resistance of
B.
graminis
f.sp.
hordei
in the UK to triazoles (Brown and Wolfe, 1991; Brown
et al.
,
1991a; Blatter
et al.
, 1998) and to morpholines (Brown
et al.
, 1991b; Brown and
Evans, 1992) -EDs calculated by probit analysis may be inaccurate (Brown, 1998).
3.3 POPULATIONS AND SAMPLES
The aim of most surveys is to understand the composition of a pathogen's
population and so to assess the value of different ways of controlling the disease.
Some surveys, such as the UKCPVS (Bayles
et al.
, 1997), surveys of cereal rusts in
the USA and Canada (Kolmer, 2005), and those of barley mildew and wheat yellow
rust in Denmark (Hovmøller, 2001, 2004; Justesen
et al.
, 2002), collect data which
are applied quite directly to disease control by indicating which resistant varieties
and fungicides are most likely to be effective. Many other pathogens have been
surveyed less frequently for similar practical purposes. Other surveys are done to
answer more general questions about the structure or evolution of a population,
often to support the medium to long-term development of disease control strategies.
When planning and conducting a survey, one should aim to be clear about what
population is being sampled and should understand how different schemes for
sampling from a population might affect the interpretation of survey data.
3.3.1 What is a population?
(a) Defining a population
There are two reasons why a pathogen population must be defined carefully before
carrying out a survey. The first is to ensure that the population sampled is
representative of the population of interest. This is especially important if a survey is
intended to support disease control measures. Secondly, one should be sure that the
samples collected are representative of the population from which they are drawn.
There may be no single definition of a population which applies to a pathogen
species in all circumstances. Indeed, it may not be easy to define a population at all.
Milgroom and Fry (1997) suggest that a useful definition of a pathogen population is
an operational one:
A fundamental concept ... is that the samples ... are so structured that the inferences
made can be applied to more general phenomena beyond the objects being observed.
The key question is to what populations will inferences be made?
In other words, the pathogen population to be sampled in a survey should be typical
of the population to which disease control is to be applied.
