Agriculture Reference
In-Depth Information
This equation was first successfully tested (
r
2
= 0.99) on a plot assessed every two
weeks after growth stage 30 (Zadoks
et al
., 1974). It was then fitted to data
concerning take-all build-up for each experimental treatment at three sites (three
regions of France), where different cultural practices (sowing date, sowing density,
total nitrogen dose, nitrogen fertiliser form, burial or removal of preceding crop
residue) were tested. The parameters
c
1
and
c
2
were estimated for each experimental
treatment at each site.
A linear model was tested to interpret
c
1
and
c
2
for each set of estimates at each
site as a function of the factors analysed and co-variables measured. Sowing date
always affected
c
1
(e.g. primary infections) whereas
c
2
(e.g. secondary infections)
was influenced by sowing date only at the most favourable sites for disease (e.g.
those with the highest infection rates, due to favourable climatic conditions). Early
sowing systematically increased
c
1
. This is consistent with previous results (Hornby
et al.,
1990) and the fact that early sowing provides a longer period for infection
before winter. The effect of early sowing on
c
2
was variable, positive for one
experimental site, negative for another.
A positive correlation was found between plants m and parameter
c
1
,
but only at
the most favourable sites. Plant density, like sowing date, had a variable effect on
c
2
.
A high plant density at early stages, when the roots are still few in number and short,
probably increases the chance of contact between the soil inoculum and living roots,
whereas it has a less predictable effect when the root system is well developed.
High levels of nitrogen application increased
c
1
and decreased
c
2
but both these
parameters were decreased by applications of nitrogen in the form of ammonium. As
reported by Sarniguet
et al.
(1992a,b), nitrogen can stimulate both the pathogenic
and the antagonistic microflora. Increases in the antagonistic microflora early in the
infection of seminal roots facilitate the development of fluorescent pseudomonads
on necrotic tissue. These pseudomonads then interfere with pathogen expansion,
particularly if nitrogen fertiliser is applied in the form of ammonium.
The hierarchy of and interactions between various factors were shown to be
important. Factors other than sowing date were generally significant only if sowing
date was also significant. Sowing date may therefore be considered the dominant
factor, and its interactions with other factors as the most important. This type of
interaction is very similar to that observed for site: several factors had a stronger
influence or were only significant if the site was favourable for disease
development. Each factor seemed to amplify the risk due to the other effects and
factors with a weak effect influenced disease only if factors with a strong effect were
also present.
This model was used to assess the efficacy of new methods of control, such as
the use of fungicidal seed treatments (Fig. 14.2). It was found that, in an early
epidemic, the fungicide significantly reduced take-all incidence during all or most of
the cropping season whereas, in late epidemics, it decreased incidence only
moderately. Seed treatment was shown to reduce incidence by delaying primary
infection (Schoeny and Lucas, 1999).
-
2
