Agriculture Reference
In-Depth Information
1986). Using much more contrasting GUAs of 0.003 m
2
and 11 m
2
with wheat-yellow
rust and wheat-leaf rust, Brophy & Mundt (1991) again determined that the smaller GUA
was more effective at reducing disease in mixtures. However, this was not always the
case, as Mundt & Browning (1985) found that increasing the scale at which component
cultivars were deployed from a GUA of 0.003-0.84 or 0.58 m
2
, had no effect on the
crown rust infection of oats. Similarly, Mundt & Leonard (1986) found no differences in
maize rust infection between GUAs ranging from 0.21 to 1.88 m
2
. Differences may be
due to particular pathogens and spore dispersal conditions and models often do not take
into account the dispersal mechanisms of the pathogen. For rhynchosporium (
R. secalis
)
on winter barley, the optimum GUA or patch size was determined at about 4 m
2
in one
study (Newton & Begg, 2008), which lead to experiments to determine whether a patchy
arrangement of component cultivars might be better than homogeneous mixing even
within fi elds. Patchy sowing was found to both reduce disease and increase yield more
than homogeneous sowings compared with the mean of monocultures sown alongside,
even in the absence of disease (Newton & Guy 2008). Controlling the spatial nature of the
'patchiness' at a fi eld scale could be problematic but this work clearly shows that crude
patchiness was easy to deploy in practice, as no pre-mixing of seed was necessary, and
whilst it didn't always result in big yield increases, it did deliver disease reduction and the
same or more yield than homogeneously pre-mixed seed. Experimental data also indicate
that for powdery mildew, in patches (rows) the selection for complex races was much less
intense than in random mixtures (which have smaller GUAs) (Huang
et al
., 1994), so a
patchy fi eld sowing should reduce selection for complexity also.
One of the biggest mixture experiments was on rice, where a diversifi cation programme
in Yunnan province demonstrated effective control of rice blast. In the fi rst year, the area
was 812 ha, expanding to 3342 ha in the second year and comprising row mixtures of
susceptible and resistant cultivars. This achieved 94% less severe rice blast than when
grown as monocultures and increased the yield by 89% (Zhu
et al
., 2000). Clearly scale
was crucial to this experiment, echoing the experience in the former German Democratic
Republic, where up to 92% of the spring barley crop was grown as mixtures (Wolfe, 1997).
Mildew declined from over 50% to less than 10%, thereby reducing the fungicide require-
ment substantially. Numerous cultivars were used, but most used the same resistance
and yet it was still a success. Other notable large scale successful uses of mixtures are
Poland, where over 90 000 ha of cereal mixtures are reported (Gacek, 1997), Denmark
with 62 000 ha in 1996 (Munck, 1997), the USA where, in Oregon, 10% of the soft white
winter wheat and 76% of club wheat (Mundt, 1994), and in Kansas 7% of wheat (Bowden
et al
., 2001)
are grown as mixtures. On a smaller scale, mixtures are supported in the
Swiss low input 'Extenso' cereal production programme (Merz & Valenghi, 1997) and
in Scotland, winter barley mixtures have been grown for some years and winter wheat
mixtures are increasing (2007) for distilling (Newton AC, unpublished data). It is also
common practice to grow cereals as mixtures in many developing countries where higher
total yield, better yield stability, better food quality, animal feed and resistance to pests
over time compared with monocultures are quoted as the rationale (Smithson & Lenne,
1996; Woldeamlak
et al
., 1998a, 1998b), and the stability of wild crop relatives from
which landraces were derived is cited as not only sources of resistance, but also indicators
of good ecological practice to be re-discovered in modern agriculture (Akem
et al
., 2000;
Wood & Lenne, 2001).
