| Because of the growing
popularity of the beat sheet method to keep track of insect numbers
in cotton, we have initiated studies to compare current insect sampling
methods, and determine conversion relationships between them.
Accurate sampling of pest and beneficial insect populations in cotton
crops is an important part of an Integrated Pest Management (IPM) strategy,
as it provides the foundation for pest management decisions. It is important
that growers and consultants have fast and effective methods of quantifying
insects so decisions are timely and accurate.
Among most cotton consultants, the beat sheet technique is gaining popularity
as a useful sampling method, particularly for monitoring beneficial
insects. But the higher counts of beneficial insects found by beat sheet
sampling compared with other methods have implications for the use of
the predator-prey ratio in pest management decisions as the ratio in
use is based on visual counts.
Studies carried out by Brad Scholz and colleagues at QDPI Toowoomba,
compared the beat sheet method with visual and d-vac suction counts
during the first nine weeks of the season (The Australian Cottongrower,
September-October 2001 p 14–17). They found that the beat sheet
gives a more realistic measure of the number of predators in cotton
at this time of the year and provides on the spot information.
The Australian Cotton CRC awarded a Summer Scholarship to Carla McKinnon
from the University of Western Sydney for a study which compared beat
sheet, visual and suction (d-vac) insect sampling methods over the whole
season, and determined conversion factors between the methods. The work
also involved examining the relative efficiency and reliability of the
various ways of sampling throughout the season.
What we did
Sampling was conducted in 12 fields on six farms in the Namoi and Macintyre
Valleys with each site being sampled fortnightly. To make data directly
comparable between the sampling methods (visual, beat sheet and d-vac),
all collections were made on a per metre basis.
During each visit to each field we sampled 12 metres visually, 12 metres
with beatsheets and 12 individual metres with a d-vac. Just over two
km of cotton row were sampled and almost 50,000 insects counted throughout
the season.
To simulate conditions that pest managers would normally experience,
samples were processed in the field and only those insects visible to
the naked eye were counted. Previous research by Mark Wade and his colleagues
(The Australian Cottongrower, November–December 2001, p 41-42)
has shown that sampling results are sensitive to the time of day.
Less insects are found during the hottest parts of the day. To avoid
this bias we completed all sampling for this study during the morning
or very late in the afternoon.
The beat sheet measured 1.5 metres by two metres and was placed in the
furrow so it extended up and over the adjacent row of cotton. A one
metre stick was used to vigorously push and shake the plants 10 times
against the plastic sheet, with the beats moving from the base to the
top of the plants. All insects dislodged onto the yellow canvas were
counted immediately. The beat sheets were washed and sterilised with
Farmcleanse between sites.
‘Pounce-net’
In addition to the three sampling methods under study, an absolute count
was carried out to try and obtain the actual numbers of insects per
metre for comparison with the relative numbers collected from the other
methods. To achieve this we developed a new technique using what we
have called a ‘Pounce-net’.
The ‘Pounce-net’ is a large elasticized gauze bag which
two people stealthily and swiftly placed over one metre of row, while
another two people cut the plants off at the base while drawing the
net closed. In the laboratory, the bags were fumigated to allow a comprehensive
count of all captured insects visible to the naked eye.
What we found
Results indicated that the beat sheet method was very effective for
sampling predatory insects and spiders. Beat sheet counts consistently
detected a higher proportion of the ‘absolute’ count of
total predators based on pounce-nets than other sampling methods (Figure
1), which confirms previous findings by Brad Scholz and colleagues.
The differences between the absolute counts and the other sampling methods
were not attributable to any particular insect group, but were spread
across all types of predatory beetles, bugs, lacewings and spiders.
At the start of the season (eight plant nodes or less), all sampling
methods produced similar results, but as the cotton grew the beat sheet
generally found twice as many predators as visual samples (Figure 2).
Our visual samples were undertaken very carefully and involved thoroughly
checking every part of every plant within each metre sampled.
Different types of insects vary in their detectability by different
sampling methods. For example, eggs tend to stick to the plant and may
not be easily dislodged or may be difficult to see on the beat sheet.
Similarly, grubs may burrow into squares and bolls and may not be easily
beaten or sucked out. Conversely, mobile insects such as green mirids
and red and blue beetles are flighty and easily dislodged from the plant.
They may also actively hide and/or escape undetected. Our results confirmed
that visual sampling is clearly superior to d-vac or beat sheet for
monitoring all Helicoverpa spp. life stages (Figure 3). Visual sampling
is also best for mites, aphids, thrips and whitefly — for which
standard sampling protocols are available.
In contrast, beat sheets are more effective for counting mobile predators
like beetles, bugs and spiders and some pests such as mirids. So sampling
methods may need to be combined in order to get an accurate assessment
of the overall insect populations.
Perfect timing
Our results showed beat sheets are faster than visual checks. They took
between three to five minutes throughout the season, while the visual
checks took increasingly longer as the crop developed (Figure 4).
Anyone can beat sheet!
A secondary study looked at the variability between the performance
of different crop scouts for beat sheet and visual sampling. We wanted
to establish which method was the most reliable regardless of the scout.
In an experiment designed by Dr Sarah Mansfield, five scouts with a
range of checking experience conducted equal numbers of visual and beat
sheet samples.
This experiment was conducted in January and repeated in February in
a field of unsprayed conventional cotton. The results clearly showed
that there was a statistically significant difference in the number
of predatory insects and spiders that different scouts were finding
using visual checks.
On the other hand, there were no significant differences between checkers
using beat sheets to count total beneficials. Beat sheets are less subject
to scout bias because they do not require the same amount of experience
and skill as visual searching.
Making decisions with beat sheet data
Beat sheets provide a fast, effective and robust method for monitoring
predatory beetles, bugs, lacewings, ants and spiders in cotton. They
are also an effective method for sampling mirids, and we will present
our mirid results and recommendations in a future Cottongrower article.
Visual checking of insect densities in cotton is still essential, particularly
for Helicoverpa and other lepidopteran pests like armyworm and tipworm.
Visual checks are also required for pests that are small or have patchy
distributions such as aphids, mites, thrips and whitefly.
For a reliable estimate of insect densities it is a good idea to regularly
beat a similar number of metres to the visual checks. Beat sheet counts
of predators should be converted to visual densities before use in management
decisions.
Our results suggest that once a crop reaches nine nodes or more, beat
sheet predator numbers can be converted to an approximate visual count
equivalent by dividing the beat sheet counts by two (Figure 2). The
converted count can then be used with the predator to pest ratio calculation
described in the IPM Guidelines.
For example, a beat sheet sample that detected 10 predators per metre
would on average be equivalent to five predators per metre counted in
a visual sample.
|
