|
Sampling and IPM decision making
Project
personnel: David
Hagstrum, Bh.
Subramanyam, JamesThrone,
Paul Flinn, Dirk Maier,
Frank
Arthur
Sampling is an integral component of
IPM. Several stored-product insect sampling
tools and statistical methods for analyzing sampling data were presented
by Subramanyam and Hagstrum (1995).
The first sequential sampling plan for rusty grain beetles
infesting wheat was developed by Subramanyam et al. (1997), and a
generic mean-variance equation for
analyzing stored-product insect sampling data was developed by
Hagstrum et al. (1997). Hagstrum et al. (1998) showed how insects in
grain samples can be estimated from
catches of insects in probe traps inserted into the grain. Information
collected from farm bins and commercial elevators in Kansas, Indiana,
and Oklahoma will be used to develop
sampling plans to estimate insect density. The current Federal
Grain Inspection Service threshold for infested grain of 2 live
insects/kg of grain will be used as
the threshold density. The sampling plan’s performance in correctly
classifying density with respect to this
threshold will be determined. The sampling plan will
be incorporated into the Stored Grain Advisor expert system.
The Electronic Grain Probe Insect
Counter (EGPIC), Patent No. 5,646,404 (Shuman
et al. 1996, Litzkow et al. 1997) is an automated system that displays
real-time data indicative of local
insect densities from infrared-beam sensors located throughout stored
commodities (Shuman and Epsky 1999, Arbogast et al. 2000, Epsky and
Shuman 2000). Automated data
collection can provide an early warning, allowing a manager increased
control options, such as the use of a minimal amount of pesticide or a
non-toxic alternative control
measure (controlled atmosphere, aeration, etc.), and it can also be used
to judge effectiveness of a treatment.
Already there is an EGPIC working group focusing on
various aspects of utilizing the sampling data for making pest
management decisions.
A near-infra red (NIR) spectrometer
has been coupled with a singulator to automatically
feed individual wheat kernels to the spectrometer. We have used this
NIRS unit to differentiate wheat
kernels containing internally-feeding insect pests (Angoumois grain
moth, lesser grain borer, rice weevil) from uninfested kernels (Dowell
et al. 1999). Accuracy is over 95%
when 3rd or
4th instars,
pupae, or adults are present inside the kernels.
We are currently conducting studies to improve accuracy of
classification of kernels containing
1st and
2nd instars.
The technology would be a good replacement for the
X-ray method for detection of internally-feeding insects because NIRS is
much faster. NIRS was also used to
detect parasitoids of rice weevils within wheat kernels. We were able
to distinguish kernels that contained internal insect pests, kernels
that contained internal parasitoids
that had attacked the internal insect pests, and uninfested kernels with
100% accuracy (Dowell et al. 1999). This
technique can be used by companies that massproduce beneficial
insects to sort parasitoids of a given stage for shipping and subsequent
release in augmentative biological control
programs. We have also been able to identify adults
of the most common Coleopteran pests of wheat using NIRS (Baker et al.
1999). Adult beetles can be
identified to genus with over 95% accuracy. We are currently developing
an automatic sorter so that a sample could be sorted based on
classification. For example, wheat
kernels could automatically be sorted into uninfested and infested
kernels. An automatic sorter could be used
for grain cleaning (e.g., removing insectinfested kernels)
or for sorting by classification (e.g., sorting kernels containing
weevil parasitoid pupae). Other
on-going studies include use of NIRS for age-grading insects and
detection of insect fragments in flour. |
|
