Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
MICRO-PLOT FIELD TRIAL SYSTEM
CROSS-REFERENCE TO RELATED APPLICATIONS
This application claims priority from United States Patent Application No.
60/948,438 filed July 6, 2007, which is herein incorporated by reference.
TECHNICAL FIELD
The present disclosure relates to seed variety testing system in particular to
small
plot research field trials.
BACKGROUND
The seed industry relies heavily on annual small plot testing to determine the
commercial merit of thousands of new plant cultivars. In the field, small
research
plots provide the only means to evaluate the agronomic and relative field
performance of new cultivars in comparison to existing cultivars. Seed
companies
rely on small incremental improvements of new cultivars in comparison with
existing
products to maintain competitiveness and progression in the marketplace.
Screening new cultivars is the largest expenditure of energy and capital in
the seed
development process. The standard method of field-testing uses conventional
two-
row replicated yield trials, which require substantial capital and labor
investment and
are prone to largely variable results.
The standard method of field-testing uses conventional two-row replicated
yield
trials, which require substantial capital and labor investment. Trials
commonly
consist of 20 to 50 entries (cultivars of interest) arranged in a statistical
design,
which is replicated 2 to 4 times within a given location. Seed is packaged
prior to
planting in small envelopes (packets). In the case of corn hybrid testing, 40
to 75
seeds of each entry are counted and put into small packets with each packet
representing one row. It is common to package 6 to 8 packets of the same entry
(2
row plot = 2 packets per plot times 3 to 4 replications per location). The
packets are
then placed in planting order by arranging the packets in trays, which are
placed on
the planter when planting. Once the packets are in planting order they are
ready for
the field. Specialized research planters are used to plant small research
plots.
-1-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Conventional small plots (experimental units) are often 2 rows wide at 76 cm
(30
inch) spacing and are 5.2m (17 feet) to 7m (23 feet) in length. Research
planters
facilitate planting over a specific row length and upon finishing each plot,
remaining
seeds are removed and seed for the next plot is planted. The process of
planting
and removing seeds is repeated across the testing location. When plants are at
the
V5-V6 growth stage (5-7 leaves), each row is shortened to the specified length
and
plants within each row of each plot are thinned to the desired population.
Plots are
often over planted due to the inaccuracy of the planting mechanism used.
Agronomic notes are taken throughout the season on each plot prior to harvest.
Specialized harvest equipment is used to harvest the plots in a conventional
manner
(collect seeds). This harvest equipment is designed to collect the seeds from
each
plot, separating them from the plant matter. The harvest equipment then
measures
the weight and moisture of the grain collected for each plot. This data is
then used
in statistical analysis to calculate the final yield of each entry within the
trial.
There is thus a need to design a system that would reduce the area required
for a
trial and thus the number of seeds required for each trial and decrease
variability.
SUMMARY
An improved seed cultivar testing system designed to replace or augment
conventional small plot research trials is provided.
The present disclosure provides a method for performing micro-plot plant
trials of a
seed cultivar, the method comprising the steps of: determining a seed
arrangement
for the seed cultivar for a seed tape using a statistical design to be tested
in a plant
trial; creating the seed tape on a water soluble planting tape, wherein the
seeds
contained by the tape are placed using a predetermined quantity and spacing
based
upon the statistical design; planting the seed tape in a micro-plot
environment;
collecting observations on the resulting plants; harvesting the plants; and
analyze
the collected plants; wherein the spacing of the seeds within the seed tape is
determined to eliminate the need to thin the plants once emerged from the
seeds,
reducing the quantity of the seeds and the trial area required for the micro-
plot.
-2-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
There is provided a method of creating a seed tape using a robotic apparatus,
the
method comprising the steps of: receiving seed tape configuration information
wherein the tape configuration defines seed type and position of the seed
within the
seed tape based on seed characteristics of the seed cultivar to the tested,
the seed
tape configuration designed to eliminate the need to thin the plants once
emerged
from the seeds, reducing the quantity of the seeds and the trial area required
for the
micro-plot; selecting a seed receptacle containing a particular seed cultivar
from a
plurality of seed receptacles; extracting a seed from the selected seed
receptacle
using a probe having a vacuum tip, wherein the seed is picked up by the probe
by a
vacuum created at the tip of the probe; placing the extracted seed into a
water
soluble tape, wherein the seed is dropped into the tape by removing the vacuum
from the probe tip; advancing the seed tape a predetermined distance based
upon
the received tape configuration; and encapsulating the seed in the tape.
There is also provided an automated apparatus for making plant seed tape
comprising: a digital controller programmed for the number of seeds and seed
spacing; a plurality of seed containers, each container contain seeds of a
particular
seed cultivar; an arm with a vacuum probe to draw a seed from a selected
container
and deposit it on a water soluble tape; a air-brush sprayer to deliver fine
water mist
moisten the water soluble tape to allow it to adhere to itself, and encasing
the seed;
a closure system to capture the seed in place the water soluble tape; a drive
for
advancing the tape in a pre-set amount to ensure the correct seed spacing on
the
tape; one or more guides for guiding the tape onto a spool; and a talc
delivery for
applying a small blast of talc powder with each rotation of the spool to
ensure easy
removal of the tape from the spool.
There is also provided a computer readable medium containing instructions for
creating a seed tape using a robotic apparatus, the instructions which when
executed by a processor perform the steps of: receiving seed tape
configuration
information wherein the tape configuration defines seed type and position of
the
seed within the seed tape and is based on seed characteristics to the tested,
the
seed tape configuration designed to eliminate the need to thin the plants once
emerged from the seeds reducing the quantity of the seeds and the trial area
-3-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
required for the micro-plot; selecting a seed receptacle from a plurality of
seed
receptacles; extracting a seed from the selected seed receptacle using a probe
having a vacuum tip, wherein the seed is picked up by the probe by a vacuum
created at the tip of the probe; placing the extracted seed into a water
soluble tape,
wherein the seed is dropped into the tape by removing the vacuum from the
probe
tip; advancing the seed tape a predetermined distance based upon the received
tape configuration; and encapsulate the seed in the tape.
BRIEF DESCRIPTION OF THE DRAWINGS
Further features and advantages will become apparent from the following
detailed
description, taken in combination with the appended drawings, in which:
FIGURE 1 shows a perspective view of a robot that mechanically loads a
predetermined number of individual seeds from each entry into a water-soluble
film;
FIGURES 2a-d shows placement of a seed into the water-soluble film;
FIGURE 3 shows top view of the robot;
FIGURE 4 shows a front view of the robot;
FIGURE 5 shows a side view of the robot;
FIGURE 6 shows a method of performing micro-plot field trials;
FIGURE 7 shows method of creating a seed tape for micro-plot field trials;
FIGURE 8 compares the yield (FIG. 8A) and moisture (FIG. 8B) of 17 corn
hybrids
from two locations of a conventional performance trial to two locations of a
micro-
plot trial;
FIGURE 9 compares the yield (FIG. 9A) and moisture (FIG. 9B) of 13 corn
hybrids
from two locations of conventional performance trial to two locations of the
micro-
plot trial system;
-4-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
FIGURE 10 provides a summary of the harvest yield (dry bushel/acre) by
treatment
(two row = 1; one row = 2; micro = 3) of five corn hybrids over four locations
in a
micro-plot field trial;.
FIGURE 11 provides a summary of the harvest yield (dry bushel/acre) by
treatment
(two row = 1; one row = 2; micro = 3) of five corn hybrids over four locations
in a
micro-plot field trial;
FIGURE 12 provides a summary of the combined grain moisture by treatment (two
row = 1; one row = 2; micro = 3) of five corn hybrids over four locations;
FIGURE 13 shows the adjusted yield (dry bushel/acre) of entries by soil type
from 8
plants per micro-plot from 9 locations;
FIGURE 14 shows the adjusted yield (dry bushel/acre) of entries by soil type
from 6
plants per micro-plot from 9 locations;
FIGURE 15 shows a processor for controlling the robot.
It will be noted that throughout the appended drawings, like features are
identified by
like reference numerals.
DETAILED DESCRIPTION
Embodiments are described below, by way of example only, with reference to
Figs.
1-15. The present disclosure provides a system and method for performing micro-
plot field trials for seed cultivar testing designed to replace or augment
conventional
small plot research trials.
There are several limitations to selecting cultivars from limited data
sources.
Selection of testing locations often does not reflect the majority of the
environments
within the targeted region. Trials are often located in high yield
environments which
create difficulties evaluating hybrid adaptation. Soil texture, soil
temperature, crop
rotation and fertility are examples of parameters varying across growing
regions.
Disease and insect pressure also vary greatly across growing regions due to
crop
rotation and weather patterns. Trial locations are often not exposed to insect
and
-5-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
08908831 WO
disease pressures due to the limited locations planted. In addition, companies
are
often limited to the number of testing locations due to the narrow optimum
planting
window. Planting on either side of this optimum date results in reduced yield
and
therefore does not accurately predict how cultivars will perform under
commercial
scale.
Increasing the environments/locations a cultivar is tested in, within its
adapted
geographic region, can greatly increase the knowledge gained on a cultivar.
The
more information collected on a cultivar increases the confidence and accuracy
of
the data collected. Using this data, breeders can make more informed decisions
on
cultivar selections advanced to commercial products.
Spatial competition between corn plants has been shown to have negative
effects
on grain yield per plant, kernel number per plant and kernel weight per plant
as the
plant population is increased. Dominate plants quickly develop within a
population
which affects the plants ability to partition resources. Industry widely
accepts that
uneven emergence is the primary reason for lower expected yields.
Border rows are often used when evaluating advanced corn hybrid yield to
account
for deviations in plant height of hybrids. The treatment of these border rows
can
influence the final outcome of the trial. A study comparing the treatment of
border
rows (i.e. the population of border rows) found significant differences in
mean grain
yield of hybrids and a significant interaction between hybrids and treatments.
Although surrounded by the identical hybrid, differences in the treatment of
border
plants have been shown to influence the outcome of harvested plots.
The disclosed micro-plot field trial system uses very small plot size and, in
the
example of corn consist of a one row plot containing 1 to 12, preferably 6 to
8 seeds
per plot spaced uniformly at 0.2m (7 inch) to simulate commercial field
scenario but
may be smaller based upon the particular seed type. In this example, entries
can be
arranged one after the other and are spaced 25cm (10 inch) from the last seed
of
the previous entry but can be varied based upon the particular seed growth
requirements. The potential exists to evaluate as few as one plant depending
on the
testing or observation required. By the term "entry" or "entries" it is meant
to refer to
-6-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
08908831 WO
an individual cultivar or seed type. Any plant seed can be used such as for
example
for field trials of corn, beans, peas, tobacco, cereal grains or other
agricultural crops.
The number of seeds and the spacing between each seed can be adjusted
according to the plant species to be tested. In the micro-plot system the plot
size is
approx. 10% the size of conventional two row plots. Depending on the testing
required , with molecular screening techniques plot size could be reduced to
as
small as one plant and less than 5 feet (1.52m) in length.
One of the significant difference between the conventional and the micro-plot
system is the ability to use conventional farm planting equipment to plant
replicated
research trials. This system eliminates the need for specialized and expensive
equipment used for planting conventional research trials. Within a few
minutes, the
conventional commercial scale planter can be equipped with a modified seed
tube
and spool holder , the trial planted and the standard equipment replaced
resulting in
minimal interruption to the cooperators' planting progress. The ability to
easily ship
or courier the equipment needed for the system allow trials to be planted
efficiently
and eliminates much of the liability of transporting conventional research
equipment.
The efficiency of the system facilitates a large number of locations to be
planted
within the limited optimum planting window (approximately 9 days). The ease of
the
micro-plot system allows individual producers to cooperate in the testing
cycle with
little or no supervision and with little interruption to their own progress.
Allowing
cooperators to plant research trials allows for grower observation and input
in the
selection process. The micro-plot system potentially allows for hundreds of
cooperator planted trials consisting of the same entries to be planted within
the
limited planting window. Using standard two row replicated trials it is common
for
companies to plant only two or three locations in any given maturity zone. The
ability to evaluate cultivars over multiple locations greatly increases the
efficiency
and accuracy of cultivar selection and segregation of the genetic by
environment
interaction (i.e. the ability of a variety to perform within a specific
growing region or
environment). Testing over more locations sooner results in less time to get
the test
seed to market and a higher commercial success rate.
-7-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
The disclosed micro-plot system increases the efficiency of the testing
program and
allows for increased testing locations utilizing a similar labor force as a
commercial
testing program. It is often difficult to acquire labor for seed preparation
prior to
spring planting. The micro-plot system eliminates the requirement of counting
and
packaging seed prior to planting commercial trials. The process is automated
with
a unique pick and place robot in a seed laboratory environment to generate
highly
accurate seed tapes. The automation of this process allows for one person to
assemble large numbers of trials in minimal time. The micro-plot system also
eiiminates the need for shortening and thinning rows within plots which is
required
using the conventional system. Reallocating labor in the testing program
allows for
increased efficiency within testing programs.
The ability to easily adjust trial design and the opportunity to basically pre-
plant the
trial in a lab environment allows for sophisticated and elaborate research
trial
layouts. For example seed spacing/plant population studies are easily
accomplished on a cultivar or bio-tech event.
The micro-plot system allows the trials to be prepared in a controlled
laboratory
environment up to the soil insertion under tight supervision. This is an
important
consideration when testing unregistered genetically engineered crops. The area
available for this type of testing is regulated and limited. By using micro-
plot testing
many more new products can be tested and advanced, allowing more rapid
commercial introduction of the crop. The micro-plot system can also be
controlled to
induce environmental stress for the purpose of selection, such as covering
large
areas to mimic drought conditions. Micro-plots would allow for much smaller
enclosures and more varieties tested.
Accurate seed spacing and the ability to have multiple sites of a specific
trial allow
for the reduction of plant number per plot to produce statistically valid data
(see the
Examples). Plant entries are arranged using statistical trial designs. Uniform
seed
spacing and small plot size allow for collection of data under ideal
conditions
reducing the variation within the trial. A reduction in plot size reduces the
overall
trial size and limits the amount of field variability, thus reducing the need
for
replicating entries at any given location. Replication of data is done by
planting
-8-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
multiple locations of the same entry list and combining the data for
statistical
analysis. A systematic check entry is included at the beginning of the trial
arrangement, within the trial and again at the end of the trial. This check
allows for a
measure of the amount of variability across the plot.
The ease and expense of the micro-plot system also allows breeding/selection
programs to evaluate many new cultivars without requiring large quantities of
seed.
Currently, trials are restricted in the number of entries that can be tested
due to the
limitations of enough area for trials or the quantity of seed available for
testing. An
entry list for a specific trial is determined first on the theoretical merit
of the hybrid
and secondly by the amount of seed available. It is not uncommon to require
400 to
3000 kernels of an entry to participate in multiple location trials. Breeders
must
make the same parental cross (this creates a new cultivar) many times to
collect
enough seed for conventional testing. With the micro-plot system, plots are
much
smaller than conventional systems and require very small quantities of seed.
For
example, the system requires only a few seeds per plot per location. Smaller
requirements of seed for a particular entry to be included within a trial
allows
breeders to create more experimental cultivars for screening, thereby
increasing the
probability of identifying a superior variety in the testing/screening
program. Seed
that would otherwise go into the testing cycle could be used to increase the
seed
available for seed production or commercialization. Reallocating resources and
energy to create crosses and develop lines from bulking up seed quantities for
testing could increase the efficiency and success of breeding and testing
programs.
Enclosing the seeds in a tape and wrapping the tape on a spool greatly reduces
the
risk of un-intentional environmental exposure to seed of non-registered bio-
tech
events. The enclosed tape also reduces the risk of seed mixing or theft of the
seed
and ensures that the seed is positioned accurately when planted.
Prior to planting, seeds of each entry are placed in a predetermined order
based
upon the trial design requirements. The seeds are then robotically placed in a
water-soluble film and rolled on a spool as shown in Figures 1 to 5. The
tape/film
provides a means of ensuring placement of the seeds in a precise manner
ensuring
results of the micro-plot trial. The robot 100 comprises an air actuated
computer
-9-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
controlled device which is engineered to place seeds from specially designed
containers 104 one seed 140 at a time in a specific order at specific spacing
onto a
water-soluble tape 112. The computer controlled robot 104 can be programmed to
place the seed 140 onto a tape 112 to reflect the research trial design and
allows for
precise placement and complicated trial designs without the opportunity of
human
error or fatigue. This allows for the manufacturing of many multiples of each
trial
design facilitating wide scale testing at multiple locations.
A supply spool 110 of water-soluble tape 112 is attached to the spool holder
and the
tape 112 is threaded through the machine 100. A digital controller 170 is
programmed for the number of seeds per plot, seed spacing on tape and the
number of plots per tape based upon the micro-plot trial design, a secondary
controller 172 may be used to control the robot movements. Seed of different
cultivars are stored in containers 104 placed on a rotating table 102 at
identified
locations 103. Each container contains a specific seed or cultivar for
selection by
the robot via one or more probes 105a and 105b for placement in the tape 112.
Container locations on the table 102 are identified for ensuring correct
placement of
the seed container 104. Not all spaces 103 on the table 102 may be utilized
for
holding seeds if not required. The containers 104 for the seeds may be of
various
shapes such as oval, rectangular or square based upon the configuration and
operation of the robot 100. The containers 104 or receptacles may form part of
the
table or may be removable containers that are place on the table. The
containers
may also have a concave shape to aid in seed extraction. The rotating table
102
requires the probes to only be moved in one direction horizontally in addition
to
vertically. However, the table 102 may be of varying shapes and the selection
probes may move in 3 coordinate space (x, y, z) to accommodate the table shape
thus not requiring movement of the table itself.
As shown in Figure 1, the circular table 102 is rotated so that the
appropriate seed
container 104 is placed under probes 105a or 105b mounted on frame 109 to
extract
seeds from the appropriate container 104. The probes 105a and 105b are
actuated
by a respective air piston 106a and 106b in the vertical direction and by 107a
and
107b in the horizontal direction. The movement and actuation of the probes can
be
-10-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
controlled by a secondary processor 172. The air piston 106a and 106b drives
the
associated vacuum probe 105a and 105b into the seed container 104. Vacuum
suction draws a seed onto the probe end, a digital negative vacuum pressure
sensor
detects the presence of a seed and will either trigger the continuation of the
process
or will cause the plunger to be deployed back into the seed container until
full
negative pressure (a seed) is achieved. The vacuum holding the seed is turned
off
dropping the seed 140 through the funnel 120 and onto the tape 112 which rolls
into
a trough as a result of the tape closure system. The movement of the probes
105a
and 105b may alternatively be controlled by motors rather than pistons.
Referring to Figure 5, the operation of the tape closure system encapsulates
the
seeds within the water soluble film/tape. Water container 122 holds water
applied to
the film to encapsulate seeds. An air-brush 124 provides a fine water mist
using
controlled pulses of air pressure to deliver water from the water container to
moisten
the film sufficiently to adhere to itself. A pulse drive motor (not shown)
rotates the
attached seed spool 160, sufficiently to advance the tape 112 the appropriate
pre-
determined amount to ensure the correct seed 140 spacing on the tape 112. The
digital sensor 200 or encoder above a roller 202 measures the linear motion of
the
tape and controls the pulse drive motor.
As shown in Figure 2, the seed 140 is dropped in to the tape 112 at a
predefined
order and location along the tape length. The tape closure system, forms a
trough
or "V" shape of the tape 112 is formed by two pairs of guides 114 to receive
the
seeds as shown in Figure 2a. A fine mist of water from reservoir 122 can then
be
provided by an air brush 124 to aid in sealing the tape 112. The tape, being
slightly
moist from the steam adheres to itself ensuring security. The tape is then
pressed
by primary rollers 130a to encapsulate the seeds as shown in Figure 2b.
Secondary
rollers 130b and tertiary rollers 130c add further pressure to the film to
adhere the
two sides of the "V" together. As shown in Figure 2c and 2d seeds are then
individually encapsulated at defined intervals. The tape 112 is guided on the
spool
by a pulse controlled piston with pulleys 180. A heat source 190 ensure that
excess
moisture is removed to mitigate any possible degradation or adherence of the
tape
to itself prior to planting. The controlled motion of the tape through the
robot aids in
-11-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
the even distribution of the tape on the spool and ensures a snag free
unwinding of
the spool 160 when planting the tape 112. A small blast of talc powder is
applied to
the tape 112 and spool 160, by the talc powder applicator 162 from reservoir
164,
with each rotation of the spool 160. The spool may be enclosed 166 to ensure
containment of the talc. This is done to ensure easy removal of the tape from
the
spool by absorbing any free moisture that may be present from the steam
delivery
system. Once the robot 100 has completed its cycle the spool 160 is removed,
an
empty spool is put into position and the cycle repeats itself.
After each seed 140 placement in the tape 112, it is advanced through a
predefined
distance for placement of the next seed. If the next seed is to be extracted
from a
different container, the table 102 is rotated by a motor 168 to move the
desired seed
container 104 to a location for the probe 105a or 105b to be moved in to
position for
extraction of the desired seed.
The water soluble tape used dissolves within minutes of being planted in the
soil
ensure precise seed placement with in the plot. A systematic check entry
(control
plot) seed can be placed at the beginning of the trial, within the trial and
again at the
end of the trial. This check measures the amount of variability across the
length of
the trial based upon know characteristics of the check entry. Identification
can also
be added to the beginning or end of the tape such as an bar code or radio
frequency
identification device (RFID) to aid in identifying the particular tape at the
planting
stage.
Figure 6 shows a method for performing the micro-plot field trial system. A
location
for the field trial is selected (step 602). The characteristics of the seeds
can be used
to define placement within the seed tape or the trial locations (step 604).
The
characteristics may be based upon desired aspects to be analyzed such growth
characteristics related to moisture requirements, spacing, light requirements,
soil
conditions, etc. The configuration of the seed tape is determined (step 606)
based
upon trial design conditions. For example, position and spacing of seeds may
be
randomized or position of different types of seeds may be randomized. One or
more
seed tapes are generated by the robot (step 608). Based upon the trial
requirements, multiple tapes may be generated either with the same seed layout
or
-12-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
with different configurations if randomization is required. The seed tape is
planted
(step 610) in selected location. Tapes of the same configuration may be
planted all
multiple locations improving the number of environments that data can be
collected
for the trial seeds.
Many seed tape planting devices are known in the art, however to utilize the
seed
tape, a modified seed tube is used to facilitate planting. This modified seed
tube
can be used with commercial planters such as for example John DeereT"",
KinzeTM
and WhiteTM planters, and therefore provides an easy adaptation to existing
equipment, eliminating the need to use specialized equipment. The planter is
lowered into the soil and the tape is covered by soil behind the planting unit
to hold it
in place. The forward motion of the planter feeds the tape off the spool, down
the
tube and into the seed trench. The tape follows the identical path of a
conventionally planted seed. Plots are identified by the larger plant spacing
between entries compared to between plants. During the growing stages of the
plans agronomic notes or observation data are collected taken on each plot
prior to
harvest (step 612).
Individual plots are then harvested (step 614) by a micro-plot harvester at
the
desired growth stage. The micro-plot harvester machine is compact enough to be
transported in the back of a small truck and narrow enough to fit between the
bordering rows of the micro-plot trial (approx 60 inches). The actual
harvesting unit
is attached to a commercial scale lawn tractor or equivalent with parallel
linkage and
a hydraulic cylinder tool provide adjustable harvest height. Once the plants
from the
plot are harvested. The plant is weighed and removed, after which the machine
moves forward to repeat the process for each consecutive plot. In the case of
corn
an adjustable stripper plates strip the ear off the corn plant as the stalk is
pulled
down through. Gathering chains with rubber covered projections every 4 inches
drag the husk covered ears up to the husking rollers. Counter rotating rubber
covered rollers pinch off and remove the husk cover from the ears and drop the
husks to the ground. A double chain drag conveyor elevate the ears up to the
sheller. A rubber covered rotating shelling drum pull the ears down and
through an
ever decreasing concave space, removing the kernels from the ear. The shelled
-13-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
kernels fall through a grated screen and the cobs roll out onto the ground.
The
shelled kernels are collected in a weighing pan, which is suspended from load
cells. A moisture probe is present in the bottom of the weighing pan. Grain
weight
and grain moisture data are measured and recorded on the data logger and the
weighing pan rotates and dumps the corn on the ground. The harvester advances
down the row and harvests the next plot. Although corn has been described any
tape of plant or grain may be harvested. The harvested plants (step 616)
analyzed
such as by weight or moisture measurements. This data is statistically
analyzed to
calculate the final yield of each entry within the trial.
Figure 7 shows a method of generating a seed tape for the micro-plot trial
system.
The tape configuration is determined either by an automated process or by a
pilot
designer (step 702). The tape configuration is based upon seed characteristics
and
trial requirements. The configuration may be received in computer readable
code
(step 704) or programmed at an input device at the robot for controlling
computer
processors 170 and 172 for operation of the robot. The particular seed and the
position within the tape is defined in the input process. In commencing
creation of
the seed tape, the defined seed is selected from the respective container by a
seed
probe (step 706) as identified in the programming code. The table is turned so
that
the container that is to be accessed by the probe is in position under the
travel path
of the probe. Position sensor 204 on the table 102 enable accurate monitoring
on
table position and can rotate the table to the appropriate container position
by
measuring rotation distance to a predefined position. The probe 106a or 106b
extracts the seed 140 from the appropriate container 104 using vacuum suction
at
the tip of the probe. The extracted seed is then placed in the tape 112 (step
708).
The tape is advanced and is sealed (step 710) and spooled onto a spool 160. If
the
seed selection of the particular seed is complete (YES at step 712) and the
tape is
complete (YES at step 714) identification can be added or inserted to the tape
(step
716) if required. The tape spool is then completed (step 718) and ready for
planting.
If additional seeds of the same type are to be placed on the tape (NO at step
712)
then the next seed is extracted from the container (step 706). If different
seed is to
be placed in the tape (NO at step 714) the appropriate seed container is
selected at
-14-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
step 704 by rotation of the table or movement of the probes. At the completion
of
the method the tape spool can be removed from the robot and provided for
planting.
The micro-plot system is an alternative to conventional replicated yield
testing. The
system provides the opportunity to plant replicated trials in multiple
locations with
commercial equipment. The system has the opportunity to increase the
efficiency
and accuracy of cultivar development and screening programs without the
capital
investment currently required. The ease and expense of the system allow seed
companies to screen and evaluate large numbers of new hybrids over many
locations, which can increase the productivity of their program.
The micro-plot system has unlimited applications. The micro-plot system has
been
evaluated as a replacement for standard conventional corn hybrid yield trials.
This
testing system can be applied to a wide range of crops and can be used for
plant
population, herbicide resistance, tillage and nutrient studies to name a few.
The following examples illustrate the improvement of the micro-plot trial
system or
conventional trial system.
EXAMPLE 1: Corn Hybrid Performance in a Micro-plot Compared to a
Traditional Hybrid Corn Performance Trial Plot
This example shows whether corn hybrid performance is relative for yield and
moisture in a micro-plot compared to the traditional hybrid corn performance
trial
plot.
Materials and Methods
Seventeen corn hybrids that were entered in table 5 of the Ontario Corn
Committee
Performance trials and 13 corn hybrids that were entered in table 6 of the
Ontario
Corn Committee Performance trials. These performance trials use commercial
research methods to evaluate parameters of corn hybrids and chosen entries
(corn
hybrids) from each table were compared to the micro-plot system.
Trials were planted using a 4 row John DeereTM 7000 Maxi-Merge planter. The
performance trial plots were planted in 76cm (30 inch) rows. Each plot was 2
rows
-15-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
wide and 5.8m long (19 feet long). 55 kernels per row were planted using a
traditional cone type seeder and thinned to 33 kernels per row resulting in a
desired
plant population of 74,000 plants per hectare (pph) (30,000 plants per acre
(ppa)).
The performance trials were analyzed as a randomized complete block design.
The
micro-plots were planted using a seed tape planting technique. A micro-plot
consisted of 9 kernels of the hybrid (entry) and one purple marker to indicate
the
end of a plot. All plots were equally spaced in a continuous row on fiberglass
drywall
tape to represent the final population of the performance trials (74,000 pph).
Modified seed tubes were installed on the planter to plant the tape. Three
consecutive plants with uniform emergence within each plot were staked and
marked for harvest. Seeds that did not emerge were planted with a purple
marker.
The micro-plot was a one replication trial with the first, last and every
fifth plot a
check hybrid. The check hybrid was MZ540, but any suitable check may have been
used. A moving means analysis was used to analyze the data. The performance
trials were harvested by machine using electronic weighing and moisture
equipment.
The tape trial was harvested by hand, shelled, weighed and moisture determined
using a Dickey John GACTM II moisture meter. Grain yields were converted to
15.5% moisture. Plots were fertilized and maintained according to provincial
recommendations.
Results
Figure 8A compares the yield of 17 hybrids from 2 locations in table 5 of the
OCC
performance trial to 2 locations from micro-plot trial. Figure 8B compares the
moistures of the same 17 hybrids. Figure 9A compares the yield of 13 hybrids
from
2 locations in table 6 of the OCC performance trial to 2 locations from micro-
plot
trial. Figure 9B compares the moistures of the same 13 hybrids.
The data from the OCC and micro-plot trials were combined and a split plot
analysis
was used to determine if there were significant differences between the two
methods of performance testing. The reliability of such an analysis is
questionable,
however, there was no significant difference for yield and moisture in the
table 5 or
table 6 comparisons. The data suggests that the hybrid performance for yield
and
moisture in a micro-plot is relative to that in the OCC performance trials.
-16-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
EXAMPLE 2: Corn Hybrid Performance in a Micro-plot Compared to a
Traditional Hybrid Corn Performance Trial Plot
This example determined whether corn hybrid performance is relative for yield
and
harvest moisture when comparing the technique of seed tape planting and cone
planting in small plot hybrid performance trials.
Materials and Methods
Four commercial corn hybrids MaizexTM MZ533, NK BrandT " N65-M7, MaizexTM
MZ540 and PrideTM K542 were used in this trial.
Corn was planted using a 4 row John DeereTM 7000 Maxi-Merge planter. All plots
were planted in 76cm (30 inch) rows. In a first trial the plots were 2 rows
wide and
19 ft. long and in second trial each plot was 4 rows wide and 5.8m (19 ft.)
long (two
center rows used for harvest). Plots were managed according to provincial
recommendations. A 3 replication, split plot design was used with seeding
technique the main plot and hybrid the split plot. To construct the seed
tapes, 33
kernels/row (30,000 ppa) were equally spaced and wrapped in fiberglass drywall
tape. Modified seed tubes were installed on the planter to plant the tape. The
cone
seeded plots were planted using a traditional 32 cell cone type seeder that is
typically used to plant hybrid performance trials. The cone seeded plots were
planted at 55 kernels/row and thinned to 33 kernels/row. The plots were
machine
harvested and yields converted to 15.5% moisture.
Results
In both first and second trials there was no significant difference in yield
between the
technique of seed tape and cone seeded plots(conventional plots) (Table I).
While
there was a significant difference for moisture between the two techniques in
the
first trial the combined data over the two trials shows no significant
difference (Table
1). Likewise, there was no significant difference between seeding technique
(main
plot) and hybrid (sub plot) for both yield and moisture (Table 2).
Table 1: Yield and moisture for main plot (tape seeded and cone seeded).
-17-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Yield (bu/ac) %H20
CV % 7.2 2.6
LSD (.05) 8 0.5
Pr>F 0.4086 0.5424
Tape 230 21.9
Cone 228 22.0
-18-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 2: Two- way table for yield and moisture data
Yield %H20
CV % 7.2 2.6
LSD (.05) 16 0.6
Pr> F 0.4315 0.1339
N65-M7
Tape 228 22.2
Cone 237 22.4
MZ 533
Tape 232 21.8
Cone 227 21.7
MZ 540
Tape 232 22.1
Cone 220 21.9
K542
Tape 229 21.4
Cone 227 22.0
-19-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Based on data over 2 years seeding technique has no effect on the performance
of
a corn hybrid for the parameters of yield and moisture.
Example 3: Corn Hybrid Performance in a Micro-plot Compared to a
Traditional Hybrid Corn Performance Trial Plot
The objective of this example was to determine if the standardized plot size
can be
reduced in corn hybrid testing when evaluating advanced corn hybrids in
performance trials.
Materials and Methods
Five locations were planted in a split plot design randomized twice within
each
location. Plots were bordered on either side by one row of a common cultivar,
MZ
540, to alleviate any border affects on the plots. Each main plot consisted of
treatments varying in experimental unit size of two rows (5.2m long, 76cm
apart), a
one row plot (5.2m long) and a micro-plot (1.4m long). All planting
populations were
at 74 400 plants/ha (30 000 plants/ac). Within each main plot, sub plots
consisted of
five commercial corn hybrids (MZ 540, MZ 4422Bt, MZ 4433Bt, MZ 4655Bt, MZ
535). Recommended fertilizer rates were applied according to provincial
recommended rates.
Two and one row plots were planted with a John Deere 7000 corn planter
equipped
with an AlmacoTM cone system. The micro-plot was planted using a modified seed
tube which mounted on the John Deere 7000 planter unit.
Prior to planting the micro-plot, eight seeds of each entry were placed within
Pval TM
film (CWS-10) at 0.2m spacing (74 400 plants/ha planting population) in the
predetermined randomized order and rolled on a wooden spool for transport to
the
field for planting. The two and one row plots were over planted with 40 seeds
per
row and later thinned to 32 plants per row at the V6 growth stage. A 0.5m to 1
m
gap was also left between plots of the two and one row plot to separate plots.
Prior to harvest, the number of plants per plot was recorded to determine the
final
population. From each plant within the micro-plot, individual ears were
shelled and
the grain weight measured. The grain samples were then combined from the eight
-20-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
plants and pooled grain moisture was measured. The remaining plots were
harvested and shelled using a modified single ear corn sheller and the
combined
grain weight and moisture was measured. Harvest yield was determined using the
following equation:
Grain Yield (bu/ac) = (((harvested grain (grams) x 1000g/kg x 2.2lbs/kg)
/(47.32 /
(100 - grain moisture)*100)) / plot area (acres)
Micro-plot harvest yield was calculated from the pooled weight of the eight
plants
and the pooled weight of the middle six plants. Harvest moisture was measured
using a LabtronicsTM moisture meter.
The trial was analyzed as two separate experiments to determine if the outside
plants within each micro-plot had any influence on the analysis. Therefore,
separate
analysis were performed using harvest data from eight plants and the middle
six
plants within each micro-plot.
Analysis of variance (ANOVA) was performed to determine significant
differences
among main plots and sub plots. Adjusted means were calculated for grain
weight
and moisture for the main plots and sub plots. Residuals were plotted against
predicted values and blocks to visually inspect for deviations from
homogeneity and
independence. Outliers were determined by visually inspected plot residual
values.
The Shapiro-Wilk statistic was calculated to test if the residuals confirmed
normality.
Using the adjusted means of the hybrids by treatment, a Spearman Rank
Correlation was performed to determine significant differences among the
hybrid
rank within each main plot. Means were also evaluated visually to determine
differences in hybrid ranking.
All statistical procedures were performed using GLM, UNIVARIATE and PLOT
procedures of SASTM v.8.2 (SAS Institute, Cary, NC). The Type One error rate
(a)
was 0.05 for all analysis unless specified.
Results
Yield analysis using 8 plants/micro-plot
-21-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
ANOVO suggests significant differences among locations, treatments and hybrids
in
the eight plants per micro-plot test. Blocks, treatment* block and
hybrid*treatment
were not significant, as shown below.
Analysis of variance results for grain yield using 8 plants.
Source df MS F-value P>F
Location 3 18868.66 65.64 0.0150
Block 1 240.72 0.84 0.4567
Treat 2 14679.21 51.07 0.0192
Treat*Block 2 287.46 0.33 0.7206
Hybrid 4 2289.72 2.62 0.0402
Hybrid*Treat 8 380.35 0.44 0.8969
R2 = 0.5620
CV = 12.09
The model was able to explain 56.2% of the total variation and had a CV of
12.09.
The Shipiro-Wilk statistic indicates the residual distribution was normal.
Visual
observation of the residuals showed a slight tendency for the residuals to get
larger
from the two-row plot to the micro-plot but otherwise appeared independent and
homogeneous.
Adjusted yield from the micro-plot was significantly different from the two-
(P=0.0103) and one-row plot (P=0.0227) (Table 3, 4). The two-row and one-row
plots were not significantly different (P=0.1644). The one-row and micro-plot
was
estimated to have 9.43bu/ac and 37.57bu/ac more yield than the two-row plot,
respectively (Table 3, 4). Visual analysis of the adjusted hybrid harvest
means by
treatment also suggests that the micro-plot yields were significantly larger
than the
two row and one row plots (Figure 8).
-22-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 3: Adjusted mean yield (bu/ac) of the two row plot, one row plot and
micro-plot
of five hybrids over four locations.
Adjusted Mean
Treatment (bu/ac) SE
Two row 227.25 4.809
One row 236.68 5.892
Micro 264.82 4.674
Table 4: Contrast and estimate of the adjusted treatment means over the four
locations.
Parameter Estimate SE T Value Pr > Itl F Value Pr > F
two vs one -9.43 7.640 -1.23 0.2206 4.63 0.1644
two vs micro -37.57 6.706 -5.6 <.0001 95.4 0.0103
one vs micro -28.14 7.521 -3.74 0.0003 42.56 0.0227
The adjusted hybrid yield was ranked by treatment. The ranking was consistent
between the 3 treatments. MZ 540 and MZ 4433Bt were not significantly
different
and had the largest yields within all treatments. All hybrids within the two-
row
treatment were not significantly different. Significant differences did occur
between
hybrids in the one row and micro-plots. There were no significant Spearman
rank
correlation between the ranking of the two-row plot and both the one row plot
(P =
0.1041) and the micro-plot (P = 0.1881).
-23-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 5: Adjusted yield (bu/ac) and standard error ranked in ascending order
of the
five hybrids by treatment.
Two Row Plot One Row Plot Micro-plot
Mean Mean Mean
Yield Yield Yield
Hybrid (bu/ac) SE Hybrid (bu/ac) SE Hybrid (bu/ac) SE
MZ535 219.47a 11.217 MZ4655Bt 221.66c 12.304 MZ4655Bt 248.39b 10.451
MZ4655Bt 224.48a 10.451 MZ535 228.09bc 11.601 MZ4422Bt 261.19b 10.451
MZ4422Bt 226.67a 11.218 MZ4422Bt 228.92bc 12.304 MZ535 263.88ab 10.451
MZ540 229.82a 10.451 MZ4433Bt 249.62ab 12.304 MZ540 265.92ab 10.451
MZ4433Bt 235.82a 10.451 MZ540 255.12a 12.304 MZ4433Bt 284.73a 10.451
Yield analysis using 6 plants/micro-plot
When analyzing the data from 6 plants within the micro-plot there were
significant
differences among locations (P=0.0152) and treatments (P=0.0213). There were
no
significant differences among blocks, treatment*block, hybrids and
hybrid*treatments, as shown below.
-24-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Analysis of variance results for grain yield using 6 plants.
Source df MS F-value P>F
Location 3 19409.70 64.84 0.0152
Block 1 779.59 2.60 0.2479
Treat 2 13770.37 46.00 0.0213
Treat*Block 2 299.36 15.17 <0.0001
Hybrid 4 2010.76 2.22 0.0737
Hybrid*Treat 8 382.75 0.42 0.9052
R2 = 0.5511
CV = 12.34
The model was able to explain 55.1% of the total variation and had a CV of
12.34.
The Shipiro-Wilk statistic indicates the residual distribution was normal.
Visual
observation of the residuals showed a slight tendency for the residuals to get
larger
from two-row plot to the micro-plot but otherwise they appeared independent
and
homogeneous.
Adjusted yield from the micro-plot was significantly different from the two
(P=0.01 14)
and one row plot (P=0.0253) (Table 6, 7). The two row and one row plot were
not
significantly different (P=0.1737). The one row and micro-plot was estimated
to
have 9.28bu/ac and 36.42bu/ac more yield than the two row plot, respectively
(Table
6, 7). Visual analysis of the adjusted hybrid harvest means by treatment also
suggests that the micro-plot yields were significantly larger than the two row
and one
row plots (Figure 9).
Table 6: Adjusted mean yield (bu/ac) of the two row plot, one row plot and
micro-plot
of five hybrids over four locations.
-25-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Treatment Adjusted Mean SE
(bu/ac)
Two row 227.27 4.901
One row 236.55 6.005
Micro 263.69 4.763
Table 7: Contrast and estimate of the adjusted treatment means over the four
locations.
Parameter Estimate SE T Value Pr > itl F Value Pr > F
two vs one -9.28 7.786 -1.19 0.2366 4.31 0.1737
two vs micro -36.42 6.834 -5.33 <.0001 86.10 0.0114
one vs micro -27.14 7.665 -3.54 0.0006 38.02 0.0253
The adjusted hybrid yield was ranked by treatment. The ranking was consistent
between the 3 treatments. MZ 540 and MZ 4433Bt were not significantly
different in
all treatments. All hybrids within the two row treatment were not
significantly
different. Significant differences did occur between hybrids in the one row
and
micro-plots. There was more similarity between significant hybrids within the
two
row plot and the micro-plot than the two row plot and the one row plot. There
were
no significant Spearman rank correlation between the ranking of the two row
plot
and both the one row plot (P = 0.1041) and the micro-plot (P = 0.5046).
Table 8: Adjusted yield (bu/ac) and standard error ranked in ascending order
of the
five hybrids by treatment.
-26-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Two Row Plot One Row Plot Micro-plot
Mean Mean Mean
Yield Yield Yield
Hybrid (bu/ac) SE Hybrid (bu/ac) SE Hybrid (bu/ac) SE
MZ535 219.54a 11.432 MZ4655Bt 221.54c 12.540 MZ4655Bt 246.87b 10.651
MZ4655Bt 224.48a 10.651 MZ535 227.90bc 11.823 MZ4422Bt 263.19ab 10.651
MZ4422Bt 226.67a 11.433 MZ4422Bt 228.80bc 12.540 MZ540 264.09ab 10.651
MZ540 229.82a 10.651 MZ4433Bt 249.50ab 12.540 MZ535 264.88ab 10.651
MZ4433Bt 235.82a 10.651 MZ540a 255.00a 12.540 MZ4433Bt 279.41a 10.651
Moisture analysis using 8 plants/micro-plot
When analyzing the moisture data, there were significant differences among
locations (P=0.0374) and hybrids (P=<0.0001). There were no significant
differences among blocks, treatments, treatment*block and hybrid*treatment ,
as
shown below:
-27-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Analysis of variance results for moisture.
Source df MS F- P>F
value
Location 3 135.95 25.90 0.0374
Block 1 0.62 0.12 0.7635
Treatment 2 6.44 1.23 0.4490
Treatment*Block 2 5.25 1.31 0.2741
Hybrid 4 30.95 7.75 <0.0001
Hybrid*Treat 8 2.54 0.64 0.7456
R2 = 0.6272
CV= 9.20
The model was able to explain 62.7% of the total variation and had a CV of
9.2. The
Shipiro-Wilk statistic indicates the residual distribution was normal
(P=0.4741).
Visual observation of the residuals showed them to be independent and
homogeneous. The micro-plot did have higher moistures than the one and two row
treatments but this was not significant (Table 9, 10 and Figure 12).
Table 9: Mean adjusted grain moisture (%) by treatment of five hybrids planted
in a
split plot design over four locations.
Treatment Adjusted SE
Moisture
N
Two Row 21.66 0.325
One Row 21.32 0.398
Micro 22.20 0.316
-28-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 10: Estimate and contrast of the treatment means of five hybrids planted
in a
split plot design over four locations.
Parameter Estimate SE T Value Pr > itl
two vs one 0.34 0.517 0.65 0.5168
two vs micro -0.54 0.453 -1.19 0.2366
one vs micro -0.88 0.509 -1.72 0.0883
The increased yield from the micro-plot in comparison to two and one row plots
occurred in both the 8 and 6 plant micro-plot. Higher yields were not
unexpected
since corn plants determine yield at a very early stage. Corn plants have been
shown to differentiate into hierarchies (dominate plants) at very early stages
(V4)
which is maximized until flowering and remains constant until senescence. Corn
yield is also determined at the 4-6 leaf stage of growth. Both the two and one
row
plots were over-planted and were thinned at the V6 stage once the growing
point
had emerged from the ground. Up to this point in growth, seedlings have
determined the competition around them and have altered their growth habits
accordingly. The increased competition may have influenced the plants to
allocate
more resources to vegetative growth at a time when reproductive growth is
determined. The micro-plot treatment was not thinned and the competition
between
plants was theoretically identical. Within commercial fields, the intra
specific
competition pressure from neighboring plants may be larger than in our micro-
plots
due to the added variation exposed to the corn on larger scales. Seeding
depth,
plant spacing, nutrient availability and soil type are a few factors which can
create
variability influencing plant growth.
Comparison between the 8 and 6 plant micro-plot determined that the 8 plant
plot
accounted for 1% more variation and had a slightly lower CV. These results may
be
explained by the decrease in the plant number sampled. Data from missing
plants
from the plot it would have a larger effect on the micro-plots of the smaller
sample
-29-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
set per experimental unit causing increased variability in the trial. In a
similar
experiment, plant height and plot population were significant covariates and
they
were able to account for more variation within the experiment.
There is little difference when comparing the ranking of yield results by
hybrid from
the 8 and 6 plants sampled from the micro-plot results compared to the one and
two
row plots. Over all the treatments, MZ 540 and MZ 4433Bt were at the top of
the
ranking and MZ 4655Bt was close to the bottom. This reflects the performance
of
these hybrids in commercial production. MZ 540 was not ranked in the top two
positions in both the micro-plot analyzed as 8 and 6 plants however it was not
significantly different from the top ranked hybrid. This may be due to this
hybrids
plant short plant height compared to the other hybrids. Measuring plant height
may
account for differences in height and may more accurately reflect the hybrids
full
potential. By eliminating the plants on either side of the micro-plot a more
consistent
ranking was obtained in comparison to the two row plot.
Harvest moisture was not significant among the treatments however there were
significant differences among hybrids. Differences among hybrids were expected
due to the range in maturities used in the experiment. The rate of grain
drying is
more a function of plant phenotype and the weather at harvest.
This example shows that the micro-plots with uniform spacing significantly
differed
from the two and one row plots however, the ranking and trend of hybrid
performance was consistent over all treatments. Sampling 8 or 6 plants within
the
micro-plot did not differ considerably. The 8 plant plot did account for more
variation
and a slightly lower CV. However, by eliminating the border rows on either end
of
the plot did result in more consistent hybrid ranking when compared to the two
row
plot. The moisture between hybrids did not differ between treatments.
Example 4: Micro-plot Analysis over Multiple Locations
The objective of this example was to determine if the standardized plot size
can be
reduced in corn hybrid testing when evaluating advanced corn hybrids in
performance trials on a multiple location testing regime.
-30-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Materials and Methods
In this example, twelve locations were planted in a modified randomized
complete
design replicated once per location. Plots were bordered on either side by a
random
hybrid chosen by the producer (often the hybrid bordering the plot was the
hybrid
planted in the entire field). A border of MZ 4422Bt was included on either end
of the
plot and as the control hybrid (check hybrid). Each plot consisted of 20
hybrids
randomized across the location. Within each location, a systematic control
hybrid
was included after every 4 plots to account for any variation that may exist
across
the trial area. Experimental unit size was one row consisting of 8 plants
equally
spaced at 0.2m spacing resulting in a planting population of 74 400 plants/ha
(30
000 plants/ac). Hybrids were included in the trial based on commercial hybrids
marketed in the geographic region and various precommercial hybrids deemed
adapted to the area.
Agronomic information was collected on all the locations and included
parameters
such as planting date, harvest date, soil type, previous crop, fertility,
herbicide used.
Prior to planting the micro-plots, eight seeds of each entry were placed
within Pval
film (CWS-10) at 0.2m spacing (74 400 plants/ha planting population) in the
predetermined randomized order and rolled on a wooden spool for transport to
the
field for planting. Planting was executed using an adaptor specifically
designed to
plant such trails.
The parameter of plant height was measured at the V6, V10 and R2 from the
ground
to the emerged collar. Plant height at the R2 stage was taken to the flag leaf
collar.
Prior to harvest, the number of plants per plot was recorded to determine the
final
population. Three locations were discarded at harvest due to poor plant
stands.
Each micro-plot was harvested on a by plant basis. Grain was collected from
all the
ears from individual plants, grain shelled and the grain weight measured. The
grain
samples were then combined from the eight plants and pooled grain moisture was
measured. Harvest yield was determined using the following equation:
-31 -
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Grain Yield (bu/ac) = (((harvested grain (grams) x 1000g/kg x 2.21bs/kg) /
(47.32 / (100 - grain moisture)*100)) / plot area (acres)
Micro-plot harvest yield was calculated from the pooled weight of the eight
plants
and the pooled weight of the middle six plants. Harvest moisture was measured
using a Labtronics moisture meter.
The trial was analyzed as two separate experiments to determine if the outside
plants within each micro-plot had any influence on the analysis. Therefore,
separate
analysis were performed using harvest data from eight plants and the middle
six
plants within each micro-plot.
Analysis of variance (ANOVA) was performed to determine significant
differences
among entries with regard to grain yield and plant height. Adjusted means were
calculated for grain weight, moisture and plant height. Residuals were plotted
against predicted values and blocks to visually inspect for deviations from
homogeneity and independence. Outliers were determined by visually inspected
plot residual values. The Shapiro-Wilk statistic was calculated to test if the
residuals
confirmed normality. Using the adjusted means of the hybrids, a Spearman Rank
Correlation was performed to determine significant differences among the micro-
plot
trial and the data set of the Ontario Corn Committee trials which use a
standard two
row experimental unit.
Agronomic data was visually evaluated to determine if the locations had
commonalities among them. If commonalities existed, the data was separated
into
those groups and the common factor was analyzed.
All statistical procedures were performed using GLM, UNIVARIATE and PLOT
procedures of SAS v.8.2 (SAS Institute, Cary, NC). The Type One error rate (a)
was 0.05 for all analysis unless specified.
Results
By plant analysis of grain yield
-32-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
ANOVA suggests significant differences among entries, blocks and the covariate
parameters of plant height and plot population. There were no significant
differences among plant position or entry x plant position. The Shipiro-Wilk
statistic
indicates the residual distribution was not normal (P=0.0001). Visual
observation of
the residuals showed them to be independent and homogeneous. The adjusted
means of plant position (Table 11) suggest that there are no significant
differences
in plant position. Significant differences among plant height of the entries
shown by
the difference of 45cm between the tallest hybrid and the shortest hybrid
(Table
12).Table 11: ANOVA analysis of the micro-plots by plant position within the
plot
measured in grams per plant.
Std.
Pltnum Estimate Error
1 220.2 13.20
2 213.7 13.20
3 213.6 13.20
4 216.5 13.22
5 214.7 13.21
6 215.9 13.20
7 218.8 13.20
8 214.1 13.22
-33-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 12: ANOVA of the micro-plot entries by yield per plant on 9 locations.
Entry Adjusted Yield by Std. Error
Plant (grams)
EX 4550Bt 241.2 14.16
MZ 4422Bt 238.9 14.13
MZ 4433Bt 232.4 14.13
MZ 4422Bt 228.8 14.04
MZ 4422Bt 225.7 14.05
MZ 4422Bt 224.6 14.02
MZ 4422Bt 224.3 14.07
MZ 4422Bt 222.9 14.00
MZ 540 222.7 14.06
EX
5469Bt/RR 219.5 14.06
EX 5462Bt 219.2 14.04
EX 45-66RR 219.0 14.06
EX 4161Bt 217.4 14.01
EX 54-55RR 215.3 14.20
EX 5466HX 214.4 14.08
EX 4565 210.8 14.21
MZ 4655Bt 210.7 14.06
-34-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
MZ 4422Bt 209.9 14.19
EX 54-71 RR 209.3 14.09
MZ 535 209.1 14.16
MZ 535HX 207.8 14.11
EX 4561 HX 200.5 14.13
MZ 366 199.9 14.12
DKC 50-18 198.9 14.03
MZ 36-66RR 195.3 14.06
EX 4562HX 195.2 14.25
Since plant position was not a significant factor, further analysis was
performed
using a pooled value for per plot yield at both 6 and 8 plants per plot.
Yield analysis using 8 plants/plot
ANOVA suggests significant differences among entries, blocks and plant
population.
There were no significant differences among the covariate entry plant height.
The
model was able to explain 79.0% of the total variation and had a CV of 9.84.
The
Shipiro-Wilk statistic indicates the residual distribution was normal
(P=0.9815).
Visual observation of the residuals showed them to be independent and
homogeneous. The adjusted means by entry are shown in Table 13.
Table 13: ANOVA adjusted means of grain yield pooled from 8 plants of the
micro-
plot over 9 locations.
-35-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Adjusted
Std. Yield
Entry Grain Yield
Error (bu/ac)
(g/plot)
EX 4550Bt 1657.9 51.71 264.7
MZ 4422Bt 1637.7 51.03 261.5
EX 5469Bt/RR 1621.7 51.39 255.7
EX 54-71 RR 1611.7 51.19 249.5
MZ 540 1608.3 51.03 244.9
EX 5466HX 1599.0 51.19 251.2
MZ 443313t 1595.3 51.61 254.1
EX 4161 Bt 1581.8 51.55 255.8
MZ 535 1569.7 51.37 251.0
EX 546213t 1553.9 51.16 243.6
MZ 4655Bt 1544.6 51.41 245.3
EX 4565 1544.4 54.34 251.2
MZ 535HX 1530.3 51.25 242.3
EX 54-55RR 1508.5 52.65 238.4
EX 45-66RR 1501.6 51.43 238.8
EX 4562HX 1468.7 51.53 233.7
EX 4561 HX 1444.6 51.34 232.5
MZ 366 1359.9 51.43 217.2
-36-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
MZ 36-66RR 1350.9 51.41 218.0
DKC 50-18 1331.0 51.93 216.3
Visual analysis of the hybrid ranking using the adjusted means of the pooled
results
do not significantly differ from the hybrid ranking results from the by plant
analysis.
Yield analysis using 6 plants/plot
ANOVA suggests significant differences among entries, blocks and plant
population.
There were no significant differences among plant height. The model was able
to
explain 75.7.0% of the total variation and had a CV of 10.79. The Shipiro-Wilk
statistic indicates the residual distribution was normal (P=0.5386). Visual
observation of the residuals showed them to be independent and homogeneous.
The adjusted means by entry are shown in table 14.
Table 14: ANOVA adjusted means of grain yield pooled from 6 plants of the
micro-
plot over 9 locations within southwestern Ontario.
Adjusted Grain Yield Std.
Entry (g/plot) Error Yield (bu/ac)
EX 4550Bt 1449.8 49.68 308.6
EX 5469Bt/RR 1441.5 49.40 303.0
MZ 442213t 1440.4 49.54 306.7
MZ 540 1437.0 49.26 291.8
EX 54-71 RR 1436.2 49.53 296.4
EX 5466HX 1435.1 49.33 300.7
MZ 443313t 1418.2 49.96 301.2
-37-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
EX 4161 Bt 1403.0 49.50 302.6
MZ 465513t 1393.5 49.37 295.1
MZ 535 1359.1 49.30 289.8
EX 4565 1342.0 52.61 291.1
EX 546213t 1339.2 49.30 279.9
EX 45-66RR 1333.1 49.36 282.7
MZ 535HX 1325.4 49.97 279.8
EX 54-55RR 1309.7 50.20 276.0
EX 4562HX 1294.4 50.00 274.6
EX 4561 HX 1275.0 49.30 273.6
MZ 366 1201.8 49.51 256.0
MZ 36-66RR 1190.7 49.77 256.1
DKC 50-18 1165.8 50.01 252.6
Hybrid ranking is similar to the results obtained from the analysis using 8
plants/plot.
Grain moisture analysis using 8 seeds/plot
ANOVA suggests significant differences among entries, blocks and plant
population.
There were no significant differences among plant height. The model was able
to
explain 81.7.0% of the total variation and had a CV of 5.79. The Shipiro-Wilk
statistic indicates the residual distribution was not normal (P=0.0001).
Visual
observation of the residuals showed them to be independent and homogeneous.
-38-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Table 15: ANOVA adjusted means of grain moisture from a pooled sample of 8
plants per plot over 9 locations.
Hybrid Maturity Adjusted
Entry (CHU) Moisture (%) Std. Error
MZ 540 3300 24.9 0.42
EX 54-71 RR 3250 23.7 0.42
EX 5462Bt 3250 22.7 0.42
EX 5466HX 3250 22.5 0.42
EX 5469Bt/RR 3250 22.2 0.42
EX 54-55RR 3250 22.1 0.43
MZ 535HX 3250 21.9 0.42
MZ 4655Bt 3150 21.7 0.42
EX 45-66RR 3150 21.6 0.42
EX 4562HX 3100 21.5 0.42
MZ 4433Bt 3100 21.4 0.43
EX 4550Bt 3100 21.3 0.42
MZ 4422Bt 3100 21.2 0.42
MZ 366 3050 21.2 0.42
MZ 535 3250 21.1 0.42
EX 4561 HX 3100 20.6 0.42
MZ 36-66RR 3050 20.4 0.42
-39-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
EX 4161 Bt 3050 20.2 0.42
DKC 50-18 3050 19.8 0.43
EX 4565 3100 19.8 0.45
Visual analysis of the hybrid ranking by moisture indicate that the moisture
of the
hybrids expectedly correspond to hybrid maturity. From Table 15, the higher
the
grain moisture the longer the hybrid maturity. This relationship is common and
generally accepted by industry.
Comparison between the Micro-plot to OCC Trials
Using the adjusted means from 8 plants per micro-plot there appears to be a
strong
relationship between the results from both testing procedures. Visual analysis
of the
hybrid ranking suggests hybrids that have high yield results in the OCC trials
also
have high yield results in the micro-plot trials. Those hybrids with low yield
results
appear also to be consistent among both testing procedures. The same
correlation
holds true for grain moisture. This is apparent in tables 16, 17, 18 and 19.
Table 16: Micro-plot data from 9 locations compared to the yield and moisture
results from the Ontario Corn Committee trials of 2 locations planted in a
replicated
complete block design with 3 replications per location.
-40-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
OCC Trials (Table 5) Micro-plot - 8 plants
Hybrid bu/ac moist Hybrid bu/ac moist
MZ 4433Bt 239.8 21.9 MZ 4433Bt 290.7 21.4
EX 4550Bt 235.3 22.0 EX 4161 Bt 290.0 20.2
MZ 4422Bt 230.9 22.1 EX 4550Bt 288.8 21.3
EX 4561 HX 228.7 20.6 MZ 442213t 288.0 21.2
MZ 465513t 226.4 22.3 MZ 465513t 272.5 21.7
EX 4562HX 222.0 21.4 EX 4562HX 264.8 21.5
EX 45-66RR 219.8 22.0 EX 45-66RR 263.5 21.6
EX 4161 Bt 217.6 21.2 EX 4561 HX 262.6 20.6
MZ 366 208.7 21.4 DKC 50-18 248.2 19.9
MZ 36-66RR 206.5 21.6 MZ 366 244.4 21.2
DKC 50-18 204.2 20.1 MZ 36-66RR 238.0 20.4
Table 17: Micro-plot data from 9 locations compared to the yield and moisture
results from the Ontario Corn Committee trials of 2 locations planted in a
replicated
complete block design with 3 replications per location.
-41-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
OCC Trials (Table 5) Micro-plot - 6 plants
Hybrid bu/ac moist Hybrid bu/ac moist
MZ 4433Bt 239.8 21.9 EX 4161 Bt 293.3 20.2287
EX 4550Bt 235.3 22.0 MZ 4433Bt 292.5 21.4169
MZ 4422Bt 230.9 22.1 EX 4550Bt 289.5 21.2905
EX 4561 HX 228.7 20.6 MZ 4422Bt 287.4 21.2461
MZ 4655Bt 226.4 22.3 EX 4561 HX 273.3 20.6412
EX 4562HX 222.0 21.4 EX 4562HX 272.4 21.512
EX 45-66RR 219.8 22.0 MZ 4655Bt 272.3 21.6773
EX 4161 Bt 217.6 21.2 EX 45-66RR 261.8 21.5773
MZ 366 208.7 21.4 DKC 50-18 247.9 19.869
MZ 36-66RR 206.5 21.6 MZ 366 240.9 21.235
DKC 50-18 204.2 20.1 MZ 36-66RR 238.5 20.444
Table 18: Micro-plot data from 9 location compared to the yield and moisture
results
from the Ontario Corn Committee trials of 6 locations planted in a replicated
complete block design with 3 replications per location.
-42-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
OCC Table 6 Micro-plot - 8 plants
Hybrid bu/ac moist Hybrid bu/ac moist
EX 5466HX 236.9 19.9 EX 5469Bt/RR 283.6 22.2
MZ 535 234.6 19.2 MZ 535 282.0 21.1
MZ 540 234.6 21.3 EX 5466HX 280.0 22.5
EX 4562HX 234.6 19.2 MZ 540 277.4 24.9
EX 546213t 234.6 19.5 MZ 535HX 273.1 21.9
MZ 535HX 232.3 19.7 EX 546213t 270.8 22.7
EX 54-55RR 223.1 20.7 EX 54-55RR 266.9 22.1
EX 5469Bt/RR 223.1 20.7 EX 4562HX 264.8 21.5
EX 45-66RR 202.4 19.9 EX 45-66RR 263.5 21.6
Table 19: Micro-plot data from 9 locations compared to the yield and moisture
results from the Ontario Corn Committee trials of 3 locations planted in a
replicated
complete block design with 3 replications per location.
OCC Table 6 Micro-plot - 6 plants
Hybrid bu/ac moist Hybrid bu/ac moist
EX 5466HX 236.9 19.9 EX 5469Bt/RR 289.1 22.235
MZ 535 234.6 19.2 MZ 535 285.5 21.121
MZ 540 234.6 21.3 EX 5466HX 283.6 22.5016
-43-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
EX 4562HX 234.6 19.2 MZ 540 277.7 24.8905
EX 5462Bt 234.6 19.5 EX 4562HX 272.4 21.512
MZ 535HX 232.3 19.7 EX 54-55RR 268.1 22.0585
EX 54-55RR 223.1 20.7 MZ 535HX 267.8 21.9169
EX 5469Bt/RR 223.1 20.7 EX 5462Bt 266.1 22.6995
EX 45-66RR 202.4 19.9 EX 45-66RR 261.8 21.5773
Data Segregation
Visual analysis of the agronomic data collected about specific management
characteristics indicate that soil type has two factors which may be
partitioned to
yield more information on individual hybrids (Table 20).
Table 20: Adjusted moisture, mean yield (bu/ac) from 8 plants/plot, mean yield
(bu/ac) from 6 plants/plot, soil type, planting date, previous crop and
tillage practice
of the microplot locations.
-44-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Adjusted Mean Mean
Yield Yield Planting Previous Tillage
Location Moisture (8 Plants) (6 Plants) Soil Type Date Crop Used
(%) (bu/ac) (bu/ac)
Tilbury 19.4 238.8 236.2 clay loam 28-Apr corn conventional
Melbourne
- A 19.9 243.8 249.5 sandy loam 3-May corn minimum
Blenhiem 19.6 259.5 258.5 clay loam 8-May soybeans no till
Dresden 20.0 270.0 271.8 sandy loam 10-May wheat conventional
PainCourt 23.5 284.9 288.4 clay loam 3-May wheat conventional
Melbourne
- B 21.9 289.4 292.1 clay loam 5-May alfalfa conventional
Chatham 24.0 291.3 288.2 sandy loam 2-May soybeans conventional
Thamesville 20.9 292.3 291.9 sandy loam 4-May corn conventional
Tilbury 26.1 313.0 314.9 clay loam 31-May wheat conventional
ANOVA using data from 8 plants / micro-plot suggests significant differences
among
soil type, soil*entry, plant population and the covariate plant height. The
Shipiro-
Wilk statistic indicates the residual distribution was normal (P=0.1691).
Visual
observation of the residuals showed them to be independent and homogeneous.
ANOVA using data from 6 plants / micro-plot suggests significant differences
among
soil type, soil*entry, plant population. The covariate plant height was not
significant.
The Shipiro-Wilk statistic indicates the residual distribution was normal
(P=0.1396).
Visual observation of the residuals showed them to be independent and
homogeneous.
Table 21: Adjusted yield (grams/plot) and moisture by soil type. Yield has
also been
expressed as bu/ac.
-45-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Plot Std. Yield
Size Parameter Effect Type Estimate Error (bu/ac)
8 plants Yield/plant soil Clayloam 1684.1 b 32.96 267.0
8 plants Yield/plant soil Sandyloam 1411.4a 45.24 229.4
6 plants Yield/plant soil Clayloam 1455.5b 30.91 307.7
6 plants Yield/plant soil Sandyloam 1241.3a 40.95 269.0
8 plants Moisture soil clayloam 21.8b 0.37
8 plants Moisture soil sandyloam 19.8a 0.47
Adjusted yield and moisture are shown in Table 21 segregated by micro-plot
size.
Clay soils tended to yield higher than sandy soil in both analysis using 6 and
8
plants per micro-plot. The clay soil also had higher harvest moisture than
sandy
soils.
The significant interaction between soil types and entries suggests a
significant
genotype by environment interaction exists. While the trend for the entries
was to
have lower yield on the sandy loam soil some varieties actually performed
better in
these environments in comparison to the trial average (Figures 11 and 12).
Identifying genotype by environment interaction is important to position
hybrids in
environments were it has the highest yield potential.
FIG. 15 shows a digital controller 1500 computing environment for executing
seed
tape creation for controlling the robot apparatus in the form of computer
readable
code for execution as implemented by processors 170 and/or 172. The computer
102 comprises central processing unit (CPU) 1502 and associated memory 1510.
The CPU(s) may be a single processor or multiprocessor system or may be
implemented by an application specific integrated circuit. In various
computing
environments, memory 1510 and storage 1540 can reside wholly on computer
environment 1500, or they may be distributed between multiple computers.
-46-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
Input devices 1530 such as a keyboard and mouse may be coupled to a bi-
directional system bus. The keyboard and mouse are for introducing user input
to a
computer and communicating that user input to processor 1502 if required.
Computer 1502 may also include a communication interface 1508. Communication
interface 1508 provides a two-way data communication coupiing via a network
link
to a network 1550 by wired or wireless connection or may provide an interface
to
other host devices by a direct radio frequency connection to enable retrieval
of data
or providing commands to remote robots 100. In any such implementation,
communication interface 1508 sends and receives electrical, electromagnetic or
optical signals which carry digital data streams representing various types of
information. Display device 1520 is provided to facilitate programming and
monitoring if required.
The CPU 1502 or similar device may be programmed in the manner of method
steps, or may be executed by an electronic system which is provided with means
for
executing for operation of the classification and search engine. The storage
device
1540 can be accessed through an input/output (I/O) interface and may include
both
fixed and removable media, such as magnetic, optical or magnetic optical
storage
systems, Random Access Memory (RAM), Read Only Memory (ROM) or any other
available mass storage technology. The storage device or media may be
programmed to provide such method steps for operation and control of robot 100
by
CPU 1502. The storage device 1540 may also store operational information
regarding the robot, such as error messages and operational and performance
data.
The storage device 1540 may also contain data regarding seed cultivars to
enable
automated generation of tape configuration if required.
Memory 420 can provide code for operation and programming of the robot 100.
The
tape configuration 1512, comprising a data file or executable code, is either
entered
by a user by an input device 1530 or retrieved from storage 1540. From the
tape
configuration 1512, commands for probe control 1514 and for controlling the
tape
1516 are generated. The commands are the sent by a data interface to the
appropriate devices for controlling the robot operation 100 to create the
desired see
tape.
-47-
CA 02692722 2010-01-06
WO 2009/006733 PCT/CA2008/001251
It will be apparent to persons skilled in the art that a number of variations
and
modifications can be made without departing from the scope of the present
disclosure as defined in the claims.
The embodiments described above are intended to be illustrative only. The
scope
of the invention is therefore intended to be limited solely by the scope of
the
appended claims.
-48-
