C.O.R.N. Newsletter 2011-03

Dates Covered: 
February 9, 2011 - February 23, 2011
Editor: 
John Yost

A Little More on Statistics, What Does Probability Mean When Reporting LSDs?

When an experiment is conducted, we select a probability level that denotes statistical significance.  This probability shows the level of confidence that we have in our conclusions regarding differences between treatments.  Scientific convention dictates that the probability level typically utilized is 95% (0.05).  For our example we used 90%, therefore there is a 90% probability that differences observed between treatments are due to the treatments, and there is only a 10% probability that the difference observed is due to random chance (or random sampling).  That might be confusing so let’s put this into practice to further the discussion.

Recall back to Scenario 1:

Treatment 1

Treatment 2

LSD 0.05

LSD 0.1

LSD 0.2

LSD 0.3

50

44

 

 

 

 

59

57

 

 

 

 

50

49

 

 

 

 

Avg. 53

Avg. 50

10.3

7.4

4.5

2.7

 The question is – is the mean for treatment 1 statistically larger than the mean of treatment 2?  Notice at a probability level of 0.05 (95%), it requires treatment 1 to be 10.3 higher than treatment 2 to be considered statistically different.  We do not meet that criterion, so our conclusion is there is no statistical difference between the two treatments at a probability level of 95%.  We would reach the same conclusion at probability levels of 90% (0.1) and 80% (0.2) based upon the calculated LSD values.

If we select 70% (0.3) as an acceptable level of probability, we can say that treatment 1 is statistically higher than treatment 2 because the difference between treatments 1 and 2 is greater than 2.7 (the LSD value).  But what are we also saying?  We are concluding that there is a 30% probability that our conclusion that treatment 1 is greater than treatment 2 is, in fact, incorrect or due to random chance.  Are you comfortable with that level of probability?  Remember the information generated is coming from a controlled experiment conducted at a small scale.  As a comparison, a probability of 50% (0.5) is essentially a coin flip.

Put another way, let’s say you had two bags (representing the two treatments above) containing 20 pieces of paper with numbers written on them.  We randomly sample 4 pieces of paper from each bag (this constitutes an experiment replicated 4 times).  We then calculate an average for the 4 samples randomly drawn.  Put the pieces drawn back into the bag and resample.  Conduct the experiment 20 times.  If the average of bag one is greater than the average of bag two 19 out of 20 times, that equates to a probability level of 95% (0.05).  Stated another way, 95% of the time when we sampled the two bags, bag 1 had a higher average than bag 2.  Thinking back to our actual experiment (scenario 1), for us to be 95% sure that treatment 1 resulted in a higher average than treatment 2 required them to have a difference of 10.3, and we did not meet that criterion.

Conduct the same experiment 20 more times.  Assume the average of bag one was only greater than the average of bag two 14 times out of 20.  This equates to a probability level of 70% (0.3).  Seventy percent of the time when we conducted the experiment bag one had a higher average than bag two.  Are you less comfortable with your conclusion?

Assume you repeated the experiment another 20 times, and found that 50% of the time the average of bag one was greater than the average of bag two.  You are now at essentially a coin flip.

Scientists want to make sure our conclusions are as correct as possible.  Using low probability levels (>0.3) to denote statistical significance increases the probability that our conclusions are incorrect.  As also displayed in the above example, the experiment may be repeated many times and only show significant differences part of the time.  The same is true for agronomic experiments.  Differences between treatments may be significant at some locations during some seasons, but that does not mean that the differences will be significant at all locations during all seasons.  That is why it is important to repeat an experiment over both time and space.

 

What We Learned About Marestail in 2010

We conducted several studies this year that expanded our knowledge on management of marestail (horseweed).  Some of the findings reinforced what we already knew and validated our current recommendations, but we did learn some new things.  It seems as though we are still trying to figure out what the most consistently effective approach is, given that some years even good marestail management programs can fail to provide adequate control.  Among the more general studies of marestail that we always conduct, we conducted the following that had more specific objectives:

1.  Where the marestail population is both ALS- and glyphosate-resistant, is there any point in using residual herbicides in the fall?

The answer in short is – no.  Our research continues to support the fact that chlorimuron (Canopy/Cloak) is the only herbicide that persists long enough into spring to have value when applied in the fall.  Even where the population is not ALS-resistant, the control from fall-applied chlorimuron is variable into the spring.  It generally won’t completely control marestail from fall through soybean canopy closure, so we recommend still using residual herbicide in spring.  Where the population is ALS-resistant, chlorimuron will no longer be effective.  All of the other residual soybean herbicides lack substantial activity into spring.  So where the population is ALS-resistant, the better approach will usually be to apply glyphosate + 2,4-D in the fall, and save the residual herbicide for spring.  It is of course still possible to go ahead and use a low rate of Canopy in the fall for residual control of other weeds, and apply the bulk of the residual herbicide in spring.

2.  What are the most consistently effective marestail burndown treatments, especially where plants are big or have survived a previous treatment?

Marestail plants are always easier to control when small, and preferably before the stem elongates more than a few inches.  We assume, though, that burndown treatments that work in difficult marestail burndown situations should also work well for relatively small plants.  We conducted studies at two sites with either very tall (single stalk up to 30 inches) plants, or plants that had survived an earlier glyphosate treatment (branched plants, up to 18 inches tall).  Results of these studies generally supported our previous recommendations.  Burndown treatments that should generally work except in extreme situations include:  glyphosate + 2,4-D; glyphosate + Sharpen; Ignite; Ignite + metribuzin; Ignite + Sharpen, or Gramoxone + metribuzin + 2,4-D.  Burndown recommendations are listed in the OSU/Purdue fact sheet, “Control of marestail in no-till soybeans”, available at https://agcrops.osu.edu/specialists/weeds

3.  What is the most consistently effective overall approach to marestail management - is there a “systems” approach that we may be missing?  See accompanying article in this issue of C.O.R.N. for more on this.

4.  What is the effect of marestail on soybean yield?  While not a direct objective of this year’s research, information on yield was a byproduct of some control studies we conducted.  In a study where we were able to roughly separate the effects of the burndown vs residual herbicide, the numbers were as follows:  ineffective burndown and no residual – 51 bu/A; effective burndown but no residual – 57 bu/A; and effective burndown and residual – 65 bu/A.  We actually found that where the burndown was ineffective, it did not matter whether residual was included since the yield was already lost.  We were working in a fairly small area with a consistently high marestail population, so this probably overestimates yield loss compared with a typical field that has a less dense and more variable population.  Nonetheless, these figures show the importance of effective marestail management, and the return on investment in the appropriate herbicides.

Do We Need a Systems Approach to Marestail Management?

In 2010, we conducted the first of several studies designed to get a better idea of what the right system, or combination of fall and spring herbicides, might be to ensure consistent marestail control.   We used LibertyLink soybeans, and worked in an area where the marestail population was resistant to glyphosate but not ALS inhibitors.  The study included all possible combinations of three fall and four spring preplant treatments, and these were followed with a postemergence application of Ignite.  Fall treatments included none, no residual (glyphosate + 2,4-D) or residual (Canopy + 2,4-D).  Spring treatments included glyphosate + 2,4-D, glyphosate + Sharpen, Ignite + metribuzin, and glyphosate + 2,4-D + Valor XLT.  We also had a treatment that consisted of fall herbicide only – Canopy + 2,4-D.  We evaluated marestail control 3 weeks after planting to measure control 3 weeks after planting to evaluate burndown herbicide effectiveness, 6 weeks after planting (at time of POST) to evaluate residual control.  The study was then treated with Ignite POST, and we evaluated again later in the season.  Here’s what we found (results also shown in the table below):

  • Where we applied Canopy + 2,4-D in the fall without a follow up spring preplant treatment, we ended up with only about 50% control of marestail in early June when the POST Ignite was applied.  These results support our recommendation not to use residual herbicide only in fall where marestail is a problem.  Even where the population is not ALS-resistant, it’s risky to assume that fall application of Canopy or another chlorimuron product can provide enough control the following year in the absence of additional residual herbicide applied in spring.
  • When evaluated 3 weeks after application, all of the spring burndown treatments effectively controlled the marestail that had emerged by April 22, even where herbicides were not applied in the fall.  This is not surprising given that the spring burndown treatments we used have a history of effectively controlling small plants in April.
  • In the absence of fall-applied herbicides, use of a spring preplant treatment that lacked substantial residual activity (including Sharpen and the low rate of metribuzin used here) resulted in less than 60% control of marestail in early June.  Where Valor XLT (4 oz/A) was included in the spring treatment, control in early June increased to 73%.  This factors in that the burndown herbicides were effective, so the difficulty here was control of marestail that emerged in May, after soybean emergence.  These results illustrate some of the problems occurring with marestail management programs in Ohio – even a relatively strong approach may not completely control marestail until soybeans canopy.
  • It was possible to improve the effectiveness of the spring residual treatments by applying herbicides the previous fall.  Because this population was not ALS-resistant, where we applied Canopy in the fall and followed with any spring preplant treatment that had at least some residual (Sharpen (1 oz/A), metribuzin (4 oz/A), or Valor XLT), we ended up with 97 to 100% marestail control in early June.
  • What was surprising from our results, however, were the apparent benefits of applying at least something in the fall, even if it lacked residual activity.  Where we applied glyphosate + 2,4-D in the fall vs where nothing was applied in fall, the control in early June increased from 73 to 92% for the spring Valor XLT treatment.  For the metribuzin and Sharpen treatments, control increased from an average of 50% to an average of 72%.  What’s fairly strange about this is that there were not any differences in control at 2 weeks after planting.  So it appears that the type of fall treatment did not influence the effectiveness of burndown treatments, but did alter the effectiveness of residual herbicides. 
  • While Ignite is generally effective for control of emerged marestail, the effectiveness of the POST Ignite application in this study was somewhat dependent upon the level of management up to that point.  Three weeks after the POST application, control of marestail ranged from 95 to 100% for any treatment that included both fall and spring preplant herbicides.  Control was reduced to 78 to 82% where fall herbicides were omitted, and the spring treatment lacked substantial residual (glyphosate+2,4-D or glyphosate+Sharpen).  The soybean canopy did suppress remaining plants well enough that control at the end of the season exceeded 90% for all treatments.  However, the results of this study show the importance of using an effective overall approach to marestail management even where the POST herbicide is capable of controlling marestail.

We will follow these articles with a third in the next issue of C.O.R.N. that summarizes viable approaches to marestail management, and lists common mistakes in marestail management.

Herbicide

Rate

Marestail control (%)

 

 

May 12

June 8

July 4

Sept 17

Fall only

 

 

 

 

 

Canopy + 2,4-D

2.2 oz + 0.5 lb

95

50

83

92

 

 

 

 

 

 

Spring only:

 

 

 

 

 

Glyphosate + 2,4-D

0.75 + 0.5 lb

100

29

80

100

Glyphosate + Sharpen

0.75 lb + 1 oz

100

41

78

100

Ignite + metribuzin 75DF

32 + 4 oz

100

60

95

99

Glyphosate + 2,4-D + Valor XLT

0.75 lb + 0.5 lb + 4 oz

98

73

98

100

 

 

 

 

 

 

Fall glyphosate + 2,4-D followed by spring:

 

 

 

 

 

Glyphosate + 2,4-D

0.75 + 0.5 lb

98

30

90

97

Glyphosate + Sharpen

0.75 lb + 1 oz

100

74

100

100

Ignite + metribuzin 75DF

32 + 4 oz

100

70

100

100

Glyphosate + 2,4-D + Valor XLT

0.75 lb + 0.5 lb + 4 oz

100

92

100

100

 

 

 

 

 

 

Fall Canopy + 2,4-D followed by spring:

 

 

 

 

 

Glyphosate + 2,4-D

0.75 + 0.5 lb

100

77

100

100

Glyphosate + Sharpen

0.75 lb + 1 oz

100

97

100

100

Ignite + metribuzin 75DF

32 + 4 oz

100

100

100

90

Glyphosate + 2,4-D + Valor XLT

0.75 lb + 0.5 lb + 4 oz

100

97

100

100

Application and planting information:

Fall – November 16

Spring – April 22

POST Ignite – June 9

Soybeans planted on April 29

May 12 rating shows effectiveness of preplant burndown

June 8 rating shows effectiveness of burndown + residual control

Northern Ohio Crops Day

February 10, 2011 is this year’s date for the Northern Ohio Crops Day that will be held at Ole Zim’s Wagon Shed, 1375 N. State Route 590, Gibsonburg, Ohio. Featured on this program will be: “Growin Good Corn: Rocket Science or Common Sense?” present by Dr. Bob Nielsen, Purdue University;  “Wheat, going for the Best Yields” by Dr. Pierce Paul OSU/OARDC;  “Insect Update Western Bean Cutworm” presented Dr. Christina DiFonzo, Michigan State University;  “Cover Crops Can we cut down on Pesticides” presented by Alan Sundermeier, OSU Extension;  and “Crop Insurance Update” from Dennis Lenhart, Hartlen Insurance. These are topics that everyone is talking about and we have some of the top specialists presenting at this meeting.

Program has been approved for Private and Commercial pesticide recertification. Participants can obtain all private recertification credits, and commercial credits are 1 hour each in 2A and 2B. CEU for Certified Crops Advisory are 3 hours in Crop Management and 1 hour in Pest Management.

The meeting starts at 9:00 a.m. and continues until 3:00 p.m. A $10.00 donation will be accepted at the door to help with expenses that includes a copy of the Corn, Soybean, Wheat, and Alfalfa Field Guide publication. Pesticide recertification credits are an additional expense.

Lunch will be provided courtesy of the Northern Ohio Crops Day Exhibitors. The program is a joint effort of Erie Basin EERA Ohio State University Extension. Please call Sandusky County office (419) 334-6340 or e-mail koenig.55@osu.edu with any questions.

Western Bean Cutworm Myth #3: All growth stages of corn are at risk for oviposition

Western bean cutworm adults begin to emerge as early as June, and adult flight can extend until mid or late September.  Peak flight, which indicates the highest point of moth activity, is usually mid to late July for Ohio, although this varies based on weather conditions.  During moth flight, adults mate, and female moths will then lay eggs in clumps ranging from 5-200 individual eggs, but usually average around 50 eggs.  Eggs are first white in color, then turn tan and pink, and then turn deep purple.  Once the eggs turn purple, eggs hatch within 24-48 hrs.  Early stage larvae are difficult to scout for, and late stage larvae enter the ear.  For this reason, we recommend scouting  egg masses, which will help time chemical applications if they are needed.  Western bean cutworms are protected from chemicals in the egg stage and once they enter the ear. 

The oviposition behavior of females will help prioritize which fields need to be scouted.  Females prefer late-stage whorl corn which has not yet tasseled.  For scouting, these fields should be the first priority, especially before and during peak flight.  As corn matures, and as the numbers of moths decline, oviposition on tasseled and later-stage corn becomes less likely.  Once tassels begin to emerge, corn becomes a less optimal host, and females will continue searching for better hosts.  Often, the preferred host is dry beans, which has caused extreme problems in Michigan and Ontario. Dry beans are often not grown in Ohio, but it is unknown what females do without good oviposition hosts.  Given that western bean cutworm adult emergence, planting date and corn growth is weather dependent, making predictions on the extent of infestation is somewhat difficult.  However, with good adult trapping and egg scouting, damage from western bean cutworm can be minimized and yield can be saved.

Conservation Tillage and Technology Conference

The Conservation Tillage and Technology Conference will be held Feb. 24-25 at Ada, in the McIntosh Center of Ohio Northern University. Complete registration and updated program information are online at http://ctc.osu.edu. Online registration is encouraged.

Elwynn Taylor, Iowa State University agricultural climatologist, will be the speaker for the opening general session, at 9:30 a.m. Starting at 11:00, concurrent sessions take over. The program includes Corn University (Thursday), Soybean School (Friday), nutrient management (both days), advanced scouting techniques (Thur), water quality (Friday), and precision agriculture (Friday). A cover crops session runs all day Thursday, starting at 8:00 a.m.

Archive Issue Contributors: 
Archive Issue Authors: 

About the C.O.R.N. Newsletter

C.O.R.N. Newsletter is a summary of crop observations, related information, and appropriate recommendations for Ohio crop producers and industry. C.O.R.N. Newsletter is produced by the Ohio State University Extension Agronomy Team, state specialists at The Ohio State University and the Ohio Agricultural Research and Development Center (OARDC). C.O.R.N. Newsletter questions are directed to Extension and OARDC state specialists and associates at Ohio State.