Developing Hazelnut Germplasm for the Upper Midwest

2011 Annual Report for LNC10-328

Project Type: Research and Education
Funds awarded in 2010: $175,000.00
Projected End Date: 12/31/2013
Region: North Central
State: Minnesota
Project Coordinator:
Dr. Lois Braun
Research Associate

Developing Hazelnut Germplasm for the Upper Midwest


Our work to develop elite hazelnut germplasm for the Upper Midwest has entered a new phase now that we have established five replicated germplasm performance trials, with up to 109 genetic lines of hybrid hazelnuts at each. The oldest of these are starting to bear nuts, so it will not be many years before we are able to make our first selections. In the meantime we are working to develop improved propagation techniques for mass production of these selections and to develop best management practices for nitrogen fertilization, weed control, and plant architecture management.

Objectives/Performance Targets

1. Plant material
–Four germplasm performance trials expanded to at least 100 plant accessions in each.
–A fifth germplasm performance trials added near Tomahawk, Wisconsin.
–Multiplication of clones with the genetic uniformity needed for future agronomic trials on N fertilization and weed control.

2. New knowledge on the following, with bulletins on each posted on the hazelnut website:
–Commercially viable methods of hazelnut propagation.
–Management of clonally-propagated planting stock.
–Weed control methods during establishment.
–N fertilization recommendations for mature hybrid hazelnuts.

3. New knowledge shared with growers and other researchers via the Upper Midwest Hazelnut Development Initiative website, summer field days and an annual regional hazelnut conference.


Objective 1. Germplasm Improvement.
We have now established five germplasm performance trials with up to 109 genetic lines of hybrid hazelnuts at each. Three of the sites are in Minnesota (St. Paul, Lake City, and Lamberton) and two are in Wisconsin (Bayfield and Tomahawk). (The Tomahawk site was added to the original proposal.) Most of the lines were selected from on-farm plantings of seed-propagated material originating from Badgersett Research Farm. In 2011 we added a few lines selected from farms in our region that had purchased their material from Grimo Nut Nursery in Ontario and St. Lawrence Nursery in New York, thereby expanding the genetic base of our collection. We consider the accession-collection phase of our project to be nearly complete for hybrid hazelnuts, because most of the exceptional plants identified by growers thus far have already been added to the trials. A few more will be added in the future as younger on-farm plantings reach maturity and are offered by growers. We will also be adding wild American hazelnuts to the trials, as described below.

Scouting for exceptional American hazelnuts in wild stands was started in Wisconsin in 2009 by Mike Demchik and Jason Fischbach. They expect to start selecting the best for propagation and inclusion in the performance trials in 2012. Scouting was delayed in Minnesota because Lois Braun failed to identify many productive sites in 2010. In 2012 however, her team identified and scouted more than 20 sites, which will be scouted again in 2012, for selections to be made in 2013.

In order to screen the accessions for resistance to Eastern Filbert Blight (EFB) we attempted to inoculate the plantings with the disease using a low-tech field technique in spring 2010 and 2011. However, because the infective period in April was not conducive to the spread of the disease in either year, we are not sure if it was successful. (Because EFB is a slow-progressing disease, we won’t know for up to two years.) Thus in 2011 we also used a more aggressive greenhouse technique to inoculate seedlings which were then transplanted amongst the healthy plants in the performance trials. These will serve as sources of disease inoculum.

In 2011 we collected our first nut yield data from the oldest (layered in 2008) and most precocious plants in the performance trials, right on schedule. We also collected data on blooming dates and practiced pollination techniques in anticipation of starting to do controlled pollinations in the near future. Although we said in our proposal that we would wait until we had five years of data on these plants before we started to make controlled crosses, we are now considering doing it sooner, and culling progeny from plants we later determine to be undesirable. This will speed up the breeding process by as many as five years, but at the expense of managing large numbers of potentially undesirable plants.

Objective 2. Propagation Methods.
Stem cuttings. Preliminary trials on rooting hardwood stem cuttings in humidity tents during winter 2009 produced plants with such higher rates of survival than we had obtained with softwood cuttings, that since then we have redirected our efforts from softwood to hardwood cuttings. In 2010 we determined that rooting was best between 1,000 and 2,000 ppm IBA. 2011 trials narrowed this range to 2,000 ppm. In 2011 we also determined that rooting is highest for suckers that are about 50 cm long, which are short enough to be placed in the humidity tents as they are. Suckers that are longer than 50 cm can be cut into two to three segments of 30 cm or more in length, thereby obtaining more than one new plant for each sucker collected. Although the segments do not root as well as the uncut suckers, especially not the medial and apical segments, enough of them do root to make this method more productive of new plants than mound layering. Rooting and subsequent survival are better with larger stem segments (~ 30 cm long) than with smaller ones (< 30 cm), presumably because they have larger stores of carbohydrates with which to grow roots and to generate new buds; new buds are key to overwinter survival. In 2012 we are focusing our efforts on trying to determine what light and temperature regimes best promote rooting.

Stock plant beds. One of the greatest challenges in doing these stem cutting trials has been the inadequate numbers of stems with uniform genetics and morphology, with which to do properly replicated trials. This was even more of a challenge for mound layering trials, for which there was exactly one experimental unit available per genotype. Thus, although I am relatively confident of the conclusions described above for our hardwood stem cutting trials, our data was not sufficiently replicated to be publishable. We look forward to remedying this problem in the near future by using stems cut from the stock plant beds established from 2009 through 2011 on the St. Paul campus of the U of M. These were established using surplus plants from the performance trials. There are now 381 stock plants, of 31 genotypes, an average of 12 each.

Originally we were hoping to start using these for either stem cutting or mound layering trials in 2011. But we decided that they were still too small then. Larger mother plants are capable of producing more suckers and more vigorous suckers. We now plan to coppice half of them in 2012 and the other half in 2013, to compare the long-term productivity of mound layering three- versus four-year old plants. We have observed that the productivity of layered plants declines with successive years of layering, so one of our objectives will be to determine how best to manage them to sustain them. Responses measured will be numbers of stems produced that are suitable for mound layering and the percentage of stems that that produce viable roots when prepared for mound layering.

Objective 3. Development of Best Management Practices for Establishment.
Transplant Timing and Type. At the time we wrote our proposal in 2009, it appeared that low survival of hazelnut transplants was a major challenge. A survey of hazelnut growers had found that survival of the 4-month old 6-inch tube-grown seedlings most commonly available at that time averaged only 28%. So one of our objectives was to figure out how improve survival. However, we soon determined that either field-grown two-year-old seedlings, or vegetatively propagated plants, not “tubelings”, would probably be the planting material of the future, and thus there was no justification in testing the tubelings. Instead, we turned the proposed trial into a comparison between fall and spring transplanting of mound layers, and between mound layers transplanted immediately after digging, layers grown out for an additional year in a nursery, and two-year old field-grown bare-root dormant seedlings.

We found no significant difference between rooted mound layers transplanted in November and those transplanted the following April: survival after one year was 91% and 89% for fall- and spring-transplanted respectively, a difference that is statistically insignificant. This is good news because it gives us flexibility in transplanting time, enabling us to better work around unpredictable weather. Although this survival rate was short of our target of 98% survival, we are satisfied with these results because the plants used for these trials were the weakest mound layers harvested that year (because they were left over after the best had been used for the performance trials and the stock plant beds). Because survival strongly correlated with quality of transplant, we are confident that if the best transplants had been used survival would have met our 98% target.

These trials also found no difference in survival between high quality 2nd year layers (layers that had been grown out in a nursery for a year before transplanting to their final locations) and high quality two-year old bare-root dormant seedlings, both of which had 100% survival. This excellent survival was probably because, with an extra year of growth, these plants were much more robust at transplanting time than the younger plants had been. Moreover, any weak plants would have been culled before inclusion in the trial. Based on these results, we recommend that planting stock be grown out in a nursery for a year before it is transplanted to the field, even though this would increase production costs.

Ground Preparation and Weed Control. These trials have not yet been planted, for several reasons. First, we did not have the extra clonal planting stock that we expected to have in 2011 because we delayed the start of the mound layering trials using our stock plant beds, because we decided that they were too small, as explained above. Secondly, we decided that this trial is a lower priority than the N fertilization trials we had planned to start later, and thus we used the limited clonal planting material that we generated in 2011 for some new N fertilization trials instead. We still plan to start the weed control trials in fall 2012 assuming we generate enough plants. We have decided to eliminate the factorial treatment no-till pre-plant ground preparation because the mound layered planting material we are using has such a broad root system that it is very difficult to dig a hole large enough for it without tilling the soil first.

N Fertilization Trials. In our proposal we discussed the need for N fertilization trials using clonal material, to simplify the statistical analysis of the results, which in earlier N trials Braun (2008) had found to be complicated by the extremely variable responses due to the genetic variability of seed-propagated plants. We proposed starting N trials after we had produced enough clonal planting material for them. But in 2011 we discovered that two growers had clonal material from some early micropropagation efforts, which they made available to us for our N fertilization research. In addition, in 2010 we started a N trial on the hazelnut planting at the Central Lakes Ag Center in Staples, Mn, where we have many years of yield data to use as a covariate in data analysis, which should make up for the genetic variability. We have also been working with Jim Mickelson, a grower who has initiated his own on-farm trial in 2010.

Preliminary results of these trials show a pattern that is consistent with Braun’s earlier N trials: leaf N responses are observed within the same growing season as the N applications but growth and yield responses may not be observed until the second year after application or later. At Staples, the leaf N a few months after application was much higher than the sufficiency threshold for all rates but the very lowest (8 g per cubic meter of plant canopy volume), suggesting that the higher rates (16, 32, and 64 g m-3) were much higher than needed. Growth responses did not appear until the second year, and pointed to 16 g N m-3 as the most effective rate. No yield responses have been observed yet.

The hazelnuts at Jim Mickelson’s were N deficient at the beginning of his trial, in which he compared urea with a coated “slow-release” urea product. He found that both produced a dramatic leaf N response within the year of application. This translated into a yield benefit the second year, which was highest following the coated urea. His trial was not designed to determine the optimal N rate, so we hope to work together with him to do that in the future.

Because of uncertainties about whether we will be able to continue the two N trials that use clonal material, we used the surplus clonal material we generated in 2011 for two future N trials, one with 40 experimental units of one genotype and the other with 30 of each of two genotypes.

Other trials. In 2011 we collected baseline data from established plants for pruning and coppicing trials to start in 2012. After about ten years plants become too large to harvest easily, either by hand or by machine, so we need to develop a system for keeping them small. Badgersett Research Farm advocates coppicing them to the ground when this happens, which takes them out of production for a few years. We are wondering whether this may be avoided with selective pruning. However, pruning may be more labor intensive than coppicing, and thus may not be as economical, even if it does not take plants out of production.

Objective 4. Increase Grower Knowledge.
Since the beginning of this grant we have had a second and third annual Upper Midwest Hazelnut Initiative Conferences. The second was held in South St. Paul, MN, March 4 and 5, 2011, and the third was in Decorah Iowa, March 2 and 3, 2012. Both were organized by Jeff Jensen of Rural Advantage. Attendance at these conferences was 48 and 57 respectively. Attendees were a mix of current and prospective growers, and a few who were merely curious. Research findings from this project were reported at both.

In addition, Rural Advantage hosts annual field days, called “Walk-n-Talks”. Three Walk-n-Talks within the grant period focused on hazelnuts:
• October 2010 – Hazelnut Valley Farm, Lake City MN – 35 attendees
• July 2011 – Hazel Acres, Fenton IA – 22 attendees
• October 2011 – Hazelnut Valley Farm, Lake City MN – 41 attendees
Norm Erickson’s field day is an annual event comprised of seminars in the morning and outdoor sessions in the afternoon.

In addition, hazelnut research was showcased as part of larger the Rosemount Research and Outreach Center’s annual open house in August, 2011, attended by approximately 2,200 people. An estimated 100 people visited the hazelnut booth staffed by hazelnut grower Norm Erickson and Lois Braun.

Finally, the Upper Midwest Hazelnut Website ( has proven to be an extremely useful channel to get research results out to growers just as soon as it is developed. The site typically has about 200 visitors per week.

Braun, L.C. 2008. Nitrogen fertilization of hybrid hazelnut in the Upper Midwest. PhD Diss.,
Dept. of Hort. Sci., Univ. Minn., St. Paul.

Coyne, C.J., S.A. Mehlenbacher, K.B. Johnson, J.N. Pinkerton, and D.C. Smith. 2000.
Comparison of two methods to evaluate quantitative resistance to eastern filbert blight in
European hazelnut. J. Amer. Soc. Hort. Sci. 125(5):603-608.

Molnar, T.J., S.N. Baxer, and J.C. Goffreda. 2005. Accelerated screening of hazelnut seedlings
for resistance to eastern filbert blight. HortScience. 40:1667-1669.

Impacts and Contributions/Outcomes

The benefits of this research to growers and consumers will mostly be realized when new germplasm is available from this work. In the meantime, growers are already benefitting from preliminary research results on best management practices, as we share them at conferences, field days, and via the Upper Midwest Hazelnut Website.


Lois Braun

[email protected]
research associate
Dept of Agronomy and Plant Genetics, University of Minnesota
1991 Upper Buford Circle, 411 Borlaug Hall
St. Paul, MN 55108
Office Phone: 6516411880
Jeff Jensen

[email protected]
Minnesota Hazelnut Foundation
3503 40th Ave., Fenton, IA 50539
Fenton , IA 50539
Office Phone: 5153206756
Dr. Don Wyse

[email protected]
Dept. of Agronomy and Plant Genetics, University of Minnesota
411 Borlaug Hall, Upper Buford Circle
St. Paul, MN 55108
Office Phone: 6126257064
Jim Mickelson

[email protected]
4817 75th St. SE
Rochester, MN 55904
Office Phone: 5072884160
Dr. Brent McCown

[email protected]
Dept of Horticulture and Center for Integrated Agricultural Systems, University of Wisconsin-Madison
1575 Linden Drive, Room 393
Madison, WI 53706
Office Phone: 6082620574
Norm Erickson

1303 NE 5th Ave.
Rochester, MN 55906
Office Phone: 5073194085
Mark Shepard

[email protected]
New Forest Enterprises LLC
P.O. Box 24
Viola, WI 54664
Office Phone: 6086271772
Roy and Teresa Cerling

[email protected]
RR1, Box 45A
Wycoff, MN 55970
Office Phone: 5073524156
Linda Meschke

[email protected]
Rural Advantage
1243 Lake Avenue, Suite 222
Fairmont, MN 56031
Office Phone: 5072385449
Jason Fischbach

[email protected]
Extension Educator
University of Wisconsin Cooperative Extension Bayfield County
P.O. Box 218
Washburn, WI 54806
Office Phone: 7153736104
Dr. Mike Demchik

[email protected]
Associate Professor of Forestry
College of Natural Resources, University of Wisconsin-Steven's Point
800 Reserve St.
Steven's Point, WI 54481
Office Phone: 7122958910