Developing Hazelnut Germplasm for the Upper Midwest

Developing Hazelnut Germplasm for the Upper Midwest

Summary

The oldest plants in our five hazelnut germplasm performance trials (in which 146 accessions are now represented) have been bearing nuts for two years now, so we are looking forward to starting to identify the best. This year we started adding the best Corylus americana accessions we identified in three years of scouting wild populations. We have been making controlled crosses for two years now, and have our first 59 F1 seedlings planted in the field and our second batch of over 1,000 seeds germinating in the greenhouse. Our micropropagation graduate student has established microcuttings from several genotypes in sterile culture. She is examining the differences between different genotypes in ability to convert IBA to IAA, which may explain differences in their ability to form adventitious roots, which would suggest alternative methods to use on the recalcitrant genotypes. In our stem cutting experiments we found that humidifiers in vented humidity tents maintained adequate humidity while avoiding overheating, which improved rooting success. In 2013 we also continued treatments and data collection on three N fertilization trials, and four pruning and coppicing trials, and also planted four new weed control trials.

Objectives/Performance Targets

1. Plant material
• Five germplasm performance trials with between 92 and 146 plant accessions in each.
• 59 seedlings of the first F1 population from crosses between local hybrids and European selections from Oregon have been planted to the field in St. Paul
• Over 1,000 seeds of the second F1 population are currently germinating in pots in the greenhouse.
• Four new plantings of clonal material have been established for N fertilization trials.
• Four new weed control trials were planted in 2013.

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

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

Accomplishments/Milestones

Objective 1. Germplasm Improvement.
Performance Trials. We now have 146 different hazelnut genotypes represented in our five germplasm performance trials (St. Paul, Lamberton, and Lake City, Minnesota, and Bayfield and Tomahawk, Wisconsin). Most, but not all, of these are replicated in triplicate at each of the five sites. These include 117 accessions selected from on-farm seedling plantings of European-American hybrids originating from Badgersett nursery in Southeast Minnesota; 10 accessions, including some named varieties, of European-American hybrids from other nurseries located in the Northeast U.S. and Canada; one European accession from an Oregon nursery; and, new this year, 18 wild C. americana accessions collected from ten locations in Central Minnesota.

The oldest plants in our trial have been producing gradually increasing amounts of nuts for three years now. We are hopeful that patterns will start to emerge when we evaluate our 2013 yield data. We also expect to start identifying accessions with resistance as a result of our Eastern Filbert Blight (EFB) field screenings, now that EFB has been successfully established at both the St. Paul and Lake City sites.

Corylus americana. The C. americana accessions added to the performance trials were selected based on two to three years of scouting for superior wild germplasm on public lands. Sixteen selections were mound layered this summer, but of these only 7 produced enough viable rooted layers to plant in triplicate at all five trials. (This poor rooting supports our impression that layering is most productive with plants growing under high fertility and with low competition, conditions which are difficult to find in wild locations. It highlights our need to develop alternative methods of propagation.) A few additional selections produced enough rooted layers to plant at the St. Paul trial only, and a few others were propagated by other methods, such as by stem cuttings or by digging suckers, making for a total of 18 C. americana accessions added to the performance trials so far.

We collected stems from those C. americana selections that did not root, which we will attempt to root them this winter with our hardwood stem cutting technique. Molly Kreiser, our graduate student, will also attempt to micropropagate them. In addition, we sent the best three rooted accessions to our colleagues in Wisconsin to micropropagate in their facilities. (Our Wisconsin colleagues are taking a slightly different approach with propagation of their 20 C. americana selections, choosing to micropropagate them first, then enter the microprogated clones into the performance trials when they are available.)

This year we also collected C. americana nuts for planting. This represents a change in strategy for identifying superior C. americana germplasm which bypasses two of the challenges with the protocol we had been following: 1) the difficulty of screening plants in the wild, due to rugged terrain, wildlife predation, and unexpected human activity, and 2) the difficulty in propagating them in the wild, for the same reasons. Instead of trying to narrow the plants at a site down to the single best one, we simply collected nuts from any and all plants that had both high yields and good quality nuts. In Minnesota we collected from about 13 sites, and also received contributions from 10 Master Gardeners. In November we planted out most of the seed in the Wisconsin DNR State Nursery at Hayward. We estimate that we planted a total of two to three thousand seeds, from 46 collections. In the spring we will inoculate the seedlings with EFB and next fall we will transplant the strongest ones out for long term evaluation. This will greatly widen the genetic basis of our work, and may help us identify some novel genes for disease resistance.

Not all of the C. americana seeds were planted at Hayward; subsamples of eight collections were set aside to use for greenhouse germination trials. Some were soaked in GA and planted in October, while the rest were put into cold moist stratification (the standard method) until late January. The latter are just now emerging. Preliminary observations suggest no difference in germination rates between the two methods, though the GA method is more labor intensive. Our objective was mostly to get a feel for how much effort it will take to grow out large numbers of seedling, for how to manage them most efficiently.

Controlled crosses. In 2013 we grew out our first group of F1 hybrid seedlings, 59 of them, first in pots in the greenhouse and now in the field in St. Paul. They were the results of eleven crosses we made in St. Paul in spring 2012 for the sake of merely practicing the technique, because we did not yet have any mother plants we felt had been adequately evaluated to be selected as parents. However, when we planted out the seed (also for the sake of practicing) and found that some of the seedlings were exceptionally vigorous, we decided we might as well find out their potential. In spring 2013 we made 39 crosses, this time with a few plants that are probable elite selections. In January 2014 we planted over 1,000 seed from these crosses in the greenhouse, which are now germinating. In about a month we will inoculate them with EFB spores for identification of genotypes with EFB resistance. We are beginning to realize that managing this many seedlings in the greenhouse and field, including keeping records on each individual plant, will be a huge job, especially as we are likely to increase the number of crosses we do each year.

Objective 2. Propagation Methods.
Micropropagation. Our collaborators at UW have reported good results micropropagating wild hazelnut germplasm, with the first plants from their efforts now planted to the field in Wisconsin. However, they still find that some genotypes propagate more readily than others. Often it is the most desirable genotypes that are hardest to propagate. Our graduate student, Molly Kreiser, is attempting to find out why in the hopes of overcoming this problem.

Kreiser has established shoots from several hazelnut genotypes in culture and is now beginning shoot multiplication with a few of these. She is looking at the natural occurrence of indole-3-butyric acid (IBA) and indole-3-carproic acid (ICapA), which has also been shown to stimulate rooting in a number of species, in some model plants. She obtained a high-resolution mass spectrum of IBA extracted from Arabidopsis seedling tissue, and is currently using similar methods to determine whether ICapA is also an endogenous compound. In the future she will try to identify these compounds in hazelnuts too. Right now she is working on developing an assay to measure conversion of IBA to IAA (indole-3-acetic acid) in hazelnuts. So far she has not been able to measure this conversion, but once she has the assay working she will use it to measure levels of IBA-to-IAA conversion in different hazelnut genotypes and test whether differences in conversion rates might explain differences in rooting ability between genotypes.

Hardwood stem cuttings. In our winter 2013 hardwood stem cutting trials, we attained the highest rooting percentages so far, 37%, but we still have not found the breakthrough we have been hoping for. In 2013 we tested a new method for maintaining humidity that we hoped would reduce excessive heat build-up within the humidity tents: instead of sealing the plastic sheeting covering the tents, we let it hang loose, and used humidifiers fed by a drip irrigation system to keep humidity high. This system was more labor intensive than our previous system, because it required routine monitoring, so the help of two undergraduates was indispensable. The new system worked: an average of 32% of stems within these unsealed humidity tents rooted, as compared to only 21% within the sealed tents. But there was significant within-treatment variability, probably due to temperature differences within the greenhouse. If we can figure out how to better control temperature we may get better results, but this might require a more technologically advanced greenhouse than we have access to.

In another experiment we attempted to root cuttings in a cooler using heating mats to warm the rooting medium, so as to stimulate root growth, at the same time as keeping stems cool, so as to inhibit shoot growth. The theory is that shoot growth before root development merely depletes stem carbohydrate stores and increases their moisture stress, whereas by forcing the stems to develop roots first their leaves would be more productive when they develop later. This was our third attempt at this approach, but the first two times we struggled to keep the rooting medium warm and moist enough within the drying environment of a cooler. This time we added insulation and watered the rooting boxes every other day. This resulted in shoot growth from buds below and near the surface of the medium, but only a few stems produced roots. Although one of the few rooted stems actually survived and resulted in a viable new plant, we decided that this method does not merit further pursuit.

In 2013 we also tested and rejected the hypotheses that 1) our use of Clorox as a sterilant was deactivating the IBA plant growth regulator, and that 2) NAA added to IBA would enhance rooting. We also confirmed that it would be possible to start stems in the greenhouse as early as January. An earlier start date means that plants are ready to be transplanted to the field earlier, thus avoiding the high maintenance costs of keeping them in pots over the summer, which also results in high mortality due to heat stress.

For our next round of stem cutting experiments, we are testing even earlier start dates. In natural settings, woody plants do much of their root growth in the fall, when they are translocating nutrients from leaves below ground for storage. We are now testing whether by collecting stems before leaf drop we might take advantage of this process. So we are comparing September, October, November, December and January start dates. We are also revisiting the question of ideal size of stem cutting.

Hardwood stem cutting research is ideal for undergraduate research projects because results are observable within just a few months. However, we do not see it as having much utility for mass production of hazelnut clones. Micropropagation is the method we expect will move a regional hazelnut industry forward. However, the stem cutting method still has utility for producing plants to use in other aspects of our research program because, as mentioned above in the section on C. americana, mound layering is not always feasible and does not always work. Five of the 18 C. americana accessions in our collection would not be there if they had not been propagated from stem cuttings. In addition, we would not have been able to plant the paired weed control trial, discussed below, if we had not generated the plants for it in our stem cutting trial.

Mound layering. 2013 was our second year of mound layering trials using plants that had themselves been produced by mound layering in 2008, testing whether N fertilization and age make a difference. Our hypothesis is that mound layering is more likely to be successful with plants that have well developed root systems, which is a function of time and fertility. We compared plants that were coppiced for layering after only three years versus four years, but obtained only weakly rooted layers from either one. It is likely that even four years were not enough time. Statistical analysis has not yet been done, but it seems that rooting was inversely related to number of stems prepared per plant, which suggests that competition between stems may reduce rooting. If this is confirmed statistically it will support the need to be more radical in our thinning of stems when we prepare them for layering.

Agronomic Research.
N fertilization. The three trials N fertilization trials all have challenges that are confounding our ability to obtain good results. At Staples, where the soil is low-organic matter sand, and thus likely to show an N response, as we observed in 2005 when the plants were young, we hoped that the high variability of the seedling population would not confound our results. But it seems that it has, in combination with high levels of EFB and a stem-boring insect. Unhealthy plants do not process nutrients normally. Thus we have not able to find any N response at Staples.

We knew that genetic variability at the Staples site could be a problem, which is why we were happy to find the other two sites, both of them with clonal plants and both on poor soils. At one of these, an on-farm planting in Iowa, with the most uniform hazelnut planting we have ever observed, both in terms of plant size and productivity, we had high hopes for obtaining useful results. At this site we observed an increase in leaf N with applied N in the first year and an even stronger response in the second year. As is typical in slow-growing woody plants, it took until the second year before a growth response was observed: in the second year plant growth was positively related to the previous year’s leaf N, but not to applied N. Also in the second year there was also a very weak but non-significant increase in yield with increased leaf N. Because yield is strongly correlated to plant size, we expected an even stronger yield response in the third year, correlated to the growth response of the second. All the responses we had observed in the first two years were linear, without the leveling off that would indicate the threshold between optimal and excessive rates, suggesting that the range of rates we were testing were too low. So in the third year we decided to increase rates. We also conceded to the grower’s request to fertilize the “buffer” plants between the treatment plants, which he observed were nutrient deficient and thus not fulfilling his goal of maximizing yield. Unfortunately, this ruined the experiment because some of the fertilizer applied to the buffer plants migrated to the treatment plants. This illustrates one of the challenges of on-farm research, which is that research goals often conflict with the goals of the grower.

At the third site, also an on-farm planting of clonal hazelnuts in Fillmore County MN, we observed a weak leaf N response to N fertilization in the first year, a moderate one in the second, and a strong one in the third, accompanied by a weak growth response in the third year. This is consistent with our previous N fertilization trials, in which we found that leaf N responses appear first, with growth responses appearing only after a year or two, and yield responses after that. These results were only observed with plants that were also fertilized with K, indicating that these plants were deficient in K according to the law of the minimum: if a plant is severely limited by deficiency of one nutrient, it will not respond to applications of other nutrients. Unfortunately, after three years of observations, we have concluded that this particular clone will never yield strongly enough to observe an N response because it is of an exceptionally unproductive genotype. Thus we do not believe it will be productive to continue this trial.

We are disappointed at the failure of these three trials to help us develop better N fertilization recommendations for growers, and thus all the more grateful that in 2011 and 2012 we installed four new plantings explicitly for future N rate trials. These clonal plantings are comprised of three plants in a row that will be fertilized the same way, but from which date will be collected only from the middle plant, to eliminate border effects. Unfortunately, we will have to wait for them to mature before starting variable rate N applications, because our objective is to develop N recommendations for bearing plants, having already established that the N requirements for young plants is very low. Our plan is to collect leaf N and nut yield data every year until they start to show signs of N limitation, and at that point initiate N fertilization treatments. This may not be for a few years, since baseline leaf N levels collected in 2013 are well within the sufficiency zone, which is typical for young plants that have not yet begun to export N through their harvested nuts.

Weed management. In spring 2013 we established three of four planned weed management trials using 2012 rooted layers, two each at Rosemount and Waseca. Two (one at each location) were on former cropland, where weed pressure is likely to be annual weeds, and two were to have been on sod, where weed pressure is likely to be perennial weeds, which present a greater challenge. Because of high commodity prices for annual crops making cropland expensive, most land converted to hazelnuts is likely to come out of sod, so we deemed it important to develop recommendations for hazelnuts planted into sod. However, the site intended for the sod planting at Waseca had to be abandoned because of seasonally waterlogged soils in spring 2013. Treatments compare hoeing, landscape fabric, woodchip mulch, pre- and post-emergent herbicides, against simple mowing as the control. These treatments are all being tested at two levels: 0.5 m2 “divots” around individual plants, versus entire 2 m2 plots. Additionally, the woodchip mulch and herbicide treatments are being tested alone and in combination. Data collected include percent ground cover by weeds and hazelnut growth parameters.

In fall 2013 we established a fourth trial to substitute for the one that was flooded out at Waseca, on a former hayfield. Roundup was sprayed to kill the vegetation in ten paired planting strips, one of which was subsequently rototilled. Since many hazelnut growers do not own tillage equipment, we deemed it useful to determine how necessary tillage is. For this trial we used 2013 rooted hardwood stem cuttings which had been generated by our greenhouse trials. Weed control treatments are the same as in the three spring-planted trials.

Tree tube trials. Fall 2012 we installed tree tubes, of two diameters and two heights, donated by Plantra Corp., around new two plantings of layers and rooted cuttings. Some growers advocate tree tubes because they facilitate use of Roundup herbicide, and protect against herbivory, but we believe that they may have harmful effects on microclimate and plant architecture. We will evaluate growth when we remove the tubes fall 2013, but thus far our impressions are that they had a positive effect on growth and survival.

Pruning and coppicing trials. In 2013 we continued data collection on the three pruning and coppicing trials we initiated in 2012 (Rosemount, Staples and Horticultural Research Center) and implemented a fourth trial at the Gibson farm near Montevideo. These plantings, three of which were planted in 2000 and one in 2003, were becoming overcrowded and were getting too tall to pick, as well as starting to show declining yields. Badgersett Research Farm advocates rejuvenating hazel bushes by coppicing them to the ground when they reach this stage. Although coppicing completely eliminates yield for a year and reduces it for a second year, subsequently yields return to previous levels or greater. But we wished to test whether annual renewal pruning might work as well, without loss of a year of yield. Our data includes time required for implementation of treatment, weight of wood removed (in case biomass energy markets are ever developed), and yield response. In addition, we are documenting patterns of regrowth with photographs taken at intervals through the seasons.

It is still too early to come to any conclusions, but we can say that regrowth of coppiced plants was vigorous on most genotypes, but coppicing knocked some others back severely. Only one of the plants coppiced in 2011 produced any nuts in 2012, so we now know that coppicing eliminates two years of yield. Pruning was much more difficult to implement than coppicing, because branches that were cut were difficult to remove from the matrix of remaining branches. In some cases pruning stimulated vigorous growth of crowding suckers at the base of the plant, but in other cases the desired vase-shape that was the objective of pruning has been maintained with little additional effort. This all suggests that response to both coppicing and pruning may be highly genotype specific, in which case results may be inconclusive until the experiment can be repeated with clonal plant material.

It may be that, like with many woody plants, pruning is best initiated when plants are young, when removal of just a few branches can determine their adult shape. For this reason, in 2013 we initiated two pruning trials on three- and four year-old clonal plantings.

Outreach.
• Our fourth Midwest Hazelnut Growers Conference was held March 1-2 in Eau Claire, WI with the proceedings posted on the Midwest Hazelnut website www.midwesthazelnuts.org. Lois Braun and Molly Kreiser both presented at the conference.
• Braun presented at one of Rural Advantage’s “Third Crops Days” in Fairmont on March 11.
• On August 20 we hosted a field day, part of Rural Advantage’s Third Crops “Walk-N-Talk” series, at the hazelnut research planting on the St. Paul campus. We invited researchers with plots of other “Forever Green” crops that were adjacent to ours to participate, making it a “Forever Green” field day.
• Kreiser and Braun presented a poster at the Green Lands Blue Waters conference in Minneapolis Nov. 21-22.

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.

Collaborators: