Macro and micronutrient density in staple grains and amaranth grown on conventionally versus sustainably managed silt loam soil

Final report for FNC19-1195

Project Type: Farmer/Rancher
Funds awarded in 2019: $9,000.00
Projected End Date: 02/28/2021
Grant Recipient: Sumner's Farm, LLC
Region: North Central
State: Missouri
Project Coordinator:
Dr. Walton Sumner, II
Sumner's Farm, LLC
Expand All

Project Information

Description of operation:

The farm is an 82 acre area on gentle hills with a creek and an artificial pond. Half of the land has been in hay production for 40 years, and the other half is forest, with trees up to 2 feet in diameter. The experimental farming area is 10,000 square feet, divided into 30 plots of 15-20' x 20', on a gently south sloping hay field. There are 5 rows of plots, separated by wood-core berms and 4-foot paths. Each row of plots is divided in half so that one half are control (conventional) soil management, and the other half are managed in pursuit of a chemically balanced soil chemistry. The control and intervention sides of each row are separated by a 15 foot area of mown grass. The top two rows have intervention plots to the west, and the bottom three rows have control plots to the west. The balanced side management also includes large amounts of composted leaves and coffee grounds, and small amounts of biochar or charcoal. Biochar is recovered from burn piles, an aerobic biochar rig that extinguishes hot coals, or anaerobic pyrolysis tubes. Biochar is mixed into compost immediately after production, and eventually the mixture is spread over intervention plots with other amendments, and tilled to a depth of 4-5 inches. The oldest intervention plots have been in development for 6 years. Cover crops of wheat, rye, or field peas grow through crop residues over the winter, and have been used for 2 years.

The farm's main crop is hay. The experimental field is planted with multiple varieties of amaranth, sweet corn, cowpeas, okra, sweet potatoes, pumpkins, sunflowers, and in the winter, rye, wheat, and peas as cover crops.


Many processes in conventional farming could cause trace element deficiencies in row crops, while crop selection and rising COlevels sustain or increase carbohydrate and calorie content. If food has a low ratio of trace elements to calories, consumers may overeat calories to acquire sufficient trace elements. The resulting nutritional imbalance would affect health, but some effects should be quickly reversible.

We are experimenting with sustainable, “balanced” fields, amended with major chemicals as guided by soil analyses, and with organic matter and trace elements from a low cost compost of tree leaves, spent coffee grounds, and charcoal made from waste debris and chaff. Compared to control fields managed with only lime and nitrogen, our crops grow faster and yield qualitatively different produce in balanced fields. In 2018 we prepared samples of multiple high yield, nutrient-dense crops for trace element analysis, including winter wheat, amaranth, okra, sweet potatoes, and pumpkins. For most crops, we are obtaining one trace element analysis in each growing condition. To prepare for human studies of nutrient dense crops on health, we sought SARE funding to conduct additional analyses of trace elements, selected organic compounds, macronutrients, and calories from staple crops in balanced and control growing conditions.

The low cost compost did not contain sufficient trace elements to correct soil imbalances, We therefore mixed mostly organic mineral amendments guided by recommendations from OrganicCalc, an online service, into compost, and spread nearly finished compost at 1/4 to 1/2 inch depth over balanced plots. This was substantially more expensive than coffee grounds and oak and maple leaves. One of these amendments was Azomite, which contains multiple essential trace elements as well as several potentially toxic elements, like lead. Conventional fields were fertilized using commercial products, typically starting with 12-12-12 NPK and adding urea for nitrogen, bone meal or superphosphate, and muriate of potash as needed to achieve recommended rates. Plots were tilled simultaneously in late spring to mid summer, depending on when the winter monoculture cover crop was ready to harvest. 

Soil measurements nearly always changed in the expected direction, but OrganiCalc targets were not met consistently. OrganiCalc limits annual repletion rates for elements that would be toxic and hard to remove if overdone, such as copper and zinc. Consequently, the program may take several years to achieve the targeted level of some elements. Nevertheless, zinc levels in the balanced plots rose to 4-7 times the levels in conventional plots. Although the extension service amendment recommendation program was capable of including zinc and sulfur, it never did. Soil organic matter rose from 2-3% in the conventional plots to 3-4% in the balanced plots, which seemed remarkably modest for the amount of labor invested in the compost piles. 

Samples were harvested, dried, ground to dust, and submitted to the Ohio State University Service Testing and Research Laboratory for elemental analysis. Several edible portions of weeds were included, as these plants’ genetics would not be directly altered by human breeding efforts, although indirect effects are possible (e.g. selection pressures from previous land management). Samples also were submitted to Midwest Laboratories for proximate analysis (e.g. caloric content, total protein) and selected vitamins, such as B2, B3, and B6. Vitamin assays are very expensive compared to elemental assays. Knowing that our soil at baseline was rich in magnesium, iron, and manganese, but relatively zinc deficient, we selected vitamins requiring at least one zinc dependent enzyme, as documented in the BRENDA enzyme database, for a step documented in the KEGG database of biosynthetic pathways. BRENDA always documents a preferred cation, but often also describes slightly slower synthesis with alternative cations, especially in microbes and fungi. This raises the possibility that “life will find a way” even in the face of seemingly stringent constraints. 

Proximate analyses were very consistent across crops and soil management strategies: soil management did not change calories and protein per gram of dry weight to any clinically significant degree within the crops that we tested. This result may not be generalizable due to the limited number of crops and samples within crop that we tested. For now we feel comfortable comparing nutrients per unit of dry weight with the expectation that calories per dry weight are fairly constant for each crop, meaning that within crop comparisons of nutrient density depend only on nutrients per dry weight.

Levels of zinc-dependent vitamins within crops were not significantly affected by soil management (ANOVA, p=0.7). Pyridoxine data are shown for wheat and rye, as an example with plenty of data points. See Pyridoxine data.

Within crops, levels of major nutrient elements were generally similar across the soil management strategies. Phosphorus was the most different across soil management strategies, but probably not enough to be clinically important. See Phosphorus graph. Below the x-axis, the dagger (†) locates the soil target level in µg/ml, and the short color-coded line segments indicated measured levels in 2 soil samples. Zinc and sulfur levels were different across crops, but not particularly different across soils. 

Within crops, levels of most trace nutritive elements were similar. Cobalt levels were essentially fixed across both crops and soil management, all values being in the remarkably narrow range between 0.83 and 0.87 µg/g, with a single outlier at 0.91, in spite of the balanced soil getting a boost from the Azomite. Molybednum levels were much higher in the balanced soil crops, presumably due to direct supplementation with Azomite and a molybdate mineral; nevertheless, this probably would have almost no impact on health in most people, due to an exquisite ability to regulate molybdenum levels across a wide range of intake values. 

Nickel and manganese were interestingly different, however. Both are essential nutrients for plants and people, but problematic when presented in excess. Manganese is massively above target in the original soil, and should have been slightly increased by Azomite additions to the balanced soil, yet the balanced soil crops had much lower levels of manganese. See Manganese graph. 

Nickel is especially interesting. Unlike Mn, Ni levels in our soil were well below Missouri and US farmland averages, as indicated by the marks below the x axis on the graph. Trace amounts of nickel were added in Azomite, again. Nickel levels in crops and wild edibles were consistently higher with conventional soil management than with balanced management, except in the case of winter rye, where nickel was always below the detection limit. See Nickel graph. What makes this so interesting is that for a decade or more dermatologists have been dabbling in low nickel diets as a treatment for people with systemic nickel allergy syndrome (SNAS). Low nickel diets typically start to worry about foods with a nickel content of 100 µg/kg (e.g. carrots), and may outright exclude foods exceeding 500µg/kg, such as legumes, with the general goal of limiting total daily nickel consumption. This is spectacularly constraining for the patient. If those foods are 90-95% water, then the range of worry starts at 1-2µg/g dry weight, and exclusion starts at 5-10µg/g of dry weight, as indicated by the small arrows under the attached graph. For people who believe that they suffer from SNAS that is responsive to dietary limitation of nickel, eating crops from balanced soil fields might allow a much more diverse diet if every food item has lower Ni concentration than predicted. Other specialties have started to look at dietary nickel as an exposure that might provoke some of their common diseases. One intriguing result is that Ni may alter gut flora and gut epithelial tight junctions, which prevent systemic exposure to large, potentially allergenic food molecules before those are digested well. 

Why did these plants accumulate nickel? We have an experiment underway trying to test two hypotheses. One is that Ni is a useful substitute for some catalytic cation missing in the conventionally managed soil, but present in the balanced soil. Another possibility is that Ni is specifically sought because the plant is deploying one of the two plant enzymes known to require nickel: those are a hydrogenase and a urease. This could mean that the plants took up nickel to metabolize the urea based fertilizer applied to the conventional plots. Serendipity. 

Levels of toxic elements were rarely elevated, but when cadmium (Cd) and lead (Pb) were elevated at all, elevations were much higher in the conventional soil crops than in the balanced soil crops. See Cadmium graph. This was particularly true for Cd in sunflower seeds and amaranth leaves. This is clinically interesting because amaranth is grown as a vegetable in west Africa and India, and is renowned for being a protein-rich crop that will grow on poor soils. This result suggests that local consumers nevertheless could pay a price in exposure to some toxic metals when vegetable amaranth grows on soils that are deficient in some respect, such as trace elements or soil organic matter. Another general way to say this is that some edible parts of these heavy metal accumulators (tobacco deserves examination, too) contain much lower concentrations of toxic metals when grown in nutrient rich soil. 

The general theme that emerges is that many crops accumulate lower amounts of heavier nutritive and non-nutritive elements when grown in chemically balanced soil with slightly increased levels of soil organic matter. Whether this trend would be still more dramatic if SOM reached 5-10% and all trace elements were fully balanced remains to be seen. 


Project Objectives:

1. Calculate ratios of trace elements and selected essential nutrients to calories in crops grown with conventional fertilizers versus balanced amendments of organic matter, major, and trace elements. The hypothesis that these ratios would be lower in conventional compared to balanced field crops was not well supported by the data collected for elements and vitamins, but could be true with further soil improvement or with other essential nutrients, such as selected amino acids. We have some evidence, not part of the original SARE proposal, that essential amino acid density increased with balanced soil management.

2. Pilot test a year-long schedule for providing diabetic human subjects with a diet based on seas, trees, and rich row crops versus conventional diabetic diet advice and row crops. This would determine resource requirements for human studies using nutrient dense foods to prevent infections in diabetic employees, replicating a similar 2003 study of multivitamins. This particular study design is probably impractical at current scale, and will require better weed control than we achieved without herbicides. However, the nickel findings suggest that some short-term experiments are possible with SNAS patients, such as a crossover trial where gut leakage is monitored following a single meal comprised of produce from balanced versus conventional soils. 


Materials and methods:

Our experimental field supports fractional factorial designs to compare long-term conventional soil management to sustainable, nutrient-balanced management. The experimental field is a fenced area of 120’ north-south by 150’ east-west, with a gentle 2’ drop from north to south, just below a forest and above the floodplain. The conventional areas are the 70×70’ southwest corner (established in 2018) and 40×70’ northeast corner (2019). These areas receive calcium carbonate to correct soil pH, and nitrogen fertilizer at labelled rates. The balanced areas are the 70×70’ southeast corner (2016) and ~30×70’ northwest corner (2019). Balanced areas receive annual major element amendments guided by soil analyses and Solomon and Reinheimer’s Organicalc (, and trace elements primarily from tree leaf and coffee ground composted with fine charcoal.

The field is divided into rows of 15-20×20’ subplots. Water capturing swales and 3” wood-core berms line the south edge of subplot rows. Crop debris is left on fields. We use shallow tilling for initial weeding and soil preparation before paper pot transplanting. Winter wheat, rye, and Austrian field peas are used as cover crops.

Preferred crops are resilient, versatile, and nutrient dense. Amaranth, okra, and pumpkin have edible leaves, fruit, and seeds; tolerate hot weather; and can be grown as microgreens or row crops, as can sunflowers. Sweet potatoes have edible leaves and tubers, and compliment okra. Winter wheat, rye, and peas tolerate cold weather and produce edible seeds. Corn and wheat are staple crops, and amaranth could be.

Identical species are grown in conventional and balanced subplots, and harvested simultaneously. Crop samples are dried to constant weight and powdered in a food processor. The OSU STAR laboratory will perform elemental analyses on 5g samples. Midwest laboratories will perform macronutrient, kilocalorie, vitamin, lipid, and amino acid analyses on powdered 50-250g samples.

Midyear update, October 2019

After submitting the original budget, it became clear that Midwest Laboratories could analyze proximates and minerals simultaneously in samples submitted for their animal feed “F9” testing, and that some vitamins could be analyzed at less expense as animal feed rather than human food. Consequently the original budget categories are revised. 

In addition, a report in the bioinformatics literature pointed to another strategy for selecting nutrients of interest. Scott-Boyer et al* reported that Mg, Zn, and vitamins B1, B2, and B6 are cofactors for many enzymes that are related to type II diabetes in GWAS studies. My review of KEGG biosynthetic pathways and BRENDA enzyme data suggested that all 3 vitamins have 2 or more synthetic steps in which Mg++ or Zn++ are preferred cofactors, and we already had some data suggesting higher Mg levels in intervention soil crops. A pilot test suggested that amaranth content of B6 might be especially sensitive to soil management, and that other B vitamins might be affected, but less dramatically. Consequently, analyses of several B vitamins were done in amaranth and rye samples. Wheat samples were collected but large amounts were lost to rodents while drying, leaving too little material to conduct the suite of analyses in triplicate for each growing condition. 

* Scott-Boyer MP, Lacroix S, Scotti M, Morine MJ, Kaput J, Priami C. A network analysis of cofactor-protein interactions for analyzing associations between human nutrition and diseases. Sci Rep 2016;6:19633.

Research results and discussion:

SARE Poster 2019 09 20 36×60

Midyear update, October 2019

The spring of 2019 saw record-breaking rainfall across the midwest, which complicated shallow tilling efforts in the new fields, especially 4 of the 6 new intervention plots, where the tractor repeatedly got stuck in mud. Eventually rows were double-dug by hand throughout the 12 new control and intervention plots. Plants grew very poorly at first, prompting a mid-summer soil analysis which was mainly remarkable for a pH of 5.2 for the 6 new intervention plots, in spite of what was expected to have been adequate lime application in the spring. In September a home garden pH meter was used to construct an extensive pH map of the entire experimental area (at least 1 sample per plot for all 30 plots, and 6 samples per plot for the 3 most suspect of the 6 new intervention plots (mean 6.5, SD 0.35), and 2 samples per plot for the the remaining 3 new intervention plots (mean 6.5, SD 0.16)), with no pH measures below 5.8, and most measures between 6.3 and 6.8. 

Sweet corn and 4 legumes had been planted in the new intervention and control fields and collected and dried for analysis, but given the tilling problems and fluctuating pH data, a late crop of corn was planted in 4 of the older plots, where rye had grown well and the soil conditions are likely more stable and the amendments better mixed. It is now tasseled and I expect to harvest ears before first frost. Among the 4 legumes, cowpeas had done very well and were planted in the remaining 2 old plots where rye had grown. 

Analyzable data have been collected for 2018 amaranth crops and 2019 winter rye. Over 100 five-gram samples of a wide variety of crops have been collected for mineral analysis by the Ohio State University STAR lab, and will be submitted when the remaining samples are collected. A large amount of okra and pumpkin fruit have been collected for analysis. Amaranth and sunflowers also were planted late and have not yet been harvested, but are producing seeds now. 

The results of 2018 amaranth and 2019 rye are summarized in the attached pdf file, which is a poster that was presented at the 2019 FMX conference in Philadelphia, arranged by the American Academy of Family Physicians (AAFP). The poster illustrates that soil nutrients like Mg++ and Zn++ are preferred divalent cations for many steps in the synthesis of vitamins related to type II diabetes, but that there are often alternative cations that work to some degree. The poster further illustrates that amaranth grown in the intervention soil had higher concentrations (lower calories per mg of nutrient) of Mg++ and Zn++, and lower concentrations of Cu++ and Zn++, compared to the control field. Vitamin synthesis was not different. Rye showed similar but not statistically different trends. While these initial results do not suggest a clinically important difference in nutritional value, and the differences may be more obvious in mineral content than vitamin content, there can be some effect from soil chemistry on nutrient density. 

Participation Summary
1 Farmer participating in research

Educational & Outreach Activities

1 Curricula, factsheets or educational tools
2 Webinars / talks / presentations

Participation Summary

2 Ag professionals participated
Education/outreach description:

So far, presentations have been made to medical professionals rather than farmers and ranchers. The promised video is nearing completion and will present the information in this report in detail, along with illustrations of composting, cover crops, polarisation for weed killing, and trace element amendments. 

The attached poster presented amaranth and rye results, and explained the relevance of soil health to type II diabetes to family physicians attending a national educational conference.

 SARE Poster 2019 09 20 36×60

Learning Outcomes

3 Farmers reported changes in knowledge, attitudes, skills and/or awareness as a result of their participation
Lessons Learned:

We suspect that urea-based fertilizers cause most plants to accumulate nickel as a cofactor for a urease, and that some of that nickel makes its way to the edible crop. Furthermore, this is likely to have a large impact on a few people’s health, and might have some impact on most people’s health, and therefore warrants further investigation. 

We found that the balanced soil amendment program is too expensive to scale very well, although we were not specifically trying to evaluate costs. Two experimental strips have been converted to no-till strategies, one using sheet mulch in a market garden simulation and one using propane burns in an open field simulation. We have designed, built, and successfully used a stab planting spear that allows very rapid direct seeding of no-till plots. 

The nickel difference and the essential amino acid changes, not described here, are interesting enough to consider pursuing a more aggressive research agenda. To that end, we are planning to establish a non-profit committed to studies of soil, food, and human health. This non-profit would work with Sumner’s Farm LLC and other farmers to study soil-food-health relationships, and disseminate findings. The video being produced for this project will likely be published on channels managed by the non-profit, if agreeable to SARE. 

Project Outcomes

1 Grant received that built upon this project
1 New working collaboration
Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.