Teff: An Alternative Crop for Oklahoma

Final Report for OS09-048

Project Type: On-Farm Research
Funds awarded in 2009: $14,948.00
Projected End Date: 12/31/2009
Region: Southern
State: Oklahoma
Principal Investigator:
Dr. Kefyalew (Girma) Desta
Montana State University
Kefyalew Desta
Oklahoma State University
Expand All

Project Information


Teff is a summer annual grass. We conducted two studies in 2009 and 2010 at various sites to evaluate performance of teff varieties and determine fertilizer recommendations. Teff grain yield ranged from 0.8 to 1.5 t ha-1 when grown in the field. We determined 67 kg N ha-1 as optimum N rate and about 25-50 kg P2O5 ha-1 as optimum P rate for teff. Additional N might be needed if teff is intended for multiple cut. Fertilizer rates should be adjusted for soil supplied N and P. Based on current costs and output prices, growers can produce teff profitably and with minimal financial risk.


Today's agricultural enterprise selection is derived by several factors including the conventional demand-supply dogma plus cultural and social forces. Unique to the US is immigrant communities from all parts of the world. With the immigrants came their favorite crops for food and herbs. Some of these crops and herbs are also well accepted by the local population for alternative uses creating a common interest and demand for those crops. One of such crop is teff [Eragrostis tef (Zucc.) Trotter]. Teff is an annual grass native to Ethiopia (Ketema, 1997). As food for human consumption, teff has unique qualities in that it contains high level of various minerals such as iron (Fe), magnesium (Mg), calcium (Ca), phosphorus (P), and thiamine ((National Research Council, 1996; Mengesha, 1965). It is an excellent source of essential amino acids, especially lysine, the amino acid that is most often deficient in common grain foods including wheat and millet (Lovis, 2003). Additionally it is low in gluten, and it can be an important component of diet for gluten intolerant people (Stallknecht et al., 1993). The crop is also excellent quality hay (Hunter et al., 2007; Nsahlai et al., 1998; Twidwell et al., 2002). There are several reasons for growing teff in Oklahoma and neighboring states as an alternative crop. Teff grows and completes its life cycle very quickly; about 90 to 100 days from emergence to maturity in normal years (Stallknecht et al., 1993). Teff gives reasonable yield when other cereals yield is depressed significantly in low or excessive moisture conditions (Hunter et al., 2007). If not for grain it can be grazed to cattle/horses or alternatively harvested for hay. The crop grows well in Veritsols such as Osage clays that have a water logging condition when precipitation is high. Teff is generally low input crop. It requires a modest amount of fertilizer contributing to reduced chemical use. Few insect and disease problems have been reported in this crop even in its native land, Ethiopia therefore no cost associated with pesticides. Teff's high grain price and niche market encourage small farmers to include it in their overall cropping system. Teff is becoming known as 'health' food among locals warranting more demand for the grain. More east African restaurants and cuisines are mushrooming everywhere in the US including Oklahoma. All these businesses depend on teff grain. Many teff distributors are losing their market and are looking for local production of teff to meet the demand for teff flour. The sustainable supply of teff flour for immigrant communities, for industries (health and baby food) and local residents for use in different recipe requires producing the crop locally rather than imports that are not reliable. The problem requires conducting on station and on-farm research on teff for good yielding and adaptable varieties with appropriate management practices. This is tremendously important as the management required to produce the crop varies with soil type and other physical conditions. In this on farm study, several varieties of teff were evaluated for adaptation and yield in Oklahoma. Verities tested in other states such as Oregon, Washington and Idaho plus some new lines stored in US Germplasm Repository were evaluated.

Project Objectives:

We hypothesized that identifying well performing teff varieties with appropriate fertility package for Oklahoma will enhance farm profitability and allow crop diversification in Oklahoma subsequently contributing to sustainability. We hypothesize that it is possible to reduce the risk of farming by increasing crop options specifically since teff is a dual purpose crop, fast in growth, and tolerant to moisture stress. The goal of this study was to add diversity of crops and create economic opportunities by adding teff in the cropping system for small farmers in Oklahoma and neighboring states. The specific objectives were to: 1. evaluate the adaptation, suitability and grain and forage yield of teff varieties, 2. develop nutrient requirement of selected teff varieties, and 3. demonstrate varieties and management package to farmers.


Click linked name(s) to expand/collapse or show everyone's info
  • Keith Boevers
  • Jerry Chandler
  • Tim Taylor
  • Yanqi Wu


Materials and methods:
Variety Study

We evaluated four to six and 10 teff varieties at two sites in 2009 and 2010, respectively. Trials were established at Hennessey (Shellabarger sandy loam fine-loamy, mixed, thermic Udic Argiustioll) and Summers (Kirkland silt loam- fine, mixed, superactive, thermic Udertic Paleustolls), and Stillwater (Kirkland silt loam- fine, mixed, superactive, thermic Udertic Paleustolls), OK in 2009. In 210, trials were conducted at Stillwater and Lake Carl Blackwell (Port silt loam-fine-silty, mixed, thermic Cumulic Haplustolls). We evaluated varieties in a Randomized Complete Block Design (RCBD) with two replications with sub-sampling. Plot size was 15 m x 6 m in 2009 and 1.5 m x 3.1 m in 2010 due to limited seed supply. We collected composite preplant soil samples from each field to determine nutrient needs of teff. Based on soil nitrogen (N), P, and potassium (K) content, only 56 kg N ha-1 and 23 kg P2O5 ha-1 was needed for all sites. We applied 1/3 of N and all P at planting. We topdressed the remaining 2/3 N as urea at booting. Teff was seeded into conventional or reduced till field with a grass seed drill at 11 kg ha-1 rate. In 2009, varieties were planted on May 13, June 2 and 5 at Stillwater, Summers and Hennessey, respectively. Emergence varied from 4 days at Stillwater to more than 3 weeks at Hennessey and Summers due to lack of rain. Broadleaf weeds were controlled with 2,4-D. Forage yield was determined from samples collected near flowering from 1 m2 area. We grew two teff varieties on a half-acre area to compare forage production (See photo 1). Grain yield (kg ha-1) was estimated from a 1.5 x 1.5 m plot for each variety. WE asked producers to rank the stand of each variety visually. Height, maturity date, and lodging were also recorded. Teff varieties were harvested in mid-September in 2009 and at different times (mid-August to late September) in 2010. One aspect of sustainable agriculture is economic profitability of a new intervention. It is important to evaluate the economic benefits of the introduction and adoption of teff as an alternative crop in Oklahoma. An economic analysis was conducted on forage and grain data.

Fertility Study

The soil fertility study was conducted at Hennessey and Summers in 2009, and at Summers and Lake Carl Blackwell (See photo 2) in 2010. In 2009, treatments include a check (no fertilizer), four N levels (0, 45, 67 and 90 kg ha-1) each at fixed levels of 25 kg ha-1 P2O5 and K2O, topdress N (45 kg N ha-1), only P (50 kg P2O5 ha-1) and K (69 kg K2O ha-1). In 2010, two additional N rates, 23 and 112 kg N ha-1 were added. The K treatment was removed from the treatment structure in 2010 since soil had adequate K and due to lack of significance in 2009. All rates were adjusted for preplant soil N and P level. The experimental design was RCBD with three replications and plot size of 3 m x 6 m. Teff variety quick-E and tiffany were planted in 2009 and 2010, respectively at a seeding rate of 11 kg ha-1 using the same planter indicated above. Teff was planted on June 2 and June 5, 2009 at Summers and Hennessey, respectively. Teff did not emerge until late June. In 2010 tiffany was planted on May 21 and 24 at Lake Carl Blackwell and Summers, respectively. We applied a third of N and all P and K (if applicable) at planting. The rest of N (including the topdress N treatment) was applied before booting. Measurements included preplant soil N, P, K, plant height, days to flowering, percent lodging, forage and grain yields, and maturity date. In 2010, we collected forage samples at booting and tested them for quality. Weed management and harvesting were performed as specified in objective 1 above. Data analysis and reporting for both objectives 1 and 2 Data was subjected to ANOVA using GLM/MIXED procedures in SAS. Following the results of omnibus ANOVA, orthogonal contrasts were developed to analyze trends and other hypotheses of interest. Significance was declared at p<0.05 probability level for all variables unless specified. The relationship between some measured variables was evaluated using correlation analysis. A linear–plateau model was fitted to grain yield data on N rate for the fertilizer rate study in 2020 (both sites). Heading ad maturity date data were not reported since comparison would not be reliable due to background interference. The economics of teff production was assessed using current input cost and output price. This analysis considered most of costs such as labor, machinery, energy, production inputs, and materials.


Field days and workshops were conducted in the past and some outreach activities are under way (see outreach part).

Research results and discussion:
Variety Study

Grain and forage yields In 2009, DZ-01-99 had the highest total forage yield for two-cut system. Varieties DZ-Cr-387, quick-E and tiffany (The latter two known hay cultivars in the US) yielded significantly lower forage than DZ-01-99. Tiffany produced significantly lower forage yield than DZ-01-1681. Forage yield at first and second cut averaged 2.4 and 8.6 t ha-1 (Fig. 1). For four of the six varieties, the two-cut system produced more forage yield than one-cut system (Fig. 1 vs. 2). Combined over sites, plant height and forage yield were positively correlated (r=0.7, p<0.01) with each other. However, grain yield was highest (1.5 t ha-1) for the shortest variety- quick-E. Grain yield of quick-E was statistically different from that of DZ-Cr-387 but not from the other two varieties (Fig. 2). The season was characterized by lack of moisture and high heat-index in June and July and heavy rain in mid-August to early September at grain filling and physiological maturity (Fig. 3) that affected growth of teff and interfered with harvesting of all trials. Average forage yield was comparable in both years. Nonetheless, it varied between the two sites in 2010; forage yield was 4 t ha-1 higher for Lake Carl Blackwell than Stillwater (Fig. 4). In 2010, DZ-01-889 at Lake Carl Blackwell had the highest forage yield although the same variety performed poorly at Stillwater (Fig. 4). At Stillwater, there was no significant forage yield difference among varieties. DZ-Cr-385 had the lowest forage yield averaged over the two sites. Grain yield did not differ between the two sites. Tiffany and DZ-Cr-387 had the highest grain yield while quick-E performed poorly in 2010 (Fig. 4). Quick-E was hit by a wave of heat at flowering in early July since it is an early variety. In 2010, at both sites average grain yield was comparable to 2009. Grain yield was not different between sites. Like 2009, grain yield did not have significant correlation with height but it was strongly correlated with forage yield (r=0.8, p<0.001). In 2010, some varieties such as DZ-01-974, DZ-Cr- 387 and tiffany performed well at both sites and had the highest grain yield. A satellite observation in a greenhouse (ideal growing conditions) showed that teff can yield up to 3.8 t ha-1 (average for DZ-Cr-387 and DZ-01-99). In another greenhouse variety study, average grain yield for 14 varieties ranged 2387 to 4227 kg ha-1 (data not shown). Almost all varieties of teff were susceptible to lodging in the field (20-70%). It is important to note that the effect of lodging on grain yield is dependent on the growth stage at which lodging occurred. If lodging occurs past milk stage, teff yield reduction could be minimal. During the course of this study, we observed that two-or more cut system can totally avoid lodging of any teff variety. Nitrogen fertilization is also key to managing lodging in teff. Yield stability over the two years (data not shown) suggested that tiffany and DZ-Cr-387 had relatively better stable forage yield over two years than other varieties. In both 2009 and 2010, DZ-01-99 and DZ-Cr-387 had better grain yield stability than any other varieties considered in the study. Teff Forage and grain quality Teff grain protein ranged 14 to 15.3%. Iron content was the only significant quality parameter among unwashed/washed or red/white varieties. Red or brown varieties had higher Fe than white varieties (Table 1). Washed teff grain (white and red) had lower Fe content than unwashed teff grain. Teff grain had the highest Ca, Fe, Mg, Zn and protein content vis-à-vis winter wheat, corn and grain sorghum. Grain sorghum had Fe content close to white teff. A comprehensive literature on teff indicted that the grain has balanced minerals and amino acids (Gamboa and van Ekris, 2008). This makes teff grain superior to some of the common grain cereals. It has minerals and proteins in an ideal balance for body uptake. For example, the phytate to iron ratio of teff ‘enjera’ is as low as 3 compared to corn and wheat (about 100). Phytate interferes with mineral uptake in the human body. Teff may compete well with other alternative crops in OK when the three purposes are designed together. Although not significant, teff forage yield ranged from 12 to 18% over site-years (data not shown).

Fertility Study

Forage and grain yields The two-site study in 2009 showed that quick-E forage and grain yields and yield components were not significantly affected by N, P and K rates. The only exception is P rate on forage yield at Summers where forage yield increased with P rate (Table 2). Grain and forage yields were not also affected by soil type x fertilizer rate interaction. Average forage yield was 7.0 and 6.6 t ha-1 at Hennessey and Summers, respectively. Grain yield at Hennessey and Summers was 753 and 638 kg ha-1 and was significantly different (p<0.05). At Hennessey, although not significant, grain yield had had a quadratic trend; yield peaked at 67 kg N ha-1 (Table 2). Application of 45 kg ha-1 topdress only resulted in 0.7 t ha-1 and 386 kg ha-1 more forage and grain yields, respectively compared with a 45 kg ha-1 split applied (preplant and topdress) at Hennessey in 2009. At Summers, this comparison did not result in a significant difference. In 2010, N rate had a significant effect on forage and grain yields, and height at Lake Carl Blackwell but only grain yield was significant at Summers. At both sites, grain yield reached maximum with the application of 67 kg N ha-1(Table 3). Average forage yield in 2010 was 11 t ha-1 across sites. Forage yield and plant height showed a linear trend as N rate increased at Lake Carl Blackwell. Height was 86 and 72 cm at Lake Carl Blackwell and Summers, respectively. In 2010, pooled over the two sites, teff grain yield data fitted well (R2=0.74) to the linear plateau model on N fertilization rates. The same critical N rate as the individual sites was obtained (Figure 5). The critical level shows that growers can add N at a rate of 4.7 kg ha-1, up to a maximum rate of 67 kg ha-1, beyond which it is no more economical to apply N for a yield goal of 1.3 t ha-1. Yield goal is of course a function of many yield limiting factors. Teff yield obtained in this study was less than its potential yield we obtained in a controlled environment and reported in other states (Norberg et al., 2007) primarily attributed to a combined effect of heat and low moisture, and lodging. Like the variety study, in both seasons, extended heat in June-July resulted in significant growth reduction and flower abortion (Figure 3 depicts the temperature and rain fall patterns in 2009). The comparison of 45 kg ha-1 topdress only versus split-N (between preplant and topdress) did not show a significant difference at both sites and for all three variables. The data suggest that topdress N would be the preferred application time. However, it is important to apply N fertilizer if soil supplied total inorganic N at planting is less than 20 kg ha-1. In 2010, teff grain yield increased linearly with an increase in P rate from 0 to 50 kg ha-1 at both sites (Table 3). Phosphorus recommendations developed for weeping lovegrass (Eragrostis curvula) in Oklahoma could be adopted for teff (Zhang and Raun, 2006). Forage quality 2010 Forage protein ranged between 9 to 17%. In 2010, sites differed in protein content; 11% at Lake Carl Blackwell and 16% at Summers, both harvested at the same stage (data not shown). Yet, other quality measurements such as acid detergent fiber (%), neutral detergent fiber (%), and total digestible nutrients (%) did not differ across sites or N fertilization regimes. Combined over sites, protein increased linearly with an increase in N rate. Phosphorus fertilization did not affect forage quality. Likewise, comparison made between topdress only and split N at N rate of 45 kg N ha-1 did not have statistically significant difference in protein. This shows that the organic matter supplied N could serve as starter N and all N can be topdressed later in the season. Particularly for dual purpose system this could be a significant saving in N fertilizer. Teff forage qualities were comparable to Timothy grass (Miller, 2009). The results of the teff fertility study suggest that for Central Oklahoma N fertilization decision should be based on soil test for N, P and K. Biological and economic optimum fertilizer N rate is about 67 kg ha-1 which corresponds with a 1.2 t ha-1 yield goal. Based on soil N, it would be better to topdress N instead of applying all N at planting. Phosphorus should be applied based on soil test index and percent sufficiency. The existing recommendation developed for weeping love-grass can be used to recommend P. Starter fertilizer is necessary if the soil is poor in N. Our experience showed that topdressing N can bring back the crop but if weather is dry, it takes the crop more energy to achieve vigorous growth. The crop is sensitive to lodging, and fertilization is at the center of all crop management practices for minimizing lodging. If conditions are ideal for multiple cut, N fertilizer must be applied following the start of regrowth.

Extension/Outreach (objective 3)

Please see publication/outreach below.

Literature Cited
    • Agegnehu, G. , A. Ghizaw, and W. Sinebo. 2006. Crop productivity and land-use efficiency of a teff/faba bean mixed cropping system in a tropical highland environment. Experimental Agriculture 42: 495–504.
    • Berhe, T., L.A. Nelson, M.R. Morris, and J.W. Schmidt. 1989. Inheritance of phenotypic traits in teff: 1. Lemma color. Journal of Heredity 80:62-70.
    • Desta, K. 2008. Teff as a potential alternative crop for Oklahoma. Plant and Sciences Extension Newsletter, Oklahoma State University, Stillwater, Oklahoma; Volume 1, no. 20.
    • Gamboa, P.A., and L. van Ekris. 2008. Survey on the nutritional and health aspects of teff (Eragrostis Tef). Red-alfa lagrotech, comunidad europea.
    • Hunter, M., P. Barney, T. Kilcer, J. Cherney, J. Lawrence, and Q. Ketterings. 2007. Teff as emergency forage. Cornell University extension, Agronomy Fact Sheet Series, FS 24. Ithaca, NY.
    • Kebede, H., R.C. Johnson, and D.M. Ferris. 1989. Photosynthetic response of Eragrostis tef to temperature. Physiology of Plant 77:262-266.
    • Ketema, S. 1997. Tef. Eragrostis tef (Zucc.) Trotter. Promoting the conservation and use of underutilized and neglected crops. Institute of plant genetics and crop plant research, Gatersleben/Int. plant genetic resources institute, (IPGRI) Rome Italy
    • Lovis, L.J. 2003. Alternatives to wheat flour in baked goods. Cereal Foods World, 48: 61-69.
    • Mengesha, M.H. 1965. Chemical composition of teff (Eragrostis tef) compared with that of wheat, barley and grain sorghum. Economic Botany 19:268-273.
    • Miller, D. 2009. Teff grass: a new alternative. In: Proceedings, 2009 California Alfalfa &amp; Forage Symposium and Western Seed Conference, Reno, NV, 2-4 December, 2009. UC Cooperative Extension, Plant Sciences Department, University of California, Davis, CA 95616.
    • Norberg, O. S., R. Roseberg, C. Shock, E. Feibert, L. Saunders, E. Eldredge, B. Charlton, and J. Smith. 2007. Teff (Eragrostis tef) yield and quality as influenced by irrigation and nitrogen. The ASA-CSSA-SSSA International Annual Meetings Abstract.
    • Nsahlai, I.V., N.N. Umunna, and M.L.K. Bonsi. 1998. The utilization of teff (Eragrotis tef) straw by sheep fed supplementary forage legumes with or without either crushed maize grain or wheat bran. Small Ruminant Research 29 (3): 303-315.
    • National Research Council. 1996. Lost crops of Africa, Volume1: grains. National Academy Press, Washington DC.
    • Spaenij-Dekking, L., Y. Kooy-Winkelaar, F. Koning. 2005. The Ethiopian cereal Tef in celiac disease. The New England Journal of Medicine 353(16): 1748 - 1749.
    • Stallknecht, G.F., K.M. Gilbertson, and J.L. Eckhoff. 1993. Teff: Food crop for humans and animals. p. 231-234. In: J. Janick and J.E. Simon (eds.), New Crops. Wiley, New York.
    • Twidwell, E.K., A. Boe, and D.P. Casper. 2002. Teff: a new annual forage grass for South Dakota. Ex 8071. Coop. Ext. Serv. South Dakota State Univ. Brookings, SD.
    • Zhang, H., and W.R. Raun. 2006. Oklahoma soil fertility handbook. Department of Plant and Soil Sciences, Oklahoma State University, Stillwater, OK.
Participation Summary

Educational & Outreach Activities

Participation Summary:

Education/outreach description:

Three field days were conducted; the first field day was conducted on August 15, 2009. A workshop was also conducted on the same day before the field day. The other two field days were conducted on September 30 and October 2, 2010. One factsheet/production Technology is in preparation. Also, a manuscript for journal publication is in preparation for submission. Two abstracts have been published. Five oral presentations were delivered to a wide range of audience (producers, Extension Educators, and other professionals). We also delivered three poster presentations. Additionally, more than 700 pamphlets were distributed at different meetings (producer, Educators, professionals etc.). A documentary about teff was aired on SUNUPTV. A webinar based seminar will be conducted in April 2011. Desta et al., Teff at 2009 ASA-CSSA-SSSA Meetings- http://acs.confex.com/crops/2009am/webprogram/Paper54361.html Mohammed et al., Teff at 2009 ASA-CSSA-SSSA Meetings- http://a-c-s.confex.com/crops/2009am/webprogram/Paper53348.html

Project Outcomes

Project outcomes:

The results of the variety and fertility studies could lead to more participation of small-scale farmers in teff production in Oklahoma. So far, several producers showed interest in using the crop as a hay crop for sheep and horse. At least one producer is considering processing and packaging of teff. More importantly, the project attracted more than dozen entrepreneurs, producers and others who are eager to see more production of teff in Oklahoma to meet their ever-increasing teff grain feedstock for flour. At least two entrepreneurs were interested in growing teff in a large scale to provide health food industry with teff grain. Although a niche market, there is considerable demand for teff grain in North America. This project also opened an opportunity to interact and share information in the High Plains of the US. The project team has been exchanging information with Kansas and Texas teff groups. A large-scale networking for common regional project and germplasm exchange is under discussion among these groups.

Economic Analysis

There is no doubt that the dual-purpose teff system will be profitable by any measurement. Cost-benefit analysis based on generous estimates of costs and conservative output prices resulted in an attractive net benefit. Accordingly, teff producer can bring net benefit of about $3300 ha-1 ($1320 ac-1) with estimated total cost of $1458 ha-1 ($583 ac-1, Table 4). If instead of hay straw is marketed at an output price of about $60 ton-1, it can result in a net benefit of more than $2400 ha-1. In this analysis we assumed current production and output costs. Although, the profit margin a producer willing to accept depends on the scale of production, the net benefits indicated here suggest that teff can be produced with minimal financial risk in Central Oklahoma.

Farmer Adoption

Results of studies and demonstrations of this project attracted producers and entrepreneurs alike. According to an informal survey, at least 10 producers tried teff since we established the studies. Additional 8 farmers and entrepreneurs out of Oklahoma (Canada, Texas, California, Minnesota, Kentucky and Colorado) contacted us based on field days and published materials. At field days and workshop sessions participants, mainly producers and others who are interested in teff evaluated varieties performance. Most participants were surprised about the grain size of teff. Some mentioned that they prefer to grow teff for hay or to graze it to their sheep and horses rather than for grain owing to the seed size. However, they are also appreciative of the nutrient rich grain. Some tried to make cookies and bread out of teff and informed the project coordinator about their success. Producers located in Oklahoma were interested in the fertility study since no recommendation has been established for teff. The pamphlets and fact sheet under preparation will provide them with the guideline for fertilization in teff.


Areas needing additional study

The project was developed to conduct experiments and demonstrations based on past results from other studies. However, growing conditions in Central Oklahoma are very different from sites where teff has been in production. Also, the results of this study lack consistence across sites and years. In light of this, further research and outreach must be planned for teff in Oklahoma. Throughout the two cropping seasons, we realized that basic studies related to soil moisture, day length and temperature must be established for a successful dual purpose teff. A 2-year project funding has been granted for $72, 000 to address this problem. The SARE sponsored project data was used to justify the grant proposal. Fertility studies and variety evaluation over years and sites will continue since the current data is restricted to selected areas. We are evaluating more than 16 teff genotypes in a greenhouse and field. A tissue culture and transformation studies are under way at Oklahoma State University with a goal to improve the lodging and stress tolerance (heat) of teff varieties. We have identified weed control methods to manage grasses in teff, however, additional studies are needed to reach at conclusive results.

Information Products

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.