Variety Trials and Production Methods for Vegetable Amaranth in the Northeast

This study investigated the production of amaranth (Amaranthus spp.) as a leafy green in the northeastern temperate climate. Amaranth is a productive and resilient crop with cultural, economic, and nutritional significance to many cultures around the world. Growing ethnic crops in the northeastern U.S. is an opportunity for growers to expand into new and diverse markets. Amaranth’s pervasiveness in global foodways and tolerance of many biotic and abiotic stresses make it a promising option for growers to engage with the ethnic produce market and diversify production. However, intensive production research for vegetable amaranth is lacking, especially in temperate climates, and amaranth varieties are underdeveloped. In 2016, ten vegetable amaranth varieties were evaluated for performance in the northeastern temperate climate. The experiment was a randomized complete block design with four replications and ten plants of each variety per replication. Yields of each plot and leaf to stem ratios of a two-plant subsample per plot were recorded. CVs were calculated for each variety as a measure of yield stability. The varieties included eight commercially-available A. tricolor varieties (‘Asia Red,’ ‘Red Garnet,’ ‘Red Callaloo,’ ‘Green Pointed Leaf,’ ‘Red Stripe Leaf,’ ‘Miriah, ‘Southern Red’), one commercially-available A. viridis variety (‘Green Callaloo’), and one heirloom A. hybridus variety from Burundi (‘Mchicha’). All plants were greenhouse-started and transplanted to a low-tunnel system, constructed with galvanized metal hoops and 0.8mil clear slitted plastic. The experiment was repeated seven times over the 2016 growing season. There was little variation between the varieties in the middle of the summer. However, two varieties that excelled in the early and late season (‘Green Pointed Leaf’ and ‘Miriah’) were also top performers all-season. Targeted production and marketing strategies have the potential to improve variety desirability. A comparison of plasticulture production systems for vegetable amaranth was also conducted during the 2016 growing season. Two varieties of A. tricolor (‘Red Stripe Leaf’ and ‘Green Pointed Leaf’) were used throughout the experiment. The four treatments were 1) gothic-style high tunnel covered in a double inflated layer of 6-mil greenhouse plastic; 2) low tunnels over raised beds with black plastic mulch, constructed with galvanized hoops covered in 0.8-mil clear slitted plastic; 3) raised beds with black plastic mulch; and 4) uncovered bare soil. A split-plot design with 10 plants in each plot and four replicates was repeated three times over the season. High tunnel plots were excluded from the second planting due to extensive Woodchuck (Marmota monax) damage. Low tunnel plots had the greatest yields in every planting. The magnitude of production system effects decreased as ambient temperatures increased throughout the season. However, rankings of the four production systems were consistent in each experiment. Yield rankings from greatest to least were: low tunnel, black plastic mulch, high tunnel (when present), and bare soil. There were occasional significant differences in leaf to stem ratios. However, the response did not follow a discernable pattern based on production system, nor was it correlated to yield. Leaf to stem ratio is likely genetic and outside the influence of production system.

environmental impact for small and beginning farms. By incorporating alternative crops, these growers may also benefit from expansion into new markets (Fritz and Meyers, 2004). One such market, which is receiving increased attention, is the ethnic crop market. As the United States population becomes more diverse, the foods and flavors that are valued by cultures around the world are becoming increasingly popular . In 2015, recent immigrants and their U.S.-born family members comprised 27% of the U.S. population, and this percentage is projected to continue to rise (U.S. Census Bureau, 2015). These shifting demographics represent an expanding market of consumers seeking their traditional produce. Additionally, diverse communities lead to an expansion of the American "food repertoire," which broadens the customer base for traditional ethnic foods .
Amaranth (Amaranthus spp.) is an attractive option for growers interested in connecting to the ethnic crop market because of its particularly wide cultural significance. Amaranth is a common green in many Asian, Latino, African, and Caribbean cultures, and competition from imports is extremely minimal. Amaranth leaves are also highly nutritious and often striking in color, qualities that lend favorably to marketing as a novel substitute for more common greens (National Research Council, 2006). However, by nature, the marketing chains and production protocols for alternative crops are not well defined. Although Amaranthus species inhabit a wide range of latitudes, cultivar-level research is sparse on suitability for intensive production in temperate climates. Studies suggest that variety sensitivities to temperature, moisture, and photoperiod are varied Wu et al., 2002). Furthermore, the shortage of regional production research means little is known about ancillary issues such as transplant tolerance and pest and disease occurrence. Amaranth varieties are less developed than more common vegetables, making replicated, regionally-focused research especially important in determining the performance and stability of available varieties.
Amaranth is a heat-loving C4 crop, and its culture will differ significantly from popular spring and fall greens (Teutonico and Knorr, 1985). Plasticulture systems, including plastic mulches, drip irrigation, and plastic crop covers, are often used to enhance the yields of warm-weather crops in temperate climates. The increased rate of heat accumulation in these systems is closely linked to increased vegetative growth in heat-loving crops; varying degrees of season extension are also possible. A wide spectrum of plasticulture designs and techniques are available to growers, and the fitness of a particular design is largely dependent on target crops, target markets, and existing production strategies . Small and beginning farmers who are interested in implementing plasticulture strategies require research-based information on the trade-offs associated with various plasticulture system-crop combinations. Given amaranth's affinity for high heat, it is likely that plasticulture production would be of some benefit in the northeastern region. Direct investigation of plasticulture amaranth production is warranted due to significant differentials in temperature requirements and light saturation points, relative to more common crops.
Studies have reported successful amaranth production with black plastic mulch and drip irrigation, but the use of crop covers has not been investigated (Meyers et al., 2010;Sciarappa, 2016).
The primary objectives of this study were the following: 1. Evaluate vegetable amaranth varieties for plasticulture production in the northeastern temperate climate.
2. Assess the benefit of three plasticulture systems for vegetable amaranth production in the northeastern temperate climate.

AMARANTH
Amaranth refers to plants of the genus Amaranthus, which contains 60-70 species of annual, mostly monoecious, plants with an upright, moderately branched growth habit. Amaranths are cultivated for ornamental, grain, or vegetable production, but most species are classified as weeds, including the well-known and troublesome pigweeds (Teutonico and Knorr, 1985). Separation of these types is not entirely distinct, taxonomically or functionally, because all Amaranthus species have edible stems, leaves, and seeds. The young leaves of grain types are commonly eaten as greens, and, although the domestication of wild amaranths began over 2,000 years ago, many more species are eaten globally than would be considered truly domesticated. Amaranths have not been the subject of modern intensive breeding efforts, and frequent hybridization between cultivated and wild populations has led to the existence of many intermediate types (National Research Council, 2006).
Amaranths have a high capacity for osmotic adjustment (Liu and Stutzel, 2002) and a C4 photosynthetic pathway that allows efficient use of CO2 in a large range of temperature and moisture stress environments, likely a major factor in their wide geographic distribution .
Amaranth is grown and eaten as a vegetable in over 50 countries worldwide, in such geographically diverse locations as South America, Nepal, China, Greece, India, and South Pacific Islands (National Research Council, 2006). Nutritional assessments of common vegetable species (A. blitum, A. cruentus, A. dubius, A. tricolor, and A. viridis) show high protein content and significant levels of essential micronutrients, including beta-carotene, iron, calcium, vitamin C, vitamin A, and folic acid (Achigan- Mziray et al., 2001;Teutonico and Knorr, 1985). High nutritional value and tolerance of many biotic and abiotic stresses have made amaranth an especially important vegetable crop in Africa, where some societies derive as much as 25% of their protein intake from amaranth leaves during the production season, and its sale by the thousands of tons annually has significant economic impact (Mandu et al., 2012;National Research Council, 2006). Nonetheless, amaranth has historically been considered a resource of the lower classes, and the phrase "not worth an amaranth" exists in multiple African languages. In contrast, in the Caribbean, amaranth has acquired symbolic cultural significance, in addition to being a dietary staple. The Creole word callaloo, which refers to both amaranth plants and a traditional stew made with amaranth, is also used with great pride colloquially to indicate the unique blend that constitutes Creole culture (National Research Council, 2006 in several changes of water, or young leaves are eaten raw . While amaranth seed production in the U.S. is around 6,000 acres, centered in the Great Plains region, commercial vegetable amaranth production is effectively nonexistent and requires increased research (Green, 2003).

ETHNIC CROP OPPORTUNITIES
Advocates for increased vegetable amaranth production and research in the U.S. cite amaranth's wide use by cultures around the world as the primary appeal for farmers (Singh and Whitehead, 1996;  . Recognition of this fact has been the impetus for extensive ethnic crop marketing studies in the ethnically diverse northeast region, focusing on Asian, African, and Latino populations Sciarappa et al., 2016). It is important to note that these studies grouped respondents into ethnic groups, rather than by immigration status. Even so,  found that 80% of survey respondents regularly purchased ethnic produce and that this sub-group spent 40-80% more on produce than those who did not regularly purchase ethnic produce, making the case for ethnic crop production even stronger than immigration numbers would suggest. Vegetable amaranth was identified as one of many relevant ethnic crops in these surveys, and ranked in the ten most frequently purchased vegetables for Indian survey respondents Sciarappa et al., 2016).
These surveys also found that ethnic produce purchasers frequently bought produce at multiple types of markets, expressed a willingness to pay a premium for their native produce, and cited freshness and availability as the main reasons for market selection. When asked specifically if they desired to buy more ethnic produce grown locally, two-thirds of respondents answered affirmatively and only 6% answered negatively, with the remaining respondents indicating they were unsure .
These key regional studies indicate that ethnic crops can offer local growers an opportunity to expand into an eager market of consumers seeking their traditional produce. Amaranth is a particularly strong candidate for local production and direct to consumer sale because sensitivities to temperature and low relative humidity make long-distance shipping challenging . Rhode Island had 38 regularly scheduled farmers' markets in 2016 (Rhode Island DEM, 2016). In these direct sales systems, ethnic crops may also appeal to high-end buyers who value produce options outside of their everyday fare.
In continued ethnic crop research at the University of Massachusetts Amherst,  have found that a vast majority of tropical ethnic vegetables are easily produced in New England and that the advantages of proximity to a diverse and densely populated market outweigh production challenges. Successful entry into the ethnic crop market, however, requires that farmers be armed with region-specific production information and an understanding of the markets for unfamiliar crops .

PRODUCTION
In comparison to the deep pool of cultural knowledge surrounding amaranth, intensive production research is lacking. Amaranths are known to tolerate marginal soils, high heat, and drought, and have been reported to display a general resilience and resistance to common pests and diseases. Studies investigating these assumptions as they relate to production protocols, however, have returned varying results. Existing literature is comprised of experiments with a large geographic and climatic range, often studying different species with little replication, so this variation is expected.
Amaranth varieties generally have not been as developed as many more common vegetables and therefore vary in their uniformity and adaptation to temperate climates (Putnam et al., 1989). Wu et al. (2002), in one of the most extensive field evaluations of Amaranthus spp. in both temperate and tropical regions of China, found plant performance varied significantly between the two climates and found a great deal of agronomic trait variation between species and between genotypes within species.
Photoperiod sensitivity and origins of genotypes were more strongly correlated to performance in different climates than was plant species. Consequently, in evaluating what information can be gleaned from existing production research, the location of and cultivars used in the studies should be given careful consideration.

Fertility
While amaranths are known to perform well in poor soils, they also respond positively to nitrogen (N) fertilization. Singh and Whitehead (1996) found that for a suggested that previous planting in legumes may provide sufficient N for commercial amaranth production (O'Brien and Price, 1983). All studies agree that amaranth fertility needs will be most dependent on pre-existing soil composition.  . Organic fertilization strategies also limit nitrate accumulation . With the positive relationship between N fertilization and amaranth yield well established, future focus should be on taking advantage of amaranth's low fertility requirements to prevent nitrate accumulation, nutrient leaching, and expenditure for farmers.

Temperature Requirements
While a general affinity for high temperatures pervades Amaranthus spp., reports of temperature requirements are mixed. Stallchknecht and  reported that amaranth seeds need soil temperatures of 15ºC for germination, while Putnam et al. (1989) and Wagoner et al. (1983) describe most amaranth species germinating when soil temperatures reach 18ºC. An A. tricolor planting date study in Georgia, however, found seeds did not germinate in the field until soil temperatures reached 25ºC (Singh and Whitehead, 1996). Optimal day-time air temperatures for growth have been reported as 30-40ºC, and night-time temperatures as 22-28ºC. This variation in reported temperature requirements is not surprising due to the range of species under consideration and the less refined (and therefore variably performing) nature of amaranth cultivars. However, it is problematic in the pursuit of amaranth production protocols for the northeastern temperate climate. For amaranth production in climates where heat tolerance is at issue, evidence of a generalized affinity for high heat may be sufficient; in the northeast, where cool night-time temperatures are at issue, temperature requirements are acutely relevant.

Pests
Unlike cool-season greens, amaranth grows in times of high disease and pest pressure , but no production studies listed prohibitive pest or disease issues. Amaranth is considered tolerant of nematodes and has been suggested to reduce nematode populations for subsequent crops in rotation cropping (O'Brien and Price, 1983). Production guides list the lygus bug (Lygus lineolarius) and European corn borer (Ostrinia nubilalis) as principle insect pests, but these insects affect amaranth inflorescences, making them more detrimental to grain production than vegetable . The amaranth weevil (Conotrachelus seniculus) and foliar insects can also feed on amaranth, but amaranths are known to be generally resilient to herbivore attacks and resistant to bacterial and fungal wilt . Furthermore, common insect pests may show a preference for wild amaranth species when present ; in fact, pigweed can serve as a companion plant for trapping leaf miners and other pests . Because it is not typically grown here, very little information on amaranth pests and diseases specifically in the northeast is available, but  reported that no pest or disease intervention was required in a three-year vegetable amaranth study in Connecticut.

Photoperiod Sensitivity
Given amaranth's adaptation to conditions from equatorial to temperate, it is not surprising that reports of photoperiod sensitivity vary. The photoperiodism of grain, ornamental, and weedy amaranths have received more attention than that of vegetable amaranths, and a wide range of responses have been reported. Most common weedy amaranths exhibit facultative short-day flowering, but it has been suggested that northern populations are more sensitive to decreasing day-length than southern populations of the same species . A. caudatus, which has been cultivated for grain production but is also eaten as a vegetable, exhibits facultative short-day flowering . Grain amaranth variety trials in Minnesota, however, reported a range of day-length flowering requirements, from short to long (Robinson, 1986).
'Pygmy Torch' amaranth, which is often listed as A. hypochondriacus or A.
hybridus is a popular day-neutral ornamental amaranth (Erwin et al., 2002). As a note of clarity, A.hybridus is the progenitor species of a group of Amaranthus species that are best-known as grain types but are also widely eaten as greens. Species within this group are genetically distinguishable from one another and from A. hybridus.
However, A. hybridus may refer to interspecific hybrids within this "A. hybridus complex," or be used cautiously in the absence of genetically precise identification (Drzewiecki, 2001 (Wu et al., 2002).  report that vegetable amaranths typically exhibit facultative short-day flowering, but the level of sensitivity varies with species and cultural practices.
It does not seem reasonable to extrapolate results of photoperiod sensitivity studies across species, nor to draw specific conclusions from the observations of Wu et al. (2002). However, the wide range of observations from these studies is noteworthy, as is the lack of evidence of any obligate photoperiod-sensitive flowering. Wu et al. (2002) suggest that matching variety origin to target production area may be useful in accommodating photoperiod sensitivities, which appear to be significant in some cases, if not entirely understood.

Cultivar Selection
Amaranth's affinity for high heat makes it an appealing crop for commercial production because it can thrive in the warmest summer months when production of comparable leafy vegetables can be challenging Singh and Whitehead, 1993). In the southern U.S., where this is of particular interest, sensory and field evaluations of vegetable amaranths have been conducted. Taste test respondents rated A. tricolor comparable to spinach  and more appealing than A. cruentus and A. dubius ). An evaluation of A. tricolor in Arkansas found six of the eight East Asian accessions yielded significantly less than the semi-savoy spinach grown for comparison. There were no significant yield differences among the A. tricolor accessions, and sensory ratings were largely favorable. Sensory scores were a composite of appearance, smell, texture, bitterness, taste, and after-taste scores of freshly cooked amaranth leaves . Five cultivars representing A. tricolor, A. dubius, and A. hybridus were compared for productivity, nutritional qualities, and resistance to damping-off in central Texas; 'Ibondwe,' a cultivar of A. dubius sourced from West Africa, excelled in all categories .
These southern U.S. cultivar selection studies are not exhaustive, however, and they do not lend themselves to comparison with one another due to inconsistent cultivar selection. Furthermore, potential sensitivities to photoperiod and temperature differentials decrease their reliability in the northeastern region. This caution is confirmed by vegetable amaranth field evaluations in Maryland and Pennsylvania.
The only A. dubius accession evaluated had the highest yield in optimum environmental conditions, but was also the most affected by cooler and wetter conditions . Amaranth cultivar selection in the northeast should proceed with the acknowledgement that much of the viable growing season could include sub-optimum conditions for amaranth.
Recent vegetable amaranth variety trials, conducted at the Connecticut Agricultural Experiment Station from 2008-2010 by , will lend most easily to comparison with this project and future studies in the region.  evaluated eight commercially-available A. tricolor cultivars at two different locations.
Some consistent patterns, but few significant yield differences, were observed over the three-year study. 'All Red' produced the greatest yields overall, but 'Red Stripe Leaf' produced the greatest yields in 2009. These two varieties were statistically equal in all locations and plantings, and often statistically equal to all but the lowest-yielding varieties. 'Bayam' and 'White Leaf' were consistently low-yielding varieties.
However, these varieties only yielded significantly less than the top one or two varieties in any planting .
The climates of Connecticut and Rhode Island are similar enough that we can expect comparable variety performance in these two locations. The common cultural practices associated with growing heat-loving crops in temperate climates are also noteworthy for the comparison of these studies. Amaranth seeds are very small and prone to crushing, so they are often directly broadcast-seeded in warmer climates . However, in more temperate climates, starting seeds in a greenhouse or other protected structure is one of the primary ways of extending the season of heat-loving crops and maximizing production.  variety trials in Connecticut used greenhouse-started transplants, whereas  direct seeded in Maryland and Pennsylvania. Transplant response of amaranth has not been directly studied, but it has been suggested in growers' guides that some varieties of A. tricolor do not respond well to transplanting . In the absence of detailed transplanting tolerance information, variety recommendations that account for response to transplanting may be most useful to growers in the northeast.
Maynard (2013) harvested varieties repeatedly, which is a common practice with amaranth and may delay flowering. However, it is possible that reporting cumulative yields of the repeated harvests masked some of the differences in variety performance. Furthermore, the varieties evaluated by Maynard (2013) encompassed a range of appearances and sizes. Heights ranged from 1.5 feet for dwarf varieties, to 3 feet; leaf colors included light green, purple, and red and green variegated. Especially for growers who are unfamiliar with amaranth, yield comparisons over an entire season may paint an oversimplified picture of variety desirability. Growers might find yield stability measures and descriptive measures, such as leaf to stem ratio, useful supplemental information.

PLASTICULTURE
The benefits of plasticulture vegetable production systems are season extension, efficient use of resources, and increased crop yield and quality. In the northeastern region, these systems are commonly used to increase the growth rate of heat-loving crops and extend the season for high value crops (Lamont, 1996;Wittwer, 1993). The basic structural components of plasticulture, including plastic mulches, plastic films and coverings, and drip irrigation, are employed to modify the microclimate surrounding the plants. These structural components can be used by small-or large-scale growers, but the ideal combination of materials, structures, and management techniques depends upon the grower's budget and scale, as well as the target production area, crops, and market .  outlines in detail how the interaction of components in a plasticulture system produces an effect greater than the sum of its parts. Plastic mulches affect the microclimate mainly by modifying the ratio of absorbitivity to reflectivity of the surface . Black plastic mulch, the most widely used color in vegetable production, generally increases soil temperatures at a 5cm depth by 2.8ºC during the day, compared to bare soil (Lamont, 1996). Energy absorbed by black plastic mulch that is not transferred to the soil will be reradiated as thermal radiation; protective structures like high tunnels and row covers can be used to reduce this energy loss. While protective structures do produce a greenhouse effect, their greatest contribution to microclimate modification is in wind protection, preventing energy gains from "mixing away" into the surrounding atmosphere and thereby amplifying the effect of plastic mulch .
Similarly, drip irrigation alone reduces water needs, relative to surface irrigation, but has the greatest effect when combined with plastic mulch. Plastic mulch both reduces evaporative water loss and prevents competition for water by controlling weed growth. The irrigated soil, in turn, provides a more conductive surface for the transfer of energy from plastic mulch to soil . Plastic mulches also reduce fertilizer leaching (especially when fertigated through drip irrigation systems), reducing input costs for growers and environmental impact (Lamont, 1996). While all plasticulture systems utilize these foundational microclimate modification concepts, the benefit of a given design will vary based on materials, structures, climate, and crops. Those potential benefits must be weighed against the capital investment and labor associated with each system. This review will focus on high tunnels and low tunnels, the two most widely used structures (after greenhouses) for enhancing yields and extending the season in temperate climates. Both of these structures are impermanent, unheated, and present relatively low-cost options for lengthening the growing season and enhancing yields . A comparison of the two systems demonstrates the spectrum of tradeoffs growers must often consider when evaluating plasticulture designs.

Low Tunnels
Low tunnels, or hoop-supported row covers, are typically made of 18-26 µm thick polyethylene sheets supported by metal hoops. The ends and sides of the plastic are secured to the ground using stakes, weights, or by burying in the soil. The first plastic row covers were solid plastic designed to minimize night-time heat loss, which required manual ventilation during the day and closure in the early evening . The development of slitted and perforated row covers alleviated this high labor cost and diminished the risk of destruction by strong winds at night . Although crop maintenance beyond irrigation still necessitates the removal of covers, slitted or perforated covers may potentially be left in place until time of harvest or until plants outgrow the structures, substantially reducing labor cost . This reduction in labor requires a compromise on night-time temperature control, but covers that can be left in place during the day accumulate more heat units over time and provide crop protection. In fact, whether solid or slitted, row covers of a thickness that will provide adequate light transmission cannot provide significant frost protection (2ºC to 3ºC maximum); instead, their benefit is mainly in these accumulated heat units, which result in large part from temperatures in the tunnels rising faster than ambient temperatures in the morning .

High Tunnels
High tunnels, conceptually, fall somewhere between low tunnels and greenhouses. High tunnels can differ somewhat in their frame shape, but all designs resemble a plastic-covered greenhouse. High tunnels are typically covered with a single or double layer of 4-mil to 6-mil thick plastic . High tunnels are passively ventilated, have no active heating system, and do not provide the same level of environmental control as a greenhouse. Consequently, high tunnels are intended for season extension, rather than the year-round production possible in greenhouses.
Ventilation is most commonly achieved by rolling up the sides of the tunnel, so tunnel length has no effect on ventilation. However, tunnel width in excess of 20ft may diminish vent effectiveness in high ambient temperatures . In contrast to low tunnels, high tunnels typically span multiple rows of crops and are tall enough for comfortable entry, making crop access and maintenance effortless; crops may grow to full maturity without adjustments. Raised beds may be constructed within high tunnels, or crops can be grown directly in tilled soil, with or without the addition of plastic mulch .
While more heat units are accumulated under low tunnels than high tunnels, that benefit ceases if and when low tunnels must be removed due to crop growth. High tunnels have a greater initial cost than low tunnels, but once constructed, have low operation cost and are much more durable . When growing high value crops, Blomgren and Fisch (2007) estimate that most high tunnel growers reclaim their high tunnel investment in one to two years; Waterer (2003)

Considerations for Amaranth Production
Within each of these systems, there is ample room for variation regarding management and selection of materials, and both high and low tunnels have been found to be potentially economically viable season extension options in the northeastern U.S. . Black plastic mulch and drip irrigation have been used successfully for intensive amaranth production studies in Maryland  and New Jersey (Sciarappa et al., 2016), but production system comparison was not the focus of these studies; the use of crop covers for amaranth has not yet been studied.
Often, in covered plasticulture systems, a balance must be struck to maximize desired heat accumulation without reaching temperatures that may damage crops . Amaranth thrives in daytime temperatures up to 40ºC , so maximum temperature increase is desirable for producing amaranth in the northeastern U.S., and excessive temperature is likely not an issue. Access of pollinators is often a consideration for crops in a protected agriculture system, but this is not an obstacle for leafy green vegetables like amaranth. Amaranth is, however, very fast-growing and harvested around three weeks after transplant . Labor input for construction and removal of low tunnels would therefore be a frequent cost, even if plants do not outgrow the tunnels and no access to the crops for maintenance is required during the growing period.
Lastly, it should be noted that amaranths are not the high value crops typically used as reference in economic analyses of plasticulture systems (Blomgren and Fisch, 2007;. Especially in the case of high tunnels, intercropping or multiple cropping with higher value crops holds potential for maximizing profits in plasticulture amaranth production. However, this strategy would likely require some compromise on temperature to ensure no damage to additional crops. seven times over the season, but the seventh planting was excluded from analyses due to frost damage. There was a significant interaction of planting date and variety on yield, but some varieties were consistently high-yielding. 'Green Pointed Leaf' and 'Miriah' had the greatest yields overall, were in the highest yielding group in every planting, and were notably high-yielding in the early-and late-season. The effect of variety on yield was reduced in the high ambient temperatures of mid-summer. 'Green Callaloo' was high yielding with poor leaf to stem ratio; a dwarf variety, 'White Leaf,' was low yielding but excelled in leaf to stem ratio. 'Red Callaloo' and 'Red Garnet' tended to have consistently low yields, and generally low but variable leaf to stem ratios. Production and marketing strategies to be considered in addition to performance measures are discussed. Ethnic crop production has received increased attention as an opportunity for farmers to expand into underserved markets. These markets are significant in the northeastern United States, and small farms with flexible, diversified production are ideally positioned to serve these markets . Amaranth (Amaranthus spp.) is eaten as a vegetable in over 50 countries worldwide, boasts a remarkable nutritional profile, and is regarded as easy to grow and resilient (National Research Council, 2006). Amaranth may offer wide familiarity to many regular buyers of ethnic produce; an appealing option for health-conscious buyers of novel produce; and a low-maintenance addition to small farms seeking sustainability through diversified production. Amaranth varieties, however, are far less developed than many more common vegetables, and little intensive production research has taken place in the United States.
The USDA Economic Research Service has identified increased diversity of the U.S.
population as one of the three most important influences on future U.S. food markets . Ethnic crop production and marketing studies in the northeastern region have found that regular ethnic produce buyers are eager to buy more of their traditional produce locally. The advantages of engagement with this market outweigh the production challenges of growing heat-loving ethnic crops in the northeastern temperate climate Sciarappa et al., 2016).
Amaranth is especially sensitive to temperature and low relative humidity, making long-distance shipping challenging . Competition from imports is therefore minimal, and amaranth is a strong candidate for fresh, direct to consumer sale. In New England, the percentage of farms that engage in direct to consumer sales, and the proportional contribution of these sales to the total agriculture market, are roughly five times that of the United States as a whole (NASS, 2012). These direct sale systems allow growers to connect with customers who tend to value both variety and high nutritional value; the potential for increased customer loyalty and valueadded pricing in these systems is especially important for the viability of small, diversified farms . Amaranth leaves are high in protein, β-carotene, iron, calcium, vitamin C, and folic acid . They have also been rated comparably to spinach in sensory evaluations . They may therefore lend well to marketing as a substitute for more common greens, even for customers who do not regularly buy ethnic produce.
Along with expansion into new markets and offering variety to existing customers, diversified production is important to mitigating risk for many small farmers . Amaranth has been reported to be resistant to many biotic and abiotic stresses, so its production in a diversified small farming system could contribute to farm resilience and sustainability. However, many ethnic crops may be unfamiliar to growers, not traditionally grown in the region, or simply understudied for intensive production. Region-specific production protocols and variety recommendations are necessary for growers to successfully realize the on-farm and market-based benefits of increased diversity .
Amaranth's wide distribution through both tropical and temperate climates is largely due to its C4 photosynthetic pathway that allows efficient use of CO2 in variable environmental conditions ); amaranth's tolerance of marginal soils, high heat, and moisture stress are well accepted, but exploration of these qualities as they relate to intensive production is lacking. Because it thrives in high heat, amaranth received some attention from researchers as a potential summer greens substitute in the Southern U.S. . However, the few evaluations of amaranth in temperate climates suggest that varieties with the highest yield potential in optimal environmental conditions may also be the most susceptible to wetter, cooler conditions Wu et al., 2002).
Because these sub-optimal conditions characterize much of the early growing season in the northeastern U.S., reliance on variety recommendations from warmer climates may not be feasible; varieties resistant to environmental fluctuation may be the best candidates for regional production. Variety recommendations in the Northeast also need to acknowledge common regional production practices, such as transplanting.
Some growers' guides suggest that not all amaranth species tolerate transplanting , but no primary studies on transplant response have been conducted.
This study evaluated ten vegetable amaranth varieties for production in the northeastern temperate climate, with a focus on the realities of regional small farms.
Plants were greenhouse-started and transplanted to a low tunnel system, two techniques commonly used to enhance yields and extend the season of heat-loving crops. Seven planting dates over the 2016 growing season allowed for observation of plant performance in both early and late growing seasons. Low fertilizer inputs and drip irrigation allowed for evaluation of these varieties as a low-cost addition to existing small farm production.  tricolor were also obtained but eliminated from the study due to poor germination rates in two initial germination tests.

Research
Culture and Design. All seeds were green-house started in 50-count cell trays, using Metro-Mix 830 soil (SunGro, Agawam, MA) covered with a thin layer of vermiculite.
The greenhouse was set to heat at 70ºF and cool by way of passive ventilation at 74ºF.
No supplemental light was used in the greenhouse. All varieties were seeded every two weeks and transplanted to the experiment site roughly two weeks later. Plots were arranged in a randomized complete block design, with ten plants of each variety and four replicates. Each block was a raised bed with double rows 12 inches apart. Plants were spaced 12 inches apart within the rows.
Directly after transplanting, low tunnels were constructed over the raised beds using galvanized metal hoops and 0.  .
Data Collection and Analysis. Plants were weighed immediately after harvest. After fresh weights were recorded, stems and leaves of a random two-plant sub-sample were separated and dried at 110ºF until they reached a constant weight. Dried leaf and stem weights were used to calculate a leaf to stem ratio, which we expressed as the leafpercentage of total dry weight.  Fig. 1. Garnet' yield comparisons to higher yielding varieties were significant more often.
The treatment effect of variety on yield was the greatest in the first planting. In the third planting, all varieties except 'Asia Red' had statistically equal yields.
Planting Date Effects. Yields of replicates within varieties were the most variable in the first planting. Combined average yields for all varieties were greatest in the fourth, fifth, and sixth plantings; the second planting had the lowest combined average yields.
There was not a significant effect of planting date on yield for 'Green Pointed Leaf,' 'Red Callaloo,' 'Southern Red,' or 'White Leaf.' Leaf to Stem Ratio. There was a significant variety by planting date interaction effect on leaf to stem ratio (p < 0.0001). There was not a strong correlation of yield and leafpercent of total dry weight (r = 0.036). Leaf to stem ratios are presented in Fig. 2.
'White Leaf' was a stand-out performer in leaf to stem ratio, with the highest leafpercent of total dry weight in every planting. The average leaf-percent of 'White Leaf' over the entire experiment was 81.99% of total dry weight. It was significantly higher than all varieties in the second and fourth plantings. However, 'White Leaf' leafpercent did not differ significantly from 'Miriah' in the first, third, and fifth plantings;   (1 kg = 2.205 lb) z square root of variety variance across planting dates, divided by the overall mean yield of the variety, and multiplied by 100.

Yield Findings Comparisons. Although our variety 'Mchicha' is not commercially
available, A. hybridus was similarly a top performing species in Mississippi , Texas , and Georgia (Singh and Whitehead, 1996).
However, these climates are markedly warmer than that of Rhode Island, and the top accessions in all three studies originated in Greece. Wu et al. (2002)  There were significant yield differences within each harvest group in every planting.
However, because these varieties represent such a range of growth habits and sizes, dissimilar yields may not exclude similar photothermal responses. There was less variability in flowering time for the first harvest group, yet more significant yield differences across plantings. There was more variability in days-to-harvest for the second harvest group, yet the yields of these varieties were not affected by planting date. It is possible that flowering for the second harvest group may be closely linked to plant growth, and given the especially long growth period in the sixth planting, it is possible that the contribution of accumulated day-light hours to growth rate is greater in these varieties. Flowering does not seem to be strongly linked to day-length for any of these varieties.

Performance Measures. Although average yield influences CV, consideration of both
yield and CV may be most useful in evaluating variety performance. Leaf to stem ratio is a valuable variety descriptor, but is not an unqualified measure of variety desirability. For an accurate reflection of variety performance, these parameters should all be weighed against one another; for ideal variety selection, production strategy and target customer base may play an equal role.
Variety CVs are useful in avoiding an artificially high estimation of varieties that exhibit high yield potential, rather than reliably high yields. This may be especially important to vegetable amaranth growers in the northeastern U.S. because temperatures for much of the growing season may be sub-optimal for amaranth. Top performing varieties in optimal conditions may be more affected by the cooler, wetter conditions of the early growing season in the northeastern U.S. . Varieties with reliable yields over the whole season may therefore be more desirable than varieties with the highest yield potential.
'Green Pointed Leaf,' was one of the two highest yielding varieties over all.
Consequently, its relatively high CV (29.90%, third highest of the varieties tested) is due to high variability across plantings (Fig. 3). However, in the third planting, when 'Green Pointed Leaf' average yield was the lowest of the season, it was still greater than two-thirds of the varieties tested and statistically equal to the top yielding varieties. Furthermore, the greatest departure from the overall mean yield of 'Green Pointed Leaf' was especially high yields in the first planting; 50% of the varieties had their lowest yields of the experiment in the first planting, and 'Green Pointed Leaf' yields were nearly double the yields of those varieties. In short, yield potential may indeed be great enough to trump stability in this case. 'Miriah,' 'Mchicha,' 'Southern Red,' and 'Green Callaloo' all had high yields and low CVs (Fig. 3), indicating they are likely desirable varieties. However, 'Green Callaloo' and 'Southern Red' both had low leaf to stem ratios, at 42.21% and 54.92% of dry weight, respectively, which may be considered less desirable (Fig. 2).  found that leaf to stem ratio was negatively correlated to yield, or that the varieties with the highest fresh weights tended to be the most stem-heavy. For the varieties tested here, no relationship between leaf to stem ratio and yield was observed (r 2 = 0.0013), but 'Red Callaloo' and 'Red Garnet' had the lowest average leaf to stem ratios and the lowest average yields overall. Growers may consider a higher planting density than was used in this study, or even polyculture with taller crops. Its dense, bushy habit lends most easily to single harvest or harvesting of individual leaves, and its tender leaves may be marketable as a substitute for more common salad components.
Conversely, 'Red Callaloo,' 'Red Garnet,' and 'Green Callaloo' had the three lowest overall leaf to stem ratios. However, the traditional Caribbean dish Callaloo makes use of both stems and leaves, as do many cooked amaranth dishes around the world.
'Red Callaloo' and 'Red Garnet' also had low yields, but their tall, stemmy habit may lend to repeat harvest, rather than single. Repeat harvest is commonly used to increase branching and delay flowering in amaranth production, and we observed increased branching in our study after accidental stem breakage from a storm in the sixth planting. 'Red Callaloo' had a substantial increase in leaf-percent for the sixth planting; 'Red Garnet's leaf-percent was roughly equal to its highest reported leafpercent of the first planting. If growers choose to grow a 'Callaloo' variety for its familiarity to a chosen customer base, they could use increased planting density and frequent harvests to increase overall yields and likely leaf to stem ratio. The magnitude of production system effects decreased as ambient temperatures increased throughout the season. Rankings of yields from the four production systems were consistent in each experiment. Yield rankings from greatest to least were: low tunnel, black plastic mulch, high tunnel (when present), and bare soil. Leaf to stem ratios of a two-plant subsample were calculated and did not appear to be influenced by production system, despite occasional significant differences.
The USDA Economic Research Service has identified the increasing diversity of the U.S. population as one of the most important influences on future U.S. food markets . from South America to Africa to Eastern Asia, and regional studies have confirmed its appeal to many of the fastest-growing ethnic populations in the northeastern U.S. Sciarappa et al., 2016).
Ethnic crop production studies have found that many tropical and sub-tropical crops are easily produced in the northeastern temperate climate. However, these studies emphasize that region-specific production and marketing research will be fundamental to the success of growers in incorporating potentially unfamiliar crops Sciarappa et al., 2016).
Amaranth has a C4 photosynthetic pathway, which is rare in dicots. The biological traits that accompany C4 photosynthesis have made amaranth an important crop in tropical and subtropical regions; amaranth is able to use C02 efficiently under moisture and heat stress . However, these traits are contradictory to those of many greens commonly grown in the northeastern U.S.
Popular vegetable amaranth species thrive in temperatures up to 40ºC (104ºF), do not tolerate temperatures below 15ºC (59ºF), and have substantially higher light saturation points than common spring and fall greens .
The components of plasticulture systems, including plastic mulches, drip irrigation, and crop covers, are valuable tools for extending the season and enhancing yields of warm-weather crops in the northeastern region . Black plastic mulch increases soil temperatures and can improve resource use efficiency; fertilizer leaching, evaporative water loss, and competition for resources from weed growth are all reduced under plastic mulch. Various forms of crop covers are used for their heating effects and provide the added benefit of enhanced crop protection from the elements and pests. When crop covers are used in addition to black plastic mulch, the effects of both components are amplified .
There is substantial room for variation in plasticulture system design; each structural and material combination results in varying levels of protection, microclimate effect, permanence, accessibility, and cost. For crop covers, low tunnels (row covers) and high tunnels represent the spectrum of these considerations. Studies suggest both systems are economically viable options for northeastern growers. However, target crops, production area, and markets should be the deciding factors in plasticulture design decisions, especially for small and beginning farmers .
The low tunnels used in this study were constructed with galvanized metal hoops covered with 0.8-mil clear slitted plastic. Early low tunnel technologies, designed to provide maximum frost protection, required daily manual ventilation. The development of breathable cover materials reduced the need for ventilation, but light transmission was also reduced. Because amaranth has a high light saturation point, reduced light transmission would limit amaranth's photosynthetic capacity. Clear slitted plastic combines high light transmission and self-ventilation. These covers cannot combat extreme temperatures at night; however, continually intact slitted tunnels accumulate more heat units over time than tunnels that must be removed during the day. Heat unit accumulation, often measured in growing degree days (GDD), is directly linked to the growth of many warm-weather crops. However, low tunnels perform this function so effectively that excessively high temperatures within tunnels becomes a concern for many crops in high ambient temperatures .
The required low tunnel materials are relatively inexpensive, but it is difficult to reuse plastic without compromising light transmission or structural integrity. A low initial investment is tempered over time by repeated material cost, as well as the corresponding labor cost of tunnel construction and removal. Any low tunnel crop maintenance beyond irrigation and fertigation requires removal of the plastic, which interrupts heat accumulation .
Although there are numerous variables associated with high tunnel design, a typical design resembles a greenhouse frame covered in 4-to 6-mil plastic. Ventilation is provided by rolling up and securing the plastic on the sides of the tunnel. High tunnels typically span multiple rows of crops and, once constructed, provide excellent protection from the elements and effortless access to crops for harvest or maintenance.
High tunnels are considered temporary structures, but plastic may be used for multiple seasons, and frames are even longer lasting . The rise in popularity of high tunnels in the U.S. is in large part due to findings that initial investments can reliably be recovered when growing high value crops (Carey et al., 2009;. Furthermore, the NRCS Environmental Quality Incentives Program (EQIP), which provides financial support for the implementation of conservation practices, has been assisting growers with high tunnel construction costs since 2010; the EQIP High Tunnel initiative is now active in all 50 states (NRCS, 2017).
Understandably, economic analyses of these systems have focused on high value crops, rather than leafy greens like amaranth. Amaranth's relatively short time to harvest and affinity for high heat also warrant special consideration in palsticulture design decisions. This study evaluated the benefit of three plasticulture systems in comparison to bare soil production of amaranth: black plastic mulch (with no crop cover), low tunnel with black plastic mulch, and high tunnel with no mulch. Black plastic mulch was selected to represent the most basic plasticulture system. Black plastic mulch and drip irrigation have been used successfully for intensive Amaranth production studies in Maryland  and New Jersey (Sciarappa et al., 2016), but production system comparison was not the focus of these studies. Low tunnels combined with black plastic mulch were selected to provide maximum heat accumulation. Given Rhode Island's temperate climate and amaranth's high temperature needs, excessive heat was not an issue. However, amaranth's short time to harvest necessitates fairly frequent tunnel construction and removal, crop maintenance notwithstanding. Acknowledging that amaranth monoculture in high tunnels is not the most profitable use of these structures for growers, the high tunnel treatment was selected to evaluate amaranth as an addition to existing high tunnel production. While there is potential for very high heat accumulation within high tunnels, growers control temperatures through ventilation. Our high tunnel system followed ventilation protocols for tomatoes, which were listed as a primary high tunnel crop in many of the 50 states in a nation-wide survey of extension agents (Carey et al., 2009). Production dates are presented in Table 1.  to the sidewall height of four feet. The ventilation procedures were those used for tomatoes, which were also growing in the high tunnel. Plants were transplanted directly into tilled soil. Weed pressure is low in high tunnels; weeds were controlled by hand-pulling when necessary.

Culture
Low Tunnel System. The low tunnel plots began with north-south oriented raised beds, covered in 1-mil embossed black plastic mulch. Tunnels were constructed immediately after transplant, using clear slitted 0.8-mil plastic and galvanized metal hoops. Hoops were placed five feet apart over raised beds with a center height of three feet. Hoops were covered with plastic, and a second set of hoops was laid over the plastic. Plastic was staked down at the ends and along the sides of the tunnels.
Sandbags were also used along the sides of the tunnels in times of high winds.
Woodchips were used to mulch between rows.
Plastic Mulch and Bare Soil Systems. Plastic mulch plots consisted of north-south oriented raised beds covered in 1-mil embossed black plastic mulch. In bare soil plots, plants were transplanted directly into tilled soil in north-south oriented beds.
Woodchips were used to mulch between rows, and weeds were controlled by handpulling within rows. Soil temperature monitors were wrapped in waterproof tape and placed at a depth of 10cm (3.9 inches). There were no plantings for which both air and both soil monitors produced data in every production system. When two datasets were available, averaged temperatures were used for analyses. The maximum variation between average temperatures recorded by replicate monitors was 0.20ºC (.36ºF).
Harvest and Data Collection. Plants were harvested 25-26 days after transplanting, by cutting the stem directly above the soil surface. The ten plants from a given plot were weighed together immediately after harvest to determine yield. After fresh weights were recorded, stems and leaves of a random two-plant subsample were separated and dried at 110ºF until they reached a constant weight. Dried leaf and stem weights were used to calculate a leaf to stem ratio, which we expressed as the leaf-percentage of total dry weight.
Data Analysis. The analysis of variance (ANOVA) functions in R Version 3.2.3 (R Core Team, Vienna Austria, 2009) were used to determine the effects and interactions of production system and variety on both yield and leaf to stem ratio. In the cases of significant effects, Fisher's LSD test was used for means separation. All tests were performed at the p < 0.05 significance level. Growing degree days were calculated by subtracting a base temperature of 50ºF from daily mean temperatures. Days with mean temperatures less than 50ºF were given a GDD value of zero. Pearson's test of correlation was used to examine the following relationships: yield and leaf to stem ratio; yield and air growing degree days; yield and soil growing degree days.

Results
Microclimate Effects. Air and soil temperature summaries are given in Table 2. The accumulation of air and soil GDD with a base temperature of 50ºF in each production system are shown in Fig. 1 and Fig. 2 respectively.
Air temperature. Low tunnels had the highest air temperatures in each planting and on average. Compared to the average bare soil air temperature, average low tunnel air temperature was 5.0ºC (9.0ºF) higher; average black plastic air temperature was 1.1ºC (2.0ºF) higher; average high tunnel air temperature was 1.6ºC (2.9ºF) higher. On average, low tunnels accumulated 169 more air GDD per planting than high tunnels; 189 more air GDD per planting than black plastic; and 242 more air GDD per planting than bare soil. The greatest range of air temperatures and air GDD for a single planting was in the first planting. These differences diminished as ambient temperatures rose in the second and third plantings.
Soil temperature. Black plastic plots had the highest soil temperatures on average and accumulated the most soil GDD. Compared to the average soil temperature in bare soil, average black plastic soil temperature was 5.1ºC (9.2ºF) higher; average low tunnel soil temperature was 3.3ºC (5.9ºF) higher; and average high tunnel soil temperature was 1.9ºC (3.4ºF) higher. On average, black plastic accumulated 247 more soil GDD per planting than bare soil and 158 more soil GDD per planting than high tunnels. Black plastic and low tunnel accumulated soil GDD differentials followed a less reliable pattern and were 127 GDD in the first planting, 59 GDD in the second planting, and 93 GDD in the third planting.   (3.9in). Growing degree days were calculated by subtracting a base temperature of 50ºF from daily mean temperatures.
Yield. High tunnel plots were excluded from the second planting due to severe Woodchuck damage, and each planting date was analyzed separately. There was no interaction nor main effect of variety, so the two varieties were grouped together in yield analyses. Average yields are shown in Fig. 3. Yields displayed a similar pattern in all three plantings: low tunnel yields were the greatest, followed by black plastic mulch; high tunnel yields ranked third when present, and bare soil plots yielded the least. Although there was no difference in rank order, the magnitude of the treatment effect decreased with each planting.
Low tunnel plots produced significantly greater yields than all other treatments in the first and second plantings. In the third planting, low tunnel yields were significantly greater than all treatments except black plastic mulch. Black plastic mulch yields were significantly greater than the bare soil yields in all three plantings, and significantly greater than high tunnel yields in the first planting. However, black plastic mulch plots did not differ significantly from high tunnel plots in the third planting. High tunnel plots yielded significantly more than bare soil plots only in the first planting.
The difference between average treatment yields was greatest in in the first planting and decreased with each planting. In the first planting, average bare soil plot yield was 4% of average low tunnel yield. High tunnels, the second-lowest yielding treatment in the first planting, yielded more than the bare soil by a factor of 3.5. Average low tunnel yield was more than double that of black plastic mulch in the first planting; in the second planting, average low tunnel yield was only about 20% greater than average black plastic mulch yield. Yields were more strongly correlated accumulated air GDD (r = 0.86) than to accumulated soil GDD (r = 0.75). (1 kg = 2.205 lb) Leaf to Stem Ratio. Leaf-percent of total dry weight was not correlated to yield (r = 0.06). There was a significant effect of variety on leaf-percent of total dry weight (p < 0.0001), but there was no significant variety by production method or variety by planting date interaction. 'Green Pointed Leaf' leaf-fraction of total dry weight was 0.04 higher than that of 'Red Stripe Leaf, with a confidence interval of 0.02-0.06 (p < 0.05). There was a significant production method by planting date interaction effect on leaf-percent of total dry weight (p < 0.0001), so the main effect of production system was analyzed for each planting date separately (Fig. 5). In the June planting, the low tunnel had significantly higher leaf-percent than all other treatments, which were statistically equal. In July, low tunnel leaf-percent was significantly lower than all other treatments, which were statistically equal. In August, there was no significant effect of production system on leaf-percent. Results are pooled from two varieties and four replicates. Bars represent ± one standard error.

Discussion
We observed a clear pattern in yield response to the production systems tested.
Amaranth is a heat-loving crop, and there is definite evidence of positive response to increases in soil and air temperatures. However, our treatments were production systems with effects and interactions outside of those quantified here. For example, low tunnels were the highest yielding plots. Average low tunnel air temperatures were 3.3 ºC (5.9ºF) higher than the second ranking high tunnel air temperatures; high tunnels ranked third in yield. Low tunnel soil temperatures, however, were second to black plastic soil temperatures by only 1.9ºC (3.3ºF); black plastic plots ranked second in yield. So, while air temperature was more strongly correlated than soil temperature to yield, it is possible that the statistical influence of very high air temperatures in the very high-yielding low tunnels masks the importance of soil temperature in this correlation comparison. Even more probable is that the combination of high air and soil temperatures in the low tunnels had a synergistic effect on yield.
Economic Perspective. It is noteworthy that the high tunnel and bare soil plots, which require the greatest and the least initial investment, respectively, produced the most similar yields. Our results suggest high tunnels are not an advisable production system for amaranth. High tunnels could be used to greater effect for amaranth with different ventilation procedures, but tailoring high tunnel temperature controls to amaranth would eliminate the possibility of polyculture with higher-value crops.
The similarity of low tunnel and black plastic plot yields is equally notable. Low tunnel yields were significantly greater than black plastic yields in all but the third planting, but black plastic yields were still roughly 50% higher than high tunnel yields and 70% higher than bare soil yields on average. Low tunnel materials, though relatively inexpensive, are an additional cost compared to black plastic mulch alone.
Crop accessibility can also become an economic consideration for low tunnels. When low tunnel plastic was removed for pesticide control in the second planting, low tunnel air temperature dropped drastically to around only 1ºC (1.8ºF) greater than the bare soil air temperature the day. The fact that crop access negates, if temporarily, the microclimate effect of the production system is unique to low tunnels and should not be overlooked in economic assessment of the system.
Nonetheless, low tunnel plastic is the only recurring material cost for low tunnels. At around $150 for 1,000 row-feet, the low tunnel plastic for each planting in this experiment cost under five dollars. The retail cost of amaranth greens, estimated by a collection of seven sources in 2017, was $2.89/lb (Gary et al., 2017). At the planting density used in this experiment, the cost of low tunnel plastic was justified. For the additional five dollars in low tunnel plastic, the difference in yields between black plastic mulch and low tunnel plots amounted to $41 in the first planting, $17 in the second planting, and $13 in the third planting. For the $0.15 for each row-foot of low tunnel plastic, yield differences amounted to $1.47/ft in the first planting, $0.42/ft in the second planting, and $0.33/ft in the third planting.
Production Timing. The treatment effect of production system diminished as the season progressed, and it is clear that planting date warrants careful consideration in amaranth production in the northeastern region. Bare soil yields in the second planting exceeded the black plastic and high tunnel yields of the first planting. Black plastic yields in the second planting exceeded the low tunnel yields of the first planting. Low tunnel yields plateaued in the second planting, and were nearly exactly equal in the third planting. Amaranth is often direct-seeded in warmer climates, but the performance of our bare soil plots suggests that intensive bare soil amaranth production in the northeastern temperate climate is only feasible in the warmest summer months, even with greenhouse-started transplants. Amaranth is fast-growing once established, but it is not a vigorous grower in early development; stand establishment has been identified as an issue in growers' guides .
Although there was 100% survival of our bare soil plants in the first planting, establishment may have been encouraged by even a minor delay in planting.
Leaf to Stem Ratio. There were significant differences in leaf to stem ratios, but there was no pattern to the effects of production method, nor a correlation with yield. These findings indicate that this measure is largely genetic, and outside the influence of production systems. Enhanced breeding would be the logical course for managing this trait. Current amaranth growers could consider production techniques such as repeated harvest to increase branching, but informed variety selection is likely the most powerful resource available.
Conclusions. Amaranth has wide appeal to ethnic produce buyers and potential as a novel substitute for more common greens. Given its short time to harvest, bare soil amaranth production is possible in the northeastern region. However, our results suggest that use of even minimal plasticulture techniques can provide considerable gains in yield and season extension. As previously stated, there is much room for variation within the systems tested here, and each produces an interaction of effects not fully quantified by this study. Nonetheless, these results can be used to guide growers in investigation of modifications that may result in promising intermediate options. One such option is the use of black plastic mulch within high tunnels. We do not recommend conforming high tunnel ventilation to amaranth temperature requirements, but the positive response of amaranth to black plastic mulch without crop covering indicates this combination would produce favorable results. Of the systems tested here, black plastic mulch raised beds represent an advantageous intersection of positive yield response, easy crop access, and polyculture potential at low cost to growers. Maximum yield and season extension, however, can be achieved through the use of low tunnels.

CONCLUSION
Amaranth is a promising alternative crop for growers in the northeastern region; these studies demonstrate the importance of informed variety selection and production protocols. The results of these variety trials support clear variety recommendations, based on the standardized production used in these trials. For reliably high yields throughout a long growing season, 'Green Pointed Leaf' and 'Miriah' are recommended varieties. These two varieties have high leaf to stem ratios and are likely well suited to single or repeat harvest. However, the varieties tested here encompass a wide range of growth habits and appearances, and further investigation of variety-specific production and marketing strategies is justified. It is likely that targeted production, including a focus on planting density and harvest technique, may increase the desirability of stem-heavy varieties like 'Green Callaloo' and lowyielding varieties like 'White Leaf.' Similarly, the production systems tested here support a clear choice for maximized amaranth yields but indicate that there is room for variation based on growers' needs. Low tunnels maximized yields, and the recurring cost of low tunnels was justified, even at our relatively thin planting density. However, black plastic mulch, whether alone or in combination with crop covers, is highly recommended. It is likely that the addition of black plastic mulch to high tunnel amaranth production would be beneficial, but higher value crops are recommended for maximizing high