Japanese vegetable soybean (edamame) is harvested at reproductive (R) growth stage six to seven (R6 to R7) when the seeds are still green [1]. At this developmental stage, the seed contains high protein, monounsaturated fatty acids, minerals elements, as well as vitamins like B1 and B2 [2,3]. Edamame, a historically Japanese delicacy has been rapidly incorporated in to American diets due to changing ethnic diversity among US population demographics and increased people's awareness of edamame nutritional qualities. Edamame is currently found in grocery stores and farmers' markets with its demand in the U.S. being estimated at 14,877 tons annually [4,5]. Edamame production is a promising opportunity to meeting the specialty crop market demands. While over 70% of the edamame sold in supermarkets across the U.S is imported from China and Taiwan, it is mainly marketed frozen and this frozen storage conditions lowers its quality drastically. Commercial frozen edamame may contain harmful bacteria such as Escherichia coli and Listeria monocytogenes [6]. In addition, consumers raise the concern of edamame source of safety and start to seek locally grown fresh edamame [7].

Lack of long-term, local supply has become a barrier for successfully marketing of edamame in the U.S because it has a very narrow field harvest window of only a few days. There is very limited information available on extending edamame harvest window though season extension production techniques in Virginia. However, growth studies under plastic film house culture and plastic film tunnel on two Japanese edamame cultivars has been done elsewhere [8]. By planting from February 15 to September 25, they were able to harvest crop from May 25 to December 3. In the US, an intensive research was conducted to extend edamame harvest window in 2003 and 2004 at University of Kentucky [9]. Four Gardensoy series edamame cultivars from MG I to IV were planted in greenhouses and transplanted into the field either covered by plastic film or with no plastic protection on April 1^st, 15^th, and 29^th in Kentucky [9]. The crop was harvested for fresh pods from June 17^th (GardenSoy 11, MG I) to August 10^th (GardenSoy 41, MG IV). In the study, GardenSoy 11 had lower marketable yield than other cultivars at all transplanting dates, and indication showed that early production by transplanted MG I cultivars may not be successful in Kentucky. Because no research has been done in Mid-Atlantic region to extend fresh edamame harvest window, growers are looking for superior edamame cultivars with high stability and adaptability. There is a possibility for increased on-farm income from edamame relatedbusinesses through off-season supply. To be able to supply fresh edamame for an extended period of time, there is a need for season extension system development.

The experiment was carried out for two growing season (2013-2014) at Randolph Research and Demonstration Farm, Virginia State University. The study was laid out as a randomized complete block design with three replications. Four edamame cultivars of different maturity groups (MG) where used: GardenSoy31 (MG III), GardenSoy41 (MG IV), Moon Cake (MG V) and Randolph (MG VI. Edamame planting in early- and mid-spring was done in a high tunnel (a polyethylene-covered semi-circular structure). Planting date varied because of unforeseen weather conditions, however, it was categorized into early-spring (Before April 21^th), mid-spring (April 21^th-May 21^th) and late spring (May 21^th-June 21^th), and late summer (Aug 21^th-Sept 21^th). During early and mid-spring planting in the high tunnel, the four edamame cultivars; Gardensoy31, Gardensoy41 and Moon cake Randolph were grown in seed beds covered with a black plastic. During mid-spring planting in the field, edamame was grown in both plastic- covered and non-covered seed beds. Another planting in the field with or without plastic-cover was carried out in late spring. Also in late August-early September, crop was established in the high tunnel to protected crop from cold temperatures later in the fall.

During harvest, all plants were harvested and pods removed. Because cultivars of different MG were planted on the same time, harvesting was sequential. Within a production system, the first to mature were those belonging to the low MG category. Total weight of harvested pods was determined and then sorted into pods with 1, 2, 3 or 4 seeds. The weight of pods in the individual categories was then obtained. Total weight of marketable pods was obtained as weight of all pods with two or more seeds pod-1. A representative sample of the marketable pods was obtained, weighed and shelled. Number of seeds was obtained, and its fresh weight was determined. About 100 gram of the shelled seeds were obtained and stored frozen to await freeze drying. The freeze dried material was ground later and analyzed to determine oil, protein, and sucrose content. Protein content was determined using the combustion method (AOAC - Official Method 990.03) with a Vario MAX CN (Elementar Americas, Inc., Mt. Laurel, NJ, USA) as described by Association of Official Analytical Chemists [10]. Protein factor of 6.25 is used for calculation. Oil was extracted from dried and grounded samples using petroleum ether in an ANKOM XT15 Extractor, Method 2, 01-30-09 (ANKOM Technology, Macedon NJ, USA) as described by AOCS [11] Official Procedure Am 5-04. Sugars were extracted from ground sample (1 g) and analyzed by HPLC following the methods optimized by Johansen et al. [12]. Sugars in the extracts were identified by comparing their retention times with standard sugars. For quantification, trehalose was used as internal standard and the sugar concentration was expressed as g/100 g dry basis. Statistical analyses were carried out using SAS procedures.

Plant growth and harvest window extension

Plants grown in high tunnel on plastic covered seed beds in both early- and mid-spring showed good growth. For mid-spring planting, while crop planted in the field on plastic-covered beds showed faster growth than those without plastic cover (Figure 1) and both were slower than that in the high tunnel. Similar use of high tunnels and other season extension techniques have been used in other horticultural crops [13,14] and its beneficial effects is a result of microclimate modification allowing crops to grow outside of their normal time schedule.

Figure 1: Growth in field plants for mid-spring planting with-(a) and without (b) plastic covered seed bed.

For crops grown in early spring, pod harvest was achieved as early as the beginning of July. In early July, Gardensoy31, a MG III vegetable soybean cultivar was harvested and Gardensoy41, MG IV was harvested about two weeks later. At the time of harvest of Gardensoy41 from early-spring planting, Gardensoy31 in the high tunnel from midspring planting was ready for harvest. Using the production systems (High tunnel, plastic covered seed- beds in the field, and conventional soybean production) and planting edamame cultivars of different maturity groups simultaneously extended the harvest window from a few days common for conventionally produced crop to several months (Table 1). Within a cultivar, harvesting occurred two weeks to a month between first and second crop in the high tunnel. For the same cultivar in the field, a crop on plastic-covered beds was harvested earlier than those without plastic, because it grew faster and matured earlier. Because of differences in maturity time attributed to planting dates, cultivar’s MG, and use of multiple production systems, pod harvest occurred from early July and continued through end of September/ early-October (Table 1). While differences in MG was responsible for harvest time differences among cultivars planted at the same time under a given production system, presence or absence of plastic cover was responsible for differences within a cultivar. In general, Gardensoy31 (MG III) and Gardensoy41 (MG IV) were harvested earlier than Mooncake (MG V) and Randolph (MG VI) when planted at the same time under similar production system. However, extremely cold temperature and animal pest destroyed edamame grown in late summer for two successive years. While this crop was not included in the result, there was a potential to extend harvest into early November.

Table 1: Planting time and harvest period for the different edamame varieties produced under given production systems. Key ___________________Approximate period within which Gadensoy31 and Gardensoy41 where harvested in that order. --------------------- Approximate period within which Mooncake and Randolph where harvested in that order.

Pod yield and yield-related traits

For a specific planting period and for a specific production system (high tunnel, field with or without plastic covered seed-bed), total harvested pods differed with edamame cultivars. For early spring planting in the high tunnel, total pod yield varied from 4300 kg ha^-1 in Gardensoy31 to over 10,000 kg ha^-1 for Randolph (Table 2). Slightly higher yields were obtained during the second planting in the high tunnel in mid- spring. The increase may be attributed to increased soil and air temperatures that may have allowed for better growth conditions since this was planted 2-3 weeks later in the spring. Yield in a crop established in August and September in the high tunnel are not included due to loss of crops from destruction by Marmota monax (ground hog), an animal pest.

Table 2: Comparison of total and marketable pod yield for edamame cultivars grown under the high tunnel at different planting dates.

In mid-spring, though not significant, marketable yield from crop established in high tunnels or in the field with plastic covered seed-bed were relatively higher than that from field without plastic cover (Table 3). Overall, Randolph produced relatively high total pod yield than other cultivars within a given production system. In open field for mid-spring planting, Randolph yield range from 14000 to 23000 kg ha^-1 fresh pods (Table 3). In late spring under conventional planting, Randolph produced the largest (P=0.05) total pod yield of 15850 kg ha^-1. The other three varieties produced similar total pod yields averaged at 9900 kg ha^-1 (Figure 2). The marketable yield for late spring planting ranged from 63 to 77 % of total pod yield depending on variety (Figure 2). The yields for same varieties in this study were lower than those obtained by others in Mississippi [5].While marketable yield for mid-spring planting range from a low of 4120 kg ha^-1 in Gardensoy31 to 6545 kg ha^-1 in Randolph under the high tunnel, it ranged from 2092 kg ha^-1 in Gardensoy31 to 10,814 kg ha^-1 in Randolph in field grown crop (Table 3 and Figure 2). These values for marketable yield is comparable to those of other vegetable soybean obtained other studies [15,16]. The difference in yield between cultivars within a production system could be attributed to differences in cultivar yield potentials. In field grown crop, plastic covered-bed increased yield in all cultivars. Under all the systems, marketable pods (≥ 2 seeds pod^-1) were between 40-75% of the total yield. The number of pods with at least two beans was proportionally higher than the other class categories and could explain the high proportion of marketable pod yield. For all varieties, there was a strong correlation between total and marketable yield (r>0.75) (data not shown). There was also strong correlation (r>0.75) between total pod yield or marketable pod yield and number of pods with one, two, or three seeds.

Table 3: Comparison of total and marketable pod yield in edamame cultivars planted in mid- spring using different production systems.

Figure 2: Total and marketable pod yield for crop under conventional planting in late spring. Values are means ± SE.

This would be expected since total yield include all pod while marketable consist of those with ≥ two seeds pod^-1 and makes sense that as number of these pods in harvest increase, total or marketable yield also increases. Similar correlation has been shown before in other vegetable soybean varieties [17].

The oil content of fresh seed harvested from different production systems and varieties were similar (P-=0.05) and averaged 158 g kg^-1 (Table 4). The sucrose content was affected by the interaction of the production system and variety (P<0.05).

Table 4: Protein and sucrose content of seeds of edamame cultivars planted at different times and in different production systems. For each seed quality attribute, values in a row followed by same lower case letter are not different (P=0.05), For each seed quality attribute, values in a column followed by same upper case letter are not different (P=0.05).

In general, sucrose content was higher for plants seeded later in the field averaging 57 g kg^-1 across varieties. Across varieties, sucrose content of seed from plants in the high tunnel averaged 43 g kg^-1. The protein and oil content of seed in this study were comparable to those reported elsewhere [16,18,19]. While sucrose content were comparable to those reported before [20,21], they are higher than those of vegetable varieties reported in other studies [22]. Seed protein content was affected by interaction of the production system and variety (P<0.05). While there was observed a general increase in seed protein content with time from mid-spring to late-spring in Gardensoy31 and Gardensoy41, there was a reduction in content for Randolph (Table 4). Across production systems, Randolph seed protein content averaged at 396 g kg^-1 was larger (P=0.05) compared to other varieties.

There is great potential to increase edamame supply in Virginia using season extension techniques. Use of high tunnel and plastic covered seed-beds when ambient conditions are unfavorable for root growth creates localized soil temperatures that allow for improved plant growth. Planted early in the high tunnel, where greenhouse effects increase air temperatures allow early plant growth and early maturity. This crop will be harvested off-season during early summer and producers may access markets when vegetable edamame supply is low and a potential for higher prices exist. A longer period of harvesting time and supply of fresh beans can be achieved with selection of appropriate mix of different MG edamame varieties and production systems. However, economic analysis should be done to determine the viability and profitability of the enterprise that incorporates multiple production techniques. Collaborative work needs to be done with food packaging entities and consumer markets to determine the edamame supply and market demand dynamics so that timely production can be done.

We would like to thank the Sustainable Agriculture Research and Extension (SARE) for funding this project.