Soybean (Glycine max) cropping in Sweden – influence of row distance, seeding date and suitable cultivars

ABSTRACT Field trials were conducted 2010–2012 in soybeans (Glycine max (L.) Merr.) to investigate cultivar choice, suitable seeding dates and row distances for cropping in Sweden. The cultivar trials showed that commercially available varieties of the 000-group can be used for cropping in Sweden. Yield typically amounted to 1500–2500 kg ha−1 of dry marketable seeds. Cultivars introduced to Sweden should be tested in the field before large-scale production as the 000-classification not fully is valid for Swedish conditions. Row distance (12.5; 25 and 50 cm) had no overall statistically significant impact on yield, but 25 or 50 cm are recommended as these distances are suitable for physical weed control such as row cultivation or torsion weeding. Seeding in the period May 15–30th is recommended to secure germination and yield. The outcome of the trials was supported by results from field demonstrations and commercial cultivar trials in 2013–2014.


Introduction
The soybean (Glycine max) is native in Asia (Singh 2010;Miladinovic et al. 2011) and has during the twentieth century become one of the most important crops in the world for animal feed and various foods.The European production of soybean is concentrated to the south and southeast regions with countries such as Ukraine, Italy, Serbia and Romania (Table 1) as major producers (FAOSTAT 2020).The production in the European part of Russia is focussed to areas west of river Volga.
The soybean is an annual short-day plant producing seeds with high content of protein (40-45%) and oil (15-20%) (Miladinovic et al. 2011).However, cultivars insensitive to photo-period have been available since mid-1940s by, e g Holmberg & Söner AB, Norrköping, Sweden.The sensitivity to photoperiods and requirement for adequate temperature for flowering and seed production has caused the division of soybeans cultivars into 13 maturity groups from X to 000.In international soybean communities, the maturity group 000 contains those soybean varieties requiring less than 2400 Crop Heat Units (CHU).They are considered as the most early maturing cultivars and can be cropped in regions with long days such as in Scandinavia.Cultivars belonging to the 000 (triple zero) and 00 (double zero) groups are cropped in countries such as Canada, Czech Republic, Austria, Poland and Germany and have been introduced to Sweden (Fogelberg 2004).
The idea of a Swedish soybean cropping has been a challenging thought since the late 1930s.The pioneer in Swedish soya breeding was Mr Sven Holmberg at the breeding company Algot Holmberg & Söner AB located in Fiskeby, Norrköping who already in 1938 began to collect local soybean varieties in Japan and northern China (Holmberg 1946;1956) for breeding of varieties suitable for Swedish conditions.
The breeding of soybeans by Holmberg and cobreeder Knut Träff (Träff 1979) during the 1940s to 1970s resulted in the well-known series of 'Fiskeby I-V' which all were adapted for cropping at latitude 58°N .'Fiskeby V' was released on the market in 1968 and was later on followed by cv.'Träff' and 'Bråvalla' before the breeding program was terminated in the 1980s in connection with a company merge.
Yields of these early Swedish varieties could be surprisingly high.Unpublished data in 1950 (Holmberg 1950) show that seed yield of some breeding lines in some cases had been as high as 2 265 kg ha −1 (Table 2).The line 201-14 was later introduced on the market as Fiskeby III.
Unfortunately, the Swedish soybean cropping never became a success.Although suitable for the climate (Holmberg 1973;Elovson 1984), the short plant height of 25-40 cm made harvest difficult (Bengtsson and Larsson 1979).The low yields and general lack of interest from feed industry resulted in a close down of breeding activities in the 1980s.
The long period of breeding activities did unfortunately not result in any major impact in Swedish agriculture why it is a common opinion that soybeans cannot be cropped in Scandinavia.In early 2000 and onwards, initiatives on soybean cropping in Sweden (Fogelberg and Wahlund 2011;Fogelberg et al. 2012;Fogelberg and Lagerberg Fogelberg 2013) partly has changed this opinion due to an increased search for domestic high-protein sources.
In contrast to Scandinavia, breeding activities in, e.g.Canada, Switzerland, Austria and Poland has continued and resulted in modern varieties suitable for cropping in Central and North-Central Europe.These cultivars are today available through a wide range of seed companies in Austria (e.g.Saatbau Linz, www.saatbau.com),Czech republic (e.g.Prograin ZIA, www.prograin-zia.com),France (e.g.RAGT, www.ragt-semences.fr)and Estonia (Estonian Crop Research Institute, ECRI, www.etki.ee).
From a Swedish perspective, a domestic production of soybeans would be valuable in many perspectives; a new legume crop is beneficial from a crop rotation point of view.Swedish agricultural production is today dominated by grains and thus there is a need for legumes such as soybeans, faba beans and peas.Moreover, the use of soya-based feed for pig meat, milk and poultry has been recognised as problematic from a consumer perspective as it is connected to climate change and deforestation in South America.However, turning from imported soybeans to domestically grown beans for production of feed requires basic knowledge of cropping methods such as cultivar choice, fertilisation, seed rate and row distances, harvest technology and production economy.
In this study, we have revisited the possibility to crop soybeans in Sweden.The aim was to (i)study effects on yield by three row distances (12.5; 25 and 50 cm, respectively); (ii) study effects on yield by seeding dates (early, mid and end of May) and (iii) in cultivar trials study and for further production suggest, suitable cultivars.Some applied aspects on cropping were also included as on-farm cropping and demonstrations for advisory service officers.

Material and methods
Field experiments were carried out in Sweden in 2010-2013.The experiments were conducted at two main geographical areas; in the region of Skåne (southern Sweden, 55 o 4´N, 14 o 18´E) and on the island of Öland (56°33´N, 16°25´E) in the Baltic Sea.The field trials were seeded and maintained by the regional Rural Economy and Agricultural Societies, an organisation specialised in private and public field trials and agricultural advisory services for farmers.Additional field trials and on-farm demonstrations were conducted on the island of Gotland and on the mainland in east Sweden.The seed company Scandinavian Seed AB (SSD) conducted in cooperation with the author, in 2012-2014 a series of field trials on suitable cultivars in Sweden.Parts of these results have been included.
In all experiments, including those carried out in cooperation with the Scandinavian Seed AB, a randomized complete block structure was used with four replications.Each plot was 2-m wide × 12-m long using 50 cm as standard row distance, except in the trial testing different row distances, where 12.5; 25 and 50 cm were used.In general, we used a seed rate of 140 kg ha −1 regardless of row distance and a seeding depth of 4-5 cm.This seed rate resulted in a plant density of about 65 plants m −2 or 30-35 plants per row meter using 50 cm row spacing; 15-20 plants using 25 cm and 8-10 plants using 12.5 cm spacing.Plant emerged after 7-10 days, flowering took place in early to mid of July followed by pod setting.Full maturity was generally obtained in late September-early October.When plants had matured a net plot of 2 × 9 m was machine harvested using a Wintersteiger plot combine.By harvesting a shorter plot than seeded, we minimised risk of mixing harvested seeds between plots.The harvest was weighed, dried and cleansed.Yield is generally presented as marketable yield in kg ha −1 .Some samples  Commercially available cultivars were used in the experiments.Two varieties were selected for use all three years, 'Bohemians' and 'Silesia', both originating from the Czech seed company Zemedelska agentura sro, ZIA (today Prograin ZIA, www.prograin-zia.com).The cultivars 'Bohemians' and 'Silesia' were chosen as the general varieties due to their earlier proven ability to mature in temperate climate zones.The 'Bohemians' is bushy type with plant height of about 90 cm, while the 'Silesia' is semi-bushy type that can reach a plant height of 130 cm.'Bohemians' is considered as a true 000-variety, while the 'Silesia' is a 00-variety approaching the 000-group, that requires a higher temperature sum (2390 UTM) compared to 'Bohemians' (2375 UTM).Cultivars used in the SSD-trials were either obtained from commercial producers or as samples from research institutes, universities and small-scale producers in the northern hemisphere.
In general seeds were untreated, but in the SSD cultivar trials, 'Merlin', 'Gallec' and 'Obelix' had been preinoculated with an unknown Bradyrhizobium japonicum product regardless of experimental year.To standardise inoculation, all seeds were prior to seeding inoculated with a commercial product (Hi-Stick) containing Bradyrhizobium japonicum obtained from Becker Underwood, today sold under the name Nodulator® by BASF.The varieties were all GM-free.
Due to lack of funding, only one cv.'Bohemians' was used in the row distance trial, whereas 'Bohemians' and 'Silesia' were used in the seeding date trials.
Seedbed preparation was done by two harrowings followed by fertilisation with 150 kg ha −1 Yara PK 7-25.Seeding took place with a Wintersteiger plot seeder.Weeds were controlled with herbicides (one spraying with Sencor + Centium 0.4 + 0.4 l ha −1 pre-emergence) and repeated inter-row hoe cultivations until canopycovered row spacing.No treatments on fungi diseases were carried out.
The harvest was conducted in late September or early October.Plots were machine harvested, generally during one working day, but in some cases, harvest continued after a 12 h break.The seeds from each plot were dried at the Torslunda Agricultural Research Station and stored until cleansing could be done by the author.

Statistical analyses
Results were analysed by the author using Tukey's HSD in a multifactor ANOVA.Software (Statgraphics Centurion XIV) was provided by Statgraphics Technologies Inc., USA.

Results and discussion
A problem noticed in the field demonstrations were damages by birds after emergence of plants up the first true leaf stage.Pigeons and crows were very attracted to the dark green colour of the seedlings even if we had taken measures to minimise bird attacks.We used nets to prevent rabbits and roe deer to enter the trials.Gas cannons, kites and vibrating plastic bands were used in the trials to prevent damages by birds.There were no visual losses of plants or yield due to animals except in 2012 when one trial had to be discarded (Table 3).

Cultivar choice, yields and maturity
The yields varied significantly between years and cultivars; in the SSD cultivar trial 2012-2014 in Skåne (Table 4) 'Bohemians' yielded between 1510 and 2900 kg ha −1 , while on Öland in 2010-2012 the yield amounted to 1140-1980 kg ha −1 .In 2014, yields were in some cases doubled compared to those achieved in 2012, an unusual wet and cold year.In 2013, Sweden experienced a quite normal year from a meteorological point of view.In the SSD trials yields of the nine tested cultivars varied between 1298 (Merlin) to 1820 (Gallec) kg ha −1 .The majority of the cultivars typically yielded about 1600 kg ha −1 .A similar trend was found in 2012 with yields of about 1400 kg ha-1 and in 2014 of 2700-3000 kg ha −1 .
In these trials yields typically amounted to 1500-2500 kg ha −1 regardless of variety.Warm season may increase yields for some cultivars to about 3000 kg ha −1 and wet, cold years lower yields to about 1400 kg ha −1 .In comparison, the average soybean yield of Europe in 2018 was 2130 kg ha-1 and in 1565 kg ha −1 in 2012 (FAOSTAT 2020).
The lower yields of Sweden compared to e g Germany (average yield of 2458 kg ha −1 in 2018; FAOSTAT 2020) is likely connected to the overall lower summer temperature and partly the long days.Soybean have a higher need of warmth than e g wheat or faba beans to germinate, develop and flower (Gibson and Mullen 1996;Miladinovic et al. 2011).A temperature above 19°C is required for flowering (Miladinovic et al. 2011) and night temperature less than 8-9°C will cause a pause in flowering (pers.comm.Krause 2013; Gagnon 2016).In order to provide optimal cropping conditions in a temperate climate zone such as Sweden, care must be taken to choose sandy fields that dry up early in spring, situated in the landscape to avoid frost in spring.
The differences in yield and maturity between the studied varieties imply a variation within the 000group that obviously is pronounced when the varieties are tested under cool climate conditions.In 2013, we measured days to maturity of a handful cultivars (Table 5) using the SSD-trial in Östergötland (latitude N 58.48;E 15,52); island of Gotland (latitude N 57.37;E18.58) and Skåne (latitude N 55.45°; E 14.16)The differences in maturity between cultivars might be an effect of the breeding material.It is well known that the Fiskeby V is an extremely early maturing variety.It has been confirmed that 'Bohemians' partly is based on breeding material of Fiskeby V (Krause, 2013, pers com.).
The chosen varieties of the maturity trial were all in the 000-group although full maturity could vary between 113 and 138 days in Skåne (southern Sweden), respectively, 129-164 days in south central Sweden (Östergötland).'Moravians', 'Paradis' and 'Bohemians' were from a maturity perspective, less affected by the geographical site compared to 'Gallec' and 'Merlin'.From an agricultural point of view, varieties that show small differences in days-to-maturity should be used.Depending on annual differences in weather, we can speculate that cultivars with a pronounced difference in days-to-maturity such as 'Merlin' or 'Gallec', can mature late even in the south part of Sweden and possibly not reach maturity in northern areas in a wet and cold season.Hence, on introducing new varieties regional trials should be carried out to establish recommendations for farmers.In a Swedish perspective, the 000-group can contain cultivars being associated with the 00-group and thus not suitable for Scandinavian conditions.

Row distances and seeding dates
There was no clear tendency on how row distance influenced yield (Table 6).Using a 12.5 cm row distance resulted in yields of 1488-2000 kg ha −1 , while 25 and 50 cm row spacing yielded 1492-1934 kg ha −1 , respectively, 1538-1944 kg ha −1 .In the Öland trial, there were no statistical differences between row distances neither in 2010 (p = .56)nor in 2011 (p = .90).The 2012 experiment was discarded due to damage by animals (roe deer).The Skåne trials showed that 12.5 cm distance yielded significantly better than 25 and 50 cm, respectively, in 2010 (p = .0002)and 2011 (p = .0006),while there were no differences in 2012 (p = .36).
Saaten Union Company (www.saaten-union.com)recommends for Germany row distances of 13.5-35 cm and wider distances of 40-50 cm in case mechanical weed    (38 cm) has proven to increase yield (Pedersen 2020).The statistical differences obtained in yields are probably more connected to the seasonal variations in rainfall and temperature, than row distance itself.There is, however, an agronomical point of view that should be considered.In Swedish grain production 12.5 cm row distance is commonly used (Hammar 1978;Boström et al. 2012) why seeding machines in general are preset to this distance.Equipment used for seeding of sugar beets, field beans and various vegetables often uses a row distance of 48-50 cm.This row distance allows for physical weed control by hoes, torsion weeders, rotating brushes et cetera, a possibility especially important in organic production and when few herbicides are available.In cool cropping regions such as the Scandinavian countries, it might also be favourable to choose row distances of 25 or 50 cm to allow more sunlight on the individual soya plants and the maturing pods.
The three seeding periods; early (May 4th-May 11th), normal (May 15th-May 31st) and late (June 7th-June 11th) represent periods when seeding of soybean took place.The early period was considered as interesting for a typical grain producer, while the normal period is in line with recommendations based on soil temperature requirements of soybean (approx.10°C).The late period would be suitable for farmers looking for the following crop after harvest of early potatoes in southern Sweden.
Yields of seeding in the early period varied between 1385 and 2017 kg ha −1 ; for the normal period between 905 and 2405 kg ha −1 , and for the late period between 680 and 1693 kg ha −1 (Tables 7 and 8).Although variations in field were recorded, there were few statistical differences in yield (Tables 7 and 8).Seeding in the late period was less favourable for yield, especially in the cold and wet year of 2012.
From a farmers' point of view, 'Bohemians' (Table 7) seemed to be a reliable variety as it yielded between 1500 and 2000 kg ha −1 regardless of seeding date and season.The variety is one of the earliest cv. on the market of Prograin origin.It is not clear if the plant type itselfthe shorter, bushy appearanceis a feature beneficial for Swedish conditions.
For 'Silesia' (Table 8) seeding in mid of May generally yielded more than early or late seeding (p = .004).The somewhat higher heat sum needed for this cultivar to mature may be an explanation; early seeding is often connected with low soil temperature and late seeding will shorten the period of growth and thus lower the total heat sum.As we noticed in the SSD trials in Skåne 2013, 'Silesia' required 138 days to full maturity compared to 118 days of 'Bohemians'.A late seeding of 'Silesia' will negatively affect the ability to reach full maturity, especially cold seasons.For the farmer, choosing 'cultivars such as 'Silesia' is risky; on one hand a warm spring can give an opportunity to seed early and use the full yield potential, while on the other hand, a cool season might result in low (700-1200 kg ha −1 ) yields.
The findings stress the importance of using true 000varities with early maturation and low heat sum requirements.An applied aspect is that future field trials should include seeding date as a parameter in evaluating new cultivars for Scandinavia.It might be better to choose varieties that produce an average yield (1500-1800 kg ha −1 ) regardless of summer temperatures instead of varieties that excel warm summers but give poor results in cold years.

Aspects on cropping practice and future development
The fact that soybean requires high soil temperatures at seeding and during development and few herbicides are allowed for use in Sweden, hamper an expansion of Swedish conventional production.As soybean pods are set close to the soil surface, flexi-headers able to cut plant stems about 30 mm above soil surface is beneficial for reduction of field losses at harvest.Flexiheaders are available for all major combine brands.The current lack of available herbicides may trigger production of organic soybeans sought after by consumers and local/ regional feed producers.Tools for physical weed control such as video or GPS-guided hoes or torsion weeders are available on the market reducing the need for herbicide use.
An interesting approach would be harvest of the fresh soybean podsedamamea product today imported either fresh or frozen from overseas.Edamame production would likely be more profitable than of mature seeds and could be an interesting alternative for small-scale vegetable farms already used to handharvest and handling of produce in a cool-chain from farm to consumers.This system, however, may require the use of plastic tunnels to secure pod development and reduce damages by birds, deer and rabbits.

Conclusions and recommendations
We conclude that soybeans can be cropped in Sweden at latitudes up to about 59°(Stockholm region).Varieties of the 000-group should be used in order to secure yield,

Notes on contributor
Fredrik Fogelberg is researcher at the RISE -Research Institutes of Sweden.He is born in Malmö, Sweden 1968, MSc Horticulture in 1993 and Dr Agricultural Engineering at the Swedish University of Agricultural Science in 1998.Fogelberg has conducted R&D in physical weed control, legumes and plantbased foods since 2000.

Table 1 .
Acreage, yields and total production of soya beans of the main production countries in Europe 2018 (FAOSTAT 2020).
Note: Original table modified by Fogelberg.

Table 3 .
Yield of soya beans in kg ha −1 corrected to 8% moisture.
Notes: x, trial is excluded due to damages by birds or roe deer.-, not included for testing.Yields with the same letters are not statistically separated at the 95% level using Tukey's HSD test.

Table 5 .
Days to maturity in 2013 at three test sites.
Note: Data from trials of Scandinavian Seed AB, Sweden carried out by the author.

Table 6 .
Yield of soya bean 'Bohemians' at two sites using three row distances.
damaged by animals, discarded.Yields with same letters are not statistically different at the 95%-level using Tukey HSD.

Table 4 .
Yield of soya beans in kg ha −1 corrected to 15% moisture.Data from Scandinavian Seed AB trials carried out by the author.-, not included for testing.Yields with the same letters are not statistically separated at the 95% level using Tukey's HSD test.control is applied.Danube Soya, a non-profit organisation for European soya producers, traders and processor recommends a row spacing of 50 cm (Rittler et al. 2020).In USA and Canada, row spacing of 30 in.(76 cm) is common although narrow spacing of 15 in.