Forty-eight genotypes of spring barley were used for this study (Table 1). Many of these genotypes have shown some level of FHB resistance in previous studies or in the literature, while Brucefield, Encore, Stander, and Viviane were susceptible checks (Choo et al., 2004, 2014). These barley genotypes originated from nine countries (China, Canada, Ecuador, Japan, Germany, Sweden, Switzerland, Russia, and the United States) and included 34 two-row, two deficiens (absence of sterile lateral florets), and 12 six-row type. Six of the genotypes were hulless and 42 were covered. The genotype Birka was specifically selected for this study as its spike was partly enclosed in the upper leaf sheath from awns emergence to maturity in the field (Anonymous, 1982), which may influence infection and disease development. The genotype Kanto Nadate Gold was included because of its cleistogamy (closed flowering) as reported by Sameri et al. (2006).

The 48 genotypes were planted in a randomized complete block design with two replicates at the Harrington Research Farm, Charlottetown Research and Development Centre, Prince Edward Island from 2009 to 2012. Each experimental unit consisted of three rows, 3 m long, and 17.8 cm between rows. Each year, a fertilizer (17-17-17) at a rate of 412 kg/ha (or 70 kg N/ha) was applied to the field before planting. The herbicide MCPA Amine 500 + Refine Extra (a.i. MCPA and thifensulfuron/ tribenuron) was sprayed at the recommended rate. Grain yield, heading date, plant height, and DON concentration was determined in each plot. For DON analysis, approximately 20 to 25 g of harvested barley seed was ground to pass through a 0.4 mm screen on a Udy-mill. To prepare extracts, 5 ml of methanol: water (1:9, v/v) were added to 1 g ground samples in 10 ml plastic tubes, which were then subjected to end-over-end mixing for one hour, centrifuged for 5 min at 2,000 rpm. DON analysis was conducted on the filtrate following the enzyme-linked immunosorbent assay (ELISA) procedures described by Sinha et al. (1995). The antibody for this ELISA was developed and produced in-house and has a high affinity for both DON and 15-ADON. The accuracy of the ELISA procedures has been reported to be comparable to that of the gas chromatography method (Sinha and Savard, 1996). The limit of quantitation was 0.1 mg/kg. DON analysis was conducted in the mycotoxin laboratory at the Ottawa Research and Development Centre, Ottawa.

The 48 barley genotypes were seeded in a randomized complete block design with three replicates at Ottawa in 2004 and 2005, and with three replicates at Hangzhou, China in 2005. At Ottawa, the experimental materials were seeded at a seeding rate of 5 g per row.

Each experimental plot consisted of two 1.5-m rows with a row spacing of 15 cm. Three F. graminearum isolates (DAOM178148, DAOM212678, and DAOM232369) from the Canadian National Mycological Herbarium, Ottawa Research and Development Centre, were used as inoculum for this study. These isolates were chosen since they are known to be aggressive in Canadian condition (Xue et al., 2004). Equal parts by volume of corn (Zea mays L.) and barley grain were mixed and soaked in water for 48 hours. Excess water was removed and the kernel mixture, which was placed in an aluminum pan (53 × 32 × 8 cm³) and covered with two layers of aluminum foil, was autoclaved for 1 h at 120°C. In each pan, 100 ml of spore suspension of individual isolates at a concentration of 5.5 × 10⁴ spores per ml was evenly poured over the surface of the kernel mixture in a laminar flow hood. The kernel mixture was incubated at room temperature for each isolate for at least 4 weeks. Kernels infected with the three isolates were mixed in the same proportion and then broadcast at a rate of 50 g per m² once in 2004 (July 5 at late stem elongation) and at a rate of 30 g per m² twice in 2005 (June 22 at early stem elongation and June 28 at booting). Shortly after inoculation, the plots were watered with a sprinkler system twice a day (at 11:00-11:30 am and 3:00-4:00 pm) for 20-30 min to increase the humidity.

Heading date, plant height, and FHB incidence (i.e., percentage of heads infected per row) were recorded. DON concentration was determined at Ottawa for all plots using the ELISA method as described earlier.

At Hangzhou, China, the experimental materials were seeded at a seeding rate of 5 g per row. Each plot consisted of one 1.5-m row with a row spacing of 30 cm. A mixture of F. graminearum isolates collected from the cultivar Giu Damai from Hailing, Zhejiang was multiplied on a potato-dextrose-agar (PDA) medium at 28°C for 10 days. Wheat kernels were soaked in water for 24 h, put in 1-l glass bottles, and then autoclaved for 30 min at 120°C. The wheat kernels were inoculated with the F. graminearum mycelium at five positions in a glass bottle and incubated at 25°C for 10 days. Infected wheat kernels were broadcast at a rate of 7.5 g per m² at the booting stage. At heading, all plots were also sprayed with a F. graminearum suspension. There was no supplementary water of the inoculated plots at this location. Heading date, plant height, and FHB incidence (i.e., percentage of heads infected per row) were recorded. DON concentration was determined for all plots at Ottawa using the ELISA method as described natural infection section of this study.

Three greenhouse trials were conducted at the Ottawa Research and Development Centre, Ottawa, from January to June each year in 2005, 2006 and 2007, using the same three isolates (DAOM178148, DAOM212678, and DAOM232369) as in grain spawn inoculation. The isolates were cultured on a modified PDA (10 g/l of dextrose amended with 34 mol/l streptomycin sulfate) and incubated at 22-25°C under mixed UV and fluorescent lighting on a 12 h light:12 h dark cycle for 14 days. The modified PDA medium was used to reduce mycelium growth, possible mutation, and poor vigour, and to increase spore production by the pathogen (Xue et al., 2004). To prepare inoculum, 0.5 ml of a concentrated macroconidial suspension (approx. 107 spores/ml), obtained from the above, was spread over the surface of the modified PDA in 9-cm Petri dishes and incubated as described above for 48 h. Ten millilitres of sterile distilled water containing 0.01% Tween 20 (polyoxyethylene sorbitan monolaurate) was then added to each dish, and the surface was scraped gently with a sterile microscope slide to dislodge spores. The resulting conidial suspension was filtered through two layers of cheesecloth and adjusted to a concentration of 5 × 10⁴ spores/ml using a hemocytometer. Separate conidial suspensions were prepared for each isolate. The final suspension used for the greenhouse consisted of 1:1:1 mixture of each of the three F. grameniarum isolates.

The seeds of 48 barley genotypes were planted in 15-cm diameter pots containing a mixture of loam soil, sand, and composted cow manure (1:1:1, v/v/v) and maintained at 23-25°C during the day and 18-20°C at night in a greenhouse. Supplemental light was provided by 300 W metal halide lamps to ensure a 16 h photoperiod and a minimum intensity of 360 mol/m²/s. Plants were fertilized with a 1% solution of 20-20-20 (N-P-K) at the fifth week after planting and once per week thereafter. The 48 barley genotypes were inoculated 10-14 days after heading. Plants were sprayed with the spore suspension at approximately 0.2 ml per spike using a DeVilbiss model 15 atomizer (The DeVilbiss Co., Somerset, PA, USA). After the inoculum dried for 30 min, plants were transferred to a polyethylene humidity chamber in a growth room for 24 h. The growth room was operated at 25°C with 12-h photoperiod at a light intensity of 250 mol/m²/s. The humidity chamber was maintained at or near 100% relative humidity by the continuous operation of two ultrasonic humidifiers. Air temperature and humidity in the chamber were monitored with a portable datalogger (model 21XL micrologger, Campbell Scientific Canada Corp., Edmonton, AB, Canada). After incubation, plants were returned to the greenhouse bench. For each barley genotype, eight replicate pots were used in 2005 and 2006 and four replicate pots in 2007. Pots were arranged in a completely randomized design in both the humidity chamber and the greenhouse after the inoculation. The severity of FHB was rated as the percentage of infected spikelets 21-23 days after inoculation when plants were at the soft dough stage.

Data were analyzed using the MIXED procedure of the SAS software (SAS Institute, 2004). The genotype was considered to be a fixed effect and all of the other terms (year, rep [year], and genotype year) to be random effects. Log-transformed DON data were used to perform the analysis of variance. Means over all of the trials in each experiment were used for calculating Pearson correlation coefficients. Pearson correlation coefficient was calculated to determine (1) if DON concentration was correlated with any of the three agronomic traits (yield, height, and heading date) in the natural infection experiment, (2) if DON concentration was correlated with FHB incidence, height, and heading date in the grain spawn experiment, (3) if DON concentration under natural infection was correlated with DON concentration and FHB incidence with grain spawn inoculation, and (4) if FHB severities with indoor spray inoculation were correlated with DON concentration under natural infection and DON concentration and FHB incidence with grain spawn inoculation.

There was a significant (P < 0.05) variation among genotypes for all four traits (Table 2). DON contamination was greatest in 2009 and 2010, intermediate in 2011, and low in 2012 and 2013. The mean yield was similar from year to year. Plant height was tall in 2009, 2010, and 2011, but short in 2012 and 2013. The heading date was early in 2009, 2011, and 2012, but late in 2010 and 2013 (Table 2). There was a significant genotype-year interaction for DON concentration, yield, plant height, and heading date (Table 3). The genotype-year interaction can be attributed to a change in magnitude of mean DON concentration in different years. All of these barley genotypes were contaminated with DON (Table 4). DON concentration of these genotypes was either positively correlated or not correlated between the 5 years. Positive correlation was detected between 2009 and 2010 (r = 0.69, P < 0.01), between 2009 and 2011 (r = 0.76, P < 0.01), between 2009 and 2012 (r = 0.41, P < 0.01), between 2010 and 2011 (r = 0.60, P < 0.01), and between 2011 and 2012 (r = 0.43, P < 0.01). Two-row genotypes, in general, contained less DON than six-row genotypes except Zhenongda 3, Zaosho 3, Mimai 114, CH9528-10, and Kanto Nakate Gold (Table 4). The two deficiens genotypes contained an average amount of DON. Hulless genotypes generally contained less DON than covered genotypes with the exception of Mimai 114. Only nine genotypes (Island, Zhedar 1, Golozernyj 1, Harbin 2r, CDC Freedom, Zhedar 2, Maja, CH9507-8, and Niedzica 1) from 2-row barley contained less than 1 mg DON kg-1, which is the recommended maximum tolerance level of DON for swine diets in Canada. These nine genotypes were generally tall (ranging from 77 cm to 96 cm) and late in the heading date (ranging from 60 to 64 days) (Table 4). Of the nine genotypes that had low DON, Island and CH9507-8 had high yield (3,700 kg/ha) while the others had low yield ( < 2,800 kg/ha) (Table 4). The correlation coefficient between DON concentration and grain yield for all genotypes was not significant at the 0.05 level (Table 5). Plant height was not correlated with DON concentration, but heading date was negatively correlated with DON concentration (Table 5). Heading date remained negatively correlated with DON concentration even when the two earliest heading dates (at least 5 days earlier under natural infection) genotypes (Zhenongda 3 and Zaoshu 3) were excluded from the correlation analysis.

On average, these 48 genotypes at Ottawa, Ontario had higher DON, higher FHB incidence, and shorter and headed earlier than those at Hangzhou, China (Table 2). Variation for all four traits among genotypes was significant, so were environment x genotype interactions for all four traits (Table 3). The nine resistant genotypes contained low DON, ranging from 1.0 to 4.3 mg/kg with grain spawn inoculation (Table 4). Two genotypes, Kanto Nakate Gold (25.8 mg/kg) and Zhenongda 3 (40.7 mg/kg), contained an extremely high concentration of DON and they were considered far outliers (Table 4). When these two outliers (Kanto Nakate Gold and Zhenongda 3) were excluded from the correlation analysis, the correlation coefficient for DON concentration was highly positive between the three environments, i.e., between Ottawa 2004 and Ottawa 2005 (r = 0.79, P < 0.01), between Ottawa 2004 and Hangzhou 2005 (r = 0.44, P < 0.01), and between Ottawa 2005 and Hangzhou 2005 (r = 0.43, P < 0.01). In this experiment, DON concentration was positively correlated with FHB incidence, but it was negatively correlated with height and heading date (Table 5). Heading date negatively correlated with DON concentration even when the two earliest heading genotypes (Zhenongda 3 and Zaoshu 3) were excluded from the correlation analysis.

FHB severity varied from year to year and from genotype to genotype (Table 2). There were significant interactions between year and genotype (Table 3). The nine resistant genotypes were relatively low in FHB severity, ranging from 24% to 39%, but early heading genotypes such as Zhenongda-3, Zaoshu-3, Mimai-114, and Kanto Nakate Gold had even lower FHB severity (8% to 18%) (Table 4). DON concentration and FHB incidence with grain spawn inoculation were positively correlated with DON concentration obtained from natural infection (Table 5). When the two outliers (Kanto Nakate Gold and Zhenongda-3) were excluded from the correlation analysis the correlation coefficient for DON concentration between natural infection and grain spawn increased from r = 0.47-0.76. FHB severity with indoor spray inoculation was not correlated with DON concentration (r = 0.12) under natural infection, and similarly, FHB severity with indoor spray inoculation was not correlated with DON concentration (r = −0.23) and FHB incidence (r = 0.19) under grain spawn (Table 5).