Plant materials and population development.

Two USDA-ARS-BARC bean breeding lines, BelFla-RR-1 and BelJersey-RR-15, both of which have the Ur-4 and Ur-11 rust resistance genes, were used as sources of resistance. BelFla-RR-1 (BF) is a fresh-market bush-type green bean with the Ur-11 gene introgressed from PI 151385 and the Ur-4 gene from ‘Early Gallatin’. BelJersey-RR-15 (BJ) is a processing bush-type green bean with the Ur-11 gene introgressed from PI 181996 and the Ur-4 gene from ‘Early Gallatin’. Both BF and BJ also have the I-gene for bean common mosaic virus resistance. Four snap bean breeding lines, HT601, HT603, HT608, and HT611, were used as sources of heat tolerance. The heat-tolerant lines were developed at Cornell University, Geneva, NY (Rainey and Griffiths, 2004, 2005c). The four heat-tolerant parents were early-maturing and had a bush-type growth habit with fleshy, straight, long, and smooth textured pods concentrated in set at the upper two-thirds of the canopy.

Eight populations were developed in Mar. and Apr. 2007 from crosses involving the two rust-resistant lines (BF and BJ) with the four heat-tolerant lines (HT601, HT603, HT608, HT611). The parents were crossed using a hooking method without emasculation (Bliss, 1980). The crosses made were: HT601 × BJ, HT603 × BF, HT608 × BF, BF × HT601, BF × HT611, BJ × HT608, BJ × HT611, and BJ × HT603. Eight F₁ hybrids were self-pollinated and subsequent generations of each of the eight populations were evaluated for rust resistance or yield components under heat stress in greenhouse comparisons.

Selection for heat tolerance.

Selection for heat tolerance was done at the F₂ and F₃ generations. Eight populations were evaluated at the F₂ generation. For each of the F₂ populations, 48 plants were grown under heat stress for preliminary selection of plants under high-temperature conditions. Eight checks that included the six parents BF, BJ, HT601, HT603, HT608, and HT611 and two snap bean cultivars, Hystyle (Harris Moran, Modesto, CA) and Juliet (Alpha Seeds, Henley on Klip, South Africa), were used during the F₂ evaluation. Eight plants were grown for each of the eight checks. Two seeds (later thinned to one plant after germination) were planted on 7 Oct. 2007 in 20-cm-diameter, 20-cm-deep round plastic pots filled with “Cornell mix” growth medium (Boodley and Sheldrake, 1972). Sets of plants of each of the eight populations and checks were randomly assigned to separate benches within the greenhouse in a completely randomized design. The plants were uniformly irrigated throughout the period of the greenhouse trial. Plants were staked and tied to ensure upright growth within the pots. Greenhouse temperature was maintained at 24/21 °C (day/night) for the first 3 weeks after which it was raised and maintained at 32/27 °C (day/night) until the end of the crop cycle to ensure heat stress started several days before anthesis.

The F₂ plants were visually selected for pod set, pod quality, and growth habit. The yield component data collected at harvest included pod number per plant, seed weight per plant (in grams), seed number per plant, seeds per pod, and single seed weight (in grams). Single-seed weight was recorded to guide selection for pod size, because seed weight and pod size are highly correlated. Fifty-five selections were made from the eight F₂ populations.

At the F₃ generation, five progeny from each of the 55 selections were evaluated in Feb. 2008 under high-temperature stress as described for the F₂ experiment. For the F₃ heat evaluation, the genotypes used as checks included ‘Bronco’ (Seminis, St. Louis, MO), ‘CT70’ (Cornell University, Geneva, NY), ‘HB1880’ (Syngenta, Boise, ID), ‘Masai’ (Syngenta), ‘Spartacus’ (Seminis), ‘Venture’ (Syngenta), and those used in the F₂ screen. Twenty-one breeding lines derived from additional crosses were also evaluated. A total of 90 genotypes was evaluated at the F₃. Plants were randomized throughout the greenhouse and maintained as previously described.

To arrive at the final selections at the F₂ and F₃ generations, means of four yield components, pod number per plant, seed number per plant, seed number per pod, and seed weight per plant (in grams), were independently ranked and mean rankings calculated for each of the genotypes. For each of the yield components, the genotypes tested were ranked in ascending order starting with the genotype with the highest value of the yield component in question. The averaging of the ranks of the four yield components giving equal weight to each minimizes possibilities of inaccuracies that may arise from selections based on only a single yield component. For instance, selections based only on pod number per plant may not accurately reflect tolerance to heat stress because some genotypes have the ability to form many pods under hot conditions but form fewer seeds, which may in addition have lower single seed weight and lower seed viability.

Selection for rust resistance.

All the 55 F₃ lines previously selected for tolerance to heat stress at the F₂ and re-evaluated for heat tolerance at the F₃ were tested at the F₄ generation for homozygosity of two dominant rust genes, Ur-4 and Ur-11. To evaluate the lines for the two rust genes, bean seedlings were inoculated with Races 67 and 108 of the bean rust pathogen under greenhouse conditions at USDA-ARS-BARC. Five checks were included in the rust evaluation: ‘Pinto 114’ (University of Idaho, Kimberly, ID) with no known rust resistance genes and susceptible to Races 67 and 108; ‘Early Gallatin’ (Gallatin Valley, Twin Falls, ID) with Ur-4 and susceptible to Race 67 and resistant to Race 108; PI 181996 with Ur-11 and resistant to Race 67 and susceptible to Race 108; and the snap bean breeding lines BF and BJ, both with Ur-4 and Ur-11 and both resistant to Races 67 and 108. The inoculations were conducted between 6 May and 18 June 2008. Additional inoculations with rust Races 49 and 108 to identify Ur-4, Race 47 to identify Ur-5, and Race 67 to identify Ur-11 were done in Aug. 2009 to determine the status (with respect to these three rust resistance genes) of selected snap bean cultivars used as checks during field trials.

To distinguish the reaction of the breeding lines to the races of the bean rust pathogen, primary leaves of each 7-d-old bean seedling inoculated with Race 67 were cut at the tip, whereas the leaf inoculated with Race 108 was left intact. The inoculum (urediniospores), at a concentration of 60,000 spores/mL, was dispersed in water using polysorbate 20. To promote pathogen development, inoculated plants were placed in the dark in a Percival dew chamber (Geneva Scientific, Fontana, WI) set at 19 °C and programmed to deliver very light dew that resulted in a relative humidity of ≈95%. The plants were incubated in the dew chamber for 16 h and then moved to a greenhouse where they were kept for 14 d before the host reaction was recorded. For each of the lines, 14 plants were scored for rust and those without rust symptoms were assumed to be fixed (homozygous) for the two rust genes and were selected.

Pustule size and type on the foliage were visually evaluated using the standard grading scale of 1 to 6 for rust evaluation (Stavely et al., 1983). The grading scale was as follows: 1 = leaves without any visible rust symptoms; 2 = leaves with necrotic spots without sporulation; 2⁺ = necrotic spots 0.3 to 1 mm in diameter; 2⁺⁺ = necrotic spots 1 to 3 mm in diameter; 3 = tiny pustules (uredinia) less than 0.3 mm in diameter; 4 = uredinia 0.31 to 0.5 mm in diameter; 5 = uredinia 0.51 to 0.8 mm in diameter; and 6 = uredinia larger than 0.8 mm. Plants with a grade of 1 to 3 (pustules absent, necrotic spots without sporulation, or tiny sporulating pustules with diameters under 0.3 mm) were considered resistant, whereas those with large sporulating pustules with grades of 4 or higher were considered susceptible. The resistance reaction conditioned by the Ur-4 rust gene in response to Race 108 or to any other race of common bean rust in which Ur-4 conditions resistance is expressed as necrotic non-sporulating spots (referred to as Grades 2, 2⁺, 2⁺⁺), also known as the hypersensitive response. The resistant reaction of the Ur-11 rust gene to Race 67, or to any other race to which Ur-11 is resistant, is expressed as faint necrotic non-sporulating spots accompanied by a few tiny sporulating uredinia (rust pustules) that is referred to as Grade 3. The susceptible reactions of Ur-4 and Ur-11 are expressed as large or very large sporulating uredinia classified as Grades 4 to 6.

Of the 55 F₃ snap bean breeding lines that were tested for rust resistance at the F₄, 15 were confirmed to be homozygous for the Ur-4 and Ur-11 genes (Table 1). From the 15 breeding lines that were fixed for the Ur-4 and Ur-11 rust resistance genes, three lines that also ranked high for heat tolerance were selected. The three selected breeding lines were (601BJ)L9, (BF601)L4, and (BF611)L11 (Table 1).

Table 1.

Mean values and rankings (R) of four yield components in 15 F₃ snap bean breeding lines determined to be homozygous for Ur-4 and Ur-11 bean rust resistance genes and a subset of five control cultivars (Bronco, Spartacus, Juliet, Venture, and Masai) in a trial where 90 snap bean genotypes were evaluated in a greenhouse under high temperature conditions, 32/27 °C (day/night).

View Table

Design of field trials in eastern Africa and Puerto Rico.

The three selected rust-resistant and heat-tolerant breeding lines (601BJ)L9 (HT2), (BF601)L4 (HT1), and (BF611)11 (HT3) were evaluated together with a rust-resistant but heat-sensitive control, (601BJ)L4 (HS1) (Table 1), and an additional set of 12 cultivars grown in different geographical regions. ‘Amy’ (Seminis), ‘PV 712’ (Pop Vriend, Andijk, The Netherlands), ‘Teresa’ (Seminis), and ‘PV 698’ (Pop Vriend) are currently grown in eastern Africa; ‘Barrier’ (Alpha Seeds) and ‘Juliet’ (Alpha Seeds) are grown in southern Africa; ‘Palati’ (Syngenta) is grown in northern Africa; ‘Opus’ (Seminis) and ‘Brio’ (Seminis) are grown in the southern United States; and ‘Bronco’ (Seminis), ‘Hystyle’ (Harris Moran), and ‘Masai’ (Syngenta) are grown in the northeastern United States. Seed for the field evaluation was increased in a climate-controlled greenhouse with temperatures maintained at 24/21 °C (day/night). Seeds were coated with the fungicide captan (Bayer Crop Science, Cary, NC) to satisfy seed import requirements and to enable a more uniform stand.

The trials were laid out in a randomized complete block design with four replicates at sites in eastern Africa and with five replicates in Puerto Rico. Six sites were planted in eastern Africa, five in Kenya, and one in Tanzania and one site in Juana Diaz, PR. For every replicate, a single row plot measuring 2.5 m in length was planted with 25 plants of each of the 16 genotypes. A planting density in which single plants were planted at a spacing of 0.5 m between rows in the sites in Africa and 1 m between rows in Puerto Rico and 0.1-m seed spacing within rows at all locations was adopted.

Location and description of field trials.

Field trials were undertaken between Mar. and June 2009 during the long rainy season at the sites in eastern Africa and from June to Sept. 2009 in Puerto Rico. The sites were selected on the basis of differences in soils and climate. Table 2 summarizes information on the locations, standard geographic coordinates, climatic conditions, and soil chemical properties at the sites of the trial. Altitude at the sites in eastern Africa ranged from 1172 m at Homabay to 1829 m at Kitale (Table 2). The field trial in Puerto Rico was conducted at the Experiment Station of the University of Puerto Rico in Juana Diaz, located in south central Puerto Rico and was the lowest of the sites at an elevation of 21 m. The Puerto Rico site differed significantly from the others in terms of latitude, soil, and climatic conditions.

Table 2.

Location, standard geographic coordinates, climatic, and soil chemical properties at six sites in eastern Africa and Puerto Rico where 16 snap bean genotypes were evaluated for yield and reaction to common bean rust during 2009 field trials.

View Table

There were temperature differences between the sites in eastern Africa, especially during the Mar. to June 2009 period of the study (Table 2). Homabay, which was at the lowest altitude of the six sites, had temperatures exceeding those at the high altitude site at Kitale by ≈3 °C and was more stressful for common bean reproductive development. Arusha had the lowest mean temperatures as a result of cooling effects of nearby Mount Meru. Puerto Rico had the highest mean temperatures among the trial sites with mean daily temperatures of 22.9 °C (minimum) and 33.8 °C (maximum) during the period that coincided with the reproductive phase of development.

Soils at the sites differed in type, chemical properties, and fertility status (Table 2). Soil pH at the sites ranged from 4.7 to 7.3. Soils at Homabay were heavy clays, high in phosphorus (P) and potassium (K) and moderate in nitrogen (N) content, and were the most fertile of the sites. Soils at Arusha were also considered fertile and were neutral in pH and had the highest K content compared with the other sites.

The sites selected for this study included Kitale, which represented the cool highland regions where most snap bean production in eastern Africa is currently concentrated. The inclusion of lower altitude sites such as Homabay and Kibos (and verification in Puerto Rico) aimed at testing and documenting potential performance of the developed snap bean genotypes at higher temperature environments in eastern Africa where snap bean production is presently low as a result of poor cultivars.

The sites were tractor ploughed and harrowed to a fine tilth before planting. Planting in eastern Africa was undertaken in Mar. and Apr. 2009 at the onset of the rainy season, whereas in Puerto Rico, the planting was completed on 25 June 2009. A compound inorganic fertilizer containing 10N–11.3P–8.3K was row-applied at planting at a rate of 200 kg·ha⁻¹ at all the sites in eastern Africa except Arusha where no fertilizer was applied because soils at the site were considered fertile. In Puerto Rico, inorganic fertilizer (containing 10N–4.4P–8.3K) was applied at 2 weeks after planting at a rate of 400 kg·ha⁻¹. Furrow irrigation supplemented rainfall at Arusha, whereas the trial plots at the other eastern Africa sites were rain-fed. The site in Puerto Rico was drip-irrigated. Standard agronomic practices were followed in managing the trial plots. The trial plots were kept free of weeds through cultivation by hand.

Field data collection and analysis.

Each of the entries was monitored for symptoms of rust and other diseases based on natural infection under field conditions. Dates on which first disease symptoms were observed at the sites were noted so as to account for possible site differences in rust severity and its effect on yield. The genotypes were scored for common bean rust at flowering (R6) and at pod filling (R8) developmental stages. To obtain information on genotypic reaction to rust, rust incidence data were obtained by counting the number of rust infected plants per plot, whereas rust severity was determined by the number of rust pustules formed per leaflet. The sets of rust incidence and severity data were used in a correlation analysis to obtain a simultaneous quantification of genotype differences in reaction to rust. In the correlation plot, genotypes that had no rust were considered highly resistant, whereas those with high rust incidence and severity were considered highly rust-susceptible. The genotypes that had high, moderate, or low rust severity on a few plants (low incidence) were assumed to be either partially rust-resistant or heterozygous for the rust resistance genes. Genotypic yield performance was estimated by the number of pods produced per plant. The total number of pods produced per plot was counted while excluding plants at the end of the rows and then divided by total number of plants examined.

The statistical model used in the data analysis was:

View Expanded

where: y_ijk = kth response at combination (i,j); μ = overall mean; α_i = main effect of genotype i; β_j = main effect of site j; αβ_ij = genotype/site interaction effect; and ɛ_ijk = random error associated with kth response at combination (i,j).

The data were statistically analyzed using regression and correlation analysis, analysis of variance, and the Tukey's test and contrasts for separating means. Regression analysis was done to determine if the yield response of the genotypes when grown at the high-temperature site in Puerto Rico could predict/confirm the response of the genotypes when grown at the high-temperature sites in eastern Africa. The statistical analyses were completed using JMP 8 software (SAS Institute, Cary, NC).

Greenhouse evaluation and selection for rust resistance and heat tolerance.

Fifteen of the 55 F₃ lines were found to be homozygous for Ur-4 and Ur-11. These lines had the typical reaction of common bean plants with the Ur-4 and Ur-11 genes when inoculated with Races 67 and 108. The 15 rust-resistant lines differed in their reaction to heat stress as shown by their mean rankings of four yield components (pod number per plant, seed number per plant, seed number per pod. and seed weight per plant) measured for each of the 90 genotypes evaluated (Table 1). Only five of the 15 rust-resistant lines were ranked among the top 50% of the 90 genotypes evaluated under heat stress. Three of the 15 rust-resistant lines were among the top 25% in the ranking of the 90 lines for heat tolerance and were selected as heat-tolerant. The three selected rust-resistant and heat-tolerant lines were HT1, HT2, and HT3 with mean rankings of 14, 7.3, and 20.5, respectively, among the 90 genotypes evaluated.

Reaction to common bean rust under field conditions.

High rust incidence and severity were observed at Arusha, Homabay, and Kitale. No rust was observed in Sabatia and Puerto Rico. No data were collected from the Kibos and Maseno site as a result of damage of trial plots. The Arusha site experienced a longer duration of lower temperatures compared with Kitale and Homabay (Table 2). The lower temperatures of 22 to 25 °C at Arusha in April to June promoted development of rust early in the season and the conditions remained favorable for rust development until the end of the crop cycle. Rust symptom development at Arusha started ≈2 weeks after planting, when the plants were at the second trifoliate leaf growth (V2) stage. At Kitale and Homabay, rust symptoms did not manifest until 6 weeks after planting when most of the genotypes were at the pod formation and pod-filling stages.

Similar trends in rust incidence on the 16 genotypes were observed across the three sites (Arusha, Homabay, Kitale), and genotypes differed significantly in the frequency of rust-infected plants (Fig. 1). Three of the four breeding lines, HT1, HT2, and HS1, all of which combine the Ur-4 and Ur-11 rust resistance genes, had no plants with visible rust symptoms at the three sites where common bean rust occurred. Approximately 25% of HT3 plants were infected with rust, indicating that this breeding line was not fixed for the targeted Ur-11 gene. The cultivars PV 698 and PV 712, both of which have the Ur-11 rust resistance gene, had no rust at the three sites. The cultivars Amy, Barrier, Brio, Bronco, Hystyle, Juliet, Masai, and Opus showed rust on more than 60% of plants at each of the three sites indicating that they were susceptible to rust at these sites. Many of these genotypes have the Ur-4 rust resistance gene, whereas others have no known rust resistance genes.

Incidence of common bean rust on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The significance bars indicate ses (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Incidence of common bean rust on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The significance bars indicate ses (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Incidence of common bean rust on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The significance bars indicate ses (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

The genotypes differed significantly in rust severity, which ranged from absence of pustules on leaves to numerous pustules of different sizes observed on plant leaves (Fig. 2). The breeding lines HT1, HT2, and HS1 and the cultivars PV 698 and PV 712 were considered highly resistant because they had no visible rust pustules present at any of the three sites (Table 3; Fig. 3). ‘Teresa’ was also highly resistant at Arusha and Kitale but not at Homabay where some plants had rust symptoms, albeit with low severity (Table 3). Palati was also considered resistant, because only a small number of plants had small pustules. There was intense rust pustule formation on some Juliet plants indicating that the cultivar is heterogeneous for rust resistance. ‘Amy’, ‘Barrier’, ‘Brio’, ‘Bronco’, ‘Hystyle’, and ‘Masai’ were highly susceptible at all three rust-infected sites (Table 3). Very large pustules were also observed on the pods of ‘Amy’, ‘Barrier’, ‘Brio’, and ‘Hystyle’ at the Arusha site, confirming that these genotypes were highly susceptible to the bean rust pathogen. For the three sites combined, rust severity was highest in ‘Amy’ followed by ‘Brio’, ‘Bronco’, ‘Barrier’, ‘Masai’, and Hystyle' (Table 3).

Table 3.

Severity of common bean rust on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites and mean severity at the three sites combined during 2009 field trials in eastern Africa.^z

View Table

Absence of common bean rust symptoms on two snap bean breeding lines, HT1 and HT2, that combine the Ur-4 and Ur-11 rust resistance genes in a heat-tolerant genetic background and severity of rust on two susceptible snap bean cultivars, Amy and Barrier, evaluated at Arusha, Tanzania, during 2009 field trials in eastern Africa.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Absence of common bean rust symptoms on two snap bean breeding lines, HT1 and HT2, that combine the Ur-4 and Ur-11 rust resistance genes in a heat-tolerant genetic background and severity of rust on two susceptible snap bean cultivars, Amy and Barrier, evaluated at Arusha, Tanzania, during 2009 field trials in eastern Africa.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Absence of common bean rust symptoms on two snap bean breeding lines, HT1 and HT2, that combine the Ur-4 and Ur-11 rust resistance genes in a heat-tolerant genetic background and severity of rust on two susceptible snap bean cultivars, Amy and Barrier, evaluated at Arusha, Tanzania, during 2009 field trials in eastern Africa.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Effect of different common bean rust resistance genes and gene combinations on rust disease incidence and severity on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Effect of different common bean rust resistance genes and gene combinations on rust disease incidence and severity on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Effect of different common bean rust resistance genes and gene combinations on rust disease incidence and severity on 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites in eastern Africa during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Rust severity correlated with rust incidence and rust was more severe on genotypes that had the highest rust incidence (Fig. 3). The correlation between rust incidence and severity underscores the contribution of different rust resistance genes and gene combinations on snap bean genotype response to rust. Genotypes that had no known rust resistance genes, including ‘Amy’ and ‘Masai’, or only had the Ur-4 gene, including ‘Opus’, ‘Brio’, ‘Barrier’, ‘Bronco’, and ‘Hystyle’, had high rust incidence and severity (Fig. 3). The high rust incidence and severity observed on Ur-4 genotypes indicates that this gene is not effective against the race(s) of the bean rust found at the three sites (Arusha, Homabay, Kitale).

‘Teresa’ and ‘Palati’, which have the Ur-5 rust resistance gene, were considered resistant to rust because they had low rust incidence and severity at the three sites (Table 3; Fig. 3). Effectiveness of the Ur-5 gene was particularly notable at Kitale and Arusha, where there were no symptoms of rust on all ‘Teresa’ plants and no symptoms on 88% to 90% of ‘Palati’ plants, indicating that the Ur-5 gene was effective against races of the rust pathogen at these sites (Fig. 1). The Ur-5 gene, which is of Middle American origin, has broad resistance against many races of the bean rust pathogen, especially races that are Andean in origin (Pastor-Corrales, 2006). Deployment of the Ur-5 gene in cultivars targeted to regions where predominantly Andean beans are grown could increase the success rate of conferring resistance against the bean rust pathogen. However, the Ur-5 conferred rust resistance in ‘Teresa’ and ‘Palati’ was not complete at the Homabay site where ≈30% of ‘Teresa’ plants and 42% of ‘Palati’ plants had mild rust symptoms (Fig. 1). This observation could be indicative of the possible presence, at the Homabay site, of a bean rust isolate that overcomes the Ur-5 gene, although there was not 100% infection of the two cultivars with the Ur-5 overcoming isolate.

‘PV 698’, which has the highly effective Ur-11 rust resistance gene and HS1, HT1, HT2, and ‘PV 712’ that have the Ur-4 and Ur-11 rust gene combination, were highly rust-resistant and had no rust at the three sites. The observation that ‘PV 698’ was rust-resistant at the sites indicates that the Ur-11 gene conferred resistance against all races of the bean rust pathogen found at these sites and that the Middle American rust Race 108, which is known to overcome the Ur-11 gene (Pastor-Corrales, 2006), was not present at the study sites.

Yield in eastern Africa and Puerto Rico.

Yield data from Puerto Rico and for three eastern Africa field sites, Arusha, Homabay, and Kitale, were analyzed. In terms of performance of all the genotypes combined at each site, pod yield was highest at Kitale followed by Arusha and Homabay, which were not significantly different from each other, followed by Puerto Rico according to Tukey's test (P ≤ 0.05). The low yield at Puerto Rico indicates that it was the most stressful of the four sites.

Bean rust infection was the most important yield-influencing factor at Arusha, whereas at Homabay, yield was largely influenced by high-temperature stress and to a lesser extent by bean rust. Rust had less impact on yield at the Homabay site because rust symptoms did not appear until 6 weeks after planting when the bean genotypes were at the pod formation and pod-filling stages. At the Puerto Rican site, high temperatures of 34/23 °C (day/night) during reproductive development resulted in significant symptoms of high-temperature stress, including bud, flower and pod abortion, poor pod fill, curved pods, and the excessive production of pin pods, thus significantly affecting yield response.

In a combined analysis of yield data from the sites, the breeding lines HS1 and HT2 were the highest yielding among the genotypes tested across the four sites, whereas ‘PV 712’, ‘Masai’, and ‘Palati’ had the lowest yields (Table 4). The high yields of the two breeding lines relative to the other genotypes tested underscores the contribution that the combined introgression of rust resistance and heat tolerance traits had in stabilizing performance of the breeding lines across the different sites.

Table 4.

Pod yield of 16 snap bean genotypes evaluated at Arusha, Homabay, and Kitale sites (in eastern Africa), Puerto Rico, eastern Africa (E. Africa) sites combined, and all four sites combined during 2009 field trials.^z

View Table

Analysis of variance of yield from Arusha, Homabay, Kitale, and Puerto Rico sites combined revealed a highly significant genotype-by-site interaction effect (P < 0.001). The genotype–by-site interaction effect remained highly significant even when data from the sites in eastern Africa alone were analyzed in a combined analysis (P < 0.001). This significant interaction effect implies that factors influencing yield were unique to each of the sites and that they differentially affected genotypic performance. Given the uniqueness of the sites, genotypic performance at the four sites was interpreted on a site-by-site basis. At Arusha, where an early outbreak of the common bean rust was the most important yield-influencing factor, ‘PV 698’, ‘Juliet’, and the breeding line HT2 were the highest yielding, whereas the rust-susceptible cultivars, Brio, Bronco, and Hystyle, were the lowest yielding (Table 4). In Kitale, where there was a late-season bean rust outbreak and higher temperatures compared with Arusha but mild compared with Homabay, the highest yield was obtained from ‘Teresa’ followed closely by ‘Masai’, ‘Bronco’, ‘Brio’, and HT2, whereas the lowest yield was from ‘Hystyle’ (Table 4). At Homabay, where there was high-temperature stress and a late-season rust outbreak, the highest yield was from HT2 followed by ‘Opus’, ‘Hystyle’, ‘HS1’, ‘Bronco’, and ‘Brio’ and the lowest yields were from ‘PV 712’ followed by ‘PV 698’, ‘Masai’, and ‘Palati’ (Table 4). In Puerto Rico, the highest yields were from HT2, ‘Bronco’, HS1, and ‘Brio’, whereas the lowest yields were from ‘PV 698’, ‘PV 712’, ‘Amy’, ‘Juliet’, and ‘Masai’ (Table 4). Pod yield of the rust-resistant and heat-tolerant breeding lines HT2, HS1, HT1, and HT3 ranked 1, 4, 5, and 6, respectively, out of the 16 genotypes evaluated in Puerto Rico (Table 4).

To understand how the different site conditions, including high rust pressure from an early-season outbreak at Arusha and high-temperature stress at Homabay and Puerto Rico, influenced genotypic performance, pod yield at the sites were compared using contrasts and regression analyses. Yield data from Arusha and Kitale were contrasted for selected rust-resistant and susceptible genotypes to test whether the difference in the duration of rust infection at the two sites significantly affected yield (Fig. 4). There was no significant yield difference between Arusha and Kitale sites for the highly rust-resistant genotypes, including HT1, HT2, and ‘PV 712’ (Fig. 4).

Effect of early-season and late-season rust disease outbreak at Arusha and Kitale sites, respectively, on pod yield of rust-resistant and rust-susceptible snap bean genotypes evaluated during 2009 field trials in eastern Africa. HT1 and HT2 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant, whereas PV712, Brio, Bronco, and Masai are cultivars. The bracketed figures appearing with the site legends respectively denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P ≤ 0.001 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Effect of early-season and late-season rust disease outbreak at Arusha and Kitale sites, respectively, on pod yield of rust-resistant and rust-susceptible snap bean genotypes evaluated during 2009 field trials in eastern Africa. HT1 and HT2 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant, whereas PV712, Brio, Bronco, and Masai are cultivars. The bracketed figures appearing with the site legends respectively denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P ≤ 0.001 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Effect of early-season and late-season rust disease outbreak at Arusha and Kitale sites, respectively, on pod yield of rust-resistant and rust-susceptible snap bean genotypes evaluated during 2009 field trials in eastern Africa. HT1 and HT2 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant, whereas PV712, Brio, Bronco, and Masai are cultivars. The bracketed figures appearing with the site legends respectively denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P ≤ 0.001 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Among the highly rust-susceptible genotypes evaluated, two types of yield responses were observed between the early, or longer duration, rust-infected Arusha site and the late (or shorter duration) rust-infected site, especially Kitale. In the first response, yield was significantly reduced at the early rust-infected site at Arusha compared with the late rust-infected site at Kitale. The cultivars that manifested this yield response were Brio, Bronco, and Masai as the highly significant contrasts of their yields at these two sites illustrate (Fig. 4). The cultivars that showed this first type of yield response illustrate the magnitude of yield loss that a rust epidemic within snap bean production regions may cause on susceptible cultivars depending on the stage in the crop growth cycle at which the disease begins. In the second yield response, in which the other highly susceptible cultivars Amy, Barrier, and Hystyle were grouped, there were no significant yield differences between the early and late rust-infected sites (Table 4). This second group of cultivars is considered rust-tolerant with respect to yield. However, there were large rust pustules on pods, which resulted in deformed pods and reduced quality even in these rust-tolerant cultivars.

Yield data from Kitale and Homabay were contrasted for all the genotypes grown to test whether the difference in temperature at the two sites significantly affected yield and whether the newly developed breeding lines were tolerant to the temperature differences. Homabay and Kitale sites differ in both altitude and average temperature with Kitale being cooler and in the highlands and Homabay hotter and at a lower altitude. The effect of rust infection on yield at these two sites was assumed to be similar but minimal, because rust outbreak at both sites was late in the season. The pod yields of ‘Amy’, ‘Teresa’, ‘PV 712’, ‘Masai’, and ‘PV 698’ were significantly reduced at Homabay, whereas the yields of HT1 and HT2 did not significantly differ at the two sites (Fig. 5). The non-significant difference in yield of the breeding lines between the higher temperature and cooler temperature site in eastern Africa indicated that they were tolerant to the higher temperatures at least to the levels experienced at Homabay.

Effect of temperature differences between Homabay and Kitale sites on pod yield of two rust-resistant and heat-tolerant snap bean breeding lines, HT1 and HT2, and four snap bean cultivars: Amy, PV 712, Teresa, and PV698 evaluated during 2009 field trials in eastern Africa. The bracketed figures appearing with the site legends, respectively, denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P < 0.001, **P ≤ 0.01 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Effect of temperature differences between Homabay and Kitale sites on pod yield of two rust-resistant and heat-tolerant snap bean breeding lines, HT1 and HT2, and four snap bean cultivars: Amy, PV 712, Teresa, and PV698 evaluated during 2009 field trials in eastern Africa. The bracketed figures appearing with the site legends, respectively, denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P < 0.001, **P ≤ 0.01 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Effect of temperature differences between Homabay and Kitale sites on pod yield of two rust-resistant and heat-tolerant snap bean breeding lines, HT1 and HT2, and four snap bean cultivars: Amy, PV 712, Teresa, and PV698 evaluated during 2009 field trials in eastern Africa. The bracketed figures appearing with the site legends, respectively, denote altitude and ranges of temperature during the period of the trial. Bars indicate se; ***significant at P < 0.001, **P ≤ 0.01 (n = 4).

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Regression analyses of data from the four sites revealed a strong positive correlation (R² = 0.71, P < 0.001) of pod yield of the 16 genotypes under high-temperature stress in Puerto Rico when compared with Homabay, the highest temperature African trial site (Fig. 6). There was a weak negative correlation of yield data between Puerto Rico and the Arusha site (R² = 0.34, P = 0.0188) and no correlation with the Kitale site (R² = 0.03, P = 0.525) (Fig. 6). The significant but weak negative correlation between Arusha and Puerto Rico yield data is attributed to contrasting responses of two sets of cultivars: 1) PV 698 and PV 712, which are rust-resistant and heat-sensitive and yielded well in Arusha but poorly in Puerto Rico; and 2) Brio and Bronco, which are rust-susceptible but heat-tolerant, yielded well in Puerto Rico and poorly in Arusha (Fig. 6). The contrasting response of these two sets of genotypes contributed significantly to the observed negative correlation.

Regression analysis of mean pod yield of 16 snap bean genotypes evaluated at three eastern Africa sites, Arusha, Homabay, and Kitale, and the eastern Africa sites combined with pod yield from Puerto Rico during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The data points are means of four replicates for the eastern Africa sites and five replicates for the Puerto Rico site. R² values followed by * and *** indicate statistical significance at P ≤ 0.05 and P ≤ 0.001, respectively. The highly significant R² value between Homabay and Puerto Rico indicates similarity in differential response of the genotypes to high temperatures at the two sites.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

View Full Size

Regression analysis of mean pod yield of 16 snap bean genotypes evaluated at three eastern Africa sites, Arusha, Homabay, and Kitale, and the eastern Africa sites combined with pod yield from Puerto Rico during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The data points are means of four replicates for the eastern Africa sites and five replicates for the Puerto Rico site. R² values followed by * and *** indicate statistical significance at P ≤ 0.05 and P ≤ 0.001, respectively. The highly significant R² value between Homabay and Puerto Rico indicates similarity in differential response of the genotypes to high temperatures at the two sites.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

Regression analysis of mean pod yield of 16 snap bean genotypes evaluated at three eastern Africa sites, Arusha, Homabay, and Kitale, and the eastern Africa sites combined with pod yield from Puerto Rico during 2009 field trials. HT1, HT2, and HT3 are breeding lines initially developed and selected as both rust-resistant and heat-tolerant. HS1 is a breeding line initially selected as rust-resistant but heat-sensitive and was included in the trial as a control. The remaining 12 genotypes are cultivars. The data points are means of four replicates for the eastern Africa sites and five replicates for the Puerto Rico site. R² values followed by * and *** indicate statistical significance at P ≤ 0.05 and P ≤ 0.001, respectively. The highly significant R² value between Homabay and Puerto Rico indicates similarity in differential response of the genotypes to high temperatures at the two sites.

Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 135, 6; 10.21273/JASHS.135.6.521

• Download Figure

At the Homabay and Puerto Rico sites ‘Amy’, ‘Masai’, ‘PV 698’, and ‘PV 712’ were the lowest yielding, whereas HT2, ‘Bronco’, HS1, and ‘Brio’ were among the highest yielding with HT2 having the highest yield at both sites (Table 4; Fig. 6). The highly significant correlation of the results for Homabay and Puerto Rico indicate that the high temperatures at the two sites similarly influenced genotype response across two very distinct agroecological zones. The lowest yielding genotypes at the two sites were therefore the most heat-sensitive, whereas the highest yielding genotypes were the most heat-tolerant. This observation confirmed that the genotype difference in yield responses at Homabay and Kitale was the result of site temperature differences.