Population development.

Six nuña bean landraces with an indeterminate growth habit and photoperiod sensitivity were obtained from the Centro Internacional de Agricultura Tropical, Cali, Colombia (Table 1).

Table 1.Common names and accession numbers of nuña landraces obtained from Centro Internacional de Agricultura Tropical (CIAT), Cali, Colombia, used to develop the temperate-adapted nuña breeding lines ‘PB24’ and ‘PB22’ and the population for genetic analysis.

View Fullscreen

Each of the six landrace accessions were grown at the University of Wisconsin, Walnut Street Greenhouses, Madison, WI, during the short-day winter months and used as females in crosses to ‘PB22’ and ‘PB24’, two temperate-adapted nuña bean breeding lines previously developed in our laboratory. ‘PB22’ and ‘PB24’ were developed using ‘Ayacucho’, a nuña landrace from Peru, as a donor parent in a cross and subsequent backcross to the recurrent parent ‘Stockbridge Indian Bean’ followed by pedigree selection among and within inbred-backcross families (Kmiecik and Nienhuis, 1997). Nuña bean lines ‘PB22’ and ‘PB24’ possess the characteristics of photoperiod insensitivity and a determinant growth habit (Type I) necessary for adaptation and production in Wisconsin; they differ primarily in seedcoat color: ‘PB24’ is uniform brown and ‘PB22’ is brown–blue (Kmiecik and Nienhuis, 1997). The F₂ populations resulting from each of the 12 crosses were field grown in the summer of 1996 at Hancock Agricultural Research Station (ARS), Hancock, WI. Individual plants exhibiting a Type I growth habit and that flowered and produced seed under long-days in Wisconsin were selected within each of the 12 populations. Within each of the 12 populations, photoperiod insensitivity appeared to segregate as a single recessive gene in the populations (data not shown). To confirm segregation for popping, a sample of 10 F₃ seeds corresponding to each selected F₂ plant was heated in hot oil and the number of unpopped, partial, and fully popped seeds were counted (data not shown). From among several hundred F₂ plants, 38 were identified that combined determinate growth habit, the ability to flower under long-days in central Wisconsin, and produced seed that popped.

The 38 F₃ families, which represent the reference population for the estimation of genetic variances, were randomly assigned with replacement among four sets of 10 lines each (five as males and five as females) and crossed using a Design II mating design (Comstock and Robinson, 1948). Successful hybridization among the 100 crosses was confirmed by planting three to 10 seeds from pods of crosspollinated flowers; the phenotype of assumed hybrid plants was compared with their corresponding male and female parents grown in adjacent rows at the Arlington ARS, Arlington, WI, in the summer of 1999. F₂ seed was harvested and bulked from three to 10 F₁ plants within each row that exhibited phenotypic characteristics of both parents.

The 25 F₃ families corresponding to each of the four sets were planted in four random blocks (sets) with two replications and evaluated over 2 years, 2002 and 2003, at the Arlington ARS, Arlington, WI. Twenty seeds were double-seeded ≈10 cm apart in 1-m rows and later thinned after the unifoliate leaves appeared to the 10 most uniformly dispersed plants in each row. Standard cultural practices were used to control weeds (Binning et al., 1995). When pods were completely dry and physiologically mature, the plants within each row were clipped off at the soil surface, placed in gunnysacks, and allowed to air dry in a greenhouse at the Arlington ARS for ≈1 month. Seed was cleaned using a mechanical thrasher followed by hand cleaning and winnowing. The seed was dried in a forced-air oven at 50 °C to ≈5% moisture (Tector Sinar™ Farmpro 6090 Moisture Analyzer; Sinar Technology, Berkshire, U.K.).

Testing and data analysis.

Fifty seeds were randomly sampled from each plot and seed weights were calculated. Seed of two plots were inadvertently popped before being weighed; thus, seed weight data for these two plots were lost. Each 50-seed sample was placed into a graduated cylinder and distilled water was added until the total volume of water and seeds equaled 25 mL. The total volume of water added was subtracted from the total volume to give the unpopped seed volume (UNPV). The seeds were drained and patted dry with paper toweling and immediately popped to minimize absorption of water into the seed. Seed was popped in a Presto Hot Air Popper model 04821 (National Presto Industries, Inc., Eau Claire, WI) for 90 s at 244 °C following recommendations for optimal popping of nuña beans (Vorwald and Nienhuis, 2008). The volume of seed after popping (PV) was obtained using a procedure similar to that used for unpopped seed. A seed was considered fully popped when the cotyledons had expanded sufficiently to shed the seedcoat. A seed was considered unpopped or partially popped if the seedcoat failed to crack or no expansion of the cotyledons was observed. The coefficient of expansion (including fully popped, partial, and unpopped seeds) was calculated as [(PV – UNPV)/UNPV] × 100 in which UNPV = unpopped volume and PV = popped volume. The popping percentage was calculated as the percentage of 50 seeds that were fully popped. For purposes of analysis of variance and covariance, the popping percentage data were transformed using arcsine⁻¹ to better approach normality; however, means are expressed as untransformed data to facilitate interpretation of the results. For analyses of variance and covariance, the data were analyzed using the standard least squares procedure of JMP statistical software, Version 7 (SAS Institute, Inc., Cary, NC). Genetic components of variance and covariance were obtained by setting observed means squares or crossproducts equal to expected values and solving for the desired component (Hallauer and Miranda, 1981). No significant deviation from homogeneity of variance between male (σ²_m) and female (σ²_f) components of variance was observed; thus, the estimate of additive genetic variance (σ²_A) for all traits was obtained by pooling mean squares (Hallauer and Miranda, 1981). Estimates of narrow sense heritability (h²) on an entry mean basis and unbiased by genotype × environment interactions and corresponding ses were calculated (Hallauer and Miranda, 1981).