MATERIALS AND METHODS

Materials: Twenty-nine advanced bean lines of different market classes and the check variety Mex-142 were used in the study. The advanced lines consisted of seven market classes that embrace two grain types; large-seeded and small-seeded. There were four large-seeded market classes: Red mottled (BCB11-130, BCB11-142, BCB11-144, BCB11-145 and BCB11-324); Red kidney (BCB11-158, BCB11-159, BCB11-162, BCB11-176 and BCB11-327); Speckled sugar (BCB11-507, BCB11-303, BCB11-386, BCB11-467, BCB11-204) and Yellow (BCB11-488) and three small-seeded market classes: Navy bean (BCB11-10, BCB11-108, BCB11-48, BCB11-62, BCB11-80 and BCB11-98); Small red (BCB11-182, BCB11-184, BCB11-202, BCB11-245 and BCB11-344); and Pinto (BCB11-512 and BCB11-515). The control variety, Mex-142 is a small-seeded white Navy bean.

Methods
Sample preparation: Bean samples were manually cleaned for any foreign materials and damaged seeds. Initial Moisture Content (MC) was determined by taking duplicate 10 g samples from each line and oven-dried at 100°C for 48 h. Then, MC (%) was calculated as follows:

Based on MC (%), 100 g bean seed solids required to fill each can were calculated (Uebersax and Hosfield, 1985) as follows:

Bean canning process: Triplicate bean samples of each line with weights equivalent to 100 g solids were accurately weighed, placed into mesh bags, soaked in water for 30 min at room temperature and then blanched for 30 min at 78°C in water containing ≈100 ppm of Ca⁺⁺ (Uebersax and Hosfield, 1985). The samples were first drained and after weighing, transferred into coded (73x110 mm) cans and covered with boiling brine containing 100 ppm Ca⁺⁺, 1.3% NaCl and 1.56% sugar. The cans were sealed with automatic can seamer (Sanitary Can Machine Co., Los Angeles, CA, USA) and cooked in an automatic retort (Barriquand Steriflow, Roanne, France) at 115.6°C for 45 min followed by instant cooling. The cans of processed beans were stored for three weeks prior to opening for evaluation of canning-quality traits (Uebersax and Hosfield, 1985).

Canning quality evaluation
Hydration Coeficient (HC): The weight gained by imbibition during soaking and blanching was used to calculate the hydration coefficient (Van Der Merwe et al., 2006) as follows:

Washed Drained Weight (WDWT) and Percentage Washed Drained Weight (PWDWT): WDWT was determined on rinsed beans drained for 2 min on a number 8 mesh (0.239 cm) screen positioned at 15° angle (Uebersax and Hosfield, 1985). PWDWT was calculated as follows (Van Der Merwe et al., 2006):

Firmness: The texture of processed beans was determined by using a Texture Analyzer TA-XT.PLUS (Stable Micro Systems, Surrey, UK) equipped with a heavy duty platform (HDP/90). A penetration force was applied to 18 individual seeds of processed beans that were randomly selected from the triplicate samples of each line. The operating condition was; pre-test speed (1.5 mm s^-1), test speed (2.0 mm s^-1), post-test speed (10 mm s^-1), trigger force (5 g) and distance (15 mm). This test was repeated five times on each line.

Visual evaluation: During draining and rinsing of the processed beans, appearance and physical traits were studied using a seven-point hedonic scale (Uebersax and Hosfield, 1985). The traits investigated were; splits (1 indicates very broken and 7 very intact), clumping (1 indicates very much clumping and 7 very little clumping), size (1 indicates very small and 7 very large), shape (1 indicates very elongated and 7 very round) and uniformity (1 indicates very variable and 7 indicate very uniform).

Statical analysis: All data was statistically analyzed using a completely randomized analysis of variance in GenStat, V.13, (VSN International, 2010). Significant differences were determined at the p<0.05 level and Duncan’s Multiple Range Test was used for mean separation where differences were obtained.

RESULTS AND DISCUSSION

Canning quality traits showed significant (p<0.05) differences in both large (Table 1) and small (Table 2) seeded genotypes. Hydration Coefficient (HC) differed significantly (p<0.05) among large and small-seeded genotypes. Among the large-seeded genotypes, BCB11-507 had the highest HC (1.63) and BCB11-145 the lowest (1.43). Mex-142, the check variety, had a higher HC than all large-seeded genotypes (1.74). The HC of the small-seeded genotypes ranged from 1.35-2.13, four lines had significantly (p<0.05) higher HC than Mex-142. In general, genotypes of navy bean market class showed higher HC than genotypes from other market classes. The hydration coefficient is an important criterion for bean processors, if HC is low, a larger quantity of beans is necessary to fill a given can volume.

Table 1:	Canning quality traits of large-seeded genotypes

Table 2:	Canning quality traits of small-seeded genotypes

Thus, a higher HC is associated with improved canning yield (Ghaderi et al., 1984). According to Balasubramanian et al. (2000), an HC of 1.8-2.0 is considered optimum by the industry and is indicative of well-soaked beans. The HC values of the small-seeded genotypes are comparable to those reported previously (Hosfield and Uebersax, 1980; Mekonnen, 2012). The generally lower HC of the large-compared to small-seeded genotypes, may be attributed to inherent differences in permeability to water associated with seed size. Del Valle et al. (1992) found that seed coats of the large-seeded red beans displayed the lowest rate of water uptake compared to small-seeded white, black and red beans. These differences have implications for the canning process, as a longer soaking or blanching period is needed for large seeded beans. Current procedures for laboratory evaluation of canning quality of beans (Uebersax and Hosfield, 1985; Balasubramanian et al., 2000; Van Loggerenberg, 2004; Van Der Merwe et al., 2006) were developed for markets dominated by the small seeded navy beans, such as in the USA and South Africa. These HC results suggest that differences in seed size should be taken into consideration during laboratory evaluation of canning quality of beans.

Differences were significant for WDWT (p<0.05) but not for PWDWT (p>0.05) among the large seeded genotypes (Table 1), BCB11-176 (294.1 g) had the highest WDWT and BCB11-303 (274.3 g) the lowest. Small-seeded genotypes differed significantly (p<0.05) in WDWT, BCB11-10 had the highest WDWT (294.4 g) and BCB11-512 the lowest (265.3 g) and in PWDWT; BCB11-10 (71.13%) had the highest and BCB11-512 the lowest (61.2%) (Table 2). Thus, in spite of their lower HC, large-seeded genotypes had higher mean WDWT and PWDWT than small seeded genotypes. This indicates that low HC led to more water uptake inside the can during the storage period of the canned beans. According to Van Loggerenberg (2004), beans seeds should achieve 80% weight increase before can-filling. This can be achieved by prolonging the soaking and blanching period until the desired HC is attained, thus avoiding excessive water uptake in the can. Based on existing canning standards, all genotypes evaluated in this study met the desired 60% PWDWT (Balasubramanian et al., 2000).

Firmness of the processed beans differed significantly (p<0.05) in both large and small-seeded genotypes. Among large seeded genotypes, the highest firmness was recorded on BCB11-130 (11.35 N) and the lowest on BCB11-386 (5.62 N). The firmness among small-seeded genotypes ranged from 14.55 N in BCB11-344 to 6.67 N in BCB11-184 with grand mean of 9.78 N. Mex-142 had firmness of 6.6 N. Differences were significant (p<0.05) among large-seeded genotypes for splits and the genotype BCB11-467 had the most broken seeds with score of 2.0 while BCB11-130 and BCB11-327 were the most intact, each with a score of 5.67. Splits also differed significantly (p<0.05) among small-seeded genotypes with the most broken seeds being recorded in BCB11-10 and BCB11-48, each with 2.33. The most intact small-seeded genotype was BCB11-245 (6.33). Mex-142 had a score of 3.33. Texture (firmness) is used as an indication of the degree of consumer acceptance of canned beans as it affects the perceived stimulus of chewing (Ghaderi et al., 1984). Consumers would prefer the bean to be soft but with fewer splits. However, soft texture is associated with seed breakdown (De Lange and Labuschagne, 2001; Van Loggerenberg, 2004). Clumping differed significantly (p<0.05) among genotypes of both grain types. Line BCB11-515 was the most clumping genotype with score of 2.67. Clumping of individual beans can be attributed to starch exudation into the canning medium (Van Loggerenberg, 2004) and indicates poor canning quality.

Significant (p<0.05) variation in size and shape was observed in both groups. Overall, seeds of most large seeded genotypes were large and elongated while small-seeded genotypes had small and rounded seeds. Beans destined for canning purposes should be uniform in size and with regular shape. Visual evaluation of bean uniformity in color and size showed significant (p<0.05) differences among genotypes. Across all genotypes, BCB11-159 (2.68) and the control variety, Mex-142 (3.67) were the least uniform. The poor uniformity of Mex-142 may indicate varietal purity deterioration since the variety has been grown since 1960s (Leakey, 1970). Furthermore, these results explain the frequent complaints by the processing industry on the purity of existing canning beans and subsequent costs of the manual cleaning of beans before the canning process.

CONCLUSION

Nearly, 50 years after Mex-142 was released as the preferred canning bean variety in eastern Africa, this study was the first attempt to evaluate new lines of dry bean genotypes for their suitability for the canning industry in Kenya. According to McCartney (1966), only soakability and hard-shell defect tests were used in the past to screen for canning quality. This study identified three small-seeded, navy bean genotypes; BCB11-80, BCB11-98 and BCB11-108 that combined good canning quality traits and which were better than the control variety, Mex-142. This study also revealed that canning quality of large-seeded bean genotypes can be improved by increasing the soaking or blanching period prior to the canning process.

ACKNOWLEDGMENTS

The authors would like to acknowledge the following persons and institutions for their invaluable support: Prof. E.G. Karuri and the staff at department of Food Science, Nutrition and Technology, the staff and management of TruFoods Ltd. And Njoro Canners Ltd.