Yield selection within Coffea arabica cv. Ruiru 11.

Aims: This study was aimed at identifying high yielding Ruiru 11 sibs in varying growing conditions. The study also intended to measure the extent to which cherry yields of Ruiru 11 are affected by the environment. Study Design: Randomized Complete Block Design with three replications. Place and Duration of Study: The study was conducted in three different agro-ecological zones in Kenya namely Mariene in Meru County, Kisii near Kisii town in Kisii county and Koru in Kericho County between November 2008 and September 2011. Methodology: Thirty four (34) Ruiru 11 sibs, all of which are resistant to Coffee Berry Disease and Coffee Leaf Rust, were evaluated in this study alongside two entries of SL28, a cultivar susceptible to the two diseases. One entry of SL28 was sprayed with copper fungicides to control, while the other SL28 entry was not sprayed with any fungicides. Planted at a spacing of 2m by 2m, each entry had 12 trees per plot per rep, giving a total of 1296 plants per experiment per site. Cherry yield recording was done during the peak harvesting period of May to July at Mariene and July to September at Koru and Kisii. The data was subjected to Analysis of Variance (ANOVA) using XLSTAT version 2012 statistical software and effects declared significant at 5% level. Results: Significant (P = .05) yield differences among Ruiru 11 sibs were obtained in all Research Article American Journal of Experimental Agriculture, 3(1): 76-88, 2013 77 years of evaluation at Koru but only in 2011 at Kisii and Mariene. There was a greater discrimination between sibs at Koru, followed by Kisii and then Mariene. Year effect was highly significant (P < .001) and equally distinguished in all sites but year x sib interactions were significant (P = .05) only at Kisii. Combined analysis for all environmental combinations showed highly significant (P < .001) differences between sibs, environments and their interaction. Environments made a greater contribution (42.6%) to the variation compared to sibs (7%). The interaction term also made a significant contribution (18.7%). The best sibs per site and those adapted to contrasting environments were identified. Conclusion: The expression of high yield variation among Ruiru 11 sibs is a sign of high potential of intra-selection within the cultivar for yield improvement. Identified sibs can be recommended to farmers and also exploited in future breeding programmes for improvement of Ruiru 11 productivity and agronomic adaptability. The occurrence of significant sib by environment (G x E) interactions was an indication that the best improvement strategy should be a multi-site selection.


INTRODUCTION
Behind oil, coffee is the second most traded commodity in the world. Its cultivation is mainly by smallholder farmers who hardly break even mainly due to low yields, high production cost and low world market prices. Increasing productivity, while reducing the cost of production is a main breeding objective of most producing countries [1]. New arabica cultivars with higher yield potential and resistance to Coffee Leaf Rust (CLR) and/or Coffee Berry Disease (CBD) have started to replace traditional varieties on a large scale in several countries [2]. The cultivar Ruiru 11 is a composite of about 60 F1 hybrid sibs each derived from a cross between a specific female and male population [3]. The cultivar was developed at the Coffee Research Station, Ruiru, Kenya, and released to growers in 1985. It combines resistance to CBD and CLR with high yield, fine quality and compact growth amenable to high density planting [3].
The economic value of Arabica coffee Coffea arabica L. is determined both by the yield potential and the bean quality [4]. Yields of 5 tons ha −1 and higher have been obtained in some close-spaced and unshaded Arabica coffee blocks e.g. in Brazil, Colombia and Kenya [5]. However, most smallholder Arabica coffee farms with no access to external inputs often produce less than 300 kg ha −1 year −1 green coffee beans, while intensively managed plantations at conventional spacing may yield an average of 2 tons ha −1 annually [6]. Data from field trials at Coffee Research Foundation, in Kenya shows that Ruiru 11 cultivar planted at a density 3300 trees/ha produces between 2.5 and 3.0 tons ha −1 year −1 [6]. Depending on conditions, coffee yields fluctuate from year to year and from location to location [7;8].
Success of a new variety depends to a great extent on its adaptability to a wide range of climatic and soil conditions [8]. Coffee can be cultivated on a wide range of soil types, provided these are at least 2 m deep, free-draining loams with a good water-holding capacity and a pH of 5-6, fertile and contain at least 2% organic matter. High-quality, acidic Arabica coffees are mostly produced on soils of volcanic origin [6]. Arabica coffee is grown in altitude ranges between 1400 and 1800 m above sea level [9]. The optimum mean annual temperature range for Arabica coffee is 18-21ºC [10]. Rainfall requirements depend on the retention properties of the soil, atmospheric humidity and cloud cover, as well as cultivation practices. The optimum annual rainfall range is 1200-1800 mm for Arabica coffee [10] with a maximum of 2500 mm [6]. Coffee plants grow and yield better if exposed to alternate cycles of wet and dry seasons [6]. Abundant rainfall throughout the year often results in scattered harvest and low yields [10]. The distribution of sunshine also has a strong influence on flowering, bean expansion and ripening. Shade decreases coffee tree productivity by about 20%, but reduces the alternate bearing pattern [11].
Knowledge of the effects of environment and genotype by environment (GxE) interaction is important to breeders in making decisions regarding the development, evaluation and release of new cultivars [7,8]. Identifying high yielding coffee genotypes is often time consuming and difficult to achieve due to the perennial nature of the crop, biennial bearing, and the large environmental component of variance for yield [1]. This study aimed at identifying high yielding Ruiru 11 sibs in varying growing conditions. The study also intended to measure the extent to which cherry yields of Ruiru 11 are affected by the environment.

Description of Study Sites
The study was conducted in three different agro-ecological zones in Kenya namely Mariene in Meru County, Kisii near Kisii town in Kisii county and Koru in Kericho County. Mariene is located at 0 0 N, 37º 35'E, at an elevation of 1524 m above sea level. The soils are andohumic acrisols, friable clays, strongly acidic, very low in bases and moderate in organic matter. Koru is located at 0º 07'S, 35º 16'E and has an elevation of 1554 m above sea level. The soils are eutric nitosols, friable clays, and weakly acidic to neutral, rich in bases, available phosphorous and moderate inorganic matter. Kisii is located at 0º 41'S, 34º47'E at 1700 m above sea level. The soils are molic nitosols, friable clays with acidic pH, low to moderate bases and are high in organic matter. The experimental plots in Koru and Kisii were established in April 1990 while Meru plot was established in April 1991. All the plots have undergone change of cycle twice and were therefore almost of the same status. Other agronomic practices were carried out as recommended. All the sites were laid out in a Randomized Complete Block Design (RCBD) with three replications.

Plant Materials and Field Layout
Thirty four (34) Ruiru 11 sibs (Table 1) were evaluated in this study alongside two entries of SL28 used as checks. One entry of SL28 was sprayed with copper fungicides to control CBD and CLR, while the other SL28 entry was not sprayed with any fungicides. All the sites were laid out in a Randomized Complete Block Design (RCBD) with three replications. Planted at a spacing of 2m by 2m, each entry had 12 trees per plot per rep, giving a total of 1296 plants per experiment per site. Cherry yield recording was done during the peak harvesting period of May to July at Mariene and July to September at Koru and Kisii. Rainfall was recorded in all the three sites for the three production seasons (years) at various berry development stages (Table 2).

Data Analysis
The data was subjected to Analysis of Variance (ANOVA) using XLSTAT version 2012 statistical software and effects declared significant at 5% level. Separate as well as combined analysis of variance was performed on data from all locations over the three production years. Least Significance Difference (LSD) was used to separate the means.

RESULTS
Cherry yield data was obtained from two locations (Koru and Mariene) over three years and two years at Kisii making a total of 8 environmental combinations. The Kisii site was omitted in 2009 as it recorded very low yields as the trees were recovering from hailstorm damage. Analysis of variance (ANOVA) obtained significant (P = .05) yield differences among Ruiru 11 sibs in all the years at Koru but only in 2011 at Kisii and Mariene. This was an indication of some genetic variation between the sibs which are considered to be closely related. Examination of the F values at each location showed that there was a greater discrimination between sibs at Koru, followed by Kisii and then Mariene. The year effect was highly significant (P < .001) and equally distinguished in all sites but year x sib interactions were significant (P = .05) only at Kisii (Table 3).  (Table 4). Multi-site analysis for the 8 environmental combinations recorded significant differences between sibs, environments and their interaction (Table 5). Further scrutiny of their contribution to total sum of squares indicated that environments made a greater contribution (42.6%) to the variation compared to sibs (7%). The interaction term also made a significant contribution (18.7%).   (Table 6). Significant yield differences were observed among the sibs in all the locations. Evaluated sibs were found to produce average yields between 3 -16 kgs ( Table 7). The high yielding but susceptible SL28 cultivar was used as a check. In all the three sites, SL28 sprayed with fungicide recorded slightly higher yields than the unsprayed SL28 in absolute terms but statistically similar. Therefore, spraying SL28 against fungal diseases had no significant effect on yield. At Kisii site, the yields of SL28 (both sprayed and unsprayed) were highly comparable to those of most Ruiru 11 sibs. The yields of sprayed SL28 were not significantly different from those of the first 30 Ruiru 11 sibs (except 143, 107, 106 and 112) while the yields of unsprayed SL28 were statistically similar to those of all Ruiru 11 sibs except R11-112. At Koru, all Ruiru 11 sibs produced better yield than SL28 in absolute terms with 17 sibs recording significantly (P = .05) higher yields than SL28. At Mariene, 8 Ruiru 11 sibs recorded significantly (P = .05) higher yields than SL28 ( Table 7).
The best performing sibs per location are shown in Table 8. The most suited sibs for Kisii site which recorded high yields in both seasons were found to be R11-131, R11-52, R11-7, R11-117, R11-6, R11-142, R11-1 and R11-41. The Koru site was found to be favourable for most of the sibs but best performing were R11-107, R11-91, R11-80, R11-117, R11-142, R11-52, R11-137, R11-11, R11-100 and R11-135. The above mentioned sibs for both Kisii and Koru sites consistently recorded high yields in varying environmental conditions. For Mariene, the best performing sibs were R11-1, R11-135, R11-11 and R11-52. The four were the only ones that yielded consistently better under all conditions and were regarded to be highly stable in terms of yields. The sibs were best discriminated at Mariene and the site was considered the best for yield selection followed by Kisii.

DISCUSSION
Although some studies have been carried out to assess variation of Ruiru 11 sibs in quality [4;12;13;14] and disease resistance [3], there is little information about their variation in yield. Ruiru 11 sibs evaluated were found to differ significantly in yields. This was an indication of high genetic variation between Ruiru 11 sibs. Similar results were obtained by Wamatu et al. [8] when evaluating related coffee clones some of which have been utilized as Ruiru 11 male parents. In Brazil, Carvalho et al. [15] observed large variability in cherry yields among F1 generation plants obtained by crossing selected coffee trees and among bourbon coffee progenies that have been harvested for 12 to 15 consecutive years. When assessing cup quality of Ruiru 11, Ojijo [12], Agwanda et al. [4], Omondi [13], Kathurima et al. [14], also reported significant variability within the cultivar.
The observed site differences indicated that the environment has a strong effect on the expression of yield potential. We attribute differences in yield to the particular edaphic and climatic conditions of each site. Wamatu et al. [8] and Anim-Kwapong and Adomako [1] also reported large environmental component of variance for yield in coffee. The three sites partly fulfilled the conditions of good selection and testing environment which include high genetic variances, high mean performance and high heritability [4]. On the basis of average performance, Koru could be the best selection site followed by Kisii as they consistently recorded the highest means which better portrayed the potential of the sibs. There was also greater discrimination between sibs at Koru, followed by Kisii and then Mariene. However, Mariene could also be a good selection site to discriminate the sibs under less favourable conditions while Koru could be the best selection site based on high genetic variances.
In our study, rainfall was taken as the first most important limiting factor and thus used to explain the observed site differences. A similar approach was also applied by Agwanda et al. Genotype by Environment (G x E) interactions is a measure of stability and adaptability of genotypes in varying environments. In this study, significant G x E interactions was observed indicating that different Ruiru 11 sibs responded differently to different environments. When evaluating related coffee clones some of which have been utilized as Ruiru 11 male parents, Wamatu et al. [8] also observed significant G x E interactions. Apart from yields, significant G x E interactions has also been reported on other desirable traits in Ruiru 11 and other types of Arabica coffee. For example, on coffee quality of Ruiru 11, Agwanda et al. [4], Omondi [13] and Kathurima et al. [14] reported G x E interactions of significant magnitude. Mawardi and Hulip [16] and Agwanda et al. [4] observed highly significant G x E interactions in bean characteristics of Arabica coffee. High G x E interactions for desirable traits have been reported as a major setback in achieving faster progress in selection [4]. These significant interactions might be to a large extent attributable to the low precision in balancing the growing conditions in the multi-site trials and may also be partly explained by trial characteristics.

CONCLUSION
The study demonstrated the existence of a high yield variation among Ruiru 11 sibs. There is therefore high potential of intra-selection within the cultivar for yield improvement. The most widely adapted Ruiru 11 sibs as well as the best sibs for the studied coffee growing areas on the basis of cherry yield were identified. These will be recommended to farmers and also be exploited in future breeding programmes for improvement of Ruiru 11 yield agronomic adaptability. The growing environment was found to have a strong effect on the expression of yield potential as portrayed by high site variations. The occurrence of significant G x E interactions was an indication that the best improvement strategy should be a multi-site selection. Future studies should therefore include many locations with more variable climatic conditions ranging from marginal to suitable coffee growing areas. Rainfall intensity and distribution especially during the early stages of berry development was also found to be critical as the highest yields were where rainfall was adequate and well distributed.