Genetic Variability and Characters Association of Hot Pepper (Capsicum annuum L.) Genotypes Tested under Irrigation in Northern Ethiopia

Hot pepper production in most areas of Ethiopia especially in Tigray region is constrained by shortage of varieties, the prevalence of fungal and bacterial as well as viral diseases. Sixty-four hot pepper genotypes were evaluated to obtain the extent of genetic variability, association among characters. The experiment was laid out using 8x8 simple lattice design at Axum Agricultural Research center in 2017/18. Data were collected for 19 agronomic characters and analysis of variance revealed significant differences (p<0.01) among the genotypes for all characters. Fruit yield ranged from 0.8 to 4.5 t ha-1 with a mean of 2.7 t ha-1. The genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) ranged from 3.57and 3.84 for days to maturity to 42.4 and 42.9% for average single fruit weight. All the traits had moderate to very high broad sense heritability while genetic advance as percent of mean (GAM) ranged from 8.34 for days to maturity to 85% for average single fruit weigh. High heritability coupled with high GAM was obtained for average single fruit weight, fruit length, dry fruit yield per plant, fruit diameter and thousand seed weight reflecting the presence of additive gene action for the expression of these traits and improvement of these characters could be done through selection. Fruit yield per hectare had positive and highly significant phenotypic and genotypic correlations with dry fruit yield per plant, average single fruit weight, fruit pericarp thickness, thousand seed weight, fruit diameter and fruit length, but it had negative and highly significant genotypic and phenotypic correlations with days to maturity. Estimates of genotypic and phenotypic direct and indirect effects of various characters on fruit yield showed that dry fruit yield per plant, fruit pericarp thickness had the highest positive direct contribution to fruit yield indicating that selection based on these characters will improve fruit yield. In conclusion, the research results showed the presence of significant variations among genotypes for agro-morphology traits. Therefore, it is recommended further evaluation of genotypes or hybrids that exhibited highest yield, quality and disease resistance in subsequent breeding programs to improve the productivity of the crop.

According to CSA (2017) the national average yields of hot pepper are 6.3 t ha-1 for green pod and 1.8 t ha-1 for the dry pod, which is far below the dry pod yield (2.5-3.7 t ha-1) of improved varieties harvested at research fields of Ethiopia (MoANR, 2016) and world average yield of 3 -4 t ha-1 (FAO, 2015). The productivity of the crop is low due to many limiting factors such as shortage of adapted high yielding varieties, using unknown seed sources and poor-quality seeds, poor irrigation system, lack of information on soil fertility, the prevalence of fungal and bacterial as well as viral diseases, lack of awareness on existing improved technologies and poor marketing system.
In the past decades, diverse pepper genotypes (>300) were introduced from different regions of the world (Fekadu et al., 2008) adding to the diversity of the crop in Ethiopia. However, Ethiopia has less benefited from research activities although some research centers are working on hot pepper variety development, which mainly focused on adaptation and release of locally adapted varieties. For efficient and effective breeding work investigation and better understanding of the variability of existing genotypes is essential. Very few studies have been conducted on hot pepper genetic variability using morphological traits (Berhanu et al., 2011;Shimeles et al., 2016;Abrham et al., 2017 andBirhanu, 2017).
Effectiveness of selection depends on the amount of variability, heritability and genetic advance, interrelations among themselves and genetic divergence present in the genetic material for yield and yield related characters. Hence, developing of varieties with the desired traits has a significant contribution to increase the yield of hot pepper in the region. Therefore, the first step in the development of varieties is assessing the genetic variability of available genotypes for the characters of interest. Similarly, information on the extent and nature of interrelationship among plant characters' help formulating efficient index selection and the relative contribution of various components traits to yield (Singh, 1993). Besides, knowledge of the naturally occurring diversity in a population helps identify diverse groups of genotypes that can be useful for the breeding program. Greater the variability in a population, greater the chance for an effective selection of desirable types (Vavilov, 1951). Therefore, assessment of variability, association and heritability of traits in hot pepper genotypes in case of Central zone of Tigray agro-ecology is essential for planning an appropriate breeding strategy for genetic improvement of the crop. Hence, the present study was undertaken with the objectives to estimate phenotypic and genotypic variations, heritability and expected genetic advance of agronomically important traits in the hot pepper genotypes, to assess the extent of associations among yield and yield related traits and to identify the most yield predicting traits.

Materials and Methods
Experimental Site: The study was conducted under irrigation at Axum Agricultural Research Center (AxARC) experimental field, Mereb Lekhe District, in the central zone of Tigray, northern Ethiopia during 2017/18 cropping season. The site is located at about 1041 kms away from Addis Ababa and 67 km to the north of Aksum town, at 14o 25'26" and 14o18'48" N latitude, and 38o 42'15" and 38o48'30" E longitude with an altitude of 1390 m.a.s.l. (Figure 1). The site is found in semi-arid tropical belt of Ethiopia with "kola" agro climatic zone and the rainy season is mono -modal concentrated in one season from late July to early September and receives from 400 -600 mm of rain fall per annum. The mean minimum and maximum temperatures ranged from 13.3 0C to 33.7 0C, respectively. The soil texture of the specific site of the study area is sandy clay loam textural class with bulk density of 1.7 g cm-3, very low in organic carbon (0.7%) with an alkaline pH of 8.2.  cropping season under irrigation. Seeds of each hot pepper genotypes were sown in seed bed of 0.6 m2 (3 rows, 0.2 m spacing between rows, 1m row length) during October 2017 to raise seedlings. Seedlings were transplanted to main field 48 days after seed sowing i.e. when the seedlings attained 15 cm height. Each genotype was planted in the main field in a plot size of 8.4m2 (2.8 m x 3 m). Each plot consisted of four rows of 3m length with inter and intra-row spacing of 0.7m and 0.3m, respectively, containing a total of 40 plants. Each incomplete block and replication was spaced 1 and 1.5 meters, respectively. The middle two rows were used for data collection leaving the two rows as borders. Fertilizer, Di-ammonium phosphate (DAP) as a source of Phosphorus was applied at the rate of 200 kg ha-1 during planting and nitrogen fertilizer was applied in the form of Urea at the rate of 150 kg ha-1 in splits, half during transplanting and the rest as side dressing at 45 days after transplanting. Watering was made following furrow irrigation at 7days interval (AxARC, 2016) was used. Weeding, hoeing and other field management and crop protection activities were done as required. Melkassa Oromiya SNNPRS = Southern Nation, Nationalities and People's Regional State, B/Gumz = Benishangul-Gumz Regional State, Acc = accession Data collected: Seventeen quantitative characters were recorded on five randomly selected plants from the two middle rows of each plot by adopting descriptors list for hot pepper (IPGRI, 1995).
Data Analysis: Data for quantitative characters were subjected to analysis of variances (ANOVA) for simple lattice design following procedures of SAS Version 9.2(SAS Institute Inc., 2010) to test the presence of significant differences among genotypes. Mean separations were estimated using Duncan Multiple Range Test (DMRT) at 5% probability levels.
Genetic Advance (GA) for selection intensity (K) at 5%was computed according to Allard (1960) as given here: GA = K*σp*H 2 Where, GA = expected genetic advance, K = the standardized selection differential at 5% selection intensity (K = 2.063), σp = is phenotypic standard deviation on mean basis and H 2 = heritability in the broad sense. The genetic advance as percentage of population means (GAM) was also estimated with the methods described by Johnson et al. (1955). Genetic advance as % of mean (GAM) was computed as: GAM = ̅ * 100 Where, GA = Genetic advance under selection and ̅ = mean of the population. According to Johson et al. (1955) genetic advance as percent of mean was classified as low (<10%), moderate (10-20%) and high (>20%).
Characters associations at genotypic and phenotypic levels were calculated from the genotypic and phenotypic and environmental covariance according to Singh and Chaundhary (1985). In Path analysis, total yield per hectare was taken as the resultant (dependent) variable while the rest of the characters were considered as casual (independent) variables. The direct and indirect effects of the independent characters on fruit yield per hectare were estimated by the simultaneous solution of the formula suggested by Dewey and Lu (1959).

Analysis of Variance (ANOVA)
There were highly significant differences (P<0.01) among the tested genotypes for all characters studied indicating presence of adequate variability among genotypes (Table 2). This significant genetic variation among genotypes suggested that the genotypes were genetically diverse and it could be a good opportunity for breeders to select genotypes for trait of interest for variety development. This finding was in agreement with the findings of Berhanu et al. (2011), Nsabiyera et al. (2013), Birhanu (2017) and Shimeles (2018).
Mean Performance of Genotypes: Genotypes had 57.5 to 76.5 days to flowering, 67.5 to 84.5 days to fruiting and 113.5 to133 days to maturity with a mean of 62.27, 75.78 and 122.6 days, respectively. The result showed a wide range of variations for days to flowering, fruiting and maturity. Similarly, Shimeles (2018) and Berhanu et al. (2011) reported the existence of wide genetic variation for those phenological characters on 49 and 20 hot pepper genotypes respectively. Acc-1 (113.5 days), Acc-49 (114.5 days) and Acc-57 (114 days) had significantly shorter days to maturity while Acc-5 (133 days), Acc-9 (130.5days) and Acc-59 (130.5days) had significantly delayed maturity. About 54.7% of the genotypes exhibited shorter number of days to maturity than the genotypes mean (122.6). Moreover, 25 genotypes were significantly earlier in maturity than the check variety (Mareko fana) that had the earliest days to maturity (Appendix Table 1). Most of the genotypes have also yield advantage over the early maturing check variety. Hence, there is an opportunity to select early maturing and high yielder genotypes better than the check variety (Mareko fana).
The minimum and maximum canopy diameter was exhibited by genotypes Acc-229697 and Acc-59, respectively. Genotypes Acc-59, Acc-15 and Acc-9101 have also yield advantage and can be used as parents in developing varieties with high canopy diameter over the check variety. However, Shimeles (2018) reported a wide range of 40.9 -76.6 cm for canopy diameter. This wide range of variability may be attributed to differences in the materials test and /or may be due to the differences of in the testing environments.

Agricultural Science
Vol. 2, No. 1; 2020 51 had thicker pericarp than the check variety. The highest fruit pericarp thickness was Acc-212912 (2.7 mm) which indicated that Acc-212912 should be given consideration for selection designed for the improvement of this trait. This is in agreement with the finding of Nsabiyera et al. (2013) and Shimeles (2018) who reported a wide range of variation for fruit pedicel length, fruit length and fruit diameter. Average single fruit weight varied from 1.55 to 7.1 gm with a mean of 3.6 g. Acc-49, Acc-212912, Acc-61, Acc-1, Acc-11, Acc-212913 and Acc-229694 depicted highest fruit weight per plant comparing to the check variety in that order.
The genotypes exhibited significant variability in fruit number per plant which ranged from 14.75 to 55.5 with a mean of 31. The lowest number of fruits per plant was depicted by genotypes Acc-57, Acc-56, Acc-229698 and Acc-49, whereas the highest number of fruits per plant by genotypes Acc-5, Acc-2, Acc-59, Acc-229700, Acc-212913, Acc-8 and Acc-229697. On the other hand, number of seeds per fruit ranged between 93 and 231 with a mean value of 139.1. The lowest number of seeds per fruit was counted for Acc-229697 and Acc-14, whereas the highest seeds per fruit for Acc-212912, Acc-3, Acc-48 and Acc-4 respectively. Accordingly, in the current study 42 genotypes scored greater than 65% number of fruits per plant and 33 genotypes scored greater than 51% number of seeds per fruit as compared with the best performing check variety (Mareko fana) ( Table 5). Similar results for number fruits per plant were reported by kadwey et al. (2015), Birhanu (2017), Kumari (2017) and Shimeles (2018).
A wide range of variation was observed for 1000 seed weight among genotypes which ranged from 4 to 7.2g with a mean of 5.6g. Genotypes Acc-212913, Acc-58, Acc-229694, Acc-212912 and Acc-11 had high seed weight of 7.2, 6.9, 6.9, 6.8 and 6.6 g while Acc-9085, Acc-9101, Acc-9099 and Acc-9086 had low 1000 seed weight of 4g each as compared to the check variety.
Marketable fruit yield per hectare ranged from 0.7-4.3 t with an overall mean of 2.5 t ha -1 . The highest marketable fruit yield per hectare was recorded for Acc-4, Acc-212913, Acc-212912, Acc-1, Acc-49, and Acc-50 while Acc-9102, Acc-52, Acc-9099, Acc-9098, Acc-230798 and Acc-230799 gave the lowest yield as compared to the check variety (Appendix Table 2). Hence, there is an opportunity to select high yielding genotypes better than the check variety. Total fruit yield per hectare ranged from 0.83 to 4.6 t ha -1 which showed wide variation with a mean value of 2.7 t ha-1. The maximum yield was obtained from Acc-4 (4.6 t ha -1 ) followed by Acc-212912 (4.5 t ha -1 ), Acc-212913 (4.3 t ha -1 ) and Acc-3 (4.3 t ha -1 ) (Appendix Table 2). Nearly, 43.8 % of the tested genotypes had fruit yields above the grand mean of genotypes. As compared with the best performing check variety (Mareko fana), 54.6% of the genotypes had yield advantages. Most of these high yielding genotypes were also earlier in maturity than the check variety (Appendix Table 2). This wide range of variability of genotypes for most traits in the study indicated the high possibility for genetic improvement of traits under consideration.

Phenotypic and Genotypic Variations:
For all characters studied, the magnitude of environmental variance was lower than the corresponding genotypic variance (Table 3). This indicates that the genotypic component of variation was the major contributor to the total variation in the studied characters. Genetic variance ranged from 0.05 for fruit pericarp thickness to 860.42 for dry fruit yield per plant while phenotypic variance values ranged from 0.07 to 870.76 for fruit pericarp thickness. The GCV ranged from 3.6% for days to maturity to 42% for average fruit weight. Similarly, PCV ranges from 3.8% for days to maturity to 42.9% for average fruit weight per plant. In general, the phenotypic coefficient of variation (PCV) was relatively higher than the corresponding genotypic coefficient of variation (GCV). The difference between PCV and GCV was narrow indicating little influence of environment on the expression of these characters and considerable amount of variation was observed for all the characters. The GCV and PCV values are normally categorized as low (<10%), moderate (10-20%) and high (>20%) as indicated by Deshmukh et al. (1986). High values of PCV and GCV indicated the existence of substantial variability for such characters and selection may be effective based on these characters. .4%) and thousand seed weight (14.5 and 15.6%). GCV and PCV were low for days to 50% flowering (7.3 and 7.8), days to 50% fruiting (6.9 and 7.3%) and days to maturity (5.2 and 5.4%), plant height (8.9 and 10.5%) and canopy diameter (6.3 and 7.2%). Similarly, Pujar et al. (2017) reported a relatively low GCV and PCV for days to flowering in 63 chilli genotypes. The high PCV and GCV are evident for the high variability that in turn offers good scope for selection. In addition, similar findings were also reported by (Sharma et al., 2010;Janaki et al., 2015;Rosmaina et al., 2016;Sahu et al., 2016).

Estimates of Heritability (h 2 ) in broad Sense:
In this study all the traits had moderate high to very high broad sense heritability in the range of 71.4 to 98.8% (Table 3) indicating that the traits studied were more influenced by genetic factors (Rosmaina et al., 2016). According to Singh (2001) heritability values greater than 80% considered as a very high, values from 60-79% as moderately high, values from 40-59% as medium and values less than 40% as low. Accordingly, the estimates of heritability of all traits in the current study were moderate to very high. The characters having very high heritability indicated relatively small contribution of the environmental factors to the phenotype and selection for such characters could be fairly easy due to high additive effect. Heritability alone provides no indication of the amount of genetic improvement that would result from selection of individual genotype. Hence, knowledge on heritability coupled with genetic advance is more useful. Genetic advance as percent of the mean (GAM) in this study ranged from 6.9% to 85% for days to maturity and average single fruit weight respectively (Table 6). According to Jonhson et al. (1955) the value of GAM is categorized as low (< 10%), moderate (10-20%) and high (> 20%). The highest GAM was recorded for average single fruit weight (85%), followed by fruit length (82.4%), number of fruits per plant (79.5%) and number of primary branches per plant (69.5%) indicating that these characters are governed by additive genes and selection will be rewarding for improvement of hot pepper for these traits. The least GAM was recorded for days to maturity (6.9 %), days to 50% fruiting (8.4%) and days to 50% flowering (8.7%). In the current result, moderately high heritability coupled with moderate GAM was observed for plant height (15.6%), canopy diameter (11.3%), stem diameter (15.1%) and fruit pericarp thickness (19.9%). These results agreed with the findings of earlier researchers (Janaki et al., 2015;Rosmaina et al., 2016;Birhanu, 2017;Kumari, 2017) who found high genetic advance as percent of mean for number of fruits per plant, average fruit weight, and number of primary branches per plant. Shimeles et al. (2016) also obtained high genetic advance as percent of mean for number of branches per plant. Similar findings were reported by earlier workers for some characters with moderate to high GCV, PCV, heritability and GAM estimates, for fruit yield per plant, fruit diameter, fruit length, average fruit weight, number of seeds per fruit and number of fruits per plant (Sharma et al., 2010;Sahu et al., 2016;Razzaq et al., 2016;Pujar et al., 2017).
Generally, for characters like dry fruits yield per plant, number of fruits per plant, number of seeds per fruit, average single fruit weight, fruit diameter, fruit length and thousand seed weight with high GCV, heritability and GAM should be considered as reliable selection criteria for crop improvement in terms of yield and its components in hot pepper.

Association of Characters:
In most cases, the genotypic correlation coefficients were higher than the phenotypic correlation coefficient which indicates that the inherent association among various characters independent of environmental influence (Table 4). Total fruit yield per hectare showed positive and highly significant (P<0.01) genotypic and phenotypic correlations with dry fruit yield per plant (r g = 0.75 and r p = 0.70), average single fruits weight (r g = 0.52 and r p = 0.49), thousand seed weight (r g = 0.47 and r p = 0.41), fruit pericarp thickness (r g = 0.44 and r p = 0.42), number of seeds per fruit (r g = 0.42 and r p = 0.39) and fruit diameter (r g = 0.37 and r p = 0.33). Total fruit yield also exhibited positive and significant (P<0.05) genotypic and phenotypic correlations with canopy diameter, stem diameter and fruit length ( Table 4). The results imply that improvement of the characters could improve capacity to synthesize and translocate photosynthesis to the organ of economic value.
This suggested that, improvement of those characters would result in a substantial increment on fruit yield that could be used in selection of genotypes for high fruit yield. Similarly, Abrham et al. (2017) and Shimeles (2018) reported higher genotypic correlation coefficients than the phenotypic ones, implying the inherent associations between various characters in Ethiopian Capsicums.
Dry fruit yield per plant had a highly significant association at both genotypic and phenotypic level with average single fruit weight (r g = 0.50, r p = 0.56), number of seeds per fruit (r g = 0.46, r p = 0.45) and thousand seed weight (r g = 0.43, r p = 0.40). These results agreed with the findings of earlier researchers (Kadwey et al., 2015;Chakrabarty and Aminul , 2017;Kumari et.al, 2017) indicating genotypic and phenotypic correlations between plant height, number of primary branches per plant, fruit length, fruit diameter, fruit pericarp thickness, fruit yield per plant, average fruit weight, number of fruit per plant, number of seeds per fruit and thousand seed weight. DFL = days to 50% flowering, DFR = days to 50% fruiting, DM = days to maturity, PH = plant height, CD = canopy diameter , NPB = number of primary branches per plant, SD = stem diameter, FPL= fruit pedicel length , FL = fruit length , FD = fruit diameter, FPT = fruit pericarp thickness, DFYP = dry red fruit yield per plant, FW = average fruit weight, NFP = number of fruits per plant, NSF = number of seeds per fruit, TSW = thousand seed weight and TFY = total fruit yield, SEM = standard error of the mean, σ 2 g = genotypic variance, σ 2 e = error variance, σ 2 p = phenotypic variance, PCV = phenotypic coefficient of variance, GCV = genotypic coefficient of variance, h 2 = broad sense heritability, GA = genetic advance, GAM = genetic advance as percent of mean.
Genotypic path coefficient analysis: In this study, the result of genotypic path coefficient analysis showed that dry red fruit yield per plant (0.46) had the highest positive direct effect on total fruit yield per hectare followed by fruit pericarp thickness (0.37), number of primary branches per plant (0.31), canopy diameter (0.2), thousand seed weight (0.18), average single fruit weight (0.16), fruit pedicel length and stem diameter (0.11), while negative direct effect was observed for days to maturity (-0.21), fruit length (-0.2), fruit diameter (-0.15) and days to 50% flowering (-0.08) while, days to 50% fruiting, plant height and number of seeds per fruit had very little positive direct effect on fruit yield per hectare though it exhibited significant and positive association with fruit yield (Table  5).This indicates the true relationship between these characters as a good contributor to fruit yield.
Similarly, Shimeles (2018) reported that direct influence of pericarp thickness on fruit yield was very high and positive and its indirect influence through fruit diameter was also positive. However, pericarp thickness showed high negative indirect effect on number of fruits per plant. Generally, based on the genotypic path analysis of agronomic characters which showed positive direct effects on fruit yield per hectare were: dry fruit yield per plant, fruit pericarp thickness, average single fruit weight, number of primary branches per plant, canopy diameter and number of seeds per fruit. This result agrees with that of Kumari (2017).
In conclusion, the present study confirmed the existence of enormous genetic variability among the hot pepper germplasm for various important morphological traits. Hence there is an opportunity to exploit these traits in order to develop genotypes that perform better than the existing varieties for the future pepper improvement program.    Residual effect = 0.50 *and ** = significant at 5% and 1% probability levels, respectively. DFL = days to 50% flowering, DFL = days to 50%fruiting, DM = days to maturity, PH = plant height, CD = canopy diameter, NPB = number of primary branches per plant, SD = stem diameter, FPL = fruit pedicel length, FL = fruit length, FD = fruit diameter, FPT = fruit pericarp thickness, DFYP = dry fruit yield per plant, FW = average single fruit weight, NFP = number of fruit per plant, NSF = number of seeds per fruit, TSW = thousand seed weight and rg = genotypic coefficient of correlation.

Conflict of Interests
The authors have not declared any conflict of interests.