Genetic Variability and Association Among Bulb Yield and Yield Related Traits of Garlic (Allium sativum L.) Genotypes at Bishoftu Agricultural Research Center, Ethiopia

Garlic (Allium sativum L.) is the second most important Allium species next to onion (Allium cepa L. in the world and from Africa, Ethiopia is ranked second most garlic produce next to Egypt. It is one of the most important vegetable crops in Ethiopia and is used as a component of food, cash and has also a tremendous use in the formulation of local medicines. Among abiotic factors affecting garlic production and productivity using garlic genotype which is susceptible for different biotic and abiotic agents contributed huge yield loss in agriculture center. Therefore, this eld experiment was conducted to assess the genetic variability in garlic genotypes, to determine association among traits and estimate the direct and indirect effects of traits on bulb yield. The eld evaluation of 34 genotypes and 2 released varieties (G-HL and Chefe) was conducted in 6x6 Triple Lattice Design at Bishoftu Agricultural Research Center during January 2020 to May 2020. Results of analysis of variance revealed the presence of signicant differences among genotypes for 18 traits. The variation observed among genotypes for bulb yield ranged from 43.80 to 147.20 kg ha-1. The six genotypes had mean performances higher than the highest performing check variety (G-HL=99.5Kg ha-1). Generally, the results of this study showed the presence of variations among genotypes for agro-morphology traits with a wide range of genetic distances that could allow selection and/or hybridization of genotypes after the results of this study are conrmed across locations and over the years. found higher values of heritability coupled with high genetic gain High heritability and high genetic gain as mentioned that where the high heritability value is accompanied by high genetic advance. In the present study, the highest estimates of heritability were observed in the case of plant height and the highest genetic advance showed in bulb yield plot. High heritability coupled with high genetic advance in per cent of the mean was recorded for number of cloves per bulb, bulb yield per plot and weight of clove, width of the leaves and length of clove. This indicates that were less inuenced by the environment.


Introduction
Garlic is a diploid species (2n = 2x = 16) is a widely cultivated in Ethiopia and preferred by most It is the most vital crop of the genus Allium next onion (Allium cepa L.) (Benke et al., 2020a andKhandagale et al., 2020) and supposed to have risen in Central Asia in India has with over 5000 years cultivation past (Benke et al., 2020a). The average annual world production of garlic is about 24,836,920 tons on 1,465,772 ha of land with the productivity of 16.9 t/ha (FAO, 2012).

Experimental Layout
The experimental material was laid out in a 6x6 Triple Lattice Design (TLD) with three replications. Cloves were planted in the eld in January, 2020. The plot size was 2m length and 1.5m width, consisting 4 rows with 5 plants per row, which comprised a total of 20 plants per plot; the clove was planted in a space of 30cm x 50cm between rows and plants, respectively. The spacing between plots, blocks and replications was 0.5m, 1m and 1.5m respectively. The standard cultural practices recommended in the "Package of Practices for Vegetable Crops" was followed for raising a healthy crop of garlic (Anonymous, 2009) and Agronomic practices recommended by Getachew et al., (2009) was followed to manage the plants.

Data Collection
All agronomic, bulb yield and yield related data were recorded from six randomly sampled plants in the middle four rows of each experimental unit/plot. However; phonological parameter was taken on plot basis. The collected quantitative traits were days to emergence (DE)

Statistical Analysis
Data collected for quantitative characters were subjected to analysis of variance (ANOVA) for triple lattice design (Table 2) using proc lattice and proc GLM procedures of SAS version 9.2 (SAS Institute Inc, 2008). The difference between treatment means were compared using TLD at 5% and 1% (P<0.05) and (P<0.01) probability levels, respectively. The genetic parameters, including the Genotypic and Phenotypic variance, Genotypic and Phenotypic coe cient of variance, Heritability (Broad Sense) and the expected Genetic advance (GA), Genetic advances as a percent of the mean (GAM) were calculated using the formula given by Falconer, (1981) and Johnson et al., (1955).
The following models were used in the following analysis of variance:

Analysis of Variance (ANOVA)
The analysis of variance for the 18 characters studied is given in (Table 3). Analysis of Variance (ANOVA) was carried out for the characters as per the procedure outlined by Gomez (1984). The mean squares due to accessions were signi cant (P<0.05) and highly signi cant (p<0.01) for all traits. There was a highly signi cant difference (P<0.01) among the tested garlic genotypes for days to emergence, days to maturity, plant height, leaves number per plant, neck diameter, bulb length, bulb diameter, bulb weight, clove length and number of cloves per bulb, bulb yield per plot, biological yield per plant and bulb yield per hectare which is due to genotype. Signi cant difference (P<0.05) was observed for leaves width and which due to genotype except for leaves length, clove diameter, clove weight and harvest index which showed non-signi cant genotype effects. Awel et al., (2011) reported the existence of genetic diversity within shallot produced in Ethiopia which is in related with present ndings. In addition, Abebech et al., (2021) found variability in garlic for tested characters which supports the present result.

Phenology and Growth Character
The results of mean performances, minimum, maximum values and CV and LSD for the 18 quantitative traits of the garlic accessions are summarized in Appendix Table 1. Similar results were also obtained by of Abdlkafer-Halmy et al., (2011), that signi cant variations exist within and among families of garlic for bulb and clove weight, number of cloves per bulb, ower stack height. Leaves length revealed non-signi cant variation among the genotypes. The results are in related with the ndings of (Abebech et al., 2021) (Table 3). It is indicated from (Appendix Table 1) that the leaves length different from each other among the maximum mean performance of leaves length was recorded in G-092-3/95(29.56cm) and minimum mean performance of leaves length (18.09cm) was recorded in G-119/06. The results are in agreement with the ndings of Ahmed et al., (2018). Leaves width showed a wide range of variation from 0.62cm for G-119/06 to 1.95cm for G-013/04 (Appendix Table 1). Maximum mean performance of leaves width was recorded in G-013/04(1.95cm. Minimum mean performance of leaves width was recorded in G-119/06 (0.62cm). The examination of data indicated that the mean value leaves width was (1.30cm). Leaves width revealed signi cant variation among the genotypes (Table 3). The results are in agreement with the ndings of Ahmed et al., (2018) and Guaray et al., (2018). The results are in related with the ndings of Bhat et al., (2017) who reported that the mean performance of number of leaves per plant were 6. 16, 7.13, 7.20, 7.52 and 7.75, respectively in garlic. Neck diameter revealed signi cant variation among the genotypes (Table 3). The ndings of Guaray et al., (2018) and Singh et al., (2015) are in close harmony with the results of the present study.

Yield and Yield Components
Bulb length revealed highly signi cant variation among the genotypes (Table 3). These ndings are in accordance with the work of Abebech et al., (2021), Panse et al., (2013) and Vatsyayan et al., (2013) who reported signi cant variation among the genotypes for this character in garlic. Highly signi cant variation was recorded for bulb diameter in the genotypes (Table 3). These ndings are similar to the work of Ahmed et al., (2018). The bulb weight exhibited highly signi cant variation among the genotypes (Table 3).
The analysis of variance indicated highly signi cant variation among genotypes for the clove length (Table 3). The results are in close conformity with Singh et al., (2015) and Kumar et al., (2017). Non-signi cant differences were observed among the genotypes for clove diameter (Table 3) and it ranged mean performance from 0.99cm to 1.92cm (Appendix Table 1). The ndings of Kumar et al., (2017) are in close conformity with the results of the present study.
Non-signi cant differences were observed among the genotypes for clove weight (Table 3).
Genotypes with late maturity, tall growing, large bulb length and bulb diameter produced more yield has been reported, Gaury et al., (2018) and Vatsyayan et al., (2013) reported wider range of variation in garlic germplasm for yield per plot. Highly signi cant variation among the genotypes for biological yield per plant was recorded (Table 3). Signi cant variation among the genotypes in bulb yield per hectare was recorded ( Table 3). The results are in close conformity with ndings for the character has been reported, Pervin et al., (2014) and Khar et al., (2015). Non-signi cant variation among the genotypes in the harvest index was recorded (Table 3).

Estimates of Genetic Parameter
Estimates of genotypic and phenotypic coe cient of variation A large amount of variability was noticed with respect to all the characters under study. Phenotypic variability is the total variability, which is observable and consists of both genotype and environmental variation. Such variation is measured in terms of phenotypic variation and Sable et al., 2020 and. In the present study, the phenotypic and genotypic variances were maximum for bulb yield per hectare (721.78) and (659.66) and minimum for neck diameter (0.054) and (0.030) (Table 5), respectively.
Low genotypic coe cients of variation were observed for plant height (9.77%), bulb diameter (8.57%), days to emergence (6.60%), days to maturity (2.70%) and harvest index (2.62%). Therefore, the larger proportion of phenotypic variance observed on these traits was contributed from the genotypic variance than the environment variance and hence, it can be exploited in breeding programs. For those traits for which the genetic variance is large relative to the environmental variance, accessions may be evaluated adequately by testing in few replicates, location and years (Yebirzaf et al., 2017). High phenotypic variations were high genetic variability for different traits and less in uence of the environment. Therefore, selection on the basis of phenotype can be effective for the improvement of these traits. Kassahun, (2006) reported high GCV and PCV estimates for bulb weight and bulb yield of garlic. Awel et al., (2011) also reported related results in that the phenotypic coe cient of variation (PCV) was higher than genotypic coe cient of variation (GCV).

Heritability estimates in broad sense
The heritability values for the 18 characters ranged from 9.89% for harvest index to 91.39% for bulb yield per hectare related with (Abebech et al., 2021) (Table   4). A broad-sense heritability assessment provides information on the relative amount of genetic and environmental variation in the population. As reported by Wright (1921), genetic advance values are helpful in predicting the expected progress to be achieved through the process of selection. For categorizing the magnitude of heritability, Johnson et al., (1955) suggested the following limits: Heritability (H) >60% -High 30-60% -Moderate and < 30% -Low.
Estimates of heritability and genetic advance in combination are more important for selection than heritability alone. Then the present study suggested that selection based on the most heritable traits such as biological yield per plant, number of leaves per plant, bulb length, bulb diameter, days to emergence, bulb yield per hectare, clove diameter, clove length, clove weight, number of cloves per bulb, neck diameter, plant height, leaves length, days to maturity, bulb yield per plant, leaves width and bulb weight would be effective for the development of garlic through breeding. This nding agreed with the ndings of the report by (Divya et al., 2021). There would be a close correspondence between the genotypes and phenotypes due to the relatively small contribution of the environmental effect on the phenotypes. This nding agreed with the ndings of the previous report by (Divya et al., 2021).
Estimate of Expected Genetic Advance (GA) -High heritability suggested the major role of genetic constitution in the expression of characters and such performance of characters is considered to be repeatable A high value was recorded for bulb yield per hectare (50.58). Moderate values were obtained for biological yield per plant (16.46) and the remaining is produced the lowest. The moderate genetic advance as a percent of mean was obtained for bulb length (15.75%), plant height (14.60%) and bulb diameter (11.77%). The lowest values were recorded for days to emergence (9.10%), days to maturity (3.60%) and harvest index (1.70%). Gupta et al., (2007) observed high heritability coupled with high genetic gain for clove diameter (63.11%) and neck diameter (60.00%).
The ndings of the present study are in relative with that of Abebech et al., (2021) and Meena et al., (2020) found higher values of heritability coupled with high genetic gain for bulb weight, cloves per bulb and bulb yield per plot. High heritability and high genetic gain for number leaf per plant and neck thickness.
High heritability and high genetic gain for bulb yield per plot relative with that of was reported by Pervin et al., (2014). Abebech et al., (2021) and Meena et al., (2020) as mentioned that where the high heritability value is accompanied by high genetic advance. In the present study, the highest estimates of heritability were observed in the case of plant height and the highest genetic advance showed in bulb yield plot. High heritability coupled with high genetic advance in per cent of the mean was recorded for number of cloves per bulb, bulb yield per plot and weight of clove, width of the leaves and length of clove. This indicates that these traits were less in uenced by the environment.

Phenotypic and Genotypic Correlation Coe cient
Results in Table 5 showed that bulb yield per plot had positive and highly signi cant genotypic and phenotypic correlations with plant height, leaves width, leaves number per plant, bulb length, bulb diameter, bulb weight, clove length, clove weight, number of cloves per bulb and biological yield per plant and showed positive signi cant with harvest index at phenotypic. Biological yield per plant showed positive and highly signi cant phenotypic correlation with all characters except with days to emergence, days to maturity, leaves length, harvest index and clove diameter (Table 5).
It showed positive and highly signi cant genotypic correlation with plant height, leaves width, leaves number per plant, neck diameter, bulb length, bulb diameter, bulb weight, clove length, clove weight, number of cloves per bulb and bulb yield per plot and showed negative and highly signi cant phenotypic correlation with bulb yield per hectare and showed positive and signi cant with clove diameter at phenotypic. Similar results have been reported by other researchers (Abebech et al., 2021).

Path Coe cient Analysis
Path coe cient analysis measures the direct and indirect contribution of independent variables on dependent variables and thus helps breeders in determining the yield components and understanding the cause of association between two variables (Baranwal et al., 2012). In this study, bulb yield was selected as a dependent variable and the others that had signi cant correlation of eighteen characters were selected as causal variables. The results of path analysis for direct and indirect effects of the characters studied both at genotypic and phenotypic level are illustrated in (Table 7 and 8) (Valter et al., 2019).

Phenotypic direct and indirect effects of traits on bulb yield per plot
Phenotypic path coe cient analysis showed that biological yield per plant followed by bulb weight, harvest index and leaves width exerted positive direct effect on bulb yield and bulb weight and leaves width had positive highly signi cant phenotypic correlation, except harvest index had positive signi cant phenotypic correlation (Table 6). Likewise, clove weight, neck diameter, days to emergence, bulb length and number of per plant exerted positive and small magnitude of direct effect but with positive and highly signi cant phenotypic association with bulb yield, except days to emergence showed positive signi cant phenotypic correlation. The residual effect in the present study was (0.0172) ( Table 6), showing that 98.28% of the variability in bulb yield was explained by the component factors. The remaining 1.72% is explained by other traits not considered in the study. This further clari ed that yield attributing traits chosen for the study of the garlic genotypes were good.

Genotypic direct and indirect effects of traits on bulb yield per plot
Biological yield had a high positive direct effect (0.998) on bulb yield per plot followed by leaves width, bulb length, clove length, clove weight and number of leaves per plant (Table 7). Biological yield showed had negative indirect via plant height, neck diameter, bulb diameter, bulb weight and number of cloves per bulb, while positive indirect effect via leaves width, number of leaves per plant, bulb length, clove length, clove weight. Besides its positive and highly signi cant correlation with bulb yield, leaves width showed direct positive low effect (0.13) on bulb yield per plot. The agreements with the result were reported by Abebech et al., (2021) and Valter et al., (2019).
The residual effect was 0.0184 (Table 7), indicating that all the traits included in the study explained a high percentage of the variation in bulb yield (98.16%), while other factors not included in the study can explain 1.84%. So, the yield components used were good. It is also suggested that further study should be made with more characters to nd out other traits which contribute the rest of the percentage of the yield. The residual effect (0.0184) indicated that most of the variability in bulb yield for the genotypes under the present has been explained by the independent variables included in the analysis.

Principal Component Analysis
The principal component analysis displayed that the gross variability experimental among the 36 test genotypes can be clari ed with four principal components with eigenvalues greater than unity (Table 8). The rst four components together accounted for about 79.67% of the total variation among the genotypes with respect to all the 18 traits evaluated and displayed the presence of substantial genetic diversity among the genotypes for most of the characters below consideration. Individually, PC1, PC2, PC3 and PC4 in that order accounted for about 56.25%, 10.70%, 7.14% and 5.59% of the gross variation among the 36 garlic genotypes evaluated for 18 traits. Therefore, PCA also helps breeders for genetic development of traits such as yield that have little heritability speci cally in primary generations via indirect selection for traits effective on yield (Rakesh et al ., 2018 andGolparvar et al., 2006).

Clustering of genotypes
The Euclidean distance matrix of garlic genotypes estimated from 18 quantitative traits was used to create dendrograms based on the Unweighted Pair-group methods with Arithmetic Means. Accordingly, the 36 garlic genotypes were grouped into three distinct clusters (Figure 1 and Table 9). Cluster II was the largest and consisted of nineteen genotypes (52.80%) of the total genotypes. Cluster I and III consisted of twelve (33.33%) and ve (13.89%) genotypes. This is because the cluster analysis sequestrates genotypes into clusters which exhibit high homogeneity within a cluster and high heterogeneity between clusters (Chatoo et al., 2018 andJaynes et al., 2003).

Cluster mean analysis
The mean values of three clusters for 18 quantitative characters are presented in (Table 10). Cluster I had mean values greater than overall mean values of genotypes for plant height, leaves length, leaves width, number of leaves per plant, neck diameter, bulb length, bulb diameter, bulb weight, clove length, clove diameter, clove weight, number of cloves per bulb, bulb yield per plot, biological yield per plant, bulb yield per hectare and harvest index (%). In addition, Cluster I had mean values lower than overall mean values of genotypes for days to emergence and days to maturity. Also, Cluster II had mean values equal with overall mean values of genotypes for neck diameter. Related with the results were reported by (Ranjitha et al., 2018). In addition, Cluster I and Cluster II had mean values greater than overall mean values of genotypes for plant height, leaves length, leaves width, number of leaves per plant, bulb diameter and harvest index while Cluster III had mean values greater than overall mean values of genotypes for days to emergence and days to maturity. This cluster had mean values lower than overall mean values of genotypes for the remaining traits. These two clusters (I and II) consisted of 31 (86.11%) with higher bulb yield and mean values greater than overall mean values of genotypes for most desirable traits suggested that selection of genotypes and/or further evaluation of members of these clusters is possible to develop varieties for the study area (Ranjitha et al., 2018).
The two clusters (I and II) consisted of thirty-one genotypes with a high mean bulb yield, but the members of these clusters were late maturing than the average yield maturity of genotypes. Cluster III consisted of ve genotypes had lower plant height, leaves length, leaves width, number of leaves per plant, neck diameter, bulb length, bulb diameter, bulb weight, clove length, clove diameter, clove weight, number of clove per bulb, bulb yield per plot, biological yield per plant, bulb yield per hectare, harvest index (%) but with higher mean values for days to emergence, days to maturity. This result is in related with Rakesh et al .,( 2018) and Panthee et al., (2006) that had found three major clusters in garlic using morphological traits.

Discussion
The highly signi cant differences indicate the existence of large variability among genotypes. There were less coe cients of variation in most of the characters indicating good precision of the experiment. These results indicate the presence of variability among the genotypes used for effective selection or vegetable improvement. Similar results were also obtained by Abdlkafer-Halmy et al., (2011), mentioned that signi cant variations exist within and among families of garlic for bulb and clove weight, number of cloves per bulb, owner stalk height, number of cloves per bulb, plant height and days to maturity. The analysis of variance indicated highly signi cant differences among the genotypes for days to emergence and days to maturity (Table 3). Variation in maturity has been reported by Panse et al., (2013).
The analysis of variance indicated signi cant variation among the genotypes for plant height with values ranging from 25.61 to 41.61cm (Table 3) and (Appendix Table 1), respectively. Variability in plant height was due to the inherent genetic makeup of the different genotypes. The variation in the number of leaves per plant in the genotypes was due to different genetic building of the genotypes. Enhanced number of leaves may be due to activated physiological processes by stimulating factors in the metabolism and growth of the plant. The number of leaves per plant revealed highly signi cant variation among the genotypes ( Table 3). Number of leaves resulted in more production of chlorophyll that ultimately led to more photosynthesis and hence high bulb weight . Present results were con rmed by ndings Gupta et al., (2007), Osman and Moustafa, Singh et al., (2015) who also reported that different garlic genotypes differed signi cantly for average bulb weight.
The variability in clove weight in garlic has been reported by Gaury et al., (2018). Highly signi cant differences among the genotypes were observed for this character (Table 3). A signi cant amount of variability for number of cloves per bulb was reported by Singh et al., (2015) and Kumar et al., (2017). The analysis of variance indicated highly signi cant differences among the genotypes for bulb yield per plot (Table 3). The phenotypic variance (σ 2 p) of all traits was higher than the genotypic variance (σ 2 g); similarly, the phenotypic coe cient of variation (PCV) was also higher than the genotypic coe cient of variation (GCV).
The phenotypic coe cient of variability was slightly higher than the corresponding genotypic coe cient of variability for all the traits. Thus, traits with a high phenotypic coe cient of variation indicating that all the characters studied had interacted with the environment. Genetic advance along with a heritability estimate provides a reliable estimate of the amount of genetic advance to be expected through phenotypic selection. High heritability estimates were obtained for bulb yield per hectare (91.39%), biological yield per plant (91.09%), number of leaves per plant (89.94%), clove length (89.10%), clove diameter (88.61%), leaves width (85.18%), clove weight (69.32%) and leaves length (65.91%). These characters may respond effectively for selection. Moderate heritability was observed for number of clove per bulb (56.36%), neck diameter (55.56%), plant height (52.69%), bulb length (52.63%), days to emergence (45.11%), bulb diameter (44.44%), bulb weight (44.04%), days to maturity (41.30%) and bulb yield per plot (40.01%), but low heritability was recorded for harvest index (9.89%). Related with the result were reported by Meena et al., (2020) and Abebech et al., (2021).
The correlation between yield per plot and these traits suggested that improvement of these traits could improve the physiological capacity of the crop to mobilize and translocate photosynthetic to the organs of economic value (the bulb), which in turn might have increased the bulb yield as observed in the study. In this study harvest index had a negative association with all traits insigni cantly except with plant height and leaves width which exhibited positive and signi cant correlations. Related with the results have been reported by Abebech et al., (2021) and Valter et al., (2019). From the correlation and path analysis of this experiment biological yield per plant and bulb weight could be considered as main components for selection for bulb yield in a garlic breeding (Rakesh et al ., 2018).
These results are in line with Bahadur and Sangeeta, (2016) the character plant height, number of leaves per plant and length of cloves had positive direct effect on bulb yield per plot. The residual effect in path analysis determines how best the component (independent) variables account for the variability of the dependent variable, which is bulb yield. The rst four components together accounted for about 79.67% of the total variation among the genotypes with respect to all the 18 traits evaluated and displayed the presence of substantial genetic diversity among the genotypes for most of the characters below consideration. Whereas, Cluster I and II consisted of two released varieties (HL and Chefe), respectively. This indicates that the crossing between superior genetic divergences of above diverse clusters might provide desirable recombinants for developing high bulb yielding garlic genotypes. Cluster II had mean values greater than overall mean values of genotypes for plant height, leaves length, leaves width, number of leaves per plant, harvest index, respectively, but for the remaining traits cluster II had mean values lower than overall mean values of genotypes.

Conclusion
Generally, the present study revealed the existence of signi cant genetic variability among the tested genotypes for different traits helpful for direct and indirect selection. Path coe cient analysis of genotypic revealed that biological yield per plant and leaves width contributes the high and low positive direct effect on bulb yield per plot, respectively. And other characters exerted positive direct effect on bulb yield via number of leaves per plant, bulb length, clove length and clove weight.
These positive direct effects indicate that other characters kept constant, increasing one of these characters was increased bulb yield, which implies that these characters are the major contributor for yield improvement under this study. On the other hand, characters exerted their negative indirect effect via plant height, neck diameter, bulb diameter, bulb weight and number of cloves per bulb. In this study, characters that are signi cant and positive correlated with bulb yield and had positive direct effect be under consideration. The principal component analysis showed the rst four principal components having eigenvalues greater than one accounted for 79.67% of the total variation of the 36 genotypes evaluated for 18 traits. The genotypes were grouped into three distinct clusters of which Cluster II, I and III consisted of 19(52.78%), 12(33.33%) and 5(13.89%) genotypes, respectively. Cluster I and II constructed by two released varieties (G-HL and Chefe) and Cluster III with 5 genotypes, respectively. Cluster I was distinguished by having mean values, highest mean values for all traits except for days to emergence and days to maturity and Cluster III characterized as consisting of genotypes with highest mean values for days to emergence (8.40) and days to maturity(110.53). The result suggested the presence of a considerable number of distant garlic genotypes to others that could be used in a crossing program to combine the desirable traits of the genotypes. Finally, the presence of variability among the genotypes, heritability and relationships in the established traits of the genotypes con rmed the option to increase garlic productivity in the target area. Hence, Selection and hybridization on those genotypes based on the trait with high GCV, heritability, genetic advance, and positive correlation coe cient and direct effect on bulb yield can be recommended for further yield improvement of garlic at respective locations.
Availability of data and material. All data sets supporting the conclusions of this article are available in the Electronic supplementary material and from the corresponding author Bayisa Kabe (bayisakabe1868@gmail.com). The achievement of plant breeding material used in this study obeys with institutional, national and international guidelines.    Table. 4. Estimates of Phenotypic (PCV) and Genotypic (GCV) coe cients of variation, Phenotypic (σ 2 p) and Genotypic (σ 2 g) variances, Heritability (H 2 ), Genetic advance (GA) and Genetic advance as percent of mean (GAM) for 18 traits in 36 garlic genotypes at BARC, Ethiopia in 2020.    (     G* Genotypes are numbered as shown in Table 1.  Dendrogram depicting similarity of 36 garlic genotypes (G1-G36 genotypes code as the description given in Table 1) by Unweighted Pair group Method with Arithmetic Means (UPGMA) clustering method for Euclidean distance matrix estimated from 18 phenology, growth traits, bulb yield and yield components.