The work was carried out at Leboudi II (OKOLA district, Lékié Department, Centre Region of Cameroon). This locality belongs to agroecological zone V or forest zone with bimodal rainfall. Rainfall is generally in the order of 1500 to 2000 mm/year with two distinct wet seasons. Over the year, the average temperature is 23°C for an altitude of 797 m [9] .

The experimental design consisted of four completely randomized blocks with three replications. The varieties studied were: Desiree, Banso, Panamera, Synergie, Maffo, Doza, and Cypira. Any variety was separated from the others by 0.5m aisles. Each block measured approximately 5m x 3m and consisted of seven (07) ridges each containing six (06) seedlings, separated from the others by a distance of twenty (20) cm.

Furthermore, the choice of the different potato collections was based on the varietal preferences of the growers and consumers, their availability and accessibility. It should be noted that each of these collections was all taken from growers in the of Bafou and Bamboutos localities of the Western Highlands of Cameroon.

Forty-eight hours (48) after the preparation of the experimental site, the potato tubers of the pre-germinated collections were planted. The maintenance consisted of weeding, followed by hoeing on the different ridges and around the experimental site. Watering was provided by rainfall in the locality.

Genetic traits of the seven genotypes of potatoes were evaluated based on morphological and yield parameters. Morphological parameters include plant height (PH), collar diameter (DC), emergence rate (TL), and viability rate (TV) evaluated 41 days after sowing. For the yield parameters, a total number of tubers (TNT), average tuber weight (AWT), average tuber volume (AVT), average tuber length (ALT), and number of tuber per plant (NTP) were determined.

The following parameters have been calculated as genetic parameters: grand mean, standard error of mean (SEM), critical difference (CD) 5%, critical difference (CD) 1%, environmental variance (VE), genotypic variance (VG), phenotypic variance (VP), environmental coefficient of variance (ECV), genotypic coefficient of variance (GCV), phenotypic coefficient of variance (PCV), broad sense heritability (H²), genetic advance (GA) and genetic advance as a percentage of mean (GAM) [10] [11] .

The data obtained were subjected to ANOVA with the software R (version 3.5.1) for the multiple comparisons of the means and their ranking, the Least Significance Test at the 5% threshold.

Principal component analysis and hierarchical cluster analysis (HCA) were performed by the variability package of the software R (version 3.5.1). The classification of the collections has been done by ascendant classification. The biplot has been constructed to bring together varieties and parameters as in [12] .

To determine the nature and degree of divergence between our collections, a PCA and hierarchical clustering (HC) were performed. Only the first two axes of

TL: Emergence rate, TV: Viability rate.

the principal component analysis (Figure 8) expressing nearly 72% were taken into account. Axis 1, which contains 39.2%, is associated with the characters TL, Yield, AWT, AVT, and ALT.

Axis 2 which counts 32.8% associated with the variables such as TNT, NFP, DC and P_High. The combination between variables and collections showed that Desiree can be selected based on parameters such as TNT, NTP, DC, and PH while Doza can be selected based on TL, Yield, AWT, AVT, ALT (Figure 8).

The variability components of potatoes namely genotypic and phenotypic variance and coefficient of variations, heritability in a broad sense, and genetic advance as a percent of mean were estimated for seven traits (Table 2).

Phenotypic variance ranged from 0.015 to 141.54 respectively for AWT and TNT, while genotypic variance ranged from 0.0091 to 80.9 respectively for the same trait. All our studies traits showed that phenotypic variance is higher than genotypic variance (Table 2)

Briefly, the result exhibited those genotypic coefficients of variance (σ² g)

varied from 12.01% to 47.55%. The length of tubers has the lowest coefficients while the DC has the highest coefficients. Phenotypic coefficients of variance ranged from 15.03% to 58.94% and the highest PCV was obtained from DC and the lowest from ALT. The phenotypic coefficient of variation (PCV) was relatively greater than the genotypic coefficient of variation (GCV) for all traits. The environmental coefficient of variance ranged from 9.02 to 34.84 for the same traits.

The Relative Difference (RD) is the estimation of the ratio of GCV in association with the respective PCV and the estimated RD values varied from 0.11% to 24.59% for AWT and AVT respectively (Table 2).

The estimated broad sense heritability (H²) varied from 49% to 65%. The highest heritability was recorded for DC (65%) followed by ATL (64%) and AWT (61%) (Table 2) and the lowest for 49% (AVT). Generally, moderate (30% ≤ h² ≤ 60%) heritability values were marked for AVT (49%), TNT (57%), and NTP (60%) whereas the rest of the traits expressed high (H² ≥ 60%) heritability values. Genetic Advance as percentage of mean (GAM) ranged from 19.8% to 79.03% respectively for ATL and DC.

TNT: total number of tubers, NTP: number of tuber per plant, AWT: average weight of tuber, AVT: average volume of tuber, ALT: average length of tuber; DC: diameter at the collar.

The Pearson correlation between the 10 traits showed that Tuber weight (AWT) was strongly and positively correlated with yield (r = 0.8) (Figure 9). A strong positive correlation was also observed between the total number of tubers (TNT) and the number of tubers per plant (TNP). Furthermore, a positive correlation was observed between tuber volume (AVT). On the other hand, DC was negatively correlated with tuber weight, tuber volume, and yield (Figure 9).

The hierarchical ascending classification gives a structuring of the 8 potato collections in the bimodal rainfall zone of Cameroon into 3 groups using 6 morphological and 04 yield parameters (Figure 10). Group II contains a single variety (Maffo) while groups I and III consist of 2 (Desirée and Sypira) and 4 (Doza, Synergire, Panamera and Banso) collections respectively.

The number of tuber plants showed a significant difference between the collections studied. The results obtained in this work are close to those obtained in the Western Region of Cameroon in which the number of tillers per plant varied

between three and seven [13] . The height of the plants, which indicates plant biomass, varies from 4.5 cm to 80 cm respectively in Maffo and Desiree. Other researchers obtained values between 34 and 51 cm [13] . In addition, the values of the diameter at the collar obtained ranged from 0.2 cm for Synergie to 0.7 cm for Desiree compared to those obtained in the West Cameroon region, which ranged from 0.65 cm to 0.91 cm [13] . This would be due to the soil and climate conditions of the two cultivation zones. The differences observed in the different growth parameters in each variety can be explained by the effect of the variety on the vegetative growth of the different plants and the yield of each variety. Indeed, in their work, they demonstrate and confirm the influence of variety on the growth and development of plants. These results are also further confirmed by [14] .

The highest number of tubers per plant (15) was observed in Desiree in contrast to the Synergie variety whose average number was 4 tubers per plant. These results are similar to those obtained on the characterization of the potato in which the number of tubers per plant varied from 2 to 16 by [15] . Furthermore, the collection with the most tubers had a non-proportional mean weight [16] and observed significant differences in 24 potato genotypes due to genetic differences. The differences would also be due to environmental factors such as soil nutrient richness. Another work showed that the available potassium had the strongest correlation with potato tuber yield which is also correlated with soil water content [17] . The low production and productivity could be attributed to the infestation of disease on standing crops [18] .

The analysis of these results showed a positive correlation between the average weight of the tubers, the number of stems/plants, and the number of tubers. Traits with such a positive correlation with yield can optimize yield as a result of their simultaneous improvement [19] .

The seven varieties studied were subdivided into three clusters. These are respectively Desiree and Cypira (cluster I), Maffo (cluster II) and Doza, Panamera, Synergire, and Banso (cluster III). As in this work, [15] clearly separated Cypira and Doza varieties into two clusters while [20] separated 146 samples into 27 groups based on 15 variables. This would be due to the environmental factors and the genotypes of the collections.

For all the variables, phenotypic coefficients of variation were higher than genotypic coefficients of variation suggesting the role of environment in the expression of traits under observation. These results are in line with those obtained by [21] . The average tuber length is intermediate, while the other parameters show GCV and PCV values above 20%. The highest GCV (47.55) and PCV (58.94) were calculated for DC. On the contrary, [22] obtained the highest value for the yield of tubers per plant and the number of tubers per plant. This highest value indicates a high potential for effective selection.

The GAM raged from 19.8 (ATL) to 79 (DC). According to [23] , the GAM was categorized as low (0 - 10%), moderate (10% - 20%), and high (>20%). Most of the traits showed high GAM (>20%) except ATL with a moderate value (19.8). The estimated values of heritability vary from 49 (AVT) to 65 (DC). These results are in line with those obtained by [11] .

A moderate heritability and high GAM were observed for AVT, TNT, and NTP. ALT has presented a high heritability and a moderate GAM, while high heritability coupled with high GAM was observed for AWT and DC suggesting selection for these traits would give a good response. The same result has been obtained by [24] for the weight of the tuber and the number of tubers per vine on sweet potato. This result suggests that there may be the presence of additive gene action and selection will be rewarding for the improvement of such traits with a simple breeding method [10] . According to [11] the effectiveness of selection is realized more quickly in those characters which have high heritability and high genetic advance.

The relative difference (RD) is the estimation of the ratio of GCV in association with the respective PCV and the estimated RD values varied from 0.11% to 24.59% for AWT and AVT respectively. High values indicate an environmental influence. All the above observations based on morphological analysis showed a strong variability among the used material in this study. This variability may be confirmed by molecular analysis as in [25] .