The material consisted of 31 IITA coloured flesh clones and 3 control varieties (Yacé, Bonoua2 and Bocou2). Bocou2 is a new improved variety introduced from IITA and selected by CNRA after a series of multi-local evaluations. It is yellow-flesh and good production potential (yields of around 32 t∙ha⁻¹ in station) and use in semolina. Yacé is a white-flesh cultivar that is widespread in the southern region around Abidjan and is used in the manufacture of semolina and placali. It has a root production level of 25 t∙ha⁻¹ in station and a dry matter content (40%) that is appreciated by users. Bonoua2 (20 t∙ha⁻¹ in station), a white-flesh cultivar, is widespread in the western region of Côte d’Ivoire and is used for the preparation of stew and foutou. It is sweet and has good culinary ability in water. The three varieties served as reference controls for yield and total carotenoid content (Bocou2), for dry matter content (Yacé), for taste and cooking (Bonoua2).

The study was conducted at the CNRA’s Station de Recherche sur les Cultures Vivrières (SRCV) in Bouaké (7˚4'N and 5˚2'W), central Côte d’Ivoire, in the Guinean savannah (Figure 1), where moderately desaturated ferralitic soils dominate. The rainfall regime is limited to two rainy seasons (April to July and September to October) and two dry seasons (August and November to March). The average annual rainfall is between 1100 and 1200 mm. The average temperature is 25.73˚C [12] and the annual insolation duration is 2200 h [13].

The assay was conducted in three replicates using the Lattice design. Each replication consisted of two blocks of 17 elementary plots that received 34 genotypes. Each plot had four lines of six plants. Spacing was 0.8 m between plants, 0.8 m between rows, 1.5 m between plots and 2 m between blocks. The land was cleared and machine plowed. The cuttings were planted horizontally to a depth at less than 10 cm. Maintenance consisted of replacing missing plants one month after planting and weeding. Subsequently, weeding was done on demand. The tuberous roots were harvested 12 months after planting.

Observations and measurements are focused on three agronomic and five technological traits.

Quantitative data (mite incidence, mite severity, yield, dry matter and total carotenoid content) were submitted to a one-way analysis of variance (ANOVA I) to test the clone effect on each trait. Before performing this test, the normal distribution according to the Shapiro-Wilk test and the homogeneity of variances were checked. When these conditions were verified and accepted, ANAOVA I was performed. Any significant ANOVA (P < 0.05) was followed by TURKEY’s highly significant difference (HSD) to rank averages. This test was used to identify objects that differed significantly from others. Data that do not follow a normal distribution according to the Shapiro-Wilk test have undergone a natural logarithm, square root and Arcsinus transformation. If the distribution remains abnormal or the variances not homogeneous after the transformations of variables, Kruskall Wallis non-parametric test was applied. Analyses were performed with Xlstat 2018.7.5 software [14]. Fiber rate, cooking and taste were scored as qualitative traits.

The tropical zone offers favorable conditions for the most damaging mite species Mononychellustanajoa on cassava during the dry season when mean temperatures vary between 24˚C and 31˚C and relative humidity values from 50% to 70% [15]. The 31 cassava coloured-flesh clones showed significant attacks as observed by [16] cited by [5] who showed that the selection of mite resistant varieties is less efficient. So, to control these attacks, the use of natural predators seems to be the most efficient method. Indeed, [16] noted that mites belonging to the family Phytoseiidae are used as the main control agents against green mites. Biological control programs already exist in some research institutes such as IITA, CIAT and EMBRAPA.

High variability in cassava yield was observed among the 31 tested clones. Differences in fresh tuberous root yields among clones were observed. This difference could be explained firstly by the number of roots per plant and, on the other hand, by the weight of tuberous roots per plant. In our study, five coloured- flesh clones had yields of at least 50 t∙ha⁻¹. These yields are higher than those recorded by [10] during the evaluation work, in which the best performing clones had yields below 40 t∙ha⁻¹. However, these yields are close (or even statistically identical) to the yields obtained by [9] which were around 60 t∙ha⁻¹ in an evaluation of coloured-flesh cassava varieties. Of the 31 coloured-flesh clones, 25 clones gave fresh tuberous root yields statistically equivalent to Bocou2 (reference control, a variety popularized in Côte d’Ivoire) which is an improved yellow-flesh, high-yielding variety. The results obtained in this study could be improved by efficient pest and disease control. Local varieties Bonoua2 and Yacé are not very productive compared to some clones that are improved varieties. These results are in agreement with [11] [17] [18] and [19] which showed the superiority of yields of improved varieties over local varieties, in other words, improved varieties give significantly higher yields than local varieties. According to [17], this situation can be explained by the recurrent selection practiced at IITA, transmitting interesting agronomic traits from previously selected high-yielding progenitors into the new varieties.

The dry matter content varied from 33.08% to 43.63% with a very highly significant difference. This is in agreement with [18] who observed highly significant differences among the dry matter rates of several improved clones. This difference can be explained by the fact that clones belong to different genotypes. The dry matter content of some coloured-flesh tested clones are comparable to those obtained by [9] and [10] which are in the range of 30% - 40% in a performance evaluation study of coloured-flesh cassava clones compared to white-flesh cassava varieties which obtained higher levels. Indeed, dry matter content is negatively correlated to total carotenoids content as reported by several authors including [20] and [21]. However, two yellow-flesh varieties (I090006 and I102480) have similar dry matter content as Yacé, the white-flesh control.

The 31 coloured-flesh clones and the control Bocou2, which is a yellow-flesh variety, had higher total carotenoid contents than the local white-flesh varieties Bonoua2 and Yacé. These results support those of [22] who showed that coloured-flesh cassava varieties have higher total carotenoid contents than white- flesh varieties. Tested clones recorded higher total carotenoid contents than the control Bocou2 (8.01 ± 0.19 µg∙g⁻¹). Clones I082425 and I084157 recorded the highest total carotenoids contents of 18.39 µg∙g⁻¹ and 17.57 µg∙g⁻¹ respectively, which is more than twofold the content of the control (8.01 µg∙g⁻¹). This could be explained by the fact that these clones have been enriched (biofortified) in carotenoids. Howerver, the highest carotenoid contents in tested clones are far lower than those observed in coloured-flesh sweet potato varieties, which range from 20 µg∙g⁻¹ to more than 100 µg∙g⁻¹ [23], but much higher than those observed in coloured-flesh cassava varieties, which range from 0.31 µg∙g⁻¹ to 10.5 µg∙g⁻¹ [24].

Coloured-flesh clones had generally bad cooking, in accordance with [25] cited by [26] and [27]. These authors showed that bad cooking of coloured-flesh cassava could be due to high carotenoid content. It could also be due to the quality of the starch i.e. the level of cooking resistant starch which could be high. Indeed, [28] found high levels of starch qualified as cooking resistant in bananas that did not cook. On the other hand, low-fiber clones could be suitable for other purposes (semolina, gari etc.).

Bad tastes are due to the content of cyanogenic glycosides that define the bitterness of cassava flesh [3]. Cassava is considered ‘bitter’, when the cyanogenic glucoside content is higher than 50 µg∙g⁻¹ of fresh matter. For bitter cassava varieties, transformation processes for the detoxification of cassava are necessary in order to make it fit for consumption. Indeed, the ingesting of large quantities of uncooked or boiled bitter roots for a very short time can be dangerous because cyanogenic glycosides can be broken down by the enzymes of the intestinal flora and release hydrocyanic acid (HCN) that is toxic to the human body [3].