Mangoes of the varieties Eldon, Haden, Paheri, Tommy Atkins, Zill and the SBMA-1 accession ( Figure 1 ) formed the plant material for the present study.

Mango varieties and Accessions were collected in the varietal collection field of the Institute de l’Environnement et de Recherches Agricoles (INERA) Banfora station located in western Burkina Faso, in the Cascades region ( Figure 2 ). Two waves of sampling were carried out.

For each sampling, 10 physiologically ripe mangoes were taken from a single plant of each mango variety, coded and placed in cartons at the Département Technologie Alimentaire (DTA), where they were subjected to ripening at room temperature in a laboratory room until taste ripeness.

Once the mangoes had reached taste maturity, they were sorted, washed, peeled and pitted. The pulp was recovered, crushed, packaged in 450 g plastic bottles and placed in a freezer to await analysis.

After peeling and pitting, the various parts of the mango were collected separately and weighed. The material balance was determined by dividing the mass of each part of the fruit by the total mass of the fruit.

$\text{\%}\left(\text{Core},\text{skin or pulp}\right)=\frac{Px}{Pt}\times 100$

where: Px = weight of each part of the mango (core, skin or pulp); Pt = total weight of mango

Water/moisture content was determined by drying 5 g of mango puree in an oven at 105˚C, in accordance with AFNOR standard [17] . Water content was calculated using the following equation:

$\text{\%}\left(\text{water}\right)=\frac{\left(M0-M\right)}{M0}\times 100$

M0 = initial sample weight;

M = final sample weight.

The color of the samples was analyzed using a colorimeter (Model NO PCE-CSM 1) based on the CIE-Lab color system, where the L* (luminosity) value ranges from black (0) to white (100), the a* value ranges from green (−60) to red (+60) and the b* value ranges from blue (−60) to yellow (+60).

Soluble solids content (SS)was determined using a digital refractometer (Bellingham + Stanley). A drop of mango pulp was placed in the refractometer’s orbit, and the Brix value was read directly from the display.

Titratable acidity was determined in accordance with French standards [18] . To do this, 5 g of mango pulp was weighed into a Falcon tube, and then suspended in 30 ml of distilled water. After 1 hour of magnetic stirring, the pH was measured directly using a pH meter (Neo-Tech.SA, SI Analytics) previously calibrated with buffer solutions pH = 4, pH = 7 and pH = 9.18. 20 ml of this solution was taken and titrated with 0.1 N NaOH in the presence of a few drops of phenolphthalein. The titratable acidity was obtained according to the equation:

$\%AT=\frac{N\times VT\times V\left(\text{NaOH}\right)}{PE\times VP}\times 0.070\times 100$

N = soda normality;

PE = test sample;

VP = volume of sample solution taken for titration;

VT = total volume of sample solution;

V= volume of NaOH solution required for titration;

0.070 = conversion factor to citric acid equivalent.

The carbohydrate/acid balance is a very important quality factor in fruit. It expresses the ratio between the rate of soluble solids (SS) and the rate of titratable acidity (AT) and can be used to determine the stage of fruit maturity or to determine the aptitude of the pulp for a given transformation. It is calculated according to the following formula:

$\text{Ratio}\left(\text{R}\right)=\frac{SS}{AT}\times 100$

where: SS = soluble solids, AT = Titratable acidity.

Total ash was determined by incineration in a muffle furnace at 550˚C in accordance with ISO 2171 [19] . For the analyses, a 5 g test sample of crushed pulp from each mango variety was weighed into a previously washed and weighed porcelain crucible. The crucible containing the pulp was then placed in a muffle furnace at 550˚C overnight, cooled in a desiccator for 1 hour and weighed again. The ash content was calculated according to the following formula:

$\%C=\frac{PF-PV}{PE}\times 100$

% C = percentage of total ash;

PE = test sample;

PF = final weight;

PV = weight of empty crucible.

The fat content (FC) of the pulp of each mango variety was determined using the Soxhlet extraction method in accordance with ISO 659 [20] . This method involves weighing the sample into a cellulose cartridge and extracting the fat continuously with a lipophilic organic solvent. After extraction, the fats are separated from the extraction solvent by distillation using a rotary evaporator. A 5 g sample of mango pulp from each variety was placed in a Soxhlet extraction cartridge. The cartridge containing the test sample was plugged with absorbent cotton and placed in the Soxhlet extraction device. The device was then placed on a 500 ml flask containing approximately 250 ml of hexane and topped with a refrigeration system connected to a cryostat to condense the solvent vapors. After 5 h of extraction, the MG was separated from the solvent by rotavapor distillation. The flask containing the fat was then placed for 30 min in an oven at 105˚C, cooled for 30 min in a desiccator and weighed. The fat content was calculated according to the equation below:

$\%MG=\frac{PF-PV}{PE}\times 100$

FP: final weight (flask + fat);

PV: empty weight of flask;

PE: test sample.

The protein content of mango cultivars and accessions was determined by the Kjeldahl method according to ISO 1871 [21] . This method is based on the determination of total nitrogen.

One gram (1 g) of crushed mango pulp was introduced into a Kjeldahl matrass, and a Djelbabs catalyst pellet and 20 ml of concentrated sulfuric acid were added. Mineralization was carried out at a progressive temperature of 100 to 550˚C for about four hours. The resulting mineralizate was cooled and made up to 50 ml with distilled water, then neutralized with a 10 N NaOH solution. The neutralized solution was distilled and the distillate was collected in an Erlenmeyer flask containing 20 ml of boric acid and a few drops of helianthine and bromocresol green. Distillation was stopped when the volume of the distillate reached 150 ml. The distillate was then titrated with 0.1 N sulfuric acid until the solution changed from green to pink. A blank was also titrated under the same conditions as the samples.

$\%P=\frac{14\times 0.1\times \left(VE-VB\right)\times {10}^{-3}}{PE}\times 6.25\times 100$

VE = Burette drop (sample);

VB = Burette drop (blank);

PE = test sample;

0.1 = sulfuric acid titer;

14 = molecular weight of nitrogen;

6.25 = nitrogen to protein conversion factor;

% P = Protein content.

Total sugars were determined by the orcinol method described by Montreuil and Spik [22] . Twenty (20) mg of pulp were taken into a beaker and 10 ml of distilled water was added. The mixture was stirred magnetically for 10 min, then transferred to a 50 ml volumetric flask, topped up with distilled water and homogenized. In test tubes, 0.5 ml homogenate was taken, 1 ml orcinol reagent and 3.5 ml H₂SO₄ (60%) were added respectively. The test tubes were homogenized and then placed in a boiling water bath for 20 min. They were then placed in the dark for 45 min and then at room temperature for 10 min. After homogenization, absorbance was measured at 510 nm using a spectrophotometer. Total sugar content was determined from a calibration curve using D-glucose as the reference sugar.

Phenolic compounds of mango cultivars and accessions were determined in extracts of pulp and powder of freeze-dried mango peel by spectrophotometry using the Folin-Ciocalteu reagent with modifications [23] . Methanol-HCl (1%) was used for extraction. A test sample of 2.5 g of mango pulp and 0.01 g of freeze-dried peel powder was placed in a 50 ml flask to which the extraction solvent was added. The flask was protected from light with aluminum foil. The mixture was placed under magnetic stirring for 10 min and then refrigerated. After 24 h maceration, the mixture was filtered through 0.45 µm Wattman filter paper.

An aliquot of 0.250 ml of the extract was taken and 1.25 ml of Folin-Ciocalteu reagent (0.2 N) was added. After incubation for 5 minutes, 1 ml of sodium carbonate solution (7.5 g/L) was added. The mixture was then placed in a water bath at 65˚C for 20 min, after which the absorbance at 760 nm was read against a blank. Measurements were performed in triplicate. The total polyphenol content was determined from a calibration curve plotted at different concentrations of gallic acid standard. Results were expressed as mg gallic acid equivalent (GAE) per 100 g of material.

The material balance for the five varieties and the mango accession ( Table 1 ), shows that the mass varied from 120.43 g to 638.10 g. The SBMA-1 accession recorded the highest value (638.10 ± 106.67) and the Paheri variety recorded the lowest value (120.43 ± 25.97). The mass of the Paheri variety is significantly different from the masses of the other varieties. According to the Codex standard [24] on mango sizing, the Zill mango variety can be classified as size A (200 - 350 g), the Haden and Tommy Atkins varieties as size B (351 - 550 g) and the Eldon mango variety and SBMA-1 accession as size C (551 - 800 g). The size of the Paheri variety is smaller than the Codex standard size A [24] . This means that this variety cannot be exported. The weights of the Eldon, Haden, Tommy Atkins and SBMA-1 accession mango varieties are close to those of the 20 Burkina Faso mango varieties studied by Kanté-Traoré, [16] , which ranged from 204 ± 11.91 to 843.43 ± 35.93 g. They are also close to those found by De Laroussilhe [25] . And Bafodé [26] and to those of mango varieties found in the European market [27] .

The proportion of the kernel of the different varieties and mango accession varied from 8.25% to 19.42%. The highest proportion was observed with the Paheri variety and the lowest value with the SBMA-1 accession. Only the kernel proportion of the Paheri variety was significantly different from those of the other varieties. The kernel proportions of the mango varieties in the present study are close to those found by Kanté-Traoré, [16] in the 20 Burkina mango varieties (5.19 ± 0.63 to 16.93 ± 1.50), except that of the Paheri variety, which is higher. They are also similar to those found in Cameroon mangoes, which ranged from 12% to 17% [28] .

The percentage of peel of different mango varieties and one accession ranged from 7.22% to 18.88%. Paheri variety had the highest value and SBMA-1 accession had the lowest value. The peel percentage of Paheri variety and SBMA-1 accession is significantly different from the skin percentage of other varieties. The peel proportion of Eldon, Haden, Tommy Atkins, Zill and the SBMA-1 accession are similar to those of the six best known and most used mango varieties in Burkina (10.50% to 13.93%) [16] . On the other hand, the skin content of the Paheri variety is close to that of the Julie (18.43 ± 2.64) and Francis (20.56 ± 0.15) varieties [16] and also to that of the Cameroonian mango varieties (17% and 19 %) reported by Kansci et al. [28] .

The pulp content of the five mango varieties and accessions studied ranged from 51.18% to 85.06%. The highest pulp value was recorded for the SBMA-1 accession, while the Paheri variety had the lowest pulp value. The pulp content of Paheri variety is significantly different from that of other mango varieties. These results are similar to those found in Burkina mangoes, which ranged from 64.55% to 80.32% [16] The flesh content of the Zill variety is close to that of the Keitt variety from Cameroon (72%) [28] . On the other hand, the pulp content of the Eldon variety is higher than that found by Kansci et al., [28] in the Keitt variety from Cameroon, but close to that found by Passannet et al., [29] in the Cœur de bœuf variety (78.68%) from Tchad.

The pulp/Kernel ratio of mango varieties and accessions ranged from 2.69 to 10.54. The highest value was recorded for the SBMA-1 accession and the lowest for the Paheri variety. These results are similar to those found in Burkina mangoes, which ranged from 3.81 to 11.67; [16] . The pulp/peel ratio ranged from 2.78 to 12.03. The SBMA-1 accession recorded the highest value and the Paheri variety the lowest. These values are similar to those found with the 14 Burkina mango varieties, which ranged from 3.16 to 9.03 [7] .

Pulp/peel ratios are a good indicator of the suitability of mango varieties for industrial processing [7] [30] .

NB: Values bearing the same letters are not significantly different (p < 0.05).

The pulp color variables of the investigated varieties and accession are presented in Table 2 below. The L* values of the five varieties and the mango accession ranged from 41.85 to 55.41. Statistical analysis showed no difference between cultivars and accessions. Haden and Tommy Atkins recorded the highest values of 55.41 ± 0.06 and 51.56 ± 0.32, respectively. Zill had the lowest value. The L* values found in this study are in the same order of magnitude as those found in Burkina mangoes (37.97 to 53.15) [7] [31] . On the other hand, they are lower than those found by Morales et al., [32] in Colombian-Caribbean mangoes (66 to 76). This difference could well be explained by the ecological difference between the two regions. The flesh of the Haden, Tommy Atkins and Eldon varieties is susceptible to browning, given the high L* value. In fact, Robles-Sanchez et al. [33] and Elsheshetawy et al., [30] reported that pulp L* values are indicators of browning during fruit processing due to contact between oxygen and polyphenol oxidases (PPO) in mango pulp. For pulp preservation,

The a* values for the five varieties and the accession ranged from 0.74 to 10.94. Haden and Paheri were significantly different from the other varieties. Haden had the highest value and Zill the lowest. The a* values found are lower than those found in Colombian-Caribbean mangoes, which ranged from 7 to 26 [32] , and in Burkina mangoes, which ranged from 4.25 to 17.75 [7] . The positive a* values obtained show that all the varieties and accessions studied have a yellow-orange color [7] .

The b* values ranged from 29.75 to 41.38. Statistical analysis showed no difference between varieties and accessions. Tommy Atkins had the highest value and Eldon the lowest. The b* values found are lower than those found in Colombian-Caribbean mangoes, which ranged from 43 to 50 [32] , but similar to those found in Burkina mangoes, which ranged from 22.94 to 56.07 [7] . Based on the a* and b* values, the pulp of the 05 cultivars and the accessions in this study were yellow-orange in color, with the Haden variety having a brighter color.

The c* values ranged from 30.10 to 41.48. Statistical analysis showed no difference between cultivars and accessions. Tommy Atkins had the highest value and Eldon the lowest. Observations show that b* and c* values are about the same. The c* values found are close to those found in Burkina mangoes, which ranged from 23.41 to 58.82 [16] .

The ΔE values ranged from 53.12 to 69.00. Statistical analysis showed no significant differences between cultivars and accessions. However, the Haden and Tommy Atkins cultivars had the highest values, while the Zill cultivar had the lowest. These values are similar to those found in Burkina mangoes, which ranged from 44.61 to 79.27 [7] .

The positive values of L*, a*, b*, c*, ∆E in the present study provide information about the yellow-orange flesh color of the varieties. This color is the majority color of most mango varieties and is appreciated by consumers. Lauricella et al. [34] reported that the mesocarp color index is considered a good quality and consumer perception parameter.

Values bearing the same letters are not significantly different (p < 0.05).

The chemical parameters of the varieties studied are presented in Table 3 below.

The pH of the 5 varieties and the mango accession ranged from 3.61 to 5.11. Paheri variety had the highest pH and SBMA-1 accession had the lowest pH. There was a significant difference among varieties and accessions. The pH of the Haden variety is similar to that of the Julie variety from Burkina [7] and that of the Mangotine variety from Tchad, which was 4.70 [28] . The pH of the Paheri variety is similar to that of postharvest ripened Indian mangoes (5.25) [35] and to that of the Alphonso variety (5.1) [36] . The Tommy Atkins and Zill varieties have pH values close to that of the Ataulfo variety (4.1) [37] .

The titratable acidity (TA) of mango ranged from 0.20% to 0.99%. The accession SBMA-1 and the varieties Eldon and Zill recorded the highest values. The TA values of SBMA-1, Eldon and Zill were significantly different. Paheri, Tommy Atkins and Haden had the lowest values. The titratable acidity of the SBMA-1 accession (0.99%) is close to that of the Dixon variety from Burkina (1.01%) [7] . The acidity of the Paheri variety is close to the average found in postharvest ripened Indian mangoes (0.19%) [35] . In terms of pH and titratable acidity, Paheri is the least acidic variety, while Eldon and SBMA-1 are the most acidic.

The soluble solids (SS) content of the 5 varieties and the mango accession ranged from 19.37% (Tommy Atkins variety) to 26.47% (Paheri variety). The SS contents of the 5 varieties and the mango accessions are in the same order of magnitude as those reported by Kanté-Traoré, [16] in the 20 varieties (11.00 to 23.10%) of Burkina mango, except that of the Paheri variety, which has a higher SS content. The SS content of the Tommy Atkins variety is close to that of the Bangui variety from Tchad [29] . While that of the Haden variety is similar to that of the Ataulfo variety (21.6%) [37] and to that of Indian mangoes, whose average was 21.35% [35] [38] . On the other hand, it is higher than that of Alphonso variety (20%) [36] . The SS content gives an indication of the sweetness of the fruit [35] . It is also an important parameter for processing. High SS levels indicate that the mango varieties and accessions studied have a very sweet taste. Paheri is the sweetest variety.

Values bearing the same letters are not significantly different (p < 0.05).

The SS/TA ratio ranged from 22.76 to 133.10. Paheri variety had the highest ratio and SBMA-1 accession had the lowest value. These results are higher than those found in the 20 Burkina Faso mango varieties, which ranged from 9.79 to 62.59 [16] . The very high SS/TA ratio of Paheri confirms its very sweet character compared to other varieties.

The Fe, K, Mg, Na and Zn contents of the five mango varieties and accession are summarized in Table 4 below.

The Fe content ranged from 7.84 to 23.23 mg. Eldon variety had the highest Fe content and Paheri variety had the lowest. These values are much higher than those found in mangoes in general, which average 0.16 mg [39] . This could be explained by the pedoclimatic conditions.

The K content ranged from 63.41 to 161.65 mg. The Zill cultivar had the highest K content and the Tommy Atkins cultivar had the lowest. The K value found in the Zill variety is close to that found in mangoes in general, which is 168 mg [39] .

Mg content ranged from 12.60 to 23.78 mg. The Zill variety had the highest value and the Eldon variety had the lowest value. The Mg content of the Eldon variety is slightly higher than that found in mangoes in general, which is 10 mg [39] .

The Na content ranged from 1.23 to 4.23 mg. These values are higher than those found in mango in general, which is 1 mg [39] . The Tommy Atkins cultivar and the SBMA-1 accession did not show any trace of sodium.

The Zn content ranged from 0.37 to 2.64 mg. The highest value was recorded in the Haden variety and the lowest in the Zill cultivar. The values found are higher than those found in mango in general, which is 0.09 mg [39] .

Table 5 shows the nutritional content of the 05 varieties and accession.

Fat content (FC) of mango varieties and accessions ranged from 0.18% (Zill variety) to 0.50% (Tommy Atkins variety). The fat content of Zill was significantly different from that of Eldon and Tommy Atkins. The fat contents obtained in the present study are higher than those found in mangoes from Cameroon, whose fat contents ranged from 0.17% to 0.33% [28] .

Total ash content ranged from 0.22% to 0.95%. The Zill variety had the highest ash, while the Tommy Atkins variety had the lowest ash content. The ash contents of Eldon, Paheri and Tommy Atkins were significantly different from those of Haden and Zill. The ash contents obtained are close to those reported by Kanté-Traoré, [16] for the 20 Burkina mango varieties (0.48% to 1.41%). However, the ash content of the Zill variety is higher than those found in mangoes from Cameroon, which ranged from 0.32 to 0.49% [28] . The ash content of the mangoes studied is lower than that found in mangoes from Congo (2.63% ± 0.14%) [40] .

Protein (P) content ranged from 0.42% (Tommy Atkins) to 2.69% (Eldon). The protein content of Paheri, Tommy Atkins and the SBMA-1 accession was significantly different from that of Eldon and Zill. The values obtained are higher than those found in mangoes from Cameroon, which ranged from 0.16% to 0.24% [28] . Also, the protein content of Eldon, Haden and Paheri varieties is higher than those found by Kanté-Traoré, [16] in the 20 mango varieties (0.21% to 0.77%).

The total sugar content of the five varieties and an accession ranged from 14.50% to 28.08%. The total sugar contents of the five varieties and the mango accession were not significantly different. These sugar contents are higher than those obtained in mangoes from Cameroon (9.43% to 15.16%) [28] and those found in mangoes in general, which were 13.66% [39] . However, they are lower than those found in 14 mango varieties from Burkina, which ranged from 68.47% to 76.79% [7] .

The energy value (EV) of the five mango varieties and accessions ranged from 66.63 Kcal/100 g to 126.88 Kcal/100 g (fresh weight). The highest value was recorded for variety Eldon and the lowest for accession SBMA-1. The energy values of the five mango varieties and accession were not significantly different. These results are higher than those found in mango in general, which is 60 Kcal/100 g [39] .

Values bearing the same letters are not significantly different (p < 0.05).

Figures 3-4 below show the total phenolic compound contents in the pulp and freeze-dried peel powder of the 05 mango cultivars and accession analyzed.

The results show that the total phenolic compound content of the fresh pulp of the five mango varieties and accession ranged from 21.79 to 30.41 mg EAG/100 g ( Figure 3 ). While the values obtained in freeze-dried peel powder ranged from 2293.02 to 5798.99 mg EAG/100 g ( Figure 4 ). The observations made show that there is no significant difference between the phenolic compound contents of the fresh flesh of the 05 mango varieties and accession. On the other hand, the phenolic compound contents of the peel powder of Haden variety and SBMA-1 accession are significantly different from those of Eldon, Paheri, Tommy Atkins and Zill varieties for P ≤ 5%. The highest levels of fresh pulp (30.41 ± 0.12 mg EAG/100 g) and freeze-dried peel powder (5798.66 ± 16.95 mg EAG/100 g) are found in the Zill variety. Regarding the content of phenolic compounds, polyphenols are more abundant in the peel than in the fresh pulp. This could be explained by the ripening process, which leads to a reduction of phenolic compounds and acids. The levels found in the fresh pulp of the 05 mango varieties and accessions in the present study are lower than those of the Ataulfo variety, which was 120 mg EAG/100 g [37] . They are also lower than those of the 20 Burkina mango varieties (48.15 and 110.41 mg EAG/100 g fresh pulp) [16] . The phenolic compound content of the Zill variety pulp is close to that of Indian mangoes, which was 32.81 ± 5.64 mg GAE/100 g [35] . Sogi et al., [41] reported phenolic compound contents ranging from 962.2 mg/100 g in vacuum-dried pulp of Tommy Atkins variety to 1725.2 mg/100 g in freeze-dried pulp of the same variety. Phenolic compounds are a group of secondary metabolites that are considered natural antioxidants with multiple biological benefits for human health [32] . Mango peel is very rich in gallic acid and could be used in the treatment of diabetes [42] . Since mango skin is also an edible part of the fruit, it should be consumed at the same time as the flesh. It could be used as a nutraceutical [42] .