The study was carried out on 2020 at Tengréla in Ivory Coast. Tengréla is a town located in the Savannas region in the extreme north of the country. Tengréla has a very dry, hot Sudanese climate. The vegetation is dominated by wooded savannah. The area produces mangoes and is ideal for livestock farming [10] .

The building used for the trial consists of a hangar topped with sheet metal and open on the west side. The other three sides are built of fixed masonry. It is subdivided into 3 pens of 10 m² each, each large enough to house a batch of 85 chickens from chick to sale. The screened enclosure provides good natural ventilation. It should also be noted that an electrical system was installed in the building to regulate temperature during the cool season. The flour of the henhouse was covered with litter at the recommended rate of 4.5 to 7.2 kg/m² depending on the raw material used [11] . The litter was changed every two weeks.

The animal material consisted of 255 COBB500 broiler chickens purchased at 01 day of age in Korhogo from Abidjan and transported to Tengréla. On arrival, the chickens were fed sugar water. A commercial starter feed was fed to all chicks without distinction for 01 week, before being replaced by the experimental rations once the batches had been formed. Three batches of 83, 83 and 84 chicks were formed, each receiving 01 control ration (CCD), an experimental ration with 10% mango seed floor (ED₁₀) and a second experimental ration incorporating 20% mango fines (ED₂₀).

Three rations, including two experimental ones (ED₁₀ and ED₂₀), were formulated to be fed chickens per batch. These rations were formulated to reduce the incorporation rate of maize, the main source of energy in intensive and/or industrial rearing, by incorporating feed and mango seed floor. Their centesimal compositions and nutritional contributions are given in Table 1. The two experimental rations were compared with a control ration, which is a commercial broiler growth feed usually sold on the Ivorian market. Feeding was carried out in 2 phases. A start-up phase (from 1st to 10th day) during which the chicks were fed a starter diet purchased on the market, and a growth-finishing phase during which the chickens were fed ED₁₀, ED₂₀ and RTC rations.

➢ Zootechnical parameters: the zootechnical parameters studied concerned:

­ Weight evolution was assessed by weekly weighing of the chickens in the morning from 6 hours before the ration was served;

­ Growth rate was measured by Average Daily Gain (ADG) calculated according to the formula: ADG (g) = FW − IW where PF (g) = final weight and PI (g) = initial weight;

­ Daily food consumption (DFC) is calculated using the formula DFC (g) = (QFD − RF) where QDF (g) = amount of feed distributed the previous day and RF (g) = feed refusal measured the following day;

­ The consumption index (CI) was calculated using the formula: CI = DFC (g)/(ADG) (g)

➢ Carcass parameters: To assess carcass characteristics, chickens were first fasted for 24 hours and weighed before slaughter. The chickens were slaughtered and the carcasses cut according to the methods recommended by [12] . After plucking, warm carcasses were weighed and then chilled at 5˚C for 24 hours. The chilled carcasses were then weighed. Each piece of the cut and each organ were weighed. The data collected from the weighting’s were used to calculate:

ED₁₀ = experimental ration incorporating mango by-products including 10% mango seed flour; ED₂₀ = experimental Ration incorporating mango by-products including 20% mango seed floor; CCD = commercial control ration without mango by-products.

­ Hot carcass yield (HCY) according to the formula: HCY (%) = (HCW/PSLW) × 100 were HCW = hot carcass weight and PSLW = pre-slaughter live weight;

­ Cold carcass yield (CCY) according to the formula: CCY (%) = (CCW/PSLW) × 100 were CCW = cold carcass weight and PSLW = pre-slaughter live weight;

­ Abdominal fat rate (AF) according to the formula: AF (%) = (AFW/CCW) × 100 were AFW = total weight of abdominal fat and CCW = cold carcass weight;

➢ Organoleptic quality of the meat: organoleptic qualities were determined by evaluating the tenderness and succulence of the meat, by tasting the salted meat fried in oil. A panel of twenty-five (25) people (men and women) carried out the tasting. The meat was served in two dishes, one identified by pink wallpaper and the other blue, containing respectively meat from the CCD ration and meat from the ED₁₀ and ED₂₀ rations. It should be noted that only the organizers could differentiate between the meats served. The carcasses were cut to ensure uniformity in the size of the pieces of meat served. After tasting, we recorded each participant’s appreciation in an event notebook for analysis.

The data collected were entered using Excel 2016 software, and the analysis was carried out with XLSTAT 2016 software using an ANOVA for the separation of means at the 5% threshold by Fisher LSD test.

The energy and nutritional composition of the rations is shown in Table 2. The levels of incorporation of mango seed floor in ED₁₀ (10%) and ED₂₀ (20%) rations resulted in a higher level of fish incorporation in the experimental diets ED₁₀ which were higher than those ED₂₀. The ED₁₀ and ED₂₀ rations were iso-energetic and iso-proteinic. The Crude Fiber (CF) contents of the RD₁₀ and ED₂₀ rations were virtually equal and higher than those of the commercial CD ration. The cost per kg of the ED₁₀ and ED₂₀ rations was lower than that of the CD ration.

Figure 1 shows that growth was continuous for all three rations from D1 to D35. The results in Table 3 show that, apart from the weights at D1 and D7, all other average weekly weights were significantly different (p < 0.05). From D7 to D35, the CCD ration was on average significantly higher than the ED₁₀ ration, which was in turn significantly higher than ED₂₀ (p = 0.000). At the seventh weighing (D49), the average weight of ED₂₀ was significantly higher (p = 0.000) than those of CCD and ED₁₀, which were homogeneous (p = 0.000).

ED₁₀ = experimental ration 1 incorporating mango by-products including 10% mango kernel; ED₂₀ = experimental; Ration incorporating mango by-products including 20% mango kernel; CCD = commercial control ration without mango by-products.

ED₁₀ = experimental ration incorporating 10% of mango seed flour; ED₂₀: experimental ration incorporating 20% of mango seed floor; CCD = commercial control ration without mango. Averages on the same line bearing the same letter are not significantly different at the significance levels indicated.

The results in Table 4 shows that ADG were significantly different and decreasing from CCD, ED₂₀ to ED₁₀ (p = 0.000). No significant difference was observed in average daily feed intake between diets. The CI for the ED₁₀ ration was significantly higher than that for CCD (p = 0.039). The average FCkgLW of the ED₁₀ and ED₂₀ experimental rations was significantly lower than that of the CCD ration (p = 0.027).

The results in Table 5 shows that average pre-slaughter live weights (PSLW), hot carcass weights (HCW) and cooled carcass weights (CCW) were significantly higher and decreased from CCD, ED₂₀ to ED₁₀ rations (p = 0.000). The average

ADG = average daily gain; CI = conversion index; FCPkgLW = feed cost of 01 kg chicken live weight production; DFC = daily feed consumption per subject; ED₁₀ = experimental ration with 10% mango seed flour; ED₂₀: experimental ration with 20% mango seed floor; CCD = commercial control ration without mango. Average means in the same line bearing the same letter are not significantly different at the significance levels in the table.

ED₁₀ = experimental ration with 10% mango seed flour; ED₂₀: experimental ration with 20% mango seed floor; CCD = commercial control ration without mango. PSLW = pre-slaughter live weight of chicken, HCW = hot carcass weight, CCW = weight of chicken after slaughter, HCY = hot carcass yield, CCY = cold carcass yield. Average means in the same line bearing the same letter are not significantly different at the significance levels in the table.

hot carcass yield (HCY) and cold carcass yields (CCY) of rations ED₂₀ and CCD were comparable (p > 0.05) and significantly higher than those of ration ED₁₀ (p = 0.009 and p = 0.003). Conversely, the mean abdominal fat content (AF) of ED₁₀ was significantly higher than that of rations ED₂₀ and CCD (p = 0.004), which were comparable with each other (p > 0.05).

Our results for carcass cuts (Table 6) show that:

- The average weights of the whole thigh, wings, chest and back of the CCD were significantly higher than those of the ED₁₀ and ED₂₀ rations (p = 0.000), which were comparable to each other;

- Mean drumstick and neck weights were significantly different and decreased from CCD, ED₂₀ to ED₁₀ (p = 0.000).

The results presented in Table 7 shows the offal and organ weights of COOB 500 broilers after slaughter, according to ration. Average total offal and gizzard weights of the RT ration were significantly higher than those of the ED₁₀ (p = 0.000) and ED₂₀ (p = 0.001) rations, which were homogeneous between them. Average gutted gizzard weights of RT and ED₂₀ were comparable and significantly higher than those of the ED₁₀ ration (p = 0.004). Average leg and head weights were significantly different and decreased from RT and ED₂₀ to ED₁₀ (p = 0.000).

The results of the taste test showed that the tenderness of the meat from chickens on the mango diet was tenderer for 60% of consumers, compared with

ED₁₀ = experimental ration with 10% mango seed floor; ED₂₀: experimental ration with 20% mango seed floor; CCD = commercial control ration without mango; Average means in the same line bearing the same letter are not significantly different at the significance levels in the table.

ED₁₀ = experimental ration with 10% of mango seed flour; ED₂₀: experimental ration with 20% of mango seed flour; CCD = commercial control ration without mango; Average means in the same line bearing the same letter are not significantly different at the significance levels in the table.

40% on the CCD diet. The succulence of meat from diets incorporating mango was also judged more succulent by 76% of consumers, compared with 24% for the CCD diet.

Our results on broiler weight development showed continuous growth from the beginning to the end of the trial. However, the growth rate of the commercial control ration (CCD) was higher than that of the ED₁₀ and ED₂₀ experimental rations. These results confirm our expectations on these two parameters. They corroborate results reported by other authors such as [6] and [7] who tested rations based on mango by-products respectively in the feed of growing Isa Brown layers and growing-finishing local breed chickens respectively. Since mango by-products are rich in fiber, they have certainly had an upward influence on the fiber content of rations, which is not conducive to good digestibility and therefore good availability of nutrients, especially energy. The consequence of this situation is that after 42 days of rearing, the CCD subjects had reached an average weight (1308 ± 251 g) acceptable for sale. On the other hand, ED₁₀ and ED₂₀ subjects waited a week longer to reach average weights of over 1200 g. At 42 days of age, the average weights of all our ED₁₀ (1081 ± 16 g), ED₂₀ (1080 ± 19 g) and CCD (1308 ± 25 g) rations are lower than those reported by [13] for 2 experimental rations based on maize espoir variety (1743.8 ± 248.0 g) and SR21 (1721.3 ± 250.2 g) compared with a commercial ration (1610.3 ± 244.9 g). In addition, the CI of the ED₂₀ ration was significantly comparable to that of the CCD, meaning that these two rations were the most highly valued compared with the experimental ED₁₀ ration. These results show that the mango by-product feeds (mango meal and mango seed flour) used in the ED₁₀ (36.75%) and ED₂₀ (45%) rations can be a good substitute for maize, which is recognized as a major source of energy. The effects of rationing trials based on maize (60.23%) as the main source of energy and groundnut meal (25.49%) as the main source of protein [14] showed the same growth rate for control rations in both cases, and double that of our experimental rations. However, the conversion levels (CI) measured weekly in the trial were close to those of our ED₂₀ ration. This performance of the ED₂₀, ration, which also doubled its weight gain compared with the ED₁₀ ration between W6 and W7, could be explained by the influence of the additional rate (10%) of incorporation of mango seed flour Contrary to weight performance, feed production costs of per kg live chicken (FCPkgLW) were lower for the ED₁₀ and ED₂₀ rations than for the CCD ration.

Our results on the average characteristics of broiler carcasses (Table 5) show the same levels of abdominal fat rate (AF) between the ED₂₀ and CCD rations, which are significantly lower than that of ED₁₀. The opposite is true for carcass yields (HCY and CCY), which are comparable for the ED₂₀ and CCD rations, but significantly higher than that of ED₁₀. These results run counter to the average of PSLW, HCW and CCW, which are significantly higher and decreasing for CCD, ED₂₀ and ED₁₀ rations. This demonstrates the ability of the ED₂₀ ration to produce more meat than the ED₁₀ ration and as much lean meat as the CCD ration. In monogastric animals, fibrous feeds favor the production of lean meat, which in principle weighs more than fatty meat. The evaluation of mango-fed pig carcasses by [15] and [5] demonstrated this assertion. Note that according to [16] mango kernels have antioxidant properties that reduce fat formation in carcasses. Our results obtained for the CCD and ED₂₀ are very close to the abdominal fat levels reported by [16] for broilers fed increasing levels of caterpillar meal and fish meal, for which it had been concluded that they were capable of producing lean meat. The significant difference in live weight at slaughter (LWS) depends on the final weights, which differed between the 3 rations. However, we note that the ED₂₀ has the same carcass yield as the CCD. As the aim of rearing chickens is to gain meat, the almost equal yields between the CCD and the ED₂₀ show that the latter, based on mango by-products, makes up for the chickens’ growth deficit through meat yield. This may be explained by the higher fat content of CCD carcasses compared with experimental rations. If we add to these good yield performances the lower production feed cost of the ED₂₀ ration than that of the CCD, we can conclude that it is technically and economically competitive with the control ration.

Meat from rations incorporating mango feed received the highest ratings for both tenderness and succulence, compared with that from the CCD mango-free control diet. This shows that, in addition to good carcass yield, the mango-based ration provides lean, tender and succulent meat, qualities that most consumers are looking for. If we add to these good organoleptic performances, good yields and low production costs, we can encourage the use of feed based on eating byproducts in broiler diets.