Zootechnical performances and profitability of local breed’s pigs fattened through the application of feed and health technological packages in traditional livestock farms in Casamance, Senegal

ABSTRACT This work aims to assess the effects of the application of feed and health technological packages on the zootechnical-economic performance of local breeds pigs in Casamance. It involved 253 young pigs, provided by 44 traditional farms. The feed and health technological package group (FHT) pigs were kept in permanent confinement and fed ad-libitum with a diet containing 15% of dried cashew apple pulp, while the controls and health technological package (HT) pigs were raised in a free-range area and fed in the usual way by the farmers. The zootechnical and economic parameters were collected and processed by ANOVA test and linear regression. The results showed a significant improvement in health status, live body weight, carcass weight of pigs, daily weight gain mean of the pigs in the FHT group compared to the HT and controls (p < 0.05). The production costs of the FHT pigs were lower than those of the HT and control pigs’ groups. They generated on average a gain for the selling in live body weight and a loss for those in carcass weight. The use of feed and health technological packages is advised in pig farming and price incentives are necessary to make the fattening activity sustainable.


Introduction
In the context of the strong demographic growth (2.7% annual growth from 2015 to 2020) experienced by most Sub-saharan African countries (Tabutin and Schoumaker 2020), the pig farming has many advantages.Indeed, it is an animal species with a short reproduction and production cycle, high prolificacy and capable to valorizing diets of various kinds (Serres 1989).In addition, pig farming provides income to low-income farmers in tropical regions (FAO 2012) and is therefore well suitable to fight poverty (Mopate et al. 2010), and diversify food and cash resources (Mopate et al. 2020).In sub-Saharan Africa, pig production is mainly extensive and the work of small family units, which account for 90% of the production of this meat, is based on the use of domestic and agro-industrial by-products (Boutonnet et al. 2000).
Thus, despite the predominance (95%) of a Muslim population in Senegal (ANSD 2020), pig production is developing with a growing herd (3.5% nationally) estimated at 451,000 head in 2018.According to Senegal's official statistics, pig meat production is 16,470 tons, and represents 14% of the total quantity of white meat produced compared to 86% for the poultry share (MEPA 2019).However, the availability of animal protein is still insufficient for consuming populations and demand is strong from Dakar and some neighboring countries (Guinea, Guinea-Bissau, Gambia), where pig occupies an important place in religious, social and festive ceremonies, to which is associated a specific market for charcuterie dedicated to imports, i.e. about 1% of production (Secka 2011;Ossebi et al. 2018;MEPA 2019).However, research conducted on pig farming in Casamance has shown that pig production is faced with certain major constraints, including management methods, nutrition and health (especially African swine fever), which hinder its development, often resulting in huge economic losses (Missohou et al. 2001;Fall 2008;Ossebi et al. 2019;OIE 2020).
Famers believe that difficulties in accessing pig feed open the door to free roaming.The majority of them doubt a possible profitability of their activity in the case of improvement of production factors, including the pigsty, feeding, health monitoring and biosecurity measures (Ossebi et al. 2018).However, the removal of these constraints, according to Santolini (2004), remains a tremendous opportunity to optimize productivity and even profitability of pig farming, as can be achieved by supplementing or fully supplementing with an improved feed based on local and less expensive ingredients (Ayssiwede et al. 2009).Therefore, it would be important for the farmers to find alternative innovations that are adapted and consensual which could improve pig breeding and production practices, hence the interest of the present study.To this end, studies have highlighted the availability in southern Senegal (Casamance) of cashew apple pulp (Anacardium occidentale, L.), an alternative feed resource relatively rich in nutrients -protein, energy, calcium, phosphorus - (Ojewola et al. 2004;Armah 2011;Donkoh et al. 2012;Okpanachi et al. 2016;Ossebi et al. 2018Ossebi et al. , 2019) ) which can be valorized up to 20% in the diet of pigs without any adverse effect on their health and performance (Oddoye et al. 2009;Armah 2011;Donkoh et al. 2012;Yao et al. 2016).
So, the objective of this study is to evaluate, through the 'learning by doing' approach, i.e. 'learning or adoption of innovation by actors through practice', the effects of the application of feed and/or health technological packages in traditional farms on the growth performance, carcass characteristics and profitability of young local pigs in real environments in Casamance, South Senegal.

Study site and period
The present study was conducted over a period of 3 months (July to October 2017) in the 3 regions of Casamance, including Kolda, Sédhiou and Ziguinchor (Figure 1).Located in southern Senegal, Casamance is bounded on the west by the Atlantic Ocean, on the east by a tributary of the Gambia River, on the north by Gambia, and on the south by Guinea Bissau.It covers 28,350 km 2 , or one out of seven of the country's area.The climate is of Sudanian type, with a rainy season from June to October and a dry season which covers the rest of the year with the alternation of three wind flows: the Maritime Trade Wind, the Harmattan and the Monsoon.Average annual temperatures are generally high, between 21 and 37°C, with an average annual relative humidity and rainfall of 69% and 1063 mm (CSE 2007;Konta et al. 2015;Ndiaye et al. 2017).

Selection and equipment of experimental pig farms
Based on the 324 pig farms previously surveyed (Ossebi et al. 2019), 44 farms, including 17 for feed and health technological package, FHT (permanent confinement of pigs fed with the experimental diet and veterinary care), 12 for health technological package, HT (free roaming of pigs with veterinary care and without experimental diet but fed by farmers as they are usually doing with household and agricultural wastes or agroindustrial by-products) and 15 controls (free roaming of pigs without both experimental diet providing and veterinary care, but fed by farmers as they are usually doing with household and agricultural wastes or agro-industrial by-products).Farms were selected in the three regions as shown in Table 1.In the FHT farms (n = 17), rearing practices were improved by the provision of a complete pig feed containing 15% of cashew draff, rearing equipment and health monitoring of the pigs.In the health technological (n = 12) and control (n = 15) farms, on the other hand, rearing practices remained unchanged.But health monitoring has been introduced in the HT farms.The selection farmers took into account the size of their herds (20 pigs with young animals aged from 2 to 3 months and weighing around 10 to 12 kg maximum), and the presence of a traditional piggery which had been improved or was likely to be refurbished.Missions to contact these farmers were organized to raise their awareness, obtain their informed consent and their support to the principles of the study according to animal welfare standards, retain them on the basis of their motivation, their commitment not to sell the animals before the end of the trial and also to provide them with breeding and maintenance equipment.Thus, in addition to maintenance equipment (brooms, portable scales with a capacity of 10 ± 0.05 and 50 ± 0.1 kg, cleaning brushes, shovels, buckets, half a plastic basin, wheelbarrow, sachet of powdered soap, rake, bleach and cresyl), the selected FHT farms received, unlike the control and HT farms, lump-sum financial support for the repair of their barns and all the experimental feed produced.

Animal material
The experimental set-up consisted of 253 local piglets (126 of which were males) between 2 and 3 months of age, supplied by the selected farms at a rate of six piglets per farmer, and weighing an average of 10 kg live body weight (Table 1).They are generally black-coated pigs with a straight muzzle and small erect or horizontal ears, sometimes thrown back (Orgeval 1997;Umutoni 2012).Their narrow body covered with long, coarse bristles is terminated by an elongated or sometimes short snout.On the other hand, we can meet black-coated pigs with a white thoracic belt; white with black or dark spots, grey, brown, sometimes white and rarely brown, then the piebald coat.They are generally small in size with short legs and sometimes with a belly close to the ground, a thigh often flat giving a poorly furnished ham.At the beginning of the trial, all pigs, whether from 'FHT', 'HT' or 'control' farms, were weighed and the males castrated in order to avoid the occurrence of possible early pregnancies during the experiment.

Habitats and sanitary technological packages in the farms
With an average surface area of 9.6 m 2 , the piggeries in the 'FHT' farms are semi-open, built of brick and mostly unpaved.They have a zinc roof with a single slope supported by a wooden frame.They have a concrete floor and a wall height of about 1.2 m.In contrast, the piggeries in the 'HT' and 'control' farms are shaped in the farmers' own way and generally correspond to traditional shelters with mostly uncemented floors as described by Ossebi et al. (2019).
In terms of health and biosecurity measures, all pigs selected as 'FHT' and 'HT' were dewormed at the beginning of the trial with a 1% Biomectin R (Vetoquinol, France) at a rate of 0.3 ml/ 10 kg live body weight, and the deworming was repeated 15 days later.They also received intramuscular Stressvitam R (Vetoquinol, France) based on amino acids (lysine and glycine), trace elements (iron and cobalt) and vitamins (A, D3, E, B1, B3, B5, B6, B12 and PP).The 'FHT' barns were cleaned, washed with a mixture of soapy water and bleach, and disinfected twice a week with cresyl until the end of the trial.During the 90 days trial, the FHT animals were all kept in permanent confinement and reared in groups of six at an average density of 1.6 m 2 per pig.Also, instructions are given to regulate access, and a mandatory disinfection footbath has been installed at the entrance to each of these FHT barns.
As for the 'HT' and 'control' pigs, they are either separated or mixed with other non-selected pigs, or kept on station or left to roam and are only locked up by the farmer during our biweekly weighing appointments.

Feed technology packages in farms
. Formulation and distribution of the pig's experimental diet Cashew apple pulp (residues from the cashew apple after extraction of the juice) and the other ingredients used (Table 2) in the manufacture of the diet were acquired on the local Casamance market, with the exception of oyster shell meal and additives (multivitamin complex -CMV, synthetic lysine and methionine), which were purchased in Dakar.
Based on the nutritional values of these ingredients, which are well known to pig farmers, and the interesting preliminary results on the digestibility of diets based on cashew meal (0-20%) obtained by Ayssiwede et al. (2017Ayssiwede et al. ( , 2021)), a diet for fattening pigs containing 15% of cashew pulp meal was formulated as a test feed for the FHT farms (Table 2).Approximately 12 tons of this diet were manually produced at the Institut Sénégalais de Recherche Agricole (ISRA) of Djibelor station in Ziguinchor, packaged at 50 kg in prolypropylene bags and distributed to the FHT pig farms concerned using a RENAULT minitruck (PTR 25 tons) in the three Casamance regions.Also, a sample of this produced feed and its ingredients were analyzed at the Laboratory of Animal Nutrition and Feeding (LANA) of EISMV in order to determine their composition in dry matter (DM), Ether Extract (EE), ash or crude mineral matter (CM), crude protein (CP), crude cellulose fibre (CF), calcium and phosphorus.The DM, ash, CP and EE contents were obtained according to the methods of the French Association of Standadization (AFNOR 1977), with respectively the Kjeldhal method (N*6.25) and the extraction under reflux by the Soxhlet apparatus for the two last (CP, EE).The CF content was obtained according to AFNOR (1993) based on the Weende method.The calcium and phosphorus contents were determined respectively according to the photometric method of atomic absorption with flame (AFNOR 1984), and the spectrophotometric method of absorption at 430 nm wavelength (AFNOR 1980), while the energy content was calculated basing on the regression equation of Noblet (2000): [ED (MJ/kg DM) = 17.37-0.051MM+ 0.01PB + 0.016MG -0.027CB]. .

Feeding practices
The pigs in the FHT farms were all fed ad-libitum with only the experimental diet produced for this purpose, which is slightly moistened before each feeding to facilitate its gripping by the animals.Instructions are given to the farmers as to the quantities of daily diet served to the pigs, which are readjusted every fortnight after the weighing of the animals.Feed is generally served in concrete, wooden or fixed plastic feeders (drums or sectioned cans); and drinking water is provided adlibitum in drinkers of the same nature as the feeders (two per piggery, with an average length of 0.4-1 m, width of 0.2-0.3m and depth of 0.15-0.2m).
The pigs in the HT and control farms were fed in the usual way of the farmer concerned.They were fed household waste, kitchen scraps, spoiled fruit, greens and agricultural waste depending on the seriousness and motivation of the farmer.

Farms monitoring and data collection
At the zootechnical level, during the 90 days of trial, environmental parameters (temperature and hygrometry) in the barns were regularly recorded three times a fortnight using thermo-hygrometers, as parameters such as live body weight, feed intake and pig mortality.Mortality and feed intake were monitored daily, while individual live body weights of the pigs were taken on an empty stomach every fortnight using a portable electronic scale with a maximum capacity of 50 kg and 100 g accuracy.At the end of the trial, 30 pigs (including 4 in FHT group, 2 in HT group and 4 in controls per region) after a 12-h water diet were weighed, slaughtered by jugular vein bleeding, dressed and completely eviscerated.The carcasses without heads and legs were then weighed.
For economic appraisal, the prices for the purchase of animals at the beginning of the trial, for feeders and drinkers, and for the construction of the pens were obtained by interviewing the breeders.The selling prices per live body weight and per carcass weight of pigs in the different regions were obtained from the farmers involved in the trial and from the local markets.For farmers with a large barn, the overall cost of building the barn was calculated per m 2 to determine the cost of the stall used in the trial.However, the prices of the raw feed materials used to manufacture the experimental diet and the rearing equipment were those applied by the various suppliers.As labour on the different farms was familybased, the cost estimate was based on the work of Ossebi et al. (2019).The annual labour cost in a semi-intensive farm obtained by them was reported per month, then per hour of daily work; the FHT farmers spent about 2 h per day on maintenance and feeding the animals.The majority of farmers used tap water for barn maintenance and watering, while some farmers used shaft water.Water use per FHT farm was estimated at two 20-litre drums per day.The stall and rearing equipment were depreciated over 10 and 3 years respectively and then related to the duration of the trial, with the exception of the plastic feeders and drinkers, which were depreciated over 1 year.The various breeding and economic data collected were recorded in the data sheets prepared for this purpose.

Determination of zootechnical and economic parameters
The collected data were entered and recorded in the Microsoft Excel spreadsheet (version 2010) which was used to calculate the different zootechnical and economic parameters.Thus, the mortality rate (MR), average live body weight (ALBW), average daily feed intake (ADFI), average daily weight gain (ADWG), feed conversion ratio (FCR), average carcass yield (ACY), production cost (PC), gross income (GI), gross margin (GM), feed cost margin (FCM) and net margin (NM) per pig were calculated per region in the same way, according to the following formulas: Due to free roaming and the lack of quantitative data on feeding and sustainable investments, calculations for the HT and control farms were based on estimates from the work of Ossebi et al. (2019).

Data processing and statistical analysis
The various data recorded and processed including zootechnical and economic parameters, and carcass characteristics calculated in Microsoft Excel 2010 spreadsheet were exported to the IBM Statistical Package for the Social Science , where the data from the FHT farms in the three regions and the comparison of the overall averages of the 'FHT', 'HT' and 'control' farms were analyzed using the one-way analysis of variance (ANOVA) test at the 5% threshold.When ANOVA shows a significant difference, it was completed by Duncan's multiple range test to locate the variations between the means of monitored farms in the three regions.
Due to the absence of some data on the HT and control farms, zootechnical (ALBW, ADWG, carcass weight and mortality) and economic (production cost per kg live body weight and kg meat, margins) parameters were used to evaluate the effects of the application of the feed and health technological package on zootechnical and economic performances.Linear regression (glm) was then used to evaluate the effectiveness of the FHT package on the zootechnical and economic performances of the FHT farms compared to the HT and control farms.The response variables were the zootechnical parameters and the mortality of the animals.The explanatory variable was the type of farm (FHT and HT).The application of the package (feed and health) was considered effective when there was a significant difference between the control farms and the HT and FHT farms in terms of zootechnical performance and mortality reduction (p < 0.05).

Nutrient composition of the diet used, environmental parameters and pig mortality
The chemical composition results showed that the fattening pig's feed used contains 83.6% dry matter, 13.5% crude protein, 13.8% ash, 1.8% ether extract and 12.5% crude fibre with a digestible energy rate of 3264 kcal/kg of feed.
Ambient temperatures in the pig farms were relatively high but stable with an overall average of 29.3°C, i.e. 29.1 ± 1°C in Ziguinchor, 29.3 ± 0.8°C in Sédhiou and 28.9 ± 1.3°C in Kolda.Humidity in the barns was volatile, with high values of 81.7 ± 3.4%, 71.5 ± 4.6% and 69.2 ± 5.4% respectively in Ziguinchor, Sédhiou and Kolda, with an overall average about 74% for the study area.
Over the entire duration of the trial, no real illnesses were recorded, besides from a few cases of temporary inappetence at the beginning of the trial and abscesses in some subjects, but these were brought under control by vitamin therapy combined with antibiotics.On the other hand, two (2) deaths were recorded in the 'FHT' farms compared to nine (9) and 14 respectively in the 'HT' and 'control' farms, i.e. a mortality rate of 2% and 14.5 and 15.5% of pigs respectively.The improvement of husbandry, health and feeding practices has therefore significantly reduced the mortality rate of pigs in the FHT farms (Table 3).

Growth-finishing performance of pigs
The evolution of the live body weight of the pigs from each farm (FHT, HT and control) during the trial is illustrated in Figure 2. It can be seen that from the first month until the end of the trial, the pigs from the 'FHT' farms showed a persistent growth superiority (p < 0.05) of almost 92 and 106% respectively compared to those from the 'HT' and 'control' farms.However, over the entire trial period, the average live body weights of the FHT pigs were similar (p > 0.05) across the Casamance region with a final overall average of 29.9 kg (Table 4).The average daily weight gain (ADWG) followed the same trend as the live body weights of the pigs from the different farms with a drastic drop in ADWG in the 'control' pigs up to two months before stabilizing (Figure 3).Thus, the overall ADWG obtained in pigs from the FHT farms was significantly higher (p < 0.05) in Sédhiou (251.6 g/day), followed by Kolda (200.3 g/day) and Ziguinchor (180.4 g/day), i.e. an overall average ADWG equal to 216.8 g/day (Table 4).With regard to the other parameters (daily feed intake, DFI, and feed conversion ratio, FCR) recorded in Table 4 during the trial, the DFI of the pigs from 'FHT' farms was similar between the three regions(p > 0.05), with an overall average of 1107 g/ day/pig.Besides the FCR obtained during the 0-15 and 31-45 d periods of the trial, where pigs from the Sédhiou farms made better use of the diet than those from the Ziguinchor and Kolda farms, the FCR for the other periods and for the entire duration of the trial in these pigs were similar between regions, with an overall average of 5.70; thus, showing that these pigs all made the same use of the diet they were given.On the other hand, there was not possible to record the feed intake of pigs in the 'HT' and 'control' farms because of the free roaming practices in most cases and the irregularity of feed distribution by some farmers.The feed used by these farmers to feed their pigs is variable and consists of kitchen scraps (collected from houses and restaurants), spoiled vegetables and fruits (especially mangoes) and greenery.

Dressing performance of pigs
The dressing performance (weight and carcass yield) of pigs from each of the farms (FHT, HT and control) studied is reported in Table 5.This table showed that the carcass weight of the pigs from the 'FHT' farms (19.1 kg) was significantly higher than that of the pigs from the 'HT' and 'control' farms (9.8 and 9.2 kg) while their carcass yields were similar.Between the pigs from the 'FHT' farms in the three regions, the carcass yields of the pigs from Ziguinchor were significantly higher than those from Kolda and Sédhiou, while the carcass weights were similar (p > 0.05) from one region to another.
In sum, the results of the linear regression (Appendix 1) showed that the feed and health technology packages generated gains in physical productivity in the FHT farms.The Health or sanitary package (improved housing, cleaning materials and products, cleaning and disinfection) significantly reduced the mortality rate of the FHT pigs (p < 0.05).The feed technology package improved live body weight, carcass weight and average daily gain of the test pigs compared to the control subjects.Despite the application of the feed package, the carcass yield of FHT, HT and control pigs remained identical (p > 0.05).

Production costs and economic profitability of pigs fattened under improved 'FHT' conditions on conventional farms
The production costs of a fattening pig, per kg live body weight and meat, the distribution of expenses in the production cost, and the profit margins obtained in the 'control', 'HT' and 'FHT' farms in the different regions are reported in Table 6 and Appendix 2. Overall, it costs CFAF 32,172 to produce a live body weight pig of about 30 kg, CFAF 17,967 for one of 15.6 kg and CFAF 17629 for one of 14.5 kg in 'FHT', 'HT' and 'control' farms respectively.In Casamance, pig is marketed by   the estimated weight and kg of meat and sometimes by live body weight.On a per unit basis, the cost per kg live body weight and per kg of meat was 62-127 FCFA and 126-150 FCFA higher in the control than in the FHT and HT farms.On the 'FHT' farms, these different production costs as well as feed costs per pig and per kg of carcass weight were similar (p > 0.05) between regions.The weak cost of production of a porker revealed that sustainable investments were extremely low (8% in the HT and control group farms versus 3% in the FHT farms) compared to operating expenses (92% versus 97%) of total production costs (Table 6).Feed costs were marginal (6 and 5%) in the HT and control farms compared to 43.3% in the FHT farms.However, in the FHT farms, the feed costs were significantly higher (p < 0.05) in Sédhiou and Kolda regions than in Ziguinchor; and, these costs combined to the purchase of piglets, constituted the most important expenditure items, whereas the latter remained the main item (82 and 85%) in the HT and control farms.
Regardless of the type of farm, the average selling price (kg live weight or kg meat) remained the same.The producer prices observed in the FHT farms were significantly (p < 0.05) different from one region to another, except for the price of pig.The selling price entire carcass of pigs on the 'FHT' farms is clearly higher than on the 'HT' and 'control' farms.The selling price per kg live body weight of pigs in the FHT farms was higher in Sédhiou (p < 0.05) than in the other two regions, while the selling price per kg meat and whole carcass was significantly higher in Ziguinchor.
Thus, the gross feed margins per pig butcher and per carcass in the FHT farms were higher than those in the HT and control farms.Between regions, these margins are significantly (p < 0.05) higher in Ziguinchor, followed by Sédhiou, then Kolda for these FHT farms.The economic appraisal shows that the sale of live pig was an interesting option only for the FHT farms in the Sédhiou region, where positive and higher margins were recorded than in the two others regions.Even though there was no significant difference (p > 0.05) between the net margins allowed by the live pig sale option in the FHT farms of the three regions, with an overall average of 137 FCFA/pig, the latter was much higher than the negative margin (−2179 FCFA and −1343 FCFA respectively) achieved in the HT and control farms.The pig carcass sale option generated significantly (p < 0.05) higher negative margins (losses) for the HT and control farms than for the FHT farms.However, it remains an advantageous option for the FHT farms in the Ziguinchor region, where positive margins were recorded, unlike in the other two regions, with losses being more pronounced in Kolda.However, an increase in the price of kg of live body weight of pig to 1250 CFA francs (15% increase of kg live body weight production cost) and of kg meat of pig to 2050 CFA francs (20% increase of kg meat production cost) generated respectively net margins of 5790 CFA francs/pig and 8549 CFA francs/whole carcass (Appendix 2).
The results of the linear regression (glm) showed that the innovation packages contributed with no significant difference to the reduction of the production cost per kg live body weight and meat and the improvement of the gross and net margins of these two types of products (p > 0.05).

Discussion
The feed used in this study in pigs, even if it has a digestible energy content (3264 kcal/kg) is similar to the one recommended (3200 kcal ED/kg) by INRA (1989).However, its crude protein content (13.5%) is still much lower than the standard recommended by the latter (15-17%) in improved breed pigs.Moreover, since the pigs used in our case are of the local breed (less efficient), this level of protein in the diet seems quite acceptable.
The average ambient temperature recorded in the study area (29.3°C) remained within the range of annual averages (21-37°C) reported by Konta et al. (2015) in this region, and below the 30°C limit below which the local pig could express its growth potential (Nonfon 2005).The high variation in humidity may be due to the fact that the trial was conducted during the so-called 'winter' period, characterized by alternating hot weather and heavy rainfall.
The low mortality rate (2%) in the FHT pigs during the trial is thought to be due to the continuous improvement of their housing, hygiene and husbandry standards, which reduced the microbial load in the barns and kept the animals free from disease and theft.The high mortality rate (14.5%) obtained despite the deworming and vitamin therapy applied in the monitored HT farms can be explained by the lack of appropriate housing, the practice of pigs free-roaming, the irregularity and questionable hygienic quality of the various feeds used (Ossebi et al. 2018).The absence of disease cases in the FHT pigs indicating that the feed used did not induce any adverse effect on the health of the pigs, is in line with observations by Oddoye et al. (2009) and Armah (2011) in Ghana.
The clear superiority of growth performance (live body weight, ADWG, carcass weight) of the FHT pigs compared to the HT and control pigs observed in this study can only be explained by the improvement of their rearing conditions, especially housing, hygiene and feeding.This observation corroborates those of Keambou et al. (2010) in Cameroon who reported that the live body weight of pigs fed an improved diet was almost 60% higher than that of piglets fed a traditional diet.Our results are also similar to those of many authors who have argued that improved sanitary and feeding practices in the traditional rearing system optimize the zootechnical performance of pigs (Agbokounou et al. 2016a(Agbokounou et al. , 2016b)).The lower average live weights at the end of the trial than those obtained by Armah (2011) in Ghana and Acero et al. (2013) in the Philippines would be due to the difference in breeds exploited, initial weight of piglets at the beginning of fattening, duration of fattening and quality of diets used.Indeed, the latter authors used higher protein diets (16.8-18.5% versus 13.5%) and whole male piglets of Large White and Landrace × Large White breeds with higher initial live body weights (13.3 and 10.9 kg) than our local breed FHT pigs.
The low average daily weight gain, ADWG, obtained in the HT and control pigs compared to the FHT pigs during this trial could be explained by the precarious rearing and feeding conditions in which these subjects were kept.Indeed, the irregular distribution of feed associated with the practice of free-roaming by the majority of farmers did not favour the good growth of the pigs.Similar observations were reported by Ayssiwede et al. (2008) and Fall et al. (2017).According to the latter, farmers in the traditional system make a minimum of investment and intervention in their activity, which exposes the animals to massive infestations and various pathologies, resulting in high mortality, slow growth and low productivity.This ADWG recorded in the HT and control pigs at the end of the study (65.4 and 54.8 g/day respectively) remains lower than that (88 g/day) reported by Nonfon (2005) in Benin in traditional farming.Furthermore, the ADWG (216.8 g/day) recorded in the FHT pigs is within the range observed (209.6-360g/day) in several studies conducted in traditional and under improved conditions (Oddoye et al. 2009;Keambou et al. 2010;Armah 2011;Acero et al. 2013) in low-income countries.In most of these studies, dried cashew pulp was fed to pigs at rates of 15-20%.The ADWG obtained in the FHT pigs is however higher than those recorded by Hêdji et al. (2015) in Benin (103.6 g/day), Mopate (2008) in Chad (145 g/ day) who used an improved feed to feed local pigs.
The similar average carcass yields in FHT, HT and control pigs (64.1%, 63.4% and 62.2% respectively) at the end of this study are consistent with that obtained by Armah (2011) in Large White pigs in Ghana.However, these yields are still lower than those obtained (80.1%) in local breed pigs in Africa (Hoffman et al. 2005).
The feed resources used by farmers to feed HT and control pigs during the trial are consistent with those described by Ossebi et al. (2018Ossebi et al. ( , 2019) ) in the same area (Casamance).These results are in agreement with the work of several authors in West Africa (Umutoni 2012;Agbokounou et al. 2016b) who also reported that the diet of pigs in traditional farming was made up of a wide range of feed resources.
The daily feed intake and feed conversion ratio of the FHT pigs remained similar between the different regions, showing that all pigs consumed and valued the feed served in the same way regardless of the experimental site that was not affected.The overall feed intake (1107 g/pig/day) and feed conversion ratio (5.7) obtained in this study are better than those (1143 g/pig/day and 11.20) reported by Mopate (2008) in Chad in local pigs.However, this feed intake is lower than those (1780 and 1350 g/pig/day) obtained by Oddoye et al. (2009) and Armah (2011) respectively in Large White pigs in Ghana in contrast to the feed conversion ratio (5 and 4.4).
The relatively low-cost price (140 FCFA/kg) of our experimental diet compared to that (200 FCFA/kg) of the commercial feed (Ossebi et al. 2018) is due to the use of cheaper local feed resources including cashew apple pulp.The average production cost of a FHT pig of 30 kg is lower (32,760 FCFA) than that obtained for the HT (34,620 FCFA) and control (36,540 FCFA) pigs of equivalent live body weight, showing that the production of pig in traditional farming is more expensive than in improved traditional farming.These results are in agreement with the work of Wadhwa et al. (2015) and Acero et al. (2013), for whom feed improvement results in a reduction in production cost.The higher production cost per kg live body weight of the FHT pigs compared to those (492-966 FCFA/kg) obtained in semi-intensive farming in West Africa (Ayssiwede et al. 2009;Tra Bi Tra 2009;Umutoni 2012;Ossebi et al. 2019) could be explained by the low number of pigs operated, which does not allow for an economy of scale.
However, this cost remains almost similar to the cost (1163 FCFA/kg) obtained by Tra Bi Tra (2009) when the farmer produce his own feed.The distribution of the cost of production shows in several studies (Ayssiwede et al. 2009;Tra Bi Tra 2009;Umutoni 2012;Ossebi et al. 2019) the hegemony of feed costs (48-90%) on the total expenses even if they are negligible in traditional farms according to our estimates (6%), the share of feed consumed during the free-roaming not being determined.Although lower, the feed cost per pig on the FHT farms followed the same trend as those of the previous authors and is justified by the integration of the costs of acquiring fattening piglets.Fixed costs, which amount to 3% of total costs, are significantly lower than those found (14.4-44.1%)by these authors, to the detriment of operating costs (97% of total costs).The difference noted is due to the low investment made by farmers in the traditional system in general and the low access to financial capital.
The gross margin of 1238 FCFA per live pig sold in the FHT farms is far higher than that of the HT farms (61 FCFA), but remains much lower than those (12,867 FCFA/pig) reported by several authors in West Africa (Ayssiwede et al. 2009;Tra Bi Tra 2009;Doumana 2011;Umutoni 2012;Ossebi et al. 2019).The selling price of pigs, the breed exploited and the development of the sector in other countries of the West African sub-region are the elements which create that difference.The high mortality rate and low performance of the control and HT farms explain the higher net margin per live pig sold in the FHT farms.However, this net margin is far lower than those obtained (5435-5902 FCFA) by Ayssiwede et al. (2009) and Ossebi et al. (2019) and the negative margin (−5597 FCFA) reported by Tra Bi Tra (2009).Pig fattening, sold by live body weight, is a profitable activity only in Sédhiou, unlike two other regions (Kolda, Ziguinchor) where this sale option is not recommended due to the significant losses incurred.In Casamance, the sale of FHT pig by carcass weight is only profitable in Ziguinchor (26.1% of the selling price), where the sale of a kg of meat is more interesting (2200 FCFA), whereas it generates significant losses, i.e. 12.8 and 39.6% of the selling price respectively in Sédhiou and Kolda.Thus, considering the acquisition price of the piglet (12,617 FCFA/pig), it is clear that the gross dietary margin generated per piglet does not allow farmers to start a new fattening cycle.There is no real organization of actors, no market place dedicated to the marketing of pigs, and producers are dominated in transactions by Guinean intermediaries who usually set the price of the animal, especially in the dry season (Ossebi et al. 2018).Low margins and even significant losses of invested capital associated with the ravages of pig diseases such as African swine fever do not encourage producers to develop this economic activity.In Casamance, selling by live weight generates investment losses in traditional livestock farming, whereas it provides modest income in improved traditional livestock farming.The low profitability in the latter case is linked to the higher production cost (1092 FCFA) compared to the selling price (1068 FCFA).Only the selling price in Sédhiou (1151 FCFA) is above the production cost (1019 FCFA/kg live body weight), thus giving positive margins.As for the sale in kg of pig, it generates losses for both the traditional rambling livestock (control) and the improved traditional livestock (FHT) because of the high production cost (1706 FCFA) compared to the sale price (1576 FCFA/kg meat).The price charged in Ziguinchor (2200 FCFA) is much higher than the cost price (1625 FCFA/kg meat) and thus allows improved traditional livestock farming to be profitable.Improved housing, health and feeding conditions for the pigs are therefore a major innovation, as they generate more income than traditional livestock farming based on free-roaming.However, the correction of the losses noted for the benefit of the breeders requires, on the one hand, incentive measures on the prices by the harmonisation of the sale per kg of live body weight and the price at the level of the regions.This strategy must be associated to the increase of 15 and 20% of the retail price to the producer respectively of the kg of live body weight and carcass weight.Such an increase would generate net margins of 6852 FCFA and 8549 FCFA (and much more) if the self-consumption of the elements of the 5th quarter (head, legs, abdominal viscera and red offal) of pigs is considered.
Finally, the results show that the innovation packages generate physical productivity gains, but they do not generate significant economic gains in the actual market context.Contrary to what farmers say (Ossebi et al. 2019), pig farming in southern Senegal is not a profitable activity due to the lack of availability of feed, which increases its price and worsens production costs.The prices charged by farmers on the markets, which are themselves set largely by Guinean intermediaries (Ossebi et al. 2018), do not take into account the level of production costs of the animal.Increasing the size of pig farms can be an option for reducing production costs and improving farmers' incomes.Hence the need, at the political level, to encourage the development of pig farming by working to make pig feed available (granting food aid for pigs) on the market to reduce its acquisition price and, in turn, the cost of the final product, to organize the stakeholders into groups to defend their interests with real power to set the price of livestock products, to support prices and to facilitate the creation of a market specific to pig or improve access to it for farmers.
The present study has some limitations.Admittedly, the research methodology adopted was based on the actionresearch approach, but it was certainly adjusted by the farmers to take the rainy season as the study period in order to have the animals ready for marketing during the major sales from October to January.This approach has the advantage of generating results that are accepted and adopted by the beneficiaries involved in the implementation of research activities in collaboration with multidisciplinary teams of researchers (Allard-Poesi and Perret 2003).Although the design took into account the study timeframes (75-120 days) already adopted in other similar scientific research (Keambou et al. 2010;Armah 2011;Mwesigwa et al. 2012;Hêdji et al., 2015), it is still necessary and desirable that this preliminary work be repeated over a longer period of time (one year) in order to confirm or refute the results obtained and generate more solid conclusions.Therefore, the analyses and conclusions of this study are at a more exploratory level.These various limitations need should be taking into account in the design and data analysis to build future studies with a periodicity of at least one year.

Conclusion
The application of feed and health technology packages significantly improved the health status (86.1 and 87.3% reduction in mortality), zootechnical parameters (92.4 and 107% improvement in body weight, 236.4 and 295.5% in average daily gain, and 94.8 and 108.2% in carcass weight) of the FHT pigs compared respectively to the HT and control pigs, but had no significant negative effect on carcass yields in FHT, HT and control pigs, feed intake and feed conversion ratio of FHT pigs in the three study regions, which remained similar.
Economically, despite similar production costs for FHT pigs in the three regions, incomes and margins were significantly better for pig sold live in Sédhiou, and for pig sold as a meat carcass in Ziguinchor; losses were significantly higher in Kolda.The commercial approach is not uniform across the regions and depends mainly on price incentives in relation to sales methods.With the exception of the Ziguinchor region, marketing pig as a meat carcass is not a sustainable option for farmers because of the losses generated and the absence of a market and processing facilities.The sale of pig by the kilogram of live body weight generates a profit of 1238 FCFA/pig, and remains an interesting option, especially for the Sédhiou region.In view of these results, the improvement of pig feed through increased use of local ingredient resources, including cashew apple pulp, and of rearing conditions (housing, hygiene and biosecurity) following the 'learning by doing' approach, constitutes a real opportunity not only to improve the productivity and profitability of traditional pig farms in this study area, but also to promote the massive adoption by farmers of the innovations whose impacts they have witnessed.The sustainability of the pig fattening activity will depend at least on reaching break-even point, and at best on raising the retail price per kg live body weight and carcass weight to 14-80% of the producer price depending on the region.

Disclosure statement
No potential conflict of interest was reported by the author(s).
Means within rows with different superscripts are significantly different at 5% level (p < 0.05).* Anova test.

Figure 3 .
Figure 3. Evolution of the Average Daily Weight Gain (ADWG) of pigs in the feed and health technological (FHT), health technological (HT) and control batches in the improved traditional farms of Casamance, Senegal..
,c: Means within rows with different superscripts are significantly different at 5% level (p < 0.05).*Data per pig estimated from Ossebi et al. (2019) without taking into account regular health monitoring, water and cleaning loads.** Anova test.

Table 1 .
Number of farms and pigs selected by region for the different allotment.

Table 2 .
Composition of different ingredients in the diet used in the pigs during the trial.

Table 3 .
Number of pigs and mortality rate by region according to type of farm.
Figure 2. Evolution of the live body weight of pigs of the feed and health technological (FHT), health technological (HT) and control batches in the improved traditional farms of Casamance in Senegal.

Table 4 .
Average daily gain (ADG), Average daily feed intake and feed conversion ratio (ADFI and FCR) of pigs according to type of farm and 'test' farms in the three regions during the trial periods of Casamance in Senegal.

Table 5 .
Carcass weights and yields of pigs from traditional 'control' and 'test' farms monitored during the trial in Casamance, Senegal.
a, b: Means within rows with different superscripts are significantly different at 5% level (p < 0.05).* Anova test.

Table 6 .
Economic analysis of pork production, kg live weight and kg meat in the Casamance region of Senegal.