24th Int. Symp. “Animal Science Days”, Ptuj, Slovenia, Sept. 21st−23rd, 2016. Acta argiculturae Slovenica, Supplement 5, 103–108, Ljubljana 2016 COBISS: 1.08 Agris category code: L01 IDENTIFICATION OF BREEDING OBJECTIVES IN COMMUNITY- BASED GOAT BREEDING PROGRAMMES IN MALAWI Wilson NANDOLO 1, 2, Maria WURZINGER 3, Gábor MÉSZÁROS 4, Curt VAN TASSELL 5, Timothy GONDWE 6, Henry MULINDWA 7, Doreen LAMUNO 8, Johann SÖLKNER 9 Identification of breeding objectives in community-based goat breeding programmes in Malawi 1 Lilongwe University of Agriculture and Natural Resources, Department of Animal Science, P.O. Box 219, Lilongwe, Malawi, e-mail: wilsonnandolo@bunda.luanar. mw 2 University of Natural Resources and Life Sciences, Division of Animal Sciences, Gregor-Mendel-Straße 33, 1180 Vienna, Austria 3 Same address as 2, e-mail: maria.wurzinger@boku.ac.at 4 Same address as 2, e-mail: gabor.meszaros@boku.ac.at 5 Animal Genomics and Improvement Laboratory, Beltsville, Maryland, 20705-2350, USA, e-mail: Curt.VanTassell@ars.usda.gov 6 Same address as 1, e-mail: tgondwe@bunda.luanar.mw 7 National Livestock Resources Research Institute, P.O Box 96, Tororo, Uganda, e-mail: mulindwaha@yahoo.com 8 Same address as 2, e-mail: doreen.lamuno@gmail.com 9 Same address as 2, e-mail: johann.soelkner@boku.ac.at ABSTRACT A study was carried out to identify farmers’ goat breeding objectives as a preparatory step towards the setting up of community based goat breeding programme in Mzimba and Nsanje Districts of Malawi. Participatory workshops were held with farmers to get a general idea of their goat breeding objectives. Farmers with more than 7 does were asked to select, with reasons, the best, average and worst does from their flocks. Productive, reproductive and morphometric data for the ranked does were recorded. There was some agreement between the farmers’ ranking and the trait measure- ments taken on the goats that were ranked, suggesting that proper measurement of those traits will assist in meeting the farmer’s objectives. The main breeding objective was to increase meat production, but the two sites had different path- ways towards this objective. Farmers in Mzimba were more interested in higher growth rates and higher mature weights, while farmers in Nsanje were more interested in improving survival traits, which are more directly related to off-take rates. It is suggested that different selection strategies should be used for each of the sites in order to realise genetic gain that meets the needs of the farmers. Key words: goats, goat breeding, breeding objectives, community-based breeding, Malawi, Africa 1 INTRODUCTION Community based breeding programs (CBBP) at- tempt to achieve genetic improvement of livestock popu- lations by direct involvement of farmers from the design of such programs to actual breeding actions. Effecting a CBBP is a process that requires a bottom-up approach (Mueller et al., 2015), which necessitates that research- ers and other development workers must fully under- stand the farmer’s circumstances before the program is set up. The process typically leads to some changes in the farmers’ farming management practices, and preferably transcends to a change in their attitudes towards animal breeding. Practically, the decisions farmers make about their animals are multifactorial (Tabbaa and Al-Atiyat, 2009), necessitating the adoption of a holistic approach that attempts to address all these factors (Abera, 2014; Ge- tachew, 2008). One of the necessary steps in this process is the identification of the farmers’ breeding objectives. A breeding objective defines the direction in which the farmer would like to go towards satisfying his demand for specific products and services from the animals (Sölkner et al., 2008), which usually translates into increase of profit. Although such desired trajectories of improve- ment may be different between individual farmers, there is generally a lot of similarity among farmers subjected Acta agriculturae Slovenica, Supplement 5 – 2016104 W. NANDOLO et al. to similar production environments and market forces, so that it is possible to have an idea of the direction in which farmers in an area would like their animals to im- prove. In conventional livestock breeding programmes, relative values of traits are derived from economic values (Goddard, 1998). Where such determination cannot be made easily, analysis of farmers’ preferred trait levels in the animals may be an indirect indication of what consti- tutes a good animal, the successful pursuit of which con- stitutes achievement of the breeding objectives (Duguma et al., 2011). This can be done through farmer interviews, farmer workshops, or ranking of own or other people’s animals (Wurzinger et al., 2006; Mirkena, 2010). This study was carried out to identify farmers’ goat breeding objectives as preparatory step towards the setting up of community based goat breeding programmes in Malawi. 2 MATERIALS AND METHODS 2.1 DESCRIPTION OF THE STUDY SITES The study was carried out in CBBP sites Mzimba and Nsanje Districts of Malawi. Mzimba District is in Northern Malawi, and the CBBP site is located on lati- tude −11.356, longitude 33.650 at an altitude of 1150 m. The CBBP site in Nsanje District is located on latitude −16.544, longitude 35.010. The area is at 70 m, and re- ceives relatively low rainfall of around 750 mm and is also prone to floods and droughts. Crop and livestock production are the most important economic activities in both sites. Based on a survey of production systems carried out prior to this study, goats are the most impor- tant livestock species, followed by chickens. Malawi had about 6,545,000 goats in 2015. The mean flock sizes in Mzimba and Nsanje are 12.9 and 9.0, respectively, and goats constitute 47 % and 38 % of the tropical livestock units (TLUs) in Mzimba and Nsanje, respectively. Goats in Nsanje are mostly kept for cash from live sales while in Mzimba the goats are kept for multipurpose reasons. Goats in the study areas are mostly of the small East Afri- can breed, managed in an extensive grazing system with tethering during the rainy season in some cases. Overall, mature bucks make up less than 10 % of the goat popula- tion, while does make up over 47 %, leading to a buck- doe ratio of no more 1 to 5. The number of young goats in both sites is lower than would be expected, indicat- ing possible attrition due to high kid mortality or sales. Mortality rates are very high in both sites, especially in young and adult male animals. Over 80 % of the mortali- ties are due to diseases (diarrhoea, pneumonia and para- sitic infestations). Nsanje has higher mortalities (35 %) than Mzimba (10 %). Majority (90 %) of the households do not control goat breeding, mostly because the animals graze together (89 %). As a result, cases of bucks mat- ing with their mothers, daughters and sisters are very common. This kind of communal goat management is common in Southern Africa (Gwaze et al., 2009). Goat off-take rates are very high, especially in Nsanje. On aver- age, a farmer in Nsanje sells about 3.5 live goats per year, compared to 1.5 in Mzimba (against mean flock sizes of 9.0 and 12.9, respectively). Farmers in Nsanje rely heavily on the goats as a safety net through times of adversities such as floods and droughts, both of which occur nearly yearly. It has been demonstrated that the goat marketing value chain in Malawi has the potential to be very profit- able (Maganga et al., 2015; Banda et al., 2011), and if the goats are genetically improved, the farmers can get more money, as the demand for goat meat is projected to keep increasing (Banda, 2008). 2.2 STUDY DESIGN Participatory farmer workshops were conducted to find out the farmers’ goat production objectives in each site. The researchers then visited individual farmers with more than 7 does in the morning when the goats were still in the pens. The total number of such farmers was 25 (14 in Nsanje and 11 in Mzimba). The farmers were asked to rank their does in terms of superiority for achieving their objectives in goat production. They were asked to identify, with reasons, their best, average and worst does. Morphometric measurements were taken on the iden- tified does. The measurements included doe body size measurements (length, heart girth and heights at with- ers and rump) and body weights. Historical records were also taken for reproduction (number of kids born, num- ber of times the doe kidded and number of kids weaned, from which twinning rate, litter size and survivability were calculated); health (body condition scores, disease incidence); production (milk yield) and economic values (the amount of money the farmer would be willing to pay for each of the identified does if the farmer wanted to buy the doe for breeding purposes). 2.3 DATA ANALYSIS Non-parametric statistics were used to describe the relationships between most qualitative phenotypes and trait level preferences. The morphometric measurements and quantitative traits of importance (twinning rate, lit- ter size and kid survivability) were analysed using the general linear model procedure in IBM SPSS (IBM Cor- poration, 2013) using a model with parity, rank and site Acta agriculturae Slovenica, Supplement 5 – 2016 105 IDENTIFICATION OF BREEDING OBJECTIVES IN COMMUNITY-BASED GOAT BREEDING PROGRAMMES IN MALAWI as the explanatory variables. For survivability, the model also included the twinning rate as a covariate. This model assumed that the rigor of ranking was the same between the farmers. 3 RESULTS AND DISCUSSION 3.1 BREEDING OBJECTIVE FROM A BROADER PERSPECTIVE Farmers in Nsanje indicated that they were mostly interested in getting more money when they sell live ani- mals. The farmers wanted animals that grow fast, with higher mature weight, and preferably with high twinning rate (in order to sell more animals per year). Farmers in Mzimba indicated that one of their major challenges is that goat body sizes have been getting smaller over the decades, and according to them, smaller animals implied less meat and money from live sales, so they were very keen to select for bigger animals, especially because the goats are also important for consumption in this area, compared to Nsanje where goats are almost exclusively used as moving banks. 3.2 DOE RANKING AND SELECTION CRITERIA Does were categorized as best or average mostly for their good mothering ability and high kid survival; big size, high twinning rate, kid sex ratio skewed towards females and ability to remain in good health condition throughout the year (Fig. 1). Kid sex ratio was not cap- tured in this study. Worst does were those with poor body condition and health and those that were troublesome in terms of management (such as by constantly wandering away from the rest of the flock). Other reasons for catego- rizing a doe in the worst category were poor mothering ability and low kid survival and single births (low twin- ning rate). The criteria for ranking does were different by site. Farmers in Nsanje were more interested in doe behav- ior, doe health, kid sex ration, kid survival, kidding in- tervals and mothering ability, while farmers in Mzimba were more interested in doe size and growth rate (Fig. 2). Doe behavior was especially important in Nsanje, where about 24.1 % of the farmers do not have goat pens, com- pared to Mzimba where the percentage of such farmers is 1.7 %. Having goats that are easy to manage is therefore important for these farmers. Figure 1: Percentages of positive and negative rankings within ranking criteria by group of doe Figure 2: Percentages of positive and negative rankings within ranking criteria by site Acta agriculturae Slovenica, Supplement 5 – 2016106 W. NANDOLO et al. 3.3 RELATIONSHIP BETWEEN DOE RANKS AND MORPHOMETRIC AND OTHER TRAIT MEAS- UREMENTS Most of the quantitative measurements (weight, body length, heart girth, height at withers, height at rump, litter size and kid survivability) were nearly nor- mally distributed within the sites, with coefficients of variation of 21 %, 10 %, 7 %, 6 %, 5 %, 30 % and 31 %. Twinning rate was highly variable, with a coefficient of variation of about 76 %. The body condition scores of the does were different between the ranks (3.1, 3.5 and 2.6 for best, average and worst does, respectively). None of the farmers who ranked their does reported milking the goats, and there was no clear relationship between incidence of diseases and rank of the doe. Similarly, there was no relationship between doe rank and doe coat colours. The mean amounts of money for which the farmers said they would be willing buy the does in each rank for breeding were MWK19,527, MWK17,157 and MWK16,463 for best, average and worst animals, respec- tively (€1 ~ MWK480), but there were no significant dif- ferences between these values. Least squares means of doe body weight, body length, heart girth, height at withers, height at rump, lit- ter size, twinning rate and survivability were, 33.41 kg, 62.43 cm, 75.99 cm, 61.14 cm, 62.98 cm, 1.56, 56.71 % and 86.95 %, respectively (Table 1). Banda et al. (1993) reported 32.4 kg, 68.9 cm, 74.9 cm and 60.5 cm for weight, body length, heart girth and height at withers, respectively for adult female East African goats on some government farms. This represents an increase of about 1 kg in body weight over 23 years. This breeding program intends to achieve 0.2 to 0.4 kg in 3 years. In comparison, female Iramba goats in Tanzania (also East African) have been reported to have 24.2 kg, 57.7 cm, 53.3 cm, 68.1 cm and 55.1 cm for body weight, height at withers, height at rump, heart girth and body length, respectively (Cheny- ambuga et al., 2012). The litter sizes seemed to have im- proved from 1.35 reported by Karua and Banda (1993) and Wilson (1988) on government and research institu- tional farms in Malawi. Heart girth (a measure of body size and weight) decreased down the ranks from best to worst, implying that farmers could indeed be interested in bigger animals. Does ranked average had the highest kid survival rates followed by those ranked best. Does in Nsanje had higher body weights, body lengths and twinning rates (Fig. 3). Although the dif- ferences in body lengths and body weights between the two sites do not shed light on the trends of body sizes of Measurement P Rank Overall Best Average Worst Body weight 0.149 35.41 ± 1.28 32.21 ± 1.47 32.60 ± 1.54 33.41 ± 0.98 Boy length 0.272 63.80 ± 1.10 62.00 ± 1.23 61.46 ± 1.14 62.43 ± 0.79 Heart girth 0.014* 78.36 ± 1.02 75.20 ± 1.14 74.40 ± 1.06 75.99 ± 0.72 Height at withers 0.159 61.94 ± 0.74 61.44 ± 0.83 60.05 ± 0.77 61.14 ± 0.52 Height at rump 0.132 63.82 ± 0.70 63.23 ± 0.79 61.91 ± 0.73 62.98 ± 0.50 Litter size 0.354 1.640 ± 0.082 1.55 ± 0.09 1.48 ± 0.08 1.56 ± 0.06 Twinning rate 0.098 68.67 ± 7.26 47.92 ± 8.13 53.5 ± 7.5 56.71 ± 5.13 Survivability 0.016* 89.92 ± 5.60 95.99 ± 5.98 75.27 ± 5.5 86.95 ± 3.74 Table 1: The distribution of weight, height at withers, body length, heart girth, height at rump, twinning rate and kid survivability * = significant at P = 0.05 Figure 3: Least squares means of body length (cm), twinning rate (%) and body weight (kg) by site Acta agriculturae Slovenica, Supplement 5 – 2016 107 IDENTIFICATION OF BREEDING OBJECTIVES IN COMMUNITY-BASED GOAT BREEDING PROGRAMMES IN MALAWI the goats in Mzimba, they do confirm the concern of the farmers in Mzimba about the small sizes of their goats. Information on growth rates, which was not cap- tured in this study, for the Malawian Small East African goats is scanty and old. Banda et al. (1993) reported pre- weaning growth rates of up to 45 grams/day on govern- ment farms. Growth rates reported for the Small Eat Af- rican goats in the region are below 50 grams/day, such as 28 grams/day reported by Safari et al. (2005). Pre-wean- ing growth rates of up to 50 grams/ day can therefore be expected. 3.4 BREEDING OBJECTIVES FROM A NARROW- ER PERSPECTIVE The expressed interest in bigger goats was only part- ly reflected by the association between the morphometric measurements, with reported worst does not being sub- stantially smaller than reported best goats. However, it is apparent from the association between doe rank and the farmers’ reasons for ranking animals that the farm- ers’ goal in both sites was to get more money from sale of meat or live goats. This implies that the farmers’ breeding objective was to produce more meat since nearly all the goats that are sold are ultimately destined for slaughter for meat (Maganga et al., 2015). More money from sale of live goats can be achieved through faster growth rates, higher twinning rates and higher survivability. However, a balance needs to be stricken between these traits, since kids born in multiples tend to have lower birth weights and lower growth rates in the pre-weaning period. In some cases, kids born in multiples may even have lower survivability, which might explain why kid mortality is higher in Nsanje (Fig. 3). Assuming that such kids born in multiples survive, they may fetch lower prices than singletons of the same age if sales are based on live weight, which is usually the case in Mzimba, so having heavier animals is important. On the other hand, sales of kids in Nsanje are generally not based on live weight, so that weight differences between individuals of the same age may not influence the prices much. Subsequently having more animals to sell is more important than having bigger animals, although in the long run, bigger and heavier animals may fetch better prices. The breeding program implemented intends to improve the goat marketing functions towards differen- tial product pricing. This points to the need to have dif- ferent selection strategies for each of the sites in order to realise genetic gain that meets the needs of the farmers. The data from the routine recording that is going to be effected for the breeding programs will establish appro- priate balances between the traits of importance in the two sites. 3.5 BREEDING GOAL AND EXPECTED SELEC- TION INTENSITIES The breeding goal of the goat farmers may be sum- marised as: Profit = Behaviour traits + Fitness traits + Productive traits + Reproductive traits The behaviour traits include the ease of manage- ment and mothering ability, both of which are not easy to measure objectively, although mothering ability can be indirectly related to kid survival. Fitness traits include health and kid survival. Productive traits include weight and growth rate. Reproductive traits include kid sex ra- tio, kidding interval and twinning rate. It is expected that the weights on the traits are going to be different between the sites to reflect the differences in the farmers’ breed- ing objectives. Selection intensity in does is expected to be very low, since farmers normally keep all females for breeding. The breeding program is expected to achieve male selection intensity (i) corresponding to selection of less than a quarter of the candidates (i > 1.27). 4 CONCLUSION The results suggest that there was some agreement between the farmers’ ranking and the measurements taken on the goats that were ranked, especially for heart girth and survivability. It is expected that with appropri- ate recording this agreement will become clearer. The main goat breeding objective was to increase meat pro- duction. Farmers in Mzimba were more interested in higher growth rates and higher mature weights, while farmers in Nsanje were more interested in improving traits aimed at increasing off-take rates. It may be neces- sary to have different selection strategies for each of the sites in order to realise genetic gain that meets the spe- cific needs of the farmers. 5 REFERENCES Abera, B. (2014). On-Farm Phenotypic Characterization of Indigenous Sheep Types in Selale Area, Central Ethiopia. Journal of Veterinary Science and Technology, 05(03). Re- trieved from http://doi.org/10.4172/2157-7579.1000180. Banda, J. (2008). Revoutionalisation of the livestock industry in Malawi. The 12th University of Malawi Inaugural Lecture, Acta agriculturae Slovenica, Supplement 5 – 2016108 W. NANDOLO et al. Crossroads Hotel, 7th November 2008, Lilongwe (p. 40). Li- longwe: The University of Malawi, Bunda College of Agri- culture. Banda, J. W., Ayoade, J. A., Karua, S. K., Kamwanja, L. A. (1993). Local malawi goat. World Animal Review, 74(75), 49–57. Banda, L.J, Dzanja, J. L. and Gondwe, T. N. (2011). Goat Marketing Systems and Channels in Select- ed Markets of Lilongwe District in Malawi. Jour- nal of Agricultural Science and Technology A1. Retrieved from http://www.davidpublishing.com/david- publishing/Upfile/3/1/2012/2012030102583455.pdf. Chenyambuga, S. W., Komwihangilo, D. M., Jackson, M. (2012). Production performance and desirable traits of Small East African goats in semi-arid areas of central Tanzania. Live- stock Research for Rural Development, 24(7). Retrieved from http://www.lrrd.org/lrrd24/7/chen24118.htm. Duguma, G., Mirkena, T., Haile, A., Okeyo, A. M., Tibbo, M., Rischkowsky, B., Sölkner, J., Wurzinger, M. (2011). Iden- tification of smallholder farmers and pastoralists’ prefer- ences for sheep breeding traits: choice model approach. Animal : An International Journal of Animal Bioscience, 5(12), 1984–92. Retrieved from http://doi.org/10.1017/ S1751731111001029. Getachew, T. M. (2008). Characterization of Menz and Afar indigenous sheep breeds of smallholders and pastoral- ists for designing community-based breeding strategies in Ethiopia. Retrieved from http://agris.fao.org/agris-search/ search.do?recordID=QM2008000148. Goddard, M. E. (1998). Consensus and debate in the definition of breeding objectives. Journal of Dairy Science, 81, Suppl 2, 6–18. Retrieved from http://doi.org/10.3168/jds.S0022- 0302(98)70150-X. Gwaze, F. R., Chimonyo, M., Dzama, K. (2009). Communal goat production in Southern Africa: A review. Tropical Ani- mal Health and Production, 41(7), 1157–1168. Retrieved from http://doi.org/10.1007/s11250-008-9296-1. IBM Corporation. (2013). IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp. Karua, S., Banda, J. (1993). The performance of the Small East African goats and their Saanen crosses in Malawi. In Small Ruminant Research and Development in Af- rica: Proceedings of the Second Biennial Conference of the African Small Ruminant Research Network, AICC. Re- trieved from http://agris.fao.org/agris-search/search. do?recordID=QM19950138574. Maganga, A. M., Chigwa, F. C. and Mapemba, L. D. (2015). Goat and goat meat markets in selected districts of Malawi: value chain, structure and efficiency. Livestock Research for Rural Development 27. Retrieved from http://www.lrrd.org/ lrrd27/5/maga27097.htm. Mirkena, T. (2010). Identifying breeding objectives of smallhold- ers/pastoralists and optimizing community-based breeding programs for adapted sheep breeds in Ethiopia. Doctoral Thesis. Universty of Natural Resources and Life Sciences, Vienna. Retrieved from https://zidapps.boku.ac.at/ab- stracts/download.php?dataset_id=8177&property_id=107. Mueller, J. P., Rischkowsky, B., Haile, A., Philipsson, J., Mwai, O., Besbes, B., Valle Zárate, A., et al. (2015). Community- based livestock breeding programmes: Essentials and ex- amples. Journal of Animal Breeding and Genetics, 132(2), 155–168. Retrieved from http://doi.org/10.1111/jbg.12136. Safari, J., Mushi, D. E., Mtenga, L. A., Eik, L. O., Kifaro, G. C., Muhikambele, V. R. M., Ndemanisho, E. E., Maeda Machang’u, A. D., Kassuku, A. A., Kimbita, E. N. and Ul- vund, M. (2005). A note on growth rates of local goats and their crosses with Norwegian goats at village level in Tan- zania. Livestock Research for Rural Development 17. . Re- trieved from http://www.lrrd.org/lrrd17/4/safa17047.htm. Sölkner, J., Grausgruber, H., Okeyo, A. M., Ruckenbauer, P., Wurzinger, M. (2008). Breeding objectives and the rela- tive importance of traits in plant and animal breeding: A comparative review. Euphytica, 161(1), 273–282. Retrieved from http://doi.org/10.1007/s10681-007-9507-2. Tabbaa, M. J., Al-Atiyat, R. (2009). Breeding objectives, selec- tion criteria and factors influencing them for goat breeds in Jordan. Small Ruminant Research, 84(1–3), 8–15. Retrieved from http://doi.org/10.1016/j.smallrumres.2009.03.007. Wilson, R. T. (1988). Small ruminant production sys- tems in tropical Africa. Small Ruminant Research, 1(4), 305–325. Retrieved from http://doi.org/10.1016/0921- 4488(88)90058-2. Wurzinger, M., Ndumu, D., Baumung, R., Drucker, A. G., Ok- eyo, A. M., Semambo, D. K., Sölkner, J. (2006): Assessing stated preferences through the use of choice experiments: Valuing (re’)production versus aesthetics in the breeding goals of Ugandan Ankole cattle breeders. In 8th World Con- gress on Genetics Applied to Livestock Production, Belo Hori- zonte, Brazil, 13.–18. 8. 2006. Instituto Prociência.