Animal Genetic Resources/Ressources Génétiques Animales/Recursos

20786336_54-0_20786336_52-0 04/06/14 3:47 PM Page 1 54 2014 CONTENTS Editorial...... i 54 Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh C.H. Ritchil, M.M. Hossain and A.K.F.H. Bhuiyan...... 1 Caractéristiques phénotypiques de la chèvre du sahel au Niger par analyse des indices de

ANIMAL GENETIC RESOURCES • RESSOURCES GÉNÉTIQUES ANIMALES • RECURSOS GENÉTICOS ANIMALES GENÉTICOS RECURSOS • ANIMALES GÉNÉTIQUES RESSOURCES • RESOURCES GENETIC ANIMAL 2014 primarité et des paramètres qualitatifs M. Mani, H. Marichatou, M. Issa, I. Chaïbou, A. Sow, M. Chaïbou and J.G. Sawadogo ...... 11 ANIMAL GENETIC ISSN 2078-6336 Caractérisation de la chèvre du sahel au Niger par analyse des indices biométriques et des paramètres phénotypiques quantitatifs RESOURCES M. Mani, H. Marichatou, M.M.M. Mouiche, M. Issa, I. Chaïbou, A. Sow, M. Chaïbou and J.G. Sawadogo ...... 21 Morphological, reproductive and productive characteristics of Sudanese native chicken an international journal C.E. Wani, I.A. Yousif, M.E. Ibrahim, H.H. Musa and K.M. Elamin ...... 33 Phenotypic characterization of indigenous chicken ecotypes in the north Gondar zone, Ethiopia Addis Getu, Kefyalew Alemayehu and Zewdu Wuletaw ...... 43 Studies on morphometrical traits of Boran bulls reared on two feedlots in Southern Ethiopia Sandip Banerjee, Mohamed Beyan Ahmed and Girma Tefere ...... 53 Analysis of the body structure of Djallonke sheep using a multideterminant approach Peter T. Birteeb, Sunday O. Peters and Michael O. Ozoje ...... 65 RESSOURCES Variation in qualitative traits in Bhutanese indigenous chickens T. Tashi and N. Dorji ...... 73 GÉNÉTIQUES ANIMALES

Genetic polymorphism of ␣S1 casein in Guéra and Sahel goat D. Traoré, Y. Sanogo, R. Fané, A. Touré, O. Cissé and A.H. Babana...... 79 un journal international Estimación de heredabilidad de la curva de crecimiento en el borrego de raza Chiapas en México A.C. Méndez-Gómez, R. López-Ordaz, M. Peralta-Lailson, R. Ulloa-Arvizu, P. Pedraza-Villagómez, F.J. Ruiz-López, J.M. Berruecos-Villalobos and C.G. Vásquez-Peláez ...... 85 Caracterización genética de la población bovina criolla de la Región Sur del Ecuador y su relación genética con otras razas bovinas L. Aguirre Riofrio, G. Apolo, L. Chalco y A. Martínez ...... 93 Genealogical and population viability analysis of a conservation nucleus of Brazilian Bergamasca sheep RECURSOS H. Carneiro, S.R. Paiva, H. Louvandini, R.M. Miranda and C. McManus ...... 103 Population structure and genealogical analysis of the Brazilian Crioula Horse GENÉTICOS ANIMALES F.C. Maciel, C.D. Bertoli, J. Braccini Neto, J.A. Cobuci and S.R. Paiva and C.M. McManus...... 115 Pedigree and population viability analyses of a conservation herd of Moura pig una revista internacional H. Carneiro, S.R. Paiva, M. Ledur, E.A.P. Figueiredo, V.H. Grings, F.C.P. Silva and C. McManus ...... 127 Evaluación del semen criopreservado de toros Curraleiro Pé Duro, perteneciente al Banco Brasilero de Germoplasma Animal I.S. Montesinos, J.O. Carvalho, J.V. Malaquias, E. Arnhold, G.E. Freneau, M.A.N. Dode, M.C.S. Fioravanti y J.R.B. Sereno...... 135 Unique cultural values of Madura cattle: is cross-breeding a threat? T.S.M. Widi, H.M.J. Udo, K. Oldenbroek, I.G.S. Budisatria, E. Baliarti and A.J. van der Zijpp ...... 141 Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia Sandip Banerjee, Mohammed Beyan and Hiwot Bekele ...... 153 Phenotypic characterization and description of production systems of autochthonous sheep breeds in Kosovo H. Bytyqi, R. Baumung, H. Mehmeti and B. Fuerst-Waltl ...... 163 Evaluation des pratiques paysannes de conservation in situ du taurin Baoulé au Sud-Ouest du Burkina Faso L.Y. Mopate, M.J-B. Kamuanga, S. Hamadou and C-Y. Kaboré-Zoungrana ...... 171 Influence of coat colour on Chamarita sheep phenotypes, behaviour, welfare and performances M. Pascual-Alonso, G.C. Miranda-de la Lama, L. Aguayo-Ulloa, M. Villarroel, S. Alierta and G.A. Maria ...... 179 Recent Publications 185 FAO Animal Genetic Resources, 2014, 54, i. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000186

Editorial

Dear reader, interested in Aichi Target 13 which refers directly to the genetic diversity of farmed and domesticated animals. The 54th volume of our journal includes 20 papers describing different aspects of animal genetic resources and their The Commission on Genetic Resources for Food and management in Africa, Asia, Europe and Latin America. Agriculture, at its Fourteenth Regular Session in April The current volume shows that our journal is truly inter- 2013, requested FAO to prepare The Second Report on national due to its international readership and authorship. the State of the World’s Animal Genetic Resources for Furthermore the papers presented cover a range of topics Food and Agriculture, as an update to the first report pub- th from characterization of breeds to the description of their lished in 2007.5 By 25 of March 2014, 100 countries had production systems. Similar to the 53rd issue, the high per- submitted country reports to FAO and another 30 countries centage of papers dealing with phenotypic and or genetic had indicated that they were in the final stages of preparing characterisation indicates the continuing importance of their reports. International organizations and regional focal the first strategic priority area of the Global Plan of points for the management of animal genetic resources6 Action for Animal Genetic Resources1. have also been invited to report on their activities supporting the implementation of the Global Plan of Action for Animal genetic resources represent an important part of Animal Genetic Resources.7 biodiversity. Therefore I would also like to draw your attention to the next edition of the Global Biodiversity The Second Report will be made available to the Eighth Outlook, the flagship publication of the Convention on Session of the Intergovernmental Technical Working Biological Diversity2. It is a periodic report that sum- Group on Animal Genetic Resources for Food and – marizes the latest data on the status and trends of biodiver- Agriculture, which is scheduled to take place 26 28 8 sity and draws conclusions relevant to the further November 2014 in Rome. The Working Group will also implementation of the Convention. The reports are based advise the Commission on whether the Global Plan of on different types on information including national Action for Animal Genetic Resources needs to be updated. ’ 9 reports, scientific literature, information from the Please encourage your country s National Coordinator to Biodiversity Indicators Partnership and supplementary attend this important meeting. You might also consider studies. The fourth edition of the Global Biodiversity providing him/her with your comments on the Second Outlook (GBO-4)3 will be prepared to provide a midterm Report, which will be published online six weeks prior 10 assessment towards the progress in the implementation to the meeting at our webpage . of the Strategic Plan for Biodiversity 2011–2020 and the Yours sincerely, 4 Aichi Biodiversity Targets. You might me be especially Roswitha Baumung

5 www.fao.org/docrep/010/a1250e/a1250e00.htm 6 http://www.fao.org/Ag/AGAInfo/programmes/en/genetics/regional_collaboration. html 7 http://www.fao.org/docrep/010/a1404e/a1404e00.htm 8 http://www.fao.org/Ag/AGAInfo/programmes/en/genetics/ 1 http://www.fao.org/docrep/010/a1404e/a1404e00.htm Intergovernmental_process.html 2 9 http://www.cbd.int/ http://dad.fao.org/cgi-bin/EfabisWeb.cgi?sid=-1,contacts 3 10 http://www.cbd.int/en/gbo4 http://www.fao.org/ag/againfo/programmes/en/genetics/Intergovernmental_process. 4 http://www.cbd.int/sp/targets/ html

i Animal Genetic Resources, 2014, 54, iii. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000198

Éditorial

Cher lecteur, Il se peut que le 13ème Objectif d’Aichi vous intéresse tout ’ ème particulièrement vu qu il se rapporte directement à la Le 54 volume de notre journal inclut 20 articles diversité génétique des animaux domestiques. décrivant différents aspects des ressources zoogénétiques et leur gestion en Afrique, Asie, Europe et Amérique À sa Quatorzième Session Ordinaire en Avril 2013, la ’ Latine. Le présent volume fait ressortir le caractère vrai- Commission des Ressources Génétiques pour l Alimentation ’ ’ ment international de notre journal vue l’internationalité et l Agriculture a demandé à la FAO d élaborer Le ’ des lecteurs et des auteurs. En outre, les articles Deuxième Rapport sur l État des Ressources Zoogénétiques ’ ’ présentés couvrent un large éventail de sujets allant de la pour l Alimentation et l Agriculture dans le Monde,comme 5 caractérisation des races à la description de leurs une mise à jour du premier rapport publié en 2007. Vers le systèmes de production. Comme dans le 53ème volume, 25 Mars 2014, 100 pays avaient soumis leurs rapports natio- ’ le haut pourcentage d’articles traitant de la caractérisation naux à la FAO et 30 autres pays avaient indiqué qu ils se trou- phénotypique et génétique indique que le premier domaine vaient dans les dernières étapes de la préparation de leurs prioritaire du Plan d’Action Mondial pour les Ressources rapports. Des organisations internationales et les centres de Zoogénétiques1 demeure toujours important. coordination régionaux pour la gestion des ressources zoogénétiques6 ont aussi été invités à informer sur leurs Les ressources zoogénétiques représentent une part impor- activitésdesoutienàlamiseenapplicationduPlan tante de la biodiversité. Ainsi, je voudrais aussi attirer d’Action Mondial pour les Ressources Zoogénétiques.7 votre attention sur la prochaine édition de Perspectives Le Deuxième Rapport sera mis à disposition à la Huitième Mondiales de la Diversité Biologique, publication phare Session du Groupe de Travail Technique Intergouvernemental 2 ’ de la Convention sur la Diversité Biologique .Ilsagit sur les Ressources Zoogénétiques pour l’Alimentation et ’ d un rapport périodique qui synthétise les plus récentes l’Agriculture, programmée pour le 26 au 28 Novembre ’ données sur l état et les tendances de la biodiversité et 2014 à Rome.8 Le Groupe de Travail fera aussi savoir à qui tire des conclusions pertinentes pour poursuivre la la Commission si, à son avis, le Plan d’Action Mondial œ mise en uvre de la Convention. Les rapports sont basés pour les Ressources Zoogénétiques doit être mis à jour. ’ sur différents types d informations, y compris des rapports Veuillez encourager le Coordonnateur National de votre ﬁ nationaux, de la littérature scienti que, des informations pays9 à assister à cette importante réunion. Vous pourrez du Partenariat sur les Indicateurs de la Diversité également transmettre à votre Coordonnateur vos com- Biologique et des études complémentaires. La quatrième mentaires sur le Deuxième Rapport, qui sera publié sur édition de Perspectives Mondiales de la Diversité notre site web10 six semaines avant la réunion. Biologique (GBO-4)3 va être préparée dans le but de four- nir une évaluation à moyen terme des progrès réalisés dans Cordialement, la mise en œuvre du Plan Stratégique pour la Biodiversité Roswitha Baumung 2011–2020 et des Objectifs d’Aichi pour la Biodiversité.4

5 http://www.fao.org/docrep/011/a1250f/a1250f00.htm 6 http://www.fao.org/Ag/AGAInfo/programmes/fr/genetics/regional_collaboration. html 7 http://www.fao.org/docrep/010/a1404f/a1404f00.htm 8 http://www.fao.org/Ag/AGAInfo/programmes/fr/genetics/ Intergovernmental_process.html 1 9 http://www.fao.org/docrep/010/a1404f/a1404f00.htm http://dad.fao.org/cgi-bin/EfabisWeb.cgi?sid=d36ed91aa3283- 2 http://www.cbd.int/ f227a3aeee1c3b28be6,contacts 3 10 http://www.cbd.int/en/gbo4 http://www.fao.org/ag/againfo/programmes/fr/genetics/Intergovernmental_process. 4 http://www.cbd.int/sp/targets/ html

iii Animal Genetic Resources, 2014, 54, iv. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000204

Editorial

Estimado lector, resultar interesante la 13ª Meta de Aichi que alude directamente a la diversidad genética de los animales domésticos. El 54° volumen de nuestra revista incluye 20 artículos que describen diferentes aspectos de los recursos zoogenéticos La Comisión de Recursos Genéticos para la Alimentación y y su ordenación en África, Asia, Europa y América Latina. la Agricultura, en su Decimocuarta Reunión Ordinaria en El presente volumen da muestras del carácter verdadera- abril de 2013, solicitó a la FAO la preparación de El mente internacional de nuestra revista dada la internaciona- Segundo Informe sobre la Situación de los Recursos lidad del público lector y de los autores. Asimismo, los Zoogenéticos Mundiales para la Alimentación y la artículos presentados cubren una amplia variedad de Agricultura, como una actualización del primer informe pub- 5 temas, desde la caracterización de las razas a la licadoen2007. Hacia el 25 de marzo de 2014, 100 países descripción de sus sistemas de producción. Como ya habían enviado sus informes nacionales a la FAO y otros 30 fi ocurrió en la 53ª entrega, el elevado porcentaje de países habían indicado que se encontraban en la fase nal de artículos que abordan la caracterización fenotípica y preparación de sus informes. También se ha invitado a orga- genética da señas de que sigue revistiendo importancia la nizaciones internacionales y a los centros de coordinación 6 primera área estratégica prioritaria del Plan de Acción regionales para la gestión de los recursos zoogenéticos a Mundial sobre los Recursos Zoogenéticos1. informar sobre sus actividades de apoyo a la aplicación del Plan de Acción Mundial sobre los Recursos Zoogenéticos.7 Los recursos zoogenéticos representan una parte impor- El Segundo Informe se hará disponible en la Octava tante de la biodiversidad. Así, me gustaría también atraer Reunión del Grupo de Trabajo Técnico Intergubernamental su atención hacia la próxima edición de la Perspectiva sobre los Recursos Zoogenéticos para la Alimentación y la Mundial sobre la Biodiversidad, publicación insignia del Agricultura, programada para los días 26–28 de noviembre 2 Convenio sobre la Diversidad Biológica . Se trata de un de 2014 en Roma.8 El Grupo de Trabajo también informe periódico que resume la información más reciente notificará a la Comisión si, en su opinión, el Plan de sobre el estado y las tendencias de la biodiversidad, y que Acción Mundial sobre los Recursos Zoogenéticos debe ser extrae conclusiones relevantes para seguir aplicando el actualizado. Anime por favor al Coordinador Nacional de Convenio. Los informes están basados en distintos tipos su país9 aasistiraesteimportanteencuentro.También de información, entre los cuales informes nacionales, podrá Usted hacer llegar a su Coordinador sus comentarios fi bibliografía cientí ca, información de la Asociación de sobre el Segundo Informe, que será publicado en nuestra Indicadores de Biodiversidad y estudios complementarios. página web10 seis semanas antes del encuentro. La cuarta edición de la Perspectiva Mundial sobre la Biodiversidad (GBO-4)3 va a ser preparada con el finde Atentamente, proporcionar una evaluación, en mitad de periodo, del Roswitha Baumung grado de avance en la aplicación del Plan Estratégico para la Diversidad Biológica 2011–2020 y de las Metas de Aichi para la Diversidad Biológica.4 Tal vez le pueda

5 http://www.fao.org/docrep/012/a1250s/a1250s00.htm 6 http://www.fao.org/Ag/AGAInfo/programmes/es/genetics/regional_collaboration.html 7 http://www.fao.org/docrep/010/a1404s/a1404s00.htm 8 http://www.fao.org/Ag/AGAInfo/programmes/es/genetics/ Intergovernmental_process.html 1 9 http://www.fao.org/docrep/010/a1404s/a1404s00.htm http://dad.fao.org/cgi-bin/EfabisWeb.cgi?sid=60d42d6f19e3c527970c5762- 2 http://www.cbd.int/ b0aeecf1,contacts 3 10 http://www.cbd.int/en/gbo4 http://www.fao.org/ag/againfo/programmes/es/genetics/Intergovernmental_process. 4 http://www.cbd.int/sp/targets/ html iv Animal Genetic Resources, 2014, 54, 1–9. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S207863361400006X

Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh C.H. Ritchil, M.M. Hossain and A.K.F.H. Bhuiyan Department of Animal Breeding and Genetics, Bangladesh Agricultural University, Mymensingh, Bangladesh

Summary The study was undertaken for a period of 30 days to investigate the phenotypic, morphological and reproduction attributes of native pigs in 14 villages of Haluaghat, Durgapur and Nalitabari upazilas in Mymensingh, Netrokona and Sherpur districts, respectively. A total of 200 experimental animals were evaluated, of which 81 were male and 119 were female. Various parameters were evaluated separately at different age groups in males and females. A pre-tested questionnaire was used to record various phenotypic information. Body weight, body length, chest girth, length from ear to tail, height at withers, height at loin region, head length, width of head, ear length, hair length, tail length and number of teats were the morphometric traits recorded. The predominant coat colour of the pigs was black, with 95 percent of them having dark skin pigmentation. The majority of pigs possessed a long, straight head (75 percent) and the rest had a shape classiﬁed as a “wide face”. All of them had erect ears with upward orientation, whereas 98.5 percent had thin, straight tails. The pigs with stocky body shape (27.5 percent) showed larger body measurements than animals with angular body shape (72.5 percent). The mean values along with the Standard Error for traits were statistically analysed. Three main independent variables (i.e. upazila, age and sex) were considered for an univariate analysis of variance and the correlations among traits were measured. The average, minimum and maximum values for age at ﬁrst farrowing, farrowing interval, litter size, gestation period and weaning period were also measured.

Keywords: Dome pig, Wakmandi pig, phenotypic traits, morphometric traits, reproductive parameters

Résumé L’étude s’est étendue sur une période de 30 jours dans le but de caractériser d’un point de vue phénotypique, morphologique et repro- ductif les porcs autochtones de 14 villages des upazilas de Haluaghat, Durgapur et Nalitabari, dans les districts de Mymensingh, Netrokona et Sherpur, respectivement. Un total de 200 animaux expérimentaux (81 mâles et 119 femelles) a été évalué. Plusieurs paramètres ont été évalués séparément, à différents âges, aussi bien pour les mâles que pour les femelles. Un questionnaire pré- testé a été utilisé pour rassembler les informations phénotypiques. Les traits morphométriques retenus ont été le poids et la longueur du corps, la circonférence thoracique, la distance des oreilles à la queue, la hauteur au garrot, la hauteur à la région lombaire, la longueur et la largeur de la tête, la longueur des oreilles, la longueur des poils, la longueur de la queue et le nombre de mamelles. La couleur de robe prédominante a été le noir, avec 95 pour cent des porcs ayant une pigmentation cutanée foncée. La plupart des porcs ont eu une tête longue à proﬁl droit (75 pour cent) alors que pour le reste, la tête avait une forme connue sous le nom de “visage large”. Dans tous les cas, les oreilles étaient dressées vers le haut et pour le 98,5 pour cent des porcs, la queue était mince et droite. Les porcs au corps râblé (27,5 pour cent) ont présenté des mesures corporelles plus grandes que celles des animaux au corps anguleux (72,5 pour cent). Les valeurs moyennes, ainsi que l’erreur standard, ont été analysées statistiquement. Trois variables indépendantes principales (i.e. l’upazila, l’âge et le sexe) ont été considérées pour l’ANOVA et les corrélations entre les paramètres ont été mesurées. Les valeurs moyenne, minimale et maximale pour l’âge à la première mise-bas, l’intervalle entre mises-bas, la taille de la portée et la durée de la gestation et de la lactation ont aussi été déterminées.

Mots-clés: porc dome, porc wakmandi, traits phénotypiques, traits morphométriques, paramètres reproductifs

Resumen El estudio se extendió sobre un periodo de 30 días con el fin de investigar los atributos fenotípicos, morfológicos y reproductivos de los cerdos autóctonos de 14 pueblos de las upazilas de Haluaghat, Durgapur y Nalitabari, en los distritos de Mymensingh, Netrokona y Sherpur, respectivamente. Se evaluó un total de 200 animales experimentales, de los cuales 81 eran machos y 119 eran hembras. Se evaluaron varios parámetros de forma separada, a distintas edades, en los grupos de machos y hembras. Se utilizó un cuestionario probado para recopilar la información fenotípica. Los rasgos morfométricos registrados fueron el peso y la longitud corporal, la circunferencia torácica, la distancia de las orejas a la cola, la altura a la cruz, la altura a la región lumbar, la longitud y la anchura de la cabeza, la longitud de las orejas, la longitud del pelo, la longitud de la cola y el número de mamas. El color de capa predominante de los cerdos fue el negro; el 95 por ciento de los ejemplares tenía pigmentación cutánea oscura. La mayoría de los cerdos tenía una cabeza larga, con perfil recto (75 por ciento), y el resto presentaba una forma conocida como “cara ancha”. Todos los ejemplares tenían orejas erectas en posición vertical y la cola era fina y recta en el 98,5 por ciento de los casos. Los cerdos con cuerpo fornido (27,5 por ciento) presentaron medidas corporales mayores que las de los animales con cuerpo anguloso (72,5 por ciento). Se analizaron estadísticamente los valores medios, junto con el error estándar, de los distintos parámetros. Se consideraron tres variables independientes principales (i.e. upazila, edad y sexo) para el ANOVA y se midieron las correlaciones entre los parámetros. También se

Correspondence to: C.H. Ritchil, Department of Animal Breeding and Genetics, Bangladesh Agricultural University, Mymensingh, Bangladesh. email: [email protected]

1 2 Ritchil et al.

determinaron los valores medio, mínimo y máximo para la edad al primer parto, el intervalo entre partos, el tamaño de la camada y la duración de la gestación y la lactación.

Palabras clave: cerdo dome, cerdo wakmandi, rasgos fenotípicos, rasgos morfométricos, parámetros reproductivos

Submitted 25 March 2013; accepted 30 January 2014

Introduction reasonably high density of pig populations. Before collec- More than 45 different ethnic groups live in Bangladesh. As a tion of data, a questionnaire (Appendix 1) was prepared in group, they are called the “indigenous people” and they raise accordance with objectives of the study. Later on, the ques- ﬁ pigs for a contribution to their livelihoods and are the main tionnaire was validated against eld conditions. consumers of pork and rearers of pig (Sus crofa domesticus) in Bangladesh. In addition, the Christians and some Hindus Study population also consume pork in large amounts. There is also a community of nomadic people called the “Dome”, who rear pigs and All the experimental animals were of the native type (i.e. sell piglets to the indigenous people. They are the main Dome or Wakmandi) pigs. All pigs were reared by the source of pigs for those people. The pig bought from Dome Garo(s) the indigenous people in the studied areas. A is called Dome’s pig. Another type of native pig is called total of 200 household pig owners from the aforemen- the “Wakmandi”. A distinctive trait of this pig is a drooping tioned sites were selected randomly and interviewed rele- mid-section, which differentiates it from the Dome. About 95 vant for the study. A total of 200 experimental animals percent of the pigs in Bangladesh are of these native types were evaluated, of which 81 were male and 119 were (Dome and Wakmandi) and the remaining 5 percent are exot- female pigs. One pig from each household was selected ic or crosses between native and exotic. There has been no randomly. thorough investigation carried out to characterize these native pigs in Bangladesh. In this context, this study was formulated Data collection to evaluate the morphological, phenotypic and reproduction attributes of native pigs, speciﬁcally in the rural areas of The study was conducted from 1 September 2012 to 30 Haluaghat, Durgapur and Nalitabari upazilas of Bangladesh. September 2012. The phenotypic traits such as coat colour, skin pigmentation, head shape, ear direction, ear orientation, tail shape and body shape were recorded based on Materials and methods subjective visual observation. Body weight of piglets was recorded by using a weighing scale, whereas body Study area weights of boars and gilts were estimated by using the following formula:- The study was undertaken at four randomly selected villages of the Haluaghat Upazila located at 25.1250°N and Chest girth2 (m) × length from ear to tail (m) × 69.3 = body 90.3500°E in Mymensingh district (Figure 1), seven vil- weight of the pig (kg) lages of the Durgapur Upazila located at 25.1250°N and where, 90.6875°E in Netrokona district (Figure 2) and three vil- M = meter lages of the Nalitabari Upazila located at 25.0833°N and 69.3 = constant 90.1954°E in Sherpur district (Figure 3). The villages kg = kilogram were selected because they were known to have a Source: The Old Farmer’s Almanac

Figure 1. Map of ﬁeld unit at Haluaghat upazila in Mymensingh district. Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh 3

Figure 2. Map of ﬁeld unit at Durgapur upazila in Netrokona district.

Figure 3. Map of ﬁeld unit in Nalitabari upazila in Sherpur district.

Body length was determined by measuring the distance Results and discussion between points of shoulder to the pin bone. Chest girth was considered as the largest circumference of the body 1. Phenotypic characterization immediately behind the shoulder. The length from ear to Coat characteristics tail was measured from the base of the ear to the base of The observed combinations of coat colour of the native pigs the tail. Height at wither was considered as vertical dis- were black (95 percent), black and white (2.5 percent), tance between the ground and the point of wither. Height greyish-black (2 percent) and brownish (0.5 percent) at loin region was considered as vertical distance between (Figure 4). The coat colour of native pig had been recorded the surface and the loin region. Head length was measured as black, white, black with white legs, white with black spot, from the distance between snout and forehead of pig, while black with white belly and brown in a previous study con- head width was the distance between the eyes. The distance ducted by Yaetsu et al. (1987) in Bangladesh, which is simi- between the base and the tip of the pinna was considered as lar to the present study. Most of the native pigs in the study the ear length. Shoulder hairs were used to determine the areas had dark skin pigmentation (95 percent). hair length. The distance between the base and the end of the tail was constituted the tail length. The number of pairs of teats was recorded by visual observation. The litter Head and ear characteristics size was based on the number born alive. Gestation and The shape of the head was long and straight in most of the ﬁ weaning period were also estimated. native pigs (75 percent) and the rest were classi ed as “wide face” (25 percent) (Figure 5). Yaetsu et al. (1987) conducted an experiment on native pigs in different Statistical analysis regions of Bangladesh and reported that a straight face The accumulated data were arranged by using Microsoft with a pointed snout and erect ears was common for native Excel 2007 and these data were compiled and analyzed pig. The present study showed that the native pigs had by using Statistical Package for Social Science (SPSS) erect ears that pointed backwards (100 percent) (Figure 5). v.11.5. Univariate analysis of variance (ANOVA) was carried out with upazila, sex, age group and sex*age as Body and tail shapes explanatory variables. The raw and residual correlations Native pigs were found to have either angular or stocky among traits were also calculated. body shapes where 72.5 percent of them belonged to 4 Ritchil et al.

Figure 4. Coat characteristics of native pig.

angular body type and rest were stocky (Figure 6). Thin, one circle. Yaetsu et al. (1987) conducted an experiment straight as well as curled tails were observed among the on native pigs in different regions of Bangladesh and native pigs. It was found that 98.5 percent of the pigs reported that some pigs with somewhat concave back had a thin, straight tail. The curled tails (1.5 percent) and pendulous belly were found. They also reported that were generally thin and showed an upward curl forming the straight tail was common for all native pigs.

Figure 5. Head and ear characteristics of native pig. Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh 5

Figure 6. Body shape of the native pig.

2. Morphological characterization of age (Table 1). In general, adult (18–24 months) males tended to be larger than females, whereas no clear pattern For morphological characterization, body measurements of was observed at younger ages. Babu et al. (2004) con- 81 male and 119 female pigs were taken. Means along ducted an experiment in the rural areas of India and with standard errors of means (SEM) of different morpho- reported that the average body weight of Large White metric characters for pigs are presented in Tables 1–3. pigs at puberty was 82.3 ± 1.8 kg (mean ± SE), which is Interactions between dependent and independent variables similar to the result of present study. The adult body length are presented in Table 4. The raw and residual correlations (7–24 months) ranged from 47–67 cm for males and 46– among the studied traits are presented at Table 5. 63 cm for females. Subalini, Silva and Demetawewa Not surprisingly, the morphological traits of native pigs (2010) reported that the average body length of Sri were found to increase (P < 0.001) with the advancement Lankan village adult male and female pig was 75.1 ± 1.9 and 69.9 ± 2.4 cm, respectively, which is 15 cm greater Table 1. Morphological characteristic of pigs affected by age than in the present study. The average chest girth for the (Mean ± SEM). adult male and female pigs was 82.8 ± 1.7 and 81.3 ± 1.1 cm, respectively. The average length from the base Parameter Age of the ear to base of the tail was higher for females at 0– 0–6 7–12 13–18 19–24 6 and 13–18 months of age. The height at wither, height months months months months at loin, head length, head width, ear length and tail length (n = 56) (n = 105) (n = 28) (n = 11) were higher for females than males at 0–6 months of age. Body weight 16.5a ± 0.9 35.3b ± 1.1 68.8c ± 2.5 89.4d ± 5.2 A total of only 20 male and 36 female piglets were mea- (kg) sured at 0–6 months of age. Hence, the small sample Body length 37.3a ± 0.9 46.9b ± 0.7 59.4c ± 1.8 65.7d ± 3.5 size might be responsible for this result. Yaetsu et al. (cm) a b c d (1987) conducted an experiment on native pigs in different Chest girth 62.3 ± 1.3 81.9 ± 0.9 104.5 ± 1.5 113.6 ± 3.2 regions of Bangladesh and reported that the average (cm) – Ear to tail 57.5a ± 1.2 73.3b ± 0.9 91.3c ± 1.3 98.6d ± 1.9 heights at wither at 13 18 months of age in males and length (cm) females were 56.0 ± 0.0 and 52.8 ± 4.8 cm, respectively, Height at 39.9a ± 0.9 49.5b ± 0.7 61.4c ± 1.1 63.0d ± 1.8 which is lower than the present study. The average wither (cm) – – – a b c d pairs of teat of the female pig at 0 6, 7 12, 13 18 and Height at loin 41.0 ± 0.9 51.5 ± 0.7 63.5 ± 1.1 65.2 ± 1.8 19–24 months of age was found to be 4.8 ± 0.1, 5.0 ± 0.0, (cm) Head length 19.6a ± 0.4 22.9b ± 0.3 26.3c ± 0.7 28.5d ±.4 5.1 ± 0.1 and 6.0 ± 0.0 pairs, respectively. Yaetsu et al. (cm) (1987) conducted an experiment on native pigs in different Head width 11.6a ± 0.2 13.8b ± 0.2 16.9c ± 0.5 19.0d ±.6 regions of Bangladesh and reported that the teat pattern of (cm) a b c d Bangladeshi pigs ranged from four to six pairs. A pattern of Ear length 7.8 ± 0.2 8.9 ± 0.2 9.1 ± 0.2 10.3 ±.3 five pairs was most frequent in all the populations exam- (cm) Hair length 6.1a ± 0.2 7.5b ± 0.2 8.4c ± 0.4 8.5d ±.6 ined, which is similar to the result of the present study. (cm) The correlations among traits were also calculated Tail length 18.7a ± 0.6 24.0b ± 0.5 28.5c ± 0.8 29.5d ± 1.5 (Table 5). The correlations were all positive and were sign- (cm) ificant at the 0.01 level (two-tailed). However, the present fi Different superscripts indicate that values are significantly different at study failed to record a signi cant difference (P < 0.05) 0.1% level of probability (P < 0.001); between the two sexes (Table 2) except for chest girth n = number of animals. and ear length at 19– 24 and 13–18 months of age group, 6 Ritchil et al.

respectively. Subalini, Silva and Demetawewa. (2010) con-

cance ducted an experiment on village pigs in Sri Lanka and ﬁ

-value) ﬁ

P found no signi cant difference (P < 0.05) between two sexes. Sudhakar and Gaur (2006) also found that effect of sex was not signiﬁcant on all body weights and measurements in Indian native pigs, except for weight at birth. The ± 4.9 0.05 ± 8.7± 7.3 0.06 ± 3.3 0.54 ± 3.6± 3.6 0.31 ± 0.5 0.21 ± 0.8 0.30 ± 0.4 0.49 ± 1.1 0.78 ± 2.8 0.44 0.72 0.50 a a b a a a a a a a a =5) ( diet for male and females was almost same though. At the n present study the differences between two populations (Dome and Wakmandi) was not done for the characteristics

24 months of age measured as the population of Wakmandi was very little. – ± 3.9± 2.8 79.0 ± 2.9 63.2 ± 2.2 107.2 ± 1.3 96.4 ± 1.4 60.4 ± 0.7 63.0 ± 0.9 28.8 ± 0.4 19.2 ± 0.8 10.0 ± 1.5 8.8 30.6 a a a a a a a a a a a =6) ( Morphological characteristics of mature pigs affected by n upazila (area) are presented in Table 3. Area contributed signiﬁcant differences (P < 0.05) between males and females in case of parameters such as wither height, loin

cance Male Female Signi height, head length, ear length and tail length. ﬁ -value) ( P

3. Reproduction attributes ± 0.2 0.05 10.5 ± 3.5± 2.4± 2.2 0.65± 1.5 0.95± 1.3 0.42± 1.3 0.34 98.0 ± 0.9 0.34 67.8 ± 0.6 119.0 0.29 100.5 0.16± 0.5 0.71 65.2 ± 1.2 67.0 0.37 28.2 0.40 18.8 8.3 In 28.5 present study, the average age at ﬁrst farrowing was a a a a a a a a b a a = 18) (

n found to be 10.43 ± 0.08 months (Table 6). The study conducted by Hossain et al. (2011) using surveyed data on indigenous pigs in hilly areas of Chittagong in 18 months of age At 19

– Bangladesh, reported that the average ages at sexual matur- ± 3.4± 2.4 69.6 ± 1.6 59.5 ± 2.6 105.4 ± 1.9 90.3 ± 1.9 60.7 ± 1.2 62.7 ± 0.7 25.6 ± 0.3 16.8 ± 0.5± 0.6 9.4 8.7 28.0 a a a a a a a a a a a =10) ( ity in Bangladeshi boar and sow were 8 and 6 months, n respectively. Naskar et al. (2007) conducted an experiment Sex on crossbred pigs under the hilly conditions of Meghalaya, India and reported that the average age at ﬁrst farrowing cance Male Female Signi

ﬁ of pig was 15.3 ± 1.2 months. Subalini, Silva and -value) ( = number of animals.

P Demetawewa (2010) conducted an experiment on village n pig in Sri Lanka and reported that the average age at ﬁrst farrowing was 9.50 ± 2.61 months. ± 1.4± 0.8± 1.1 0.24± 1.1 0.65± 0.9 0.43± 0.9 0.09 67.2 ± 0.5 0.77 59.3 ± 0.3 102.8 0.98± 0.4 0.48 93.0 ± 0.3 0.16 62.8 ± 0.7 0.24 65.1 0.98 27.7 0.86 17.2 8.6 8.0 29.4 a a a a a a a a a a a = 60) ( The average farrowing interval of native pigs was found to n be 6.09 ± 0.02 months. Babu et al. (2004) conducted an experiment in the rural areas of India on Large White pig and reported that the average farrowing interval was 12 months of age At 13 ± 1.9± 1.1 34.1 ± 1.7 46.7 ± 1.3 81.3 ± 1.1 72.1 ± 1.2 49.7 ± 0.5 51.5 ± 0.4 23.1 ± 0.2 13.5 ± 0.3 9.2 ± 0.7 7.5 24.1 – a a a a a a a a a a a = 45) ( 180.19 ± 1.03 days. Subalini, Silva and Demetawewa n (2010) conducted an experiment on village pig in Sri Lanka and reported that the average farrowing interval was 8.91 ± 2.49 months. cance Male Female Signi

ﬁ The average litter size of native pigs in this study was 6.71 -value) ( P ± 0.15. Hossain et al. (2011) reported that the average litter cantly between sexes have different superscripts;

ﬁ size of indigenous pig at hilly region of Bangladesh was 9.5 ± 0.28. Prakash et al. (2008) reported that the average ± 1.1± 1.2± 1.5 0.73± 1.3 0.08± 1.0 0.75± 1.1 36.8 0.43± 0.6 47.3 0.70± 0.3 82.8 0.70± 0.2 74.9 0.45± 0.3 49.3 0.28± 0.7 51.5 0.13 22.6 0.44 14.2 0.09 8.6 7.5 23.9

a a a a a a a a a a a litter size of Indian indigenous pig was 6.78 ± 0.11. Nath =36) ( n

4.8 ± 0.1and Deka (2003), conducted 5.0 ± 0.0 an experiment on growth 5.1 ± per- 0.0 6.0 ± 0.0 formance of large black pig in Assam, India and reported that the average litter size at birth and weaning were 9.70 ± 6 months of age At 7 – ± 1.7± 1.2 16.7 ± 2.5 38.4 ± 2.4 62.6 ± 1.7 58.2 ± 1.7 40.1 ± 0.7 41.3 ± 0.4 19.9 ± 0.4 11.8 ± 0.4 8.0 ± 0.9 5.9 19.4

a a a a a a a a a a a 0.29 and 8.40 ± 0.32, respectively. = 20) ( Male Female Signi n At 0 ( 7.5 6.3 35.2 61.8 39.4 40.6 19.2 11.3 17.5 56.3 16.1 The average gestation period and the average weaning period of these native pigs were 115.35 ± 0.50 and 44.88 ± 0.37 days, respectively. Rajiv and Pandey (2000) conducted an experiment on the economics of pig rearing in Haryana, India and reported that the average gestation Morphological characteristics of pigs as a function of sex (Mean ± SEM). period was 112–114 days. Hossain et al. (2011) reported that the average weaning period of indigenous piglets in = Means from the same age group that differ signi Body length (cm) Chest girth (cm) Height at wither (cm) Height at loin (cm) Head length (cm) Head width (cm) Ear length (cm) Hair length (cm) Tail length (cm) Pair of teat Ear to tail length (cm) Table 2. Parameter a,b Body weight (kg) Bangladesh was 42 days. Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh 7

Table 3. Morphological characteristics of mature pigs as a function of upazila (Mean ± SEM).

Parameter Upazila

Haluaghat Durgapur Nalitabari

Male Female Significance Male Female Significance Male Female Significance (n = 15) (n = 26) (P-value) (n = 34) (n = 19) (P-value) (n = 12) (n = 38) (P-value)

Body weight (kg) 37.1a ± 2.8 39.9a ± 2.9 0.53 49.10a ± 4.2 53.2a ± 5.1 0.54 57.5a ± 7.9 43.3a ± 3.6 0.07 Body length (cm) 48.5a ± 1.4 50.7a ± 1.1 0.24 51.6a ± 1.7 52.7a ± 3.1 0.73 53.6a ± 4.0 49.1a ± 1.6 0.22 Chest girth (cm) 82.4a ± 1.9 85.0a ± 2.1 0.41 90.8a ± 2.8 93.1a ± 3.5 0.62 95.0a ± 5.3 87.5a ± 2.4 0.16 Ear to tail length (cm) 77.0a ± 1.5 76.0a ± 1.8 0.71 80.5a ± 2.2 83.8a ± 2.9 0.39 84.2a ± 4.4 75.3b ± 2.0 0.04 Height at wither (cm) 51.6a ± 1.2 49.6a ± 1.7 0.43 51.3a ± 1.7 58.2b ± 1.5 0.01 59.6a ± 2.8 52.0b ± 1.1 0.00 Height at loin (cm) 53.8a ± 1.3 51.2a ± 1.8 0.34 53.5a ± 1.7 60.2b ± 1.5 0.01 61.6a ± 3.0 54.0b ± 1.2 0.00 Head length (cm) 22.2a ± 0.6 25.5b ± 0.7 0.00 24.2a ± 0.6 25.6a ± 0.8 0.19 25.4a ± 1.3 22.1b ± 0.6 0.01 Head width (cm) 14.5a ± 0.4 14.6a ± 0.5 0.88 15.3a ± 0.5 16.4a ± 0.6 0.21 15.3a ± 1.0 13.6a ± 0.4 0.08 Ear length (cm) 8.5a ± 0.2 9.0a ± 0.1 0.07 8.7a ± 0.2 9.7b ± 0.2 0.01 9.0a ± 0.4 9.1a ± 0.6 0.95 Hair length (cm) 7.6a ± 0.5 8.1a ± 0.4 0.43 7.7a ± 0.3 8.8b ± 0.4 0.04 7.5a ± 0.6 7.0a ± 0.2 0.40 Tail length (cm) 25.3a ± 1.2 25.4a ± 1.1 0.94 24.1a ± 0.7 28.2b ± 0.9 0.00 28.4a ± 1.0 23.7b ± 0.8 0.00 Pair of teat 5.0 ± 0.0 5.2 ± 0.1 5.0 ± 0.0 a,b = Means from the same age group that differ signiﬁcantly between sexes have different superscripts; n = number of animals.

4. Management system of native pigs and some unconventional feeds such as cauliflowers, boiled arum, grass or other vegetables for pigs. Pigs are Girth tethering was found to be the most popular and wide- fed thrice a day. Most prevalent diseases are diarrhoea, ly used housing system for pigs in this study. Pigs were fed coccidiosis, pneumonia and haemorrhagic septicaemia. thrice a day. Pig owners used to supply boiled rice, rice Easy management, adaptability, cost-effectiveness and dis- polish, rice bran, etc. as conventional feeds. Some uncon- ease tolerance are the main advantages of native pig farm- ventional feeds were also fed to them. Tube-well or natural ing while malnutrition, lack of scientific knowledge, lack reservoir was the source of water for pigs. of veterinary and extension services are the major According to Ritchil et al. (2013), pigs in Bangladesh are constraints. reared mostly by poor and landless people (52 percent). Ritchil et al. (2013) also found that rearing native pig is very The majorities of farmers are housewife (55 percent) or easy for those people who rear and consume it. They believe day labour (17 percent). The literacy level is satisfactory. that it is very much cost effective, available, adaptable and fi Rearing systems are different and the mean gure is 97 merely affected by diseases. The socio-economic status of percent for girth tethering. The sanitation condition is the pig rearers is appreciable. Although many constraints not satisfactory. The pig owners use to supply boiled are taking part in the management of pig, the meat from rice, rice polish, rice bran, rice husk, fermented rice, etc. pig is the main source of protein for those people who rear and consume it. Therefore, according to Ritchil et al. (2013) lack of mature breeding boars, technical knowhow Table 4. Interactions between dependent and independent and preventive healthcare services were noticed which variables. need to be extended among the pig rearers for better liveli- Parameter Significance (P-value) Error (mean hood and welfare of their keepers. Hence, the observed square) body weight and reproduction attributes of the native pigs Area Sex Age Sex*age may serve as a base for designing any pig improvement pro- Body weight (kg) 0.06 0.05 0.00 *(0.030) 118.69 gramme in Bangladesh. Body length (cm) 0.14 0.90 0.00 NS (0.287) 54.48 Chest girth (cm) 0.02 0.27 0.00 NS (0.146) 89.18 Ear to tail length 0.00 0.38 0.00 NS (0.452) 68.19 (cm) Conclusion Height at wither 0.19 0.35 0.00 NS (0.420) 45.56 (cm) Native pigs in Bangladesh showed diversity in their pheno- Height at loin 0.21 0.38 0.00 NS (0.578) 50.97 type and morphology and positive reproductive perform- (cm) Head length (cm) 0.00 0.76 0.00 NS (0.112) 10.91 ance, which indicates their potential for economic Head width (cm) 0.00 0.68 0.00 NS (0.586) 4.33 sustainability through purebred genetic improvement. Ear length (cm) 0.69 0.34 0.01 NS (0.808) 4.21 However, these native pigs are usually raised in conditions Hair length (cm) 0.00 0.43 0.00 NS (0.707) 3.74 with less than optimal husbandry and hence, this aspect Tail length (cm) 0.06 0.33 0.00 NS (0.436) 21.29 needs immediate attention to ensure their conservation Pair of teat 0.01 0.00 0.00 ***(0.000) 0.03 through utilization for the food and livelihood of indigen- Non-significant (NS) = P > 0.05; * = P < 0.05; *** = P < 0.001 ous, Christian and some Hindu people in Bangladesh. This 8 Ritchil et al.

Table 5. Raw (on the diagonal) and residual (below the diagonal) correlations among traits.

Body Body Chest Ear to tail Height at Height at Head Head Ear Hair Tail weight length girth length wither loin length width length length length

Body 1 0.82** 0.96** 0.92** 0.81** 0.81** 0.65** 0.75** 0.32** 0.50** 0.57** weight Body 0.03 1 0.80** 0.80** 0.69** 0.69** 0.59** 0.67** 0.22** 0.42** 0.49** length Chest 0.07 0.01 1 0.91** 0.83** 0.84** 0.66** 0.73** 0.35** 0.54** 0.60** girth Ear to tail 0.04 0.01 0.03 1 0.81** 0.82** 0.64** 0.71** 0.29** 0.54** 0.62** length Height at −0.03 −0.05 −0.02 −0.02 1 0.99** 0.59** 0.69** 0.37** 0.43** 0.62** wither Height at −0.03 −0.06 −0.02 −0.02 0.18 1 0.59** 0.69** 0.42** 0.45** 0.63** loin Head −0.06 −0.05 −0.06 −0.07 −0.11 −0.11 1 0.77** 0.30** 0.49** 0.63** length Head −0.04 −0.03 −0.06 −0.06 −0.06 −0.07 0.13 1 0.33** 0.40** 0.59** width Ear −0.01 −0.07 −0.01 −0.03 −0.01 0.03 −0.08 −0.02 1 0.28** 0.38** length Hair −0.05 −0.06 −0.01 0.00 −0.09 −0.09 0.02 −0.09 −0.03 1 0.41** length Tail −0.08 −0.09 −0.07 −0.03 −0.02 −0.03 0.04 −0.005 −0.05 −0.04 1 length

**Correlation is signiﬁcant at the 0.01 level (two-tailed). Residuals are computed between observed and reproduced correlations. There are 30 (45.0%) non-redundant residuals with absolute values >0.05.

Table 6. Reproductive performance of native pig. in the rural areas of Rangamati and Khagrachari districts of Bangladesh. Bangladesh J. Anim. Sci., 40(1–2): 28–33. Parameter Mean ± Minimum Maximum SEM Naskar, S., Khan, M.H., Anubrata, D. & Bordoloi, R.K. 2007. Performance of T & D (Tamworth × Desi) crossbred pigs under Age at ﬁrst farrowing 10.43 ± 0.08 8 12 hilly condition of Meghalaya. Environ. Ecol., 25(2): 308–311. (months) Nath, D.R. & Deka, D. 2003. Litter trait and preweaning growth per- Farrowing interval (months) 6.09 ± 0.02 6 7 formance of large black pig in Assam. Indian Vet. J., 80(3): 287–289. Litter size (Number) 6.71 ± 0.15 2 12 Gestation period (months) 3.8 ± 0.0 3.3 4 Prakash, M.G., Ravi, A., Kumari, B.P. & Srinivas Rao, D. 2008. Weaning period (months) 1.5 ± 0.0 1 1.8 Reproductive and productive performance of crossbred Pigs. Indian J. Anim. Sci., 78: 1291–1297. Rajiv, J. & Pandey, U.K. (2000). Economics of pig rearing in Haryana. – can be done through a well-designed national pig improve- Indian J. Anim. Sci., 70(12): 1268 1271. ment programme, which may include training of the pig Ritchil, C.H., Faruque, M.O., Tabassum, F., Hossain, M.M. & rearers on general pig husbandry, supply of preventive Bhuiyan, A.K.F.H. 2013. The socio-economic status of pig rearers vaccines and production and dissemination of quality and the management system of native pigs in Bangladesh. Indian J. Anim. Sci., 83(11): 1226–1228, November 2013. (heavier but proliﬁc) boars for use in the subsistence farms by the government, non-government organization Subalini, E., Silva, G.L.L.P. & Demetawewa, C.M.B. 2010. Phenotypic characterization and production performance of village Pigs in Sri or organized pig farmers of the country. Lanka. Trop. Agric. Res., 21(2): 198–208. Sudhakar, K. & Gaur, K. 2006. Pre-weaning growth in indigenous pigs of eastern region. Indian J. Anim. Sci., 3(4): 25–30. References Website: http://www.almanac.com/content/all-right-reasons-raise-pigs; http://www.thepigsite.com/articles/541/weighing-a-pig-without-a-scale Babu, G.N., Naidu, K.V., Rao, A.S. & Vijay, S. 2004. Certain repro- Yaetsu, K., Takashi, A., Ikuo, O., Katuaki, O., Takao, N., ductive parameters in Large White pigs maintained with garbage feed- Yoshizane, M., Hasnath, M.A., Mostofa, K.G., Faruque, O.M. ing in rural areas. Indian J. Anim. Sci., 74(4): 438–440. & Majid, M.A. 1987. Morphological studies of native pigs in Hossain, M.E., Chakma, S., Khatun, M.M., Hasanuzzaman, M., Bangladesh. Genet. Stud. Breed Differ. Native Domest. Anim. Miah, M.Y. & Biswas, M.A.A. 2011. Production systems of swine Bangladesh, 2:47–58. Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh 9

Appendix 4. Production & Reproduction

Appendix 1. An interview schedule for Parameters Time baseline survey on native pig farming in Age at ﬁrst farrowing Bangladesh Farrowing interval Litter size ...... Gestation period Sl. No . Weaning period 1. Primary Information

5. Pig morphology information Name : Address : Father’s Name/Husband’s Name : Piglet ’ Mother s Name : Character Male Female Gilt Boar Sow Age : Household land : Body weight (kg) Land for cultivation : Body length (cm) Experience in pig farming : Chest girth (cm) Family Member : Length from ear to tail (cm) Primary Occupation : Height at withers (cm) Secondary Occupation : Height at loin region (cm) Level of education : Head length (cm) Mobile No. : Head width (cm) Monthly Income (Family) : Ear length (cm) GPS Location : Alt.: Lat.: Hair length (cm) Tail length (cm) 2. Type of pig Pair of teats (number)

Type Breed and no. Age 6. Phenotypic characters Pure Crossbred Native breed Trait Male Female Dome Wakmandi Coat colour Milking Skin pigmentation Pregnant Head shape Sow Ear direction Gilt Ear orientation Boar Tail shape (Castrated) Body shape Boar (Not Castrated) Piglet (male/ Date: Name & Signature of enumerator female)

3. Feed

Age class Feed Purchased Amount feed Water type feed (kg/day) (litre/day)

Boar Gilt Sow Piglet (male/ female) Animal Genetic Resources, 2014, 54, 11–19. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000507

Caractéristiques phénotypiques de la chèvre du sahel au Niger par analyse des indices de primarité et des paramètres qualitatifs

M. Mani1, H. Marichatou1, M. Issa2, I. Chaïbou3, A. Sow4, M. Chaïbou1 and J.G. Sawadogo4 1Faculté d’Agronomie, Université Abdou Moumouni, BP 10960, Niamey, Niger; 2Faculté des Sciences et Techniques, Université Abdou Moumouni, BP 10662, Niamey, Niger; 3Faculté d’Agronomie, Université de Maradi, Niger; 4Ecole Inter-Etats des Sciences et Médecines Vétérinaires, Dakar, Sénégal

Résumé Au Niger, le troupeau caprin est composé de trois races dont deux majoritaires: la chèvre rousse de Maradi (17,5%) qui a fait l’objet de nombreuses investigations et la chèvre du sahel (80%) qui est peu ou pas connue. Ainsi, afind’établir les caractéristiques phénotypiques de la chèvre du sahel au Niger, des investigations sur les paramètres qualitatifs (port des appendices et les patrons pigmentaires) par observation directe des animaux ont été conduites dans la partie Nord-Ouest du pays (régions de Tillabéry, Tahoua et Niamey). Sur la base de l’adhésion volontaire des éleveurs, 443 caprins (77% de femelles, 39% ayant 4 paires d’incisives permanentes) répartis dans 145 exploitations et 60 sites (7,38 ± 6,87 têtes/site) ont été caractérisés. L’analyse des indices de primarité et l’analyse multivariée (ACM) font ressortir qu’il n’existe pas de différence significative (P < 0,05) entre sexe pour tous les paramètres qualitatifs àl’exception de la présence de la barbiche (mâle 55%, femelle 41%). La distribution des paramètres est significativement différente selon les régions à l’exception du type des cornes (majoritairement ibex dans toutes les régions). En effet, les oreilles sont majoritairement tombantes à Tillabéry, pédonculées à Tahoua et Niamey et les patrons pigmentaires sont dominés par eumélanique noir à Tahoua, eumélanique chocolat à Niamey et phaeomélanique à Tillabéry. En définitif, les caprins des trois régions sont significativement différents sur le plan phénotypique qualitatif. Mais, dans l’ensemble, cette population caprine présente les caractéristiques d’une population primaire, ce qui offre de grandes possibilités d’amélioration génétique.

Mots-clés: caractéristiques phénotypiques, qualitatifs, chèvre du sahel, Niger

Summary In Niger, there are three goat breeds including two main ones: the Maradi red goat (17.5%) which has been the subject of numerous investigations and the Sahel goat (80%) which is little studied or unknown. Thus, to establish the phenotypic characteristics of the Sahel goat breed in Niger, investigations into qualitative parameters (appendices and pigmentation) by direct observation of animals were conducted in the northwestern part of the country (Tillabery, Tahoua and Niamey regions). Based on the voluntary participation of farmers, 443 goats (77% female, 39% with 4 pairs of permanent incisors and therefore considered adult) in 145 farms and 60 sites (7,38 ± 6,87 animals/site) were characterized. Analysis of primarity indices and multivariate analysis (ACM) emphasizes that there is no significant difference (P < 0.05) between sexes for all quality parameters except the carrying of a beard (male 55%, female 41%). The parameters distribution is significantly different between regions except the type of horn (mostly ibex in all regions). Indeed, ears are mostly drooping in Tillabéry, stalked in Tahoua and Niamey and pigmentation is dominated by black eumélanique in Tahoua, chocolate eumélanique in Niamey and phaeomélanique in Tillabéry. Goats of the three regions are significantly different based on phenotypic qualitative parameters. But, generally, the goat study population has the characteristics of a primary population which offers great opportunities for genetic improvement.

Keywords: phenotypical features, qualitative features, Sahel goat, Niger

Resumen En Níger, la cabaña caprina se compone de tres razas, de las cuales dos son mayoritarias: la cabra roja de Maradi (17,5 por ciento), que ha sido objeto de muchas investigaciones, y la cabra del Sahel (80 por ciento), de la cual se sabe poco o nada. Así, con el ﬁn de caracterizar fenotípicamente la cabra del Sahel en Níger, se han investigado parámetros cualitativos (apéndices y patrones de pigmentación), por observación directa de los animales, en la parte Noroeste del país (regiones de Tillabéry, Tahoua y Niamey). Contando con la participación voluntaria de los ganaderos, se caracterizaron 443 ejemplares de ganado caprino (77 por ciento hembras, 39 por ciento con 4 pares de incisivos permanentes) repartidos entre 145 explotaciones y 60 emplazamientos (7,38 ± 6,87 cabezas/ emplazamiento). El análisis de los índices de primariedad y el análisis multivariante (ACM) muestran que no existen diferencias sign- iﬁcativas (P < 0,05) entre los sexos para ninguno de los parámetros cualitativos, exceptuando la presencia de barba (machos 55 por ciento, hembras 41 por ciento). Los parámetros estudiados, con la excepción del tipo de cornamenta (principalmente de tipo ibex

Faculté d’Agronomie, Université Abdou Moumouni, BP 10960, Niamey, Niger. Tel: (00227) 93 91 65 31. Fax: (00227) 20 31 66 12. Email: [email protected] / [email protected]

11 12 Mani et al.

en todas las regiones), se distribuyen de forma estadísticamente diferente según las regiones. Así, las orejas están mayoritariamente caídas en Tillabéry, son pedunculadas en Tahoua y Niamey, y los patrones de pigmentación dominantes son de diseño eumelánico negro en Tahoua, eumelánico chocolate en Niamey y feomelánico en Tillabéry. En deﬁnitiva, existen diferencias signiﬁcativas, en el plano fenotípico cualitativo, entre la ganadería caprina de las tres regiones. Sin embargo, esta población caprina presenta, en con- junto, características de una población primaria, lo cual ofrece grandes posibilidades de mejora genética.

Palabras clave: Palabras clave: características fenotípicas, cualitativo, cabra del Sahel, Níger

Soumis: 27 Le juin 2013; admis: 16 Le décembre 2013

Introduction malgré son importance numérique et son omniprésence Au Niger, les caprins représentent 36% du cheptel (Niger, dans toutes les zones climatiques du Niger. 2007) et occupent une place de choix dans la vie Cette étude vise à établir les caractéristiques phénotypiques socioéconomique et culturelle de la population. Ils con- qualitatives de la chèvre du sahel au Niger. En effet, si la stituent une spéculation importante en matière de commer- caractérisation phénotypique vise à identifier et établir la cialisation de bétail sur pied, mais aussi de production. diversité inter et intra races à travers des traits physiques, les C’est également une source principale de protéines ani- paramètres qualitatifs, partie intégrante de cette males en milieu rural. caractérisation phénotypique, renseignent autrement sur les caractères d’adaptation des races à leur milieu d’élevage mais ’ L espèce caprine se présente sous trois races au Niger: la aussi, peuvent traduire les préférences des acteurs (éleveurs et chèvre du sahel (79,54%) qui est présente dans toutes les consommateurs) ou des aspects culturels (FAO, 2012). régions du pays, la chèvre rousse (17,48%) prédominante dans la bande sud du pays et la chèvre naine d’Afrique de l’Ouest (chèvre Djallonké) dite localement chèvre « gourma » (2,98%) qui est présente dans la partie ouest de la région Matériel et Méthodes de Tillabéry à la frontière avec le Burkina Faso (Rhissa, 2010). Echantillon d’étude Comparativement à la chèvre rousse de Maradi qui a fait L’étude a été conduite dans la partie Nord-Ouest du Niger, l’objet de nombreuses investigations sur les plans ethnologi- notamment, les régions de Tahoua (départements de que et zootechnique (Robinet, 1967; Marichatou et al.,2002; Tahoua, Abalak et Tchintabaraden), Tillabéry Hamidou,1995; Saadou, 2005; Mani, 2009; etc.), il existe (département de Téra et Ouallam) et la Commune peu ou pas de données scientifiques sur la chèvre du sahel, Urbaine de Niamey (figure 1) de juillet 2011 à mai

Figure 1. Zone d’étude Caractéristiques phénotypiques de la chèvre du sahel 13

Tableau 1. Nombre de sites et de caprins par région de présentée dans la note technique de Danchin-Burge (lien l’échantillon d’étude électronique: idele.fr/?eID = cmis_download&oID = work- Régions Tahoua Tillabéry Niamey Total space://SpacesStore/415d632d-8485-495c-9e02-87c53ab30 fcc) a été utilisée. Cette échelle distingue neufs (9) patrons N sites(*) 15 18 27 60 pigmentaires dont: N caprins 103 238 102 443 Moyennes/site 6,87 ± 2,92 13,22 ± 8,13 3,78 ± 4 7,38 ± 6,67 ➢ deux patrons unis: eumélanique (noir ou chocolat) et Maximum/site 10 27 16 27 phaeomélanique; Minimum/site 2211 ➢ cinq patrons présentant des variations de mélanine au N = nombre niveau dorsal, ventral et de la face: joue rouge, (*) Site = village ou quartier eumélanique et feu, eumélanique uni et feu à ventre clair, chamoisé (ou blaireau, ou badger face) et patron sauvage; ’ 2012. Les échantillons de sites et d individus ont été ➢ Deux patrons mantelés: mantelé antérieur et mantelé calculés à partir de la formule de proportion suivante: postérieur.

n = (Z/e)2 × p(1 − p) Sur ces patrons, il peut y avoir des dépigmentations de la phaeomélanine mais aussi, des panachures. Les dépigmentations peuvent s’étendre sur le corps entier de –“” ’ n est la taille de l échantillon; l’animal se traduisant par un mélange de poils blancs –“ ” fi Z est le niveau de con ance, en général, Z = 1,96 à un soit sur un animal de couleur entièrement noir ce qui fi intervalle de con ance de 95%; donne le gris, soit sur un animal de couleur entièrement –“ ” P est le niveau initial des indicateurs; rouge, ce qui donne le rouan ou au niveau de la tête (à –“” ’ e est la marge d erreur. la base des oreilles, des yeux ou du nez) donnant l’impres- Du fait de l’importance de la chèvre du sahel dans la zone sion des mouchetures, on parle de « frosting ». d’étude, les valeurs de “P” et “e” choisis pour les calculs Les panachures quant à elles sont des altérations pigmen- du nombre de sites et d’individus sont respectivement taires qui, se caractérisent par la présence des tâches de 80% et 10% et 50% et 5%. En outre, s’agissant de blanches à contour irrégulier sur une ou plusieurs parties l’échantillon d’individus, un forfait supplémentaire de du corps allant de la simple tâche (panachure localisée), 15% de la valeur calculée a été ajouté. à la grande panachure étendue à la quasi-totalité du En raison du manque de données sur les effectifs caprins corps (panachure généralisée), rendant souvent même au niveau villageois d’une part et en fonction du mode difficile la lecture du patron pigmentaire (patron illisible). ’ ’ d élevagedanslazoneéchantillondautre part, le nombre A partir des données des paramètres visuels, les indices de ’ ’ d individus dans les sites n a pas été pondéré en fonction de primarité (ou de traditionalité) tels que définis par la densité des caprins. Machado, Lauvergne et Souvenir zafindrajaona (1992)et Au total 443 caprins (77% des femelles) appartenant à 145 Lauvergne et al. (1993) ont été calculés. exploitations sur 60 sites choisis au hasard en fonction de ➢ Indice de primarité «loci à effet visible en l’adhésion volontaire des éleveurs ont été caractérisés. Le ségrégation» (IPs): c’est la fréquence (ou pourcentage) tableau 1 présente le nombre de sites et de caprins par du nombre des loci à effet visible en ségrégation (n ) région de l’échantillon d’étude. Selon l’âge, apprécié à s rapporté au nombre total de loci à effet visible (Ns) partir de l’échelle de dentition (Hamito, 2009), identifiés dans l’espèce. l’échantillon est composé de 22,1% de caprins de 0 – 1 – – an; 16,5% de 1 2 ans; 22,3% de 2 3 ans; 18% de 3 IPs = ns/Ns – 4 ans et 20,3% de plus de 4 ans. ➢ Indice de primarité allèles au locus agouti (IPa): Soit Na le nombre total d’allèles en agouti considéré dans Indicateurs de caractérisation une population et na le nombre d’allèles identifiés Plusieurs critères de classification des races caprines afri- dans cette même population, l’indice de primarité caines ont été relatés par Bouchel et Lauvergne (1996). allèles au locus agouti se définit comme suit: Les indicateurs de caractérisation considérés dans cette = − / − étude, extraits de FA0 (1986), Lauvergne et al. (1993)et IPa (na 1) (Na 1) FAO (2012) sont les ports des oreilles et des cornes, la présence/absence de la barbiche et de pampilles (pendeloques), la structure (longueur et type) de pelage et les patrons pigmentaires. Ces paramètres ont été appréciés par observa- Analyses statistiques tion directe et les informations sont répertoriées sur une fiche Les différents indicateurs étant codifiés, une maquette de caractérisation élaborée à cet effet. Cependant, pour la élaborée au logiciel SPSS 17.0 a été utilisée pour enre- détermination des patrons pigmentaires, l’échelle visuelle gistrer toutes les informations recueillies. Ces dernières 14 Mani et al.

ont fait l’objet d’une analyse descriptive préliminaire et de prédominance des patrons phaeomélanique (30% en moy- test de comparaison (Test de Khi carré) au logiciel SPSS et enne avec 28% des mâles et 31% des femelles), une analyse en composante multiple (ACM) au logiciel eumélanique chocolat (29% en moyenne dont 28% des XLSTAT 2012. mâles et 29% des femelles) et eumélanique noir (27% en moyenne dont 24,4% des mâles et 28% des femelles). Il faut cependant noter que la majorité de patrons observés (68,84%) est panachée (tache blanche) dont 65% pana- Résultats chure généralisée (64% des mâles et 67% des femelles); le pourcentage de dépigmentation est de 11,5% dont les Analyse descriptive des paramètres qualitatifs plus dominantes sont le rouan avec en moyenne 45,5% étudiés (11,1% des mâles et 52,2% des femelles) et le frosting qui représente 40% (66,7% des mâles et 34,8% des Le tableau 2 présente les fréquences observées selon le femelles). La figure 2 présente quelques patrons observés. sexe des paramètres qualitatifs étudiés. Il ressort de ce tableau que le port d’oreilles est tombant (60%) aussi bien En considérant les variables (tableau 2) dont les modalités chez les mâles que chez les femelles, les cornes sont de ont des effectifs supérieurs à 5 (ports d’oreilles et de type ibex dans plus de 85% des cas (chez les deux cornes, présence/absence de barbiche et des pampilles, sexes). La barbiche et les pendeloques sont observées sur types de poils), la répartition des fréquences selon le les deux sexes en moyenne respectivement sur 44,2% sexe est statistiquement non significative (P > 0,0 5) à l’ex- (54,9% des mâles et 41,1% des femelles) et 52,3% ception de la présence/absence de barbiche (P = 0,014). (56,4% des mâles et 51% des femelles) des animaux. Le S’agissant de la répartition selon les régions, le test de pelage chez les deux sexes est entièrement court à ras Khi carré indique qu’elle est statistiquement significative (plus de 97%) et lisse à 56,1 % (51% des mâles et pour les paramètres présence/absence de barbiche et des 57,6% des femelles). Il ressort en outre que tous les pampilles, port d’oreilles, type de poils et non significative patrons pigmentaires ont été observés avec une pour port de cornes (tableau 3).

Tableau 2. Répartition des fréquences des paramètres qualitatifs selon le sexe

Paramètres Modalités Mâle Femelle Tout sexe

N% N % N%

Port d’oreilles Dressé 7 6,9a 25 7,3a 32 7,2 Pédonculé 34 33,3a 113 33,1a 147 33,2 Tombant 61 59,8a 203 59,5a 264 59,6 Port de cornes Ibex 87 86,1a 287 84,7a 374 85,0 Autre 14 13,9a 52 15,3a 66 15,0 Présence/absence de barbiche Présence 56 54,9a 140 41,1b 196 44,2 Absence 46 45,1a 201 58,9b 247 55,8 Présence/absence de pendeloques Présence 57 56,4a 174 51,0a 231 52,3 Absence 44 43,6a 167 49,0a 211 47,7 Longueur de poils Court à ras 99 97,1 335 98,2 434 98,0 Mi-long 3 2,9 6 1,8 9 2,0 Types de poils Lisse 50 49,0a 143 42,1a 193 43,7 Non lisse 52 51,0a 196 57,6a 248 56,1 Patrons pigmentaires Eumélanique noir 20 24,4 82 27,9 102 27,1 Eumélanique chocolat 23 28,0 84 28,6 107 28,5 Phaeomélanique 23 28,0 91 31,0 114 30,3 Joue rouge 2 2,4 4 1,4 6 1,6 Eumélanique et feu 0 0,0 1 0,3 1 0,3 Eumélanique uni et feu ventre claire 0 0,0 4 1,4 4 1,1 Chamoisé ou blaireau 4 4,9 10 3,4 14 3,7 Sauvage 1 1,2 3 1,0 4 1,1 Mantelé antérieure 5 6,1 1 0,3 6 1,6 Mantelé postérieur 1 1,2 6 2,0 7 1,9 Invisible 3 3,7 8 2,7 11 2,9 Dépigmentation Frosting 6 66,7 16 34,8 22 40,0 Rouan 1 11,1 24 52,2 25 45,5 Gris 2 22,2 6 13,0 8 14,5 Panachure Localisée 26 36,1 78 33,5 104 34,1 Générale 46 63,9 155 66,5 201 65,9

Les fréquences selon le sexe portant des lettres différentes sur la même ligne sont statistiquement différentes. Caractéristiques phénotypiques de la chèvre du sahel 15

Figure 2. Quelques patrons pigmentaires observés [eumélanique noir Téra (a), eumélanique noir Niamey (b), eumélanique chocolat Téra (c), phaeomélanique Niamey (d), patron illisible Niamey (e), patron illisible Téra (f), chamoisé Niamey (g et h), sauvage Niamey (i et j), mantelé postérieur Niamey (k), mantelé antérieur Niamey (l), grise Niamey (m), chèvre rouanée Téra (n), frosting Niamey (o), panachure Téra et Oullam ( p et q)].

Indices de primarité est la plus importante sur ces axes, mais aussi, ces axes représentent la plus grande proportion des cosinus carrés Les tableaux 4 et 5 présentent respectivement les différents élevés des modalités. Le premier plan factoriel (F1-F2) loci à effet visible en ségrégation et les allèles au locus peut cependant être utilisé pour l’interprétation. La carte fac- agouti identifiés. Dans les départements d’Abalak et de torielle des modalités (figure 3) indique que toutes les Tchintabaraden (1ers sites d’enquête), les données pour modalités sont bien représentées. Il ressort que: les indices de primarité n’ont pas été abordées. Il ressort (tableau 4) que sur la population totale, IPS = 0,92. En – le port d’oreilles est majoritairement pédonculé dans les effet, un seul locus à effet visible en ségrégation pressenti régions de Tahoua et Niamey et tombant dans la région notamment «polled» caractérisant les chèvres mottes n’a de Tillabéry; le port d’oreilles dressées bien que faible- pas été identifié. La valeur de l’IPS varie selon les ment représenté dans les trois régions est majoritaire départements, elle est de 0,83 dans le département de dans la région de Tahoua; Tahoua et à Niamey et 0,67 dans les départements de – les cornes conformément au test de Khi carré, sont Téra et Ouallam (région de Tillabéry). Cependant, le tab- majoritairement de type ibex dans les trois régions leau 5 indique que les allèles au locus agouti pressentis et avec une forte proportion dans la région de Tillabéry; recherchés ont tous été identifiés dans la zone d’étude, – la barbiche et les pampilles sont présentes sur la majorité IPA = 1. Selon les départements, l’IPA varie de 0,57 à 0,86. de chèvres de la région de Tillabéry et majoritairement absentes dans les populations caprines de la région de Niamey et Tahoua; Analyse multi-variée des paramètres qualitatifs – les poils, de types majoritairement lisses dans la région Au-delà de l’étude de la traditionalité, une Analyse en de Tahoua et non lisses dans les régions de Tillabéry Composante Multiple (ACM) a été conduite afin de mieux et Niamey, sont presque exclusivement courts à ras apprécier la dispersion des paramètres phénotypiques visi- dans les trois régions; bles. Les deux premiers axes factoriels F1 et F2 totalisent – S’agissant des patrons pigmentaires, la figure 3 montre 51,47% de l’inertie cumulée (tableau 6). La proportion des que le patron phaeomélanique est majoritairement contributions élevées des modalités des différentes variables observé à Tillabéry (majorité des chèvres), 16 Mani et al.

Tableau 3. Paramètres qualitatifs selon les régions

Variables Modalités Tahoua Tillabéry Niamey Total

N% N % N % N%

Port d’oreilles Dressé 21 20,6a 2,8b 9 8,7c 32 7,2 Pédonculé 44 43,1a 24 10,1b 79 76,7c 147 33,2 Tombant 37 36,3a 212 89,1b 15 14,6c 264 59,6 Port des cornes Ibex 83 82,2a 208 88,1a 83 80,6a 374 85,0 Autre 18 17,8a 28 11,9a 20 19,4a 66 15,0 Barbiche Présence 46 45,1a 123 51,7b 27 26,2c 196 44,2 Absence 56 54,9a 115 48,3b 76 73,8c 247 55,8 Pampille Présence 32 31,7a 168 70,6b 31 30,1c 231 52,3 Absence 69 68,3a 70 29,4b 72 69,9c 211 47,7 Type de poils Lisse 90 88,2a 92 38,8b 11 10,7c 193 43,7 Non lisse 12 11,8a 144 60,8b 92 89,3c 248 56,1 Longueur de poils Mi-long 8 7,8 1 ,4 0 ,0 9 2,0 Cours à ras 94 92,2 237 99,6 103 100,0 434 98,0 Patrons pigmentaires Eumélanique noir 22 43,1 47 21,2 33 32,0 102 27,1 Eumélanique chocolat 3 5,9 58 26,1 46 44,7 107 28,5 Phaeomélanique 7 13,7 98 44,1 9 8,7 114 30,3 Joue rouge 1 2,0 0 ,0 5 4,9 6 1,6 Eumélanique et feu 1 2,0 0 ,0 0 ,0 1 ,3 Eumélanique uni et feu ventre claire 2 3,9 2 ,9 0 ,0 4 1,1 Chamoisé ou blaireau 2 3,9 8 3,6 4 3,9 14 3,7 Sauvage 0 ,0 2 ,9 2 1,9 4 1,1 Mantélé antérieure 4 7,8 0 ,0 2 1,9 6 1,6 Mantelé postérieure 4 7,8 1 ,5 2 1,9 7 1,9 Illisible 5 9,8 6 2,7 0 ,0 11 2,9 Dépigmentation Frosting 1 33,3 7 29,2 14 50,0 22 40,0 Rouan 1 33,3 13 54,2 11 39,3 25 45,5 Gris 1 33,3 4 16,7 3 10,7 8 14,5 Panachure Localisée 11 32,4 64 32,2 29 40,3 104 34,1 Générale 23 67,6 135 67,8 43 59,7 201 65,9

Les fréquences sur la même ligne, portant des lettres différentes sont signiﬁcativement différentes selon les régions.

eumélanique noir est observé dans les trois régions avec observé à l’ouest (Niamey et Tillabéry), joue rouge est des proportions beaucoup plus importantes à Tahoua et majoritairement identiﬁé à Niamey et absent dans la Niamey, eumélanique chocolat est majoritairement région de Tillabéry, eumélanique et feu est exclusivement signalé dans la région de Tahoua, eumélanique et Tableau 4. Indice de primarité «loci à effet visible en ségrégation» (IPs) Tableau 5. Indice de primarité allèles au locus agouti (IPa)

Loci à effet visible en Départements Total Allèles au locus Départements Total ségrégation identiﬁés agouti identiﬁés Tahoua Téra Ouallam Niamey Tahoua Téra Ouallam Niamey 52 81 157 103 52 81 157 103

Agouti 1 1 1 1 1 Eumélanique (noir, 11 1 11 Brown 1 1 1 1 1 non agouti) Roan 1 1 1 1 1 Chamoisé 1 1 1 1 1 Frosting 1 1 0 1 1 Eumélanique et feu 01 0 01 Beard 1 1 1 1 1 ventre clair Ear Length 1 0 0 1 1 Eumélanique et feu NR NR NR NR NR Ear Carriage 1 1 1 1 1 ventre coloré Ear Curling NR NR NR NR NR Joue rouge 1 0 0 1 1 Polled 0 0 0 0 0 Mantelé antérieur 1 0 0 1 1 Horn Reduction 1 0 0 1 1 Mantelé postérieur 1 0 1 1 1 Hair Length 0 0 1 0 1 Phaeomélanique 11 1 11 Wattles 1 1 1 1 1 (fauve) Spotting 1 1 1 1 1 Sauvage fauve 0 1 1 1 1 ns 10 8 8 10 11 Na 6 5 5 7 8 Ns 12 12 12 12 12 NA 8 8 8 8 8 Ips 0,83 0,67 0,67 0,83 0,92 IPa = (na – 1)/(Na – 1) 0,71 0,57 0,57 0,86 1

NR = Non recherché Ns: nombre total de loci à effet visible identiﬁés, ns: NR = Non recherché; NA: nombre total d’allèles en agouti considéré dans nombre des loci à effet visible en ségrégation rapporté. une population; na : na le nombre d’allèles identiﬁé. Caractéristiques phénotypiques de la chèvre du sahel 17

Tableau 6. Valeurs propres et pourcentages d’inertie selon les axes

F1 F2 F3 F4 F5 F6 F7 F8 F9

Valeur propre 0,242 0,201 0,185 0,176 0,153 0,152 0,147 0,146 0,143 Inertie (%) 9,413 7,799 7,206 6,848 5,965 5,906 5,718 5,684 5,556 % cumulé 9,413 17,212 24,419 31,267 37,232 43,138 48,856 54,539 60,095 Inertie ajustée 0,013 0,005 0,002 0,002 0,000 0,000 0,000 0,000 0,000 Inertie ajustée (%) 38,459 13,006 7,042 4,319 0,433 0,317 0,068 0,042 0,000 % cumulé 38,459 51,466 58,508 62,827 63,261 63,578 63,646 63,688 63,688

feu ventre clair est observé à Tahoua et Tillabéry, Traoré et al., 2008; Hamito, 2009; Yakubu, 2010; chamoisé est indiqué dans les trois régions et majori- Halima et al., 2012). tairement dans la région de Tillabéry, sauvage à ’ Niamey et Tillabéry, mantélé antérieur et postérieur Le port d oreilles est à environ 60% tombant et 33% majoritairement à Tahoua et les chèvres totalement pédonculé. Pareils constats ont été rapportés sur la panachées (patron invisible) sont identiﬁées dans les chèvre du sahel au Burkina et au Tchad (Traoré et al., trois régions. 2008; Dumas, 1980). En effet, selon Dumas (1980), les oreilles de la chèvre du sahel tchadienne sont longues, larges, pendantes sur les joues, dépassant le museau lors- Discussion qu’elles sont rabattues vers l’avant. Seuls ou associés à d’autres, les indicateurs choisis dans Tous les animaux (100%) sont cornus (aucun animal cette étude sont classiques pour la caractérisation des petits motte a été rencontré). Cela est conforme aux résultats ruminants en Afrique et à travers le monde (Lauvergne rapportés par Yakubu, Raji et Omeje (2010) sur les et al., 1993; Lanari et al., 2003; Traoré et al., 2006; chèvres locales du Nigéria, Lauvergne et al.(1993)au

Figure 3. Carte factorielle des modalités des paramètres phénotypiques qualitatifs 18 Mani et al.

Nord Cameroun. Les cornes sont l’un des appendices African Long-Legged) et la chèvre mossi (Wilson, 1991; importants en milieu tropical car intervenant dans le Meyer, 2001; Traoré et al., 2006; Traoré et al., 2008; mécanisme de thermorégulation (Robertshow, 2006; Yakubu, Raji et Omeje, 2010). FAO, 2012). Dans le cadre de cette étude, les dimen- La robe est multicolorée, tous les patrons pigmentaires ’ sions des cornes varient en fonction de l âge et du identifiés chez l’espèce caprine ont été observés avec dom- sexe. Elles sont longues chez les animaux adultes, inance chez les deux sexes des patrons Phaeomélanique fi nes chez les femelles et épaisses et souvent torsadées (Brown), Eumélanique chocolat et Eumélanique noire chez les mâles. Les cornes sont de type ibex (plus de (30,3%, 28,5% et 27,1% respectivement). Cependant, 84% chez les deux sexes), conformément aux résultats dans la plupart des cas (plus de 68%) la robe est panachée de Traoré et al. (2008) chez les chèvres locales du et le pourcentage de dépigmentation est non négligeable Burkina en général et particulièrement sur la chèvre du (11,5%). Ces résultats sont comparables à ceux de sahel, et de Dumas (1980) qui précise pour la chèvre Lauvergne et al. (1993) au Nord Cameroun (tous les patrons du sahel tchadienne «les cornes sont assez longues observés avec prédominance des patrons phaeomélanique, chez le mâle, elles sont dirigées en arrière et en haut sauvage et eumélanique respectivement) et de Traoré en divergeant». Par contre ces résultats diffèrent de et al. (2008) chez la chèvre du sahel. En outre, ces ceux rapportés par Lauvergne et al. (1993)auNord résultats se distinguent de ceux de Yakubu, Raji et Omeje Cameroun (11,7% type ibex). (2010) sur les races WAD et Red Sokoto goats au Nigéria La présence de la barbiche a été identifiée chez les ani- avec prédominance respective de eumélanique noir et maux des deux sexes (mâles 55% et femelles 41%). La phaeomélanique (red). L’importance des panachures ’ présence/absence de la barbiche est le seul paramètre qua- d une part et le pourcentage non négligeable de litatif dont la différence entre sexe a été statistiquement dépigmentation d’autre part peuvent trouver une explication significative. Ce dimorphisme sexuel quant au port de la dans le cadre de la thermorégulation. En effet selon Acharya barbe a été rapporté par plusieurs auteurs avec des propor- et al. (1995), les caprins à poils courts sont peu résistants tions variables selon le sexe: Wilson (1991) sur la chèvre aux radiations solaires et consomment moins que ceux à West African Long-Legged (chèvre du sahel), Traoré poils longs. De même, les caprins noirs sont moins et al. (2006) sur la chèvre locale mossi du Burkina Faso, résistants que les rouges, à leur tour moins résistants que Yakubu, Raji et Omeje (2010) sur les races caprines les blancs. L’émergence de la panachure et de locales du Nigéria. Traoré et al. (2008) ont constaté un dépigmentation sur les autres patrons, autrement dit la ’ taux d’absence de barbe de plus de 75% sur les chèvres présence des taches blanches d une part et le mélange des locales du Burkina. La présence de la barbe chez les poils blancs aux poils rouges et noirs d’autre part, rehaus- deux sexes avec précellence chez le mâle tient d’une sera probablement la résistance aux radiations solaires et part, du fait que la barbe constitue un caractère sexuel sec- peut donc traduire une certaine adaptation au milieu. ondaire chez le mâle et d’autre part, le gène codant pour la L’ACM indique que tous les paramètres qualitatifs étudiés barbe existe à l’état récessif chez les femelles et dominant sont régulièrement répartis dans toutes les régions de la chez les mâles (Yakubu, Raji et Omeje, 2010). zone d’étude. Cependant, elle fait ressortir que les caprins dans les trois régions se distinguent par le port des oreilles Comme autre appendice, la présence de pendeloques a été (majoritairement tombant dans la région de Tillabéry, observée sur plus de 50% chez les deux sexes. Ces pédonculé à Niamey et pédonculé et dressé dans la résultats se rapprochent de ceux de Traoré et al. (2008) région de Tahoua), la présence de la barbe et des pendelo- qui rapportent que bien que la fréquence du port de pende- ques (majoritairement présents dans la région de Tillabéry loques dans la population caprine locale soit relativement et absents dans la région de Tahoua et Niamey) et le patron faible (30%), la chèvre du sahel spécifiquement recouvre pigmentaire (majoritairement eumélanique noir à Tahoua, plus de 75%. Cependant, des fréquences inférieures ont eumélanique chocolat à Niamey et phaeomélanique à été rapportées par Yakubu, Raji et Omeje (2010)au Tillabéry). Cette hétérogénéité selon les régions reflète Nigéria sur les chèvres de races West Africain Dwarf l’originalité de la race caractérisée par une diversité de Goat (mâle 31,64% et femelle 5,78%) et Red Sokoto forme dans sa zone d’expansion (Wilson, 1991; Meyer, Goat (mâle 1,69% et femelle 3,78%). Selon Ricordeau et 2001). Bouillon (1967), dans l’espèce caprine, les pendeloques sont présentes avec une fréquence variable dans les différentes races et il est bien difficile de dire si c’est la présence ou l’absence de pendeloques qui constitue le Conclusion caractère normal ou sauvage. Ces résultats font ressortir une hétérogénéité sur le plan Le pelage est presque entièrement court à ras, les animaux phénotypique qualitatif des caprins dans la zone d’étude. à poils mi-longs sont rarement rencontrés (2%) et cela En effet, selon les régions, la répartition des paramètres conformément aux races caprines locales de la sous visuels étudiés a été statistiquement différente à l’excep- région, notamment la chèvre rousse (red sokoto goat), la tion du type des cornes. Visiblement la chèvre dans la chèvre Djallonké (WAD goat), la chèvre du sahel (West zone d’étude est multicolorée. La fréquence des Caractéristiques phénotypiques de la chèvre du sahel 19

panachures est élevée (plus de 68%) d’où une proportion Hamito, D. 2009. Goat breeds of Ethiopia: A guide for identification and importante des robes pies (pie rouge dans la région de utilization. Ethiopia Sheep and Goat Productivity Improvement Tillabéry et la CUN et pie noire dans la région de Program (ESGPIP).Technical Bulletin 23: 11. Tahoua). Les animaux des deux sexes portent des cornes Lanari, M.R., Taddeo, H., Domingo, E., Perez-Centeno, M. and (majoritairement de type ibex), une proportion importante Gallo, L. 2003. Phenotypic differentiation of exterior traits in local tous sexes confondus porte la barbiche (proportion statisti- Criollo Goat Population in Patagonia (Argentina). Archiv fur Tierzucht Dummerstorf 46 (4): 347–356. quement plus élevée chez le mâle) et des pendeloques. Le pelage est court à ras et les oreilles sont majoritairement Lauvergne, J.J., Bourzat, D., Souvenir, Z.P., Zeuh, V. et Ngo Tama, ’ A.C. 1993. Indices de primarité de chèvres au Nord Cameroun et au tombantes. L analyse des indices de primarité indique Tchad. Revue d’Elevage et de Médecine Vétérinaire des Pays que cette population présente globalement, les Tropicaux 46 (4): 651–665. ’ caractéristiques d une population primaire. Cependant, fi ’ fi Machado, T.M., Lauvergne, J.J. et Souvenir Za ndrajaona, P. l analyse multivariée fait ressortir quelques spéci cités 1992. Peuplement caprin du Brésil: Scénario du peuplement caprin selon les régions. du Brésil depuis la découverte. Archivos de Zootecnia 41 (extra): – Ces résultats doivent être complétés par des données 455 466. biométriques afind’affiner l’image phénotypique type de Mani, M. 2009. Le cycle sexuelde lachèvre rousse de Maradi: Etude descrip- la chèvre du sahel dans la zone d’étude. tive et progestéronomie. Mémoire de Master II. EISMV Dakar 5: 31. Marichatou, H., Mamane, L., Banoin, M. et Baril, G. 2002. Performances zootechniques des caprins au Niger: étude comparative Remerciements de la chèvre rousse de Maradi et de la chèvre à robe noire dans la zone Les auteurs remercient le partenaire financier qui est de Maradi. Revue d’Elevage et de Médecine Vétérinaire des Pays – l’Agence Universitaire de la Francophonie (AUF), Tropicaux 55 (1): 79 84. précisément son Bureau Afrique de l’Ouest et le CNF de Meyer, C. 2001. Races d’animaux d’élevage en Afrique intertropicale et Niamey. Ils remercient également le Programme de méditerranéenne: Les caprins. Mémento de l’Agronome. Productivité Agricole en Afrique de l’Ouest (PPAAO) CIRAD-EMVT. France, 22. qui a contribué matériellement et financièrement aux NIGER. Recensement Général de l’Agriculture et du Cheptel fi activités sur les caprins du Sahel. (RGAC). 2007. Résultats dé nitifs (Volet cheptel) (2): 170. Rhissa, Z. 2010. Revue du secteur de l’élevage au Niger. Ministère de l’Elevage, des Pêches et des Industries Animales. Niger, 115. Références bibliographiques Ricordeau, G. et Bouillon, J. 1967. Hérédité des pendeloques en race saanen: différences de fécondité entre les génotypes avec et sans pendeloques. Annales de Zootechnie 16 (3): 263–270. Acharya, R.M., Gupta, U.D., Sehgal, J.P. et Singh, M. 1995. Coat characteristics of goats in relation to heat tolerance in the hot tropics. Robertshaw, D. 2006. Mechanisms for the control of respiratory evap- Small Ruminant Research. 18: 245–248. orative heat loss in panting animals. Journal of Applied Physiology 101: 664–668. Bouchel, D. et Lauvergne, J.J. 1996. Le peuplement de l’Afrique par la chèvre domestique. Revue d’Elevage et de Médecine Vétérinaire des Robinet, A.H. 1967. La chèvre rousse de Maradi son exploitation et sa Pays Tropicaux 49 (1): 80–90. place dans l’économie et l’élevage de la République du Niger. ’ Danchin-Burge, C. [en ligne]. La couleur chez les caprins. Institut de Revue d Elevage et de Médecine Vétérinaire des Pays Tropicaux 20 – l’Elevage Dept. Génétique. (Disponible à l’adresse électronique: (1): 129 186. idele.fr/?eID = cmis_download&oID = workspace://SpacesStore/ Saadou, M. 2005. Performance de reproduction et de production de la 415d632d-8485–495c-9e02–87c53ab30fcc). chèvre rousse de Maradi en milieu rural au Niger. Thèse de Dumas, R. 1980. Contribution à l’étude des petits ruminants du Tchad. Médecine Vétérinaire Dakar 16: 82. Revue d’Elevage et de Médecine Vétérinaire des Pays Tropicaux 33 Traoré, A., Hamidou, H.T., Adama, K., Luis, J.R., Ivan, F., Isabel, – (2): 215 233. A., Sangaré, M., Bouchel, D., Jean Paul, P., Dominique, F., FAO. 1986. Animal genetic resources data banks. 2. Descriptor lists for Sawadogo, L. and Goyache, F. 2008. Multivariate analyses on mor- cattle, buffalo, pigs, sheep and goat. FAO. Rome, Italia. Animal phological traits of goats in Burkina Faso. Archiv fur Tierzucht Production and Health 59 (2): 151. (available at http://www.fao. Dummerstorf 51 (6): 588–600. org/docrep/009/ah760e/ah760e00.HTM) Traoré, A., Tamboura, H.H., Kaboré, A., Yaméogo, N., Bayala, B. et FAO. 2012. Phenotypic characterization of animal genetic resources. Zaré, I. 2006. Caractérisation morphologique des petits ruminants Food and Agriculture Organization of the United Nations. Rome, (ovins et caprins) de race locale “Mossi” au Burkina Faso. GRI Italia. Animal Production and Health Guidelines 11: 144. (available 2006 39: 39–50. at http://www.fao.org/docrep/015/i2686e/i2686e00.htm) Wilson, T.R. 1991. Small ruminant production and the small ruminant Halima, H., Michael, B., Barbara, R. and Markos, T. 2012. genetic resource in tropical Africa. FAO. Rome. Animal Production Phenotypic characterization of Ethiopian indigenous goat populations. and Health Paper, 194. African Journal of Biotechnology 11 (73): 13838–13846. Yakubu, A., Raji, A.O. and Omeje, J.N. 2010. Genetic and phenotypic Hamidou, I. 1995. Contribution à l’analyse des paramètres de reproduc- differentiation of qualitative traits in Nigerian indigenous goat and tion de la chèvre Rousse de Maradi. Thèse de Médecine Vétérinaire sheep populations. ARPN Journal of Agricultural and Biological Dakar 11: 83. Science 5 (2): 58–66. Animal Genetic Resources, 2014, 54, 21–32. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000046

Caractérisation de la chèvre du sahel au Niger par analyse des indices biométriques et des paramètres phénotypiques quantitatifs

M. Mani1, H. Marichatou1, M.M.M. Mouiche2, M. Issa3, I. Chaïbou4, A. Sow5, M. Chaïbou1 and J.G. Sawadogo5 1Faculté d’Agronomie, Université Abdou Moumouni, BP 10960, Niamey, Niger; 2ESMV de Ngaoundéré, Cameroun; 3Faculté des Sciences et Techniques, Université Abdou Moumouni, BP 10960, Niamey, Niger; 4Faculté d’Agronomie de l’Université de Maradi, Niger; 5Ecole Inter-Etats des Sciences et Médecines Vétérinaires, Dakar, Sénégal

Résumé Très répandue en Afrique sèche, la chèvre du sahel représente environ 80% du troupeau caprin Nigérien. Conformément aux orienta- tions nationales au Niger et dans la perspective d’une meilleure gestion de cette ressource génétique aux niveaux local et national des travaux de caractérisation phénotypique de la race ont été conduits au Nord-Ouest du pays de juillet 2011 à mai 2012. Sur la base de l’adhésion volontaire des éleveurs, des investigations de caractérisation des paramètres quantitatifs (hauteur au garrot et à la croupe, tour de poitrine, profondeur du thorax, largeur du bassin et longueurs du corps, des cornes et des oreilles) ont concerné 443 caprins (77% de femelle, 39% ayant 4 paires d’incisives permanentes) repartis dans 145 exploitations, 60 sites (7,38 ± 6,87 têtes/site), dans les régions de Tillabéri, Tahoua et Niamey. L’analyse des indices biométriques couplée à l’analyse des variances (ANOVA) et des analyses multivariées (ACP, AFD) ont fait ressortir que le cheptel caprin dans cette zone comporte quatre sous types morphologiques (Tillabéri, Tahoua, Niamey et un sous type intermédiaire Niamey-Tillabéri). Du fait de la coexistence à Niamey de plusieurs races (locales et exotiques) et son corolaire le brassage génétique, la chèvre du sahel correspondrait à deux sous types: celui de Tillabéri de grande taille (64,98 ± 6,77 cm au garrot et 68,27 ± 5,91 cm à la croupe) à oreilles et cornes longues (20,3 ± 3,08 et 14,75 ± 4,43 cm respectivement) et de Tahoua également de grande taille (60,71 ± 10,81 cm au garrot et 62,46 ± 10,65 cm à la croupe) à oreilles et cornes courtes (12,32 ± 3,62 cm et 9,95 ± 4,01 cm respectivement).

Mots-clés: caractéristiques phénotypiques, quantitatifs, chèvre du sahel, Niger

Summary Found throughout the dry regions of Africa, the Sahel goat represents about 80% of Niger’s goats. In accordance with national breeding policy in Niger and with a view to better planning management of that genetic resource at local and national levels, its phenotypic characterization has been conducted in the northwestern part of the country from July 2011 to May 2012. Based on the voluntary participation of farmers, investigations to characterize the quantitative parameters (height at withers and rump, body length, chest girth, chest depth, hip width and length of horns and ears) focused on 443 goats (77% female, 39% adult) distributed in 145 farms and 60 sites (7,38 ± 6,87 animals/site) in the Tillabéri, Tahoua and Niamey regions. Biometric indexes analysis coupled with analysis of variance and multivariate analysis (ACP, AFD) showed that the study area goat has four morphological subtypes (Tillabéri, Tahoua, Niamey and an intermediate Niamey-Tillabéri). Due to the coexistence of several breeds (local and exotic) in Niamey and consequences like genetic mixing, the Sahel goat can correspond to two subtypes: a large one (64.98 ± 6.77 and 68.27 ± 5.91 cm at withers and rump respectively), long ears and horns (20.3 ± 3.08 and 14.75 ± 4.43 cm) in Tillabéri and another large one (60.71 ± 10.81 and 62.46 ± 10.65 cm at wither and rump), shorter ears and horns (12.32 ± 3.62 and 9.95 ± 4.01 cm) in Tahoua.

Keywords: Phenotypic characteristics, Niger, Sahel goat, quantitative features

Soumis: 27 Le juin 2013; admis: 7 Le janvier 2014

Introduction sahel, au Sénégal, Mauritanie, Mali, Burkina Faso, Ressource animale caractéristique des zones sèches, la Niger, Tchad, Nigéria (Dumas, 1980; Wilson, 1991; chèvre du sahel est très répandue dans tous les pays du Meyer, 2001). Sa présence est également signalée dans les régions humides d’Afrique occidentale et centrale au Togo, en Côte d’Ivoire, au Cameroun (Lauvergne et al., Correspondence à envoyer à: H. Marichatou Faculté d’Agronomie, Université 1993; Meyer, 2001). Abdou Moumouni, BP 10960, Niamey, Niger. Tel: (00227) 93 91 65 31,. Fax: (00227) 20 31 66 12,. Adresse électronique: [email protected] / marimani_m@ Au Niger, la chèvre du Sahel représente environ 80% du yahoo.fr cheptel caprin (Rhissa, 2010) et est élevée dans toutes les

21 22 H. Marichatou et al.

régions du pays. Cependant, à l’image des autres pays voisins (département de Téra et Ouallam) et la Commune Urbaine tels que le Tchad, le Burkina Faso (Dumas, 1980; Lauvergne de Niamey (Figure 1). et al., 1993;Gnanda,2008;Traoré,2010), la chèvre du sahel Les travaux étant conduits en milieu réel, l’adhésion des est peu connue au Niger. L’objectif visé dans ce travail est éleveurs a constitué un facteur déterminant pour la collecte d’établir les caractéristiques phénotypiques quantitatives de des données. Ainsi, au niveau des exploitations où l’accord la chèvre du sahel au Niger. En effet, une étape importante de principe a été acquis, un certain nombre de têtes est dans la connaissance des ressources génétiques animales, la aléatoirement isolé pour les mensurations nécessaires. Au caractérisation phénotypique est selon FAO (2012), essen- total, 145 exploitations et 443 caprins (dont 77% de tielle pour la planiﬁcation de la gestion des ressources femelles, 22,1% âgés de 0 à 1 an, 16,5% de 1 à 2 ans, zoogénétiques aux niveaux local, national, régional et 22,3% de 2 à 3 ans, 18% de 3 à 4 ans et 20,3% de plus mondial. Les paramètres phénotypiques quantitatifs étant de 4 ans) répartis en 60 sites, ont constitué l’échantillon directement corrélés aux paramètres de production de caractérisation. Le Tableau 1 présente le nombre de (FAO, 2012), leur connaissance est importante pour la sites et de caprins par région de l’échantillon de l’étude maîtrise et l’amélioration des performances de production et le Tableau 2 la répartition selon les tranches d’âge de des animaux. l’échantillon d’étude.

Indicateurs de caractérisation Matériel et Méthodes En Afrique, la classiﬁcation des races caprines a été évolutive, ’ plusieurs combinaisons de critères, généralement adoptées Echantillon d étude chez les bovins, ont été utilisées, une bibliographie Inscrite dans le cadre des investigations de caractérisation conséquente en ce sens a été établie par Bouchel et phénotypique de la chèvre du sahel au Niger, l’étude a Lauvergne (1986). Pour atteindre les objectifs assignés à été conduite cumulativement à celle de la caractérisation notre étude, les indicateurs quantitatifs retenus de la FAO phénotypique qualitative (Mani et al., sous presse), de (FAO, 1986;FAO,2012) sont: la hauteur au garrot (HG), juillet 2011 à mai 2012 dans la partie nord-ouest du la longueur du corps (LC), le tour de poitrine (TP), la largeur Niger, notamment, les régions de Tahoua (départements du bassin (LB), la hauteur à la croupe (HC), la profondeur du de Tahoua, Abalak et Tchintabaraden), Tillabéri thorax (PT), la longueur des oreilles (LO) et la longueur des

Figure 1. Zone d’étude. Caractérisation de la chèvre du sahel au Niger 23

Tableau 1. Nombre de sites et de caprins par région de L’ACP vise à représenter graphiquement les relations entre l’échantillon d’étude. individus par l’évaluation de leurs ressemblances, ainsi ’ Régions Niamey Tillabéri Niamey Total que les relations entre variables par l évaluation de leurs liaisons (Martin, 2004). Ces principes sont également Nombre de sites(*) 15 18 27 60 pris en compte en AFD (Baccini et Besse, 2007). Selon Nombre de caprins 103 238 102 443 Martin (2004), deux individus se ressemblent, ou sont Moyennes/site 6,9 ± 2,9 13,2 ± 8,1 3,78 ± 4 7,38 ± 6,67 ’ ’ Maximum/site 10 27 16 27 proches, s ils possèdent des valeurs proches pour l ensem- Minimum/site 221 1 ble des variables. Cette déﬁnition sous entend une notion de proximité qui se traduit par une distance. Ainsi, la dis- (*) Site = village ou quartier. tance entre deux individus Xi et Yj est déﬁnie par:

p 2 2 cornes (Lc). Les mensurations (Figure 2) sont effectuées à d (Xi, Yj) = (Xik − Y jk ) l’aide d’une règle, d’un mètre ruban et d’un compas. k=1 L’âge des animaux est déterminé par examen de la denti- ’ Par ailleurs, deux variables sont liées si elles ont un tion, l échelle présentée par Hamito (2009) a été utilisée fort coefficient de corrélation linéaire. Sur un échantillon à cet effet tout en considérant que chez l’espèce caprine, de taille n, le coefficient de corrélation linéaire (dit l’animal atteint l’âge adulte à partir de 4 paires d’incisives de Bravais-Pearson ou de Pearson) entre deux variables (Bouchel et al., 2006). X(x1,x2, ....xi)etY(y1,y2, ...yi) est donné par Les données morpho-biométriques ont permis d’établir les l’équation suivante rapportée par Martin (2004)et indices biométriques tels que définis par Lauvergne et al. Rakotomalala (2012). (1993) et Bourzat et al. (1993). n (x − x) . (y − y) ➢ = i=1 i i Indice de gracilité sous-sternal (IGS): rapport du vide r n − 2 n − 2 sous-sternal [différence entre hauteur au garrot (HG) et i=1 (xi x) i=1(yi y) profondeur du thorax (PT)] sur la profondeur du thorax (PT). Plus le coefficient de corrélation est proche de 1 (en valeur absolue), plus la relation est forte (négativement si r < 0 et IGS = (HG − PT)/PT positivement si r > 0). Si r = 0, il ya absence de corrélation.

➢ Indice auriculaire thorax (IAT): longueur de l’oreille Graphiquement, les variables sont représentées dans l’espace (LO) sur la profondeur du thorax (PT). (de dimension n) par des vecteurs. On considère les données centrées réduites par ligne afin de faciliter l’analyse du nuage IAT = LO/PT d’individu car permettant de ramener l’origine des axes au centre de gravité G correspondant à l’individu moyen de la population (Martin, 2004). Analyses statistiques X (x − x, ...... , x − x); Les données recueillies ont fait l’objet d’analyse descriptive 1 n et d’une comparaison des moyennes (ANOVA) au logiciel Y(y1 − y, ...... , yn − y); SPSS (version 17.0). Des analyses multivariées notamment l’Analyse en Composante Principale (ACP) et l’Analyse Le cosinus de l’angle α entre ces vecteurs n’est autre que le Factorielle Discriminante (AFD) ont été effectuées au logi- coefficient de corrélation, Cos(α) = r (formule ci-haut). ciel XLSTAT 2012. Il faut cependant noter que la synthèse des informations de l’analyse descriptive sous forme de tableaux et figures a été faite au tableur Excel. Résultats Tableau 2. Configuration selon l’âge et le sexe de l’échantillon d’étude. Analyse descriptive des données Tranches d’âge Mâle Femelle Total Paramètres biométriques selon la région, le sexe et les ’ N%N%N %tranches d âges Le Tableau 3 présente les moyennes et écarts types des 0–1 an 52 11,7% 46 10,4% 98 22,1% différents paramètres biométriques étudiés selon les – 1 2 ans 20 4,5% 53 12,0% 73 16,5% régions et le sexe. Il ressort trois classements 2–3 ans 19 4,3% 80 18,1% 99 22,3% 3–4 ans 6 1,4% 77 17,4% 83 18,7% décroissants selon les régions: >4 ans 5 1,1% 85 19,2% 90 20,3% – Tillabéri, Tahoua, Niamey pour HG et HC; Total 102 23,0% 341 77,0% 443 100,0% – Tillabéri, Niamey, Tahoua pour TP, PT, LO et Lc; N = Nombre de caprins caractérisé par tranche d’âge. – Tahoua, Tillabéri, Niamey pour LC et LB. 24 H. Marichatou et al.

Figure 2. Mensurations réalisées [hauteur au garrot (a), hauteur à la croupe (b), profondeur du thorax (c), tour de poitrine (d), largeur du bassin (e), longueur des cornes (f)].

Cependant, l’analyse des variances suivie de test post hoc, de Tillabéri par le TP et de plus de celles de Niamey par la ont montré que les différences entre les populations PT et les chèvres de Tillabéri se distinguent statistiquement caprines des trois régions sont statistiquement signiﬁcatives de celles de Tahoua et Niamey par la LC. (P < 0,05) pour HG, LC, LB, HC et LO. En outre, les Le Tableau 4 présente les moyennes et écarts types des chèvres de Tahoua se distinguent statistiquement de celles différents paramètres biométriques selon les tranches Caractérisation de la chèvre du sahel au Niger 25

d’âge. Il résulte que tous les paramètres varient avec l’âge. ’

± 4,01 ± 4,43 ± 4,1 ± 5,49 ± 4,75 L analyse des variances fait ressortir des différences statisti- a a x y b Lc quement significatives à certaines tranches d’âges. Les variations d’HG, HC LC, TP et Lc d’une part et LO et LB d’autre part sont apparues statistiquement semblables. En effet, les différences sont significatives entre les tranches d’âge à l’exception des tranches d’âge successives à partir ± 3,62 9,95 ± 3,08 14,75 ± 2,42± 4,36± 10,85 4,68 11,4 13,18 c a x y b LO de 2 ans pour HG et Lc. Comparativement à ces derniers, les variations sont significatives pour TP entre 2–3 ans et 3–4 ans et non significatives pour LC entre 1–2 ans et 3–4 ans puis, 2–3 ans et > 4 ans et pour HC entre 2–3 ans et > 4 ans. S’agissant de LB et LO cependant, les var-

± 4,77 12,32 ± 4,76 20,3 ± 4,76± 3,97± 14,64 5,62 16,34 17,46 iations sont non signiﬁcatives à l’exception de la tranche a x y b b PT d’âge 0–1 an et les tranches > 3 ans pour LB et en plus entre 1–2 ans et > 4 ans pour LO. La Figure 3 présente l’évolution des différents paramètres biométriques selon le sexe et les tranches d’âge. Il ressort

± 10,65 24,10 ± 5,91 30,09 ± 8,23± 10,32± 7,86 29,60 27,05 29,08 que chez le mâle, les valeurs maximales sont atteintes à c HC a x y b l’âge de 2–3 ans pour les paramètres tels que HG, LC, LB, HC, et la PT, les autres paramètres (TP, LO et Lc) con- tinuant leur croissance jusqu’àl’âge de quatre ans et au delà. Chez la femelle par contre, tous les paramètres étudiés augmentent selon l’âge jusqu’au-delà de 4 ans. ± 5,36 62,46 ± 4,34 68,27 ± 2,76± 5,06± 5,34 59,32 61,58 65,83 c a x x b LB En termes de comparaison selon le sexe, la Figure 3 permet de noter selon les tranches d’âge: – les mâles sont supérieurs aux femelles à toutes les tranches d ’âge pour la Lc; – les mâles sont supérieurs aux femelles pour HG, TP, LC, ± 17,47 20,8 ± 8,18 14,09 ± 9,2±14,46± 12,94 12,29 14,3 15,43 LC c a x y b HC, PT, et inférieurs aux femelles pour LO pour la tranche d’âge 0–1 an et inversement pour la tranche d’âge 1–2 ans; – à2–3 ans, au moment où les paramètres sont globalement à leur optimum chez les mâles, ces derniers sont supérieurs aux femelles pour tous les paramètres; ± 14,85 69,51 ± 9,1 61,82 ± 10,18± 10,58± 10,85 48,17 56,94 61,46 a b x y TP ab – au-delà de trois ans, les mâles sont supérieurs aux femelles pour HG et LO et inférieurs pour TP, LC et PT. Aﬁn de mieux apprécier la répartition selon les régions d’une part et les paramètres directeurs de cette répartition d’autre part, il a été abordé l’analyse des indices ± 10,81 62,2 ± 6,77 67,34 ± 8,98± 10,56± 8,21 65,90 61,03 67,26 c a x y HG b biométriques mettant en exergue la hauteur au garrot, la

61,9 ± 8,27 62,73 ± 9,24 58,18 ± 7,23 13,21 ± 2,27 65,78 ± 7,38profondeur 29 ± 3,25 19,46 ± 3,57 de la 14,02 ± 5,31 poitrine et la longueur des oreilles et 65,78 ± 6,152,08 ± 8,4158,65 ± 8,62 68,54 ± 8,69 57,56 ± 7,03 68,58 ± 9,59 62,76 ± 8,17 42,16 ± 5,63 50,10 ± 9,31 14,32 ± 4,71 10,12 ± 1,71 12,98 68,92 ± ± 2,67 5,3 54,64 ± 8,18 60,81 ± 7,72 30,37 ± 5,05 26,64 ± 3,83 30,55 ± 4,66 20,51 ± 2,91 13,8 ± 2,31 14,91 ± 2,4 14,94 ± 4,17 8,67 ± 4,35 11,56 ± 5,02 59,34 ± 13,1761,22 ± 9,83 61,15 ± 14,38 62,59 ± 15,10 67,95 ± 18,16 70,10 ± 17,29 20,13 ± 5,67 21,05 ± 5,25 60,45 ± 12,82 63,22 ± 9,69 23,99 ± 3,23 24,15 ± 5,26 13,02 ± 2,94 12,05 ± 3,83 9,3552 10,18 ± 3,69 57,05 60,71 64,98 l’analyse multivariée mettant en jeux tous les paramètres morpho-biométriques en relation avec les paramètres qua- 49 25 78 28 74 189 238 103 102 litatifs dont la région et le sexe. Indices biométriques Les indices biométriques moyens sont: ➢ Indice de gracilité sous sternal IGS = 1,21 ± 0,34 Sexes N FTout sexe 443 341 62,15 ± 8,99 63,16 65,82 ± 11,09 60,42 ± 13,42 15,17 ± 5,29 64,85 ± 8,66 28,61 ± 5,35 17,2 ± 4,63 12,77 ± 4,98 M F Total M F Total M F Total ➢ Indice Auriculo-thorax IAT = 0,44 ± 0,14 Les valeurs moyennes indiquent que les caprins dans la zone d’étude sont globalement de type longipes (IGS > 1), aux oreilles longues. Cependant, ces moyennes cachent de Moyennes et écarts types moyens (cm) des paramètres biométriques selon les régions grandes disparités. La Figure 4, présente les valeurs moyennes de ces indices selon les départements de la zone de l’étude. Il ressort que ces indices varient selon les HG: hauteur au garrot,Les TP: moyennes tour des de régions poitrine, portant LC: des longueur lettres du différentes corps, (a, LB: b largeur ou du c) bassin, sur HC: la hauteur même à colonne la sont croupe, statistiquement PT: dfférentes profondeur (P du < 0,05). thorax, LO:r Longueu des oreilles, Lc: Longueur des cornes. Tillabéri Niamey Moyennes générales M 102 58,79 Tahoua Tableau 3. Régions départements. L’IGS varie approximativement avec la 26 H. Marichatou et al.

longueur des pattes (HG – PT) alors que les variations de (a) (bc) (cd) (de) (e) l’IAT sont imputables à l’évolution des deux facteurs (LO

Lc et PT). Ainsi, dans la zone d’étude, les chèvres sont de

9,18 ± 4,53 16,1 ± 4,68 type brèvipes àoreillesmoyennesdanslarégionde 11,66 ± 3,38 12,78 ± 4,48 14,57 ± 4,41 Niamey, longipes à oreilles longues dans la région de Tillabéri (Téra et Ouallam) et longipes à oreille relativement courtes dans la région de Tahoua (Tahoua, Abalak). (a) (ac) (bc) (c) (ab)

LO Analyse multi-variée des paramètres

15,63 ± 3,84 16,14 ± 4,3 17,33 ± 4,69 18,40 ± 5,18 18,54 ± 4,44 biométriques Analyse en composante principale L’ACP fait ressortir (Tableau 5), les corrélations entre les

(a) (ab) (b) (bc) (c) paramètres morpho-biométriques étudiés en variables principales mais aussi entre eux et la région et le sexe en vari- PT ables supplémentaires. Conformément au principe de ’

25,57 ± 4,54 27,91 ± 3,49 28,77 ± 4,89 29,75 ± 4,87 31,27 ± 6,54 Kaiser (1960), les deux premiers axes F1 et F2 de l ACP sont les seuls dont les valeurs propres sont supérieures à 1, ils résument en outre environ 67,8% des observations. Pour cela, et du fait que l’essentiel des variables princi- (a) (bc) (cd) (d) (d) pales mais aussi supplémentaires ont leurs cosinus carrés

HC élevés sur ces axes, le premier plan factoriel (F1xF2) s’adapte bien pour l’interprétation des résultats de l’ana- 57,14 ± 8,94 63,24 ± 5,95 66,60 ± 7,49 68,39 ± 7,78 69,37 ± 5,99 lyse. Il ressort sur ce plan factoriel (Figure 5) que toutes les variables (principales et supplémentaires) sont bien représentées. Cette ﬁgure permet en outre de constater

(a) (ab) (ab) (b) (b) une forte corrélation positive entre HG, HC, Lc, TP, LC, PT et LO. Ce groupe de variable fortement corrélé est en LB corrélation approximativement nulle avec LB. Il ressort

13,32 ± 5,42 15,09 ± 4,82 15,62 ± 4,37 15,93 ± 4,39 16,09 ± 6,65 que ce groupe de variable est bien caractéristique des chèvres de la région de Tillabéri alors que celles de Tahoua sont bien caractérisées par LB, LC, HG, HC et celles de Niamey par PT, LO, Lc et TP. En outre, cette (a) (bd) (bd) (d) (b) ﬁgure permet de lire que globalement à tout âge compris

LC tous les paramètres sont en corrélation positive avec le sexe femelle et négative avec le sexe mâle. 50,87 ± 14,08 59,23 ± 9,08 62,47 ± 12,04 64,39 ± 12,77 65,85 ± 12,31

Analyse Factorielle Discriminante ’

(c) L Analyse Factorielle Discriminante (AFD) contribue à (a) (d) (d) (bc) afﬁner les résultats de l’ACP. En considérant comme vari-

TP ables dépendantes la région et le sexe et comme variables explicatives l’âge et les huit paramètres morpho-

57,16 ± 8,72 61,93 ± 7,9 66,11 ± 10,73 70,63 ± 9,81 73,66 ± 9,11 biométriques, les deux premiers axes factoriels expliquent 100% de la variabilité. Il ressort de cette analyse âge des paramètres biométriques ’ (Figure 6), que les facteurs les plus discriminants mieux

(bc) (cd) (de) (e) représentés sur les deux axes, autrement dit qui permettent (a) de mieux caractériser les chèvres, sont HG, HC, LC, Lc, HG LO, LB et PT. La matrice de confusion (Tableau 6) montre que les individus sont bien classés en moyenne dans 92,55%, trente trois (33) individus soit 7,45% étant reclassés. Il ressort de la répartition des individus dans le plan factoriel (Figure 7) 739983 60,21 ±90 5,82 63,66 ± 7,43 66,21 ± 7,62 67,44 ± 6,42 98 53,79 ± 9,2 que dans les régions de Tahoua et Tillabéri, les individus 443 62,15 ± 8,99 65,82 ± 11,09sont 60,42 ± 13,42 bien discriminés 15,17 ± 5,29 sur 64,85 ± 8,66 les deux 28,61 ± 5,35 axes contrairement 17,2 ± 4,63 12,77à ± 4,98 la : Tableau 4: Variation selon l région de Niamey où les individus sont représentés principalement sur l’axe F1. Cette répartition fait également ressortir que les populations de chèvres dans les trois régions ne sont 2 ans 3 ans 4 ans 1an Ages N – – – – Tableau 4. 1 2 3 > 4 ans Total Sur la même colonne, une moyenne portant au moins une lettre (a, b, c ou d) portée par une autre, ne lui est pas statistiquement différente (P < 0,05). 0 pas totalement distinctes. Il existe en faible proportion des Caractérisation de la chèvre du sahel au Niger 27

Figure 3. Evolution selon l’âge et le sexe des paramètres morpho – biométriques étudiés ( M: mâle; F: femelle). sous groupes intermédiaires entre les régions. Cependant, Tahoua bien discriminées par l’importance de LC et LB, conformément à l’ACP, l’AFD (Figures 6et7) montre que cellesdelarégiondeTillabériparl’importance de HG et dans la majorité, les populations caprines se distinguent HC, PT, Lc et LO, tandis que celles de Niamey sont d’une région à une autre, les chèvres dans la régions de mieux représentées par TP et PT.

Figure 4. Répartition des indices biométriques selon les départements de la zone d’étude. 28 H. Marichatou et al.

Tableau 5. Matrice de corrélation (Pearson (n))

Variables Âges HG TP LC HC PT LB L0 Lc Tah Til Ny F M

Âges 1 HG 0,48 1 TP 0,52 0,72 1 LC 0,36 0,62 0,56 1 HC 0,44 0,94 0,75 0,66 1 PP 0,33 0,48 0,49 0,17 0,50 1 lB 0,16 0,18 −0,10 0,41 0,12 −0,05 1 LO 0,22 0,52 0,37 0,22 0,58 0,42 −0,23 1 Lc 0,47 0,63 0,53 0,40 0,65 0,58 0,13 0,54 1 Tah −0,07 −0,09 −0,18 0,37 −0,15 −0,46 0,56 −0,55 −0,31 1 Til 0,09 0,34 0,15 0,11 0,43 0,30 −0,22 0,73 0,44 −0,59 1 Ny −0,03 −0,31 0,00 −0,50 −0,35 0,10 −0,30 −0,30 −0,21 −0,30 −0,59 1 F 0,35 0,20 0,24 0,14 0,21 0,16 0,09 0,10 0,15 −0,06 0,06 −0,02 1 M −0,35 −0,20 −0,24 −0,14 −0,21 −0,16 −0,09 −0,10 −0,15 0,06 −0,06 0,02 −1,00 1

Les valeurs en gras sont différentes de 0 à un niveau de signiﬁcation alpha = 0,05.

Figure 5. Cercle de corrélation ACP. Caractérisation de la chèvre du sahel au Niger 29

Figure 6. Cercle de corrélation AFD.

Discussion femelles (77%) conformément à l’image générale des troupeaux enquêtés (les mâles sont minoritaires et Plusieurs études de caractérisation des caprins (Lauvergne généralement jeunes), mais aussi traduit l’importance et al., 1993; Zeuh et al., 1997; Lanari et al., 2003; Traoré socioéconomique de l’élevage caprin dans la zone et al., 2006; Bouchel et al., 2006; Traoré et al., 2008; d’étude. En effet, les mâles sont abattus à bas âges pour Verma et al., 2010; Ebegbulem et al., 2011; Halima des circonstances de cérémonies (mariage, baptême, et al., 2012) ont fait usage seuls ou associés à d’autres, tabaski ou autres fêtes) mais aussi de sacriﬁce. Ils constitu- des indicateurs utilisés dans le cadre de cette étude. ent également une épargne facilement mobilisable en cas L’échantillon étudié est composé majoritairement des de besoin pécuniaire et les femelles sauf en cas de forces majeures sont gardées pour le renouvellement du troupeau. Ce constat a été relaté au Nigéria, au Burkina, au ’ ’ Tableau 6. Matrice de confusion pour l échantillon d estimation Botswana, en Ouganda (Samuel et Salako, 2008; Traoré (région) et al., 2006; Katongole et al., 1996; Semakula et al., de \ Vers Niamey Tahoua Tillabéri Total % correct 2010).

Niamey 87 0 16 103 84,47% Les moyennes des données biométriques obtenues sont Tahoua 4 93 5 102 91,18% inférieures à celles présentées par Wilson (1991) pour la Tillabéri 5 3 230 238 96,64% chèvre « Arabe » du Tchad et les chèvres « Maure » du Total 96 96 251 443 92,55% Mali pour HG, PT, LO et Lc, les chèvres « Toureg » et 30 H. Marichatou et al.

Figure 7. Répartition des individus dans le plan factoriel (F1-F2).

« Sahel » du Mali pour HG et PT, mais semblable à la entre les deux sous populations du sud Tchad (Zeuh et al., chèvre « Sahel » du Mali pour la LO. Par ailleurs, ces 1997). Egalement, cet IGS moyen se rapproche de celui moyennes biométriques obtenus sont supérieures à celles de la population caprine toutes races confondues du des chèvres sahéliennes, soudanienne (chèvre Djallonké) Burkina Faso (Traoré et al., 2008), mais spéciﬁquement et sahélo-soudanienne (Mossi) pour les paramètres HG, inférieur à celui de la chèvre sahélienne et supérieur à PT, LC, LO, Lc et HC (Traoré et al., 2008; Samuel ceux des chèvres sahélo-soudanienne et soudanienne. Cela et al., 2008; Abdulmojeed, Adebowale et Ikhide, 2010; suppose que la population caprine de la zone d’étude se Ebegbulem et al., 2011) mais se rapprochent des rapproche du point de vue longueur des pattes à celles du données de la chèvre « Red Sokoto Goat » obtenues par Burkina Faso et du Nord Cameroun. L’IAT quant à lui, est Abdulmojeed et al. (2010). inférieur à ceux des populations caprines de ces deux ’ Les indices biométriques varient selon les localités entités. Cette faiblesse de l IAT ne traduit pas que la LO (régions et départements), l’IGS variant de 0,94 à 1,67 et et PT sont inférieures bien au contraire, mais tient au ’ ’ l’IAT de 0,35 à 0,52. Etant donné que l’IGS évolue avec fait que l évolution de l IAT est apparue indépendante ’ la longueur des pattes et l’IAT dépend des deux grandeurs d un seul des deux facteurs du fait des écarts entre les conjugués, LO et PT, les variations des indices valeurs des deux paramètres (LO approximativement biométriques indiquent des distinctions entre les popula- moitié de PT). tions caprines des trois régions étudiées par la longueur Les résultats de l’analyse en composante principales font des pattes et des oreilles et la profondeur du thorax. ressortir qu’indépendamment de l’âge, le sexe femelle Les variations de l’IGS observées sont semblables aux var- est en corrélation positive avec tous les paramètres iations rapportées sur les populations caprines du Sénégal biométriques étudiés. Cela peut être inhérent aux écarts (Bouchel et al., 2006). Cependant, l’IGS moyen obtenu est d’effectifs (femelle 77%). L’analyse des variances a semblable a l’IGS des populations caprines du Nord indiqué que les paramètres biométriques varient statisti- Cameroun mais inférieur à celui des populations caprines quement selon l’âge et l’analyse selon les tranches du Nord Tchad (Bourzat et al., 1993) et intermédiaire d’âges a fait ressortir globalement que les mâles sont Caractérisation de la chèvre du sahel au Niger 31

supérieurs aux femelles avec quelques exceptions selon les Conclusion tranches d’âges et les paramètres. En effet, selon Vigne, Peters et Helmer (2002), le sexe est le facteur le plus Aux termes de cette étude, il ressort que l’espèce caprine influant sur les dimensions chez les chèvres, les mâles dans la zone d’étude est diversement repartie selon les âgés de plus de 12 mois sont toujours plus grands que régions en fonction des paramètres quantitatifs étudiés. les femelles; ensuite vient l’environnement se traduisant Trois écotypes spécifiques aux trois régions et un type par une diminution des dimensions des chèvres des intermédiaire Niamey-Tillabéri ont été distingués. Les zones froides et humides vers les zones arides, alors que valeurs moyennes à faibles des paramètres morpho- l’âge semble avoir peu d’impact sur les dimensions des biométriques des caprins de Niamey est la résultante de chèvres. Samuel et Salako (2008) et Semakula et al. l’hétérogénéité zoo-génétique de ce troupeau avec la (2010), étudiant respectivement les caractéristiques présence de la diversité génétique de la sous région et biométriques de la chèvre Djallonké (West African son corolaire le brassage génétique. De ce fait, mise à Dwarf Goat) au Nigéria et la chèvre Mubende en part la population hétéro-génétique de Niamey, l’on retien- Ouganda, arguent que l’âge et le sexe sont des facteurs dra que la chèvre du sahel se présente sous deux types influençant significativement le poids corporel et les phénotypiques dans la zone d’étude. Une population type paramètres biométriques. Les femelles chez la chèvre de la région de Tillabéri et celle type de la région de Djallonké sont trouvées supérieures aux mâles dans les Tahoua. Les caractéristiques sur le plan morpho- tranches d’âge de 0 à 2 ans selon les résultats de Samuel biométrique sont: grande taille (en moyenne 64,98 ± et Salako (2008) et inversement pour toutes les tranches 6,77 cm et 60,71 ± 10,81 cm au garrot et 68,27 ± 5,91 cm d’âge chez la chèvre Mubende selon les résultats de et 62,46 ± 10,65 cm à la croupe respectivement à Semakula et al. (2010). Cependant, les résultats Tillabéri et Tahoua), oreilles et cornes longues dans la d’Ebegbulem et al. (2011) indiquent chez la chèvre région de Tillabéri (en moyenne respectivement environ Djallonké au Nigeria que les mâles sont supérieurs aux 20,3 ± 3,08 cm et 14,75 ± 4,43 cm) et courtes dans la femelles pour HG, HC, TP, LO, LC et LB. région de Tahoua (environ respectivement 12,32 ± 3,62 cm et 9,95 ± 4,01 cm en moyenne). Le tour de poi- ’ ’ fi L AFD et l ACP, con rment les résultats des tests statis- trine, la longueur du corps, la largeur du bassin et la pro- ’ tiques et de l analyse des indices biométriques, autrement fondeur du thorax sont en moyenne respectivement 62,2 ± dit la distinction entre les populations caprines de trois 14,85 et 67,34 ± 9,1 cm; 69,51 ± 17,47 et 61,82 ± 8,18 cm; ’ régions. L AFD révèle une quasi distinction et 20,8 ± 5,36 et 14,09 ± 4,34 cm et 24,10 ± 4,77 et 30,09 ± homogénéisation respectivement inter et intra région 4,76 cm respectivement à Tahoua et Tillabéri. pour les populations caprines étudiées. Cependant, afind’affiner la caractérisation, il est fonda- Les tests statistiques et les analyses multivariées (ACP et mental de savoir qu’en est-il sur les plans génétiques et AFD) permettent de distinguer dans la zone d’étude quatre zootechniques entre ces deux sous types? En outre, eu (4) sous populations distinctes de chèvres en fonctions des égard à l’omniprésence de la chèvre du sahel dans toutes paramètres morpho-biométriques étudiés. Une sous popu- les Régions du Niger, il s’avère important d’étendre les tra- lation de Tillabéri caractérisée par de HG, HC, LO, TP, vaux de caractérisation sur les autres régions du pays. et Lc élevés mais LB et LC faibles par rapport aux autres sous populations, une sous population de Tahoua avec LB et LC élevées, HG et HC intermédiaires et TP, PT, LO, Remerciements Lc faibles par rapport aux autres, une sous population de Niamey avec PT, LO moyens et des valeurs faibles Les auteurs remercient le partenaire financier qui est pour les autres paramètres et une sous population l’Agence Universitaire de la Francophonie (AUF), intermédiaire entre la sous population de Tahoua et celle précisément son Bureau Afrique de l’Ouest et le CNF de des Niamey avec TP et Lc moyens. Des différences statis- Niamey. Ils remercient également le Programme de tiquement significatives sur le plan phénotypique qualitatif Productivité Agricole en Afrique de l’Ouest (PPAAO) ont également été rapportées (Mani et al., sous presse) qui a contribué matériellement et financièrement aux entre les populations des trois régions. activités sur les caprins du Sahel. Les valeurs moyennes à faibles observées pour les chèvres de Niamey sont inhérentes au caractère urbain de son élevage, qui fait d’elle un melting-pot où est rencontré Références bibliographiques une diversité de races notamment, les races locales (chèvre Rousse, chèvre du sahel), sous régionale (chèvre Abdulmojeed, Y., Adebowale, E.S. and Ikhide, G.I. 2010. Comparative multivariate analysis of biometric traits of West African Dwarf and Djallonké, chèvre Mossi) et leurs croisés. Les individus Red Sokoto goats. Tropical Animal Health and Production, Springer. intermédiaires peuvent être inhérents à la proximité, en Baccini, A., et Besse, P. 2007. Exploration Statistique. Institut de effet, Niamey est circonscrite dans la région de Tillabéri Mathématiques de Toulouse—UMR CNRS C5219. Laboratoire de qui, elle-même partage une longue frontière avec la Statistique et Probabilités. Institut National des Sciences Appliquées région de Tahoua. de Toulouse—31077 – Toulouse cedex 4., 122. 32 H. Marichatou et al.

Bouchel, D., Lauvergne, J.J. 1996. Le peuplement de l’Afrique par la Mani, M., Marichatou, H., Issa, M., Chaïbou, I., Sow, A., Chaïbou, chèvre domestique. Revue d’Elevage et de Médecine Vétérinaire des M., Sawadogo, J.G. Caractéristiques phénotypiques de la chèvre du Pays Tropicaux 49 (1): 80–90. sahel au Niger par analyse des indices de primarité et des paramètres qualitatifs., Animal Genetic Resources. (sous presse). Bouchel, D., Sow, R.S., Bibe, B., Tixier-Boichard, M., Lauvergne, J.J., Poivey, J.P., Rognon, X. 2006. Caractérisation et cartographie Martin, A., 2004. L’analyse de données. ENSIETA – Réf.: 1463. (En ligne) des ressources génétiques caprines du Sénégal à l’aide d’indices Accès internet: http://www.arnaud.martin.free.fr/Doc/polyAD.pdf. phanéroptiques, d’indices morphobiométriques et de marqueurs Meyer, C. 2001. Races d’animaux d’élevage en Afrique intertropicale moléculaires: méthodologie et résultats préliminaires. Rencontres et méditerranéenne: Les caprins. Mémento de l’Agronome. Recherches Ruminants 13: 257. CIRAD-EMVT, France, 22. Bourzat, D., Souvenir, P.Z., Lauvergne, J.J. et Zeuh, V. 1993. Rakotomalala, R. 2012. Analyse de corrélation: Étude des dépendances - Comparaison morpho-biométrique de chèvres au Nord Cameroun et Variables quantitatives. Version 1.0. Université Lumière Lyon 2. (En ’ au Tchad. Revue d Elevage et de Médecine Vétérinaire des Pays ligne) Accès internet: http://eric.univ-lyon2.fr/~ricco/cours/cours/ – Tropicaux 46 (4): 667 674. Analyse_de_Correlation.pdf. ’ Dumas, R. 1980. Contribution à l étude des petits ruminants du Tchad. Rhissa, Z. 2010. Revue du secteur de l’élevage au Niger. Ministère de ’ Revue d Elevage et de Médecine Vétérinaire des Pays Tropicaux 33 l’Elevage, des Pêches et des Industries Animales, Niger, 115. (2): 215–233. Samuel, F.O.K. and Salako, A.E. 2008. Body measurement character- Ebegbulem, V.N., Ibe, S.N., Ozung, P.O. and Ubua, J.A. 2011. istics of the West African Dwarf (WAD) Goat in deciduous forest Morphometric trait characteristics of West African dwarf goats in zone of Southwestern Nigeria. African Journal of Biotechnology 7 Abia state, south east Nigeria. Continental Journal of Agricultural (14): 2521–2526. Science 5 (2): 1–6. Semakula, J., Mutetikka, D., Kugonza, R.D. and Mpairwe, D. 2010. FAO. 1986. Animal genetic resources data banks. Descriptor lists for cat- Variability in Body Morphometric Measurements and Their tle, buffalo, pigs, sheep and goat. FAO. Rome, Italia. Animal Application in Predicting Live Body Weight of Mubende and Small Production and health. 59 (2): 151. East African Goat Breeds in Uganda. Middle-East Journal of Scientiﬁc Research 5 (2): 98–105. FAO. 2012. Phenotypic characterization of animal genetic resources. Food and Agriculture Organization of the United Nations. Rome, Traoré, A. 2010. Caractérisation des ressources génétiques caprines du Italia. Animal Production and Health Guidelines 11: 142. (accessible Burkina Faso a l’aide d’indices morpho-biométriques et de marqueurs at http://www.fao.org/docrep/015/i2686e/i2686e00.htm). moléculaires. Thèse de Doctorat Unique. Université de Ouagadougou, Burkina Faso, 110. Gnanda, I.B. 2008. Importance socio-économique de la chèvre du Sahel burkinabé et amélioration de sa productivité par l’alimentation. Thèse Traoré, A., Hamidou, H.T., Adama, K., Luis, J.R., Ivan, F., Isabel, de Doctorat Unique. Université Polytechnique de Bobo-Dioulasso, A., Sangaré, M., Bouchel, D., Jean Paul, P., Dominique, F., Burkina Faso, 198. Sawadogo, L. and Goyache, F. 2008. Multivariate analyses on morphological traits of goats in Burkina Faso. Arch. Tierz, Dummerstorf Halima, H., Michael, B., Barbara, R. and Markos, T. 2012. 51 (6): 588–600. Phenotypic characterization of Ethiopian indigenous goat populations. African Journal of Biotechnology 11 (73): 13838–13846. Traoré, A., Tamboura, H.H., Kaboré, A., Yaméogo, N., Bayala, B. et Zaré, I. 2006. Caractérisation morphologique des petits ruminants ﬁ Hamito, D. 2009. Goat breeds of Ethiopia: A guide for identi cation and (ovins et caprins) de race locale “Mossi” au Burkina Faso. GRI utilization. Ethiopia Sheep and Goat Productivity Improvement 2006 39: 39–50. Program (ESGPIP). Technical Bulletin 23: 11. Verma, N.K., Dixit, S.P., Aggarwal, R.A.K., Dangi, P.S. and Kaiser, H.F. 1960. The application of electronic computers to factor ana- Joshi, B.K. 2010. Phenotypic and genetic characterization of – lysis. Educational and Psychological Measurement 20, 141 151. Sangamneri goat breed. Indian Journal of Animal Sciences 80 – Katongole, J.B.D., Sebolai, B. and Madimabe, M.J. 1996. (11): 1109 1114. Morphological Characterization of the Tswana goat. In: S.H.B. Lebbie Vigne, J-D., Peters, J. and Helmer, D. 2002. First Steps of and E. Kagwini (eds.), Small Ruminant Research and Development in Animal Domestication: New archaeozoological approaches. 9th Africa. Proceedings of 3rd Biennial Conference of the African Small ICAZ Conference, Durham 2002. The First Steps of Animal Ruminant Research Network, UICC, Kampala, Uganda, 5–9. Domestication (eds J-D Vigne, J. Peters and D. Helmer), 125–146. Lanari, M.R., Taddeo, H., Domingo, E., Perez-Centeno, M. and Wilson, T.R. 1991. Small ruminant production and the small ruminant Gallo, L. 2003. Phenotypic differentiation of exterior traits in local genetic resource in tropical Africa. FAO. Rome. Animal Production Criollo Goat Population in Patagonia, Argentina. Archiv fur and Health Paper, 194. Tierzucht Dummerstorf 46 (4): 347–356. Zeuh, V., Lauvergne, J.J., Bourzat, D. et Minvielle, F. 1997. Lauvergne, J.J., Bourzat, D., Souvenir, Z.P., Zeuh, V. et Ngo Tama, Cartographie des ressources génétiques caprines du Tchad du A.C. 1993. Indices de primarité de chèvres au Nord Cameroun et au Sud-Ouest: Hauteur au garrot (HG), profondeur de thorax (PT) et Tchad. Revue d’Elevage et de Médecine Vétérinaire des Pays indice de gracilité sous-sternale (IGs). Revue d’Elevage et de Tropicaux 46 (4): 651–665. Médecine Vétérinaire des Pays Tropicaux 50 (3): 250–260. Animal Genetic Resources, 2014, 54, 33–41. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000137

Morphological, reproductive and productive characteristics of Sudanese native chicken

C.E. Wani1, I.A. Yousif2, M.E. Ibrahim3, H.H. Musa4 and K.M. Elamin5 1Department of Animal Production, Faculty of Veterinary Science, University of Bahr Elgazal, Wau, South Sudan; 2Department of Genetics and Animal Breeding, Faculty of Animal Production, Khartoum North, University of Khartoum, Sudan; 3Institute of Endemic Diseases, University of Khartoum, Sudan, Khartoum; 4Department of Microbiology, Faculty of Medical Laboratories, University of Khartoum, Sudan, Khartoum; 5Department of Poultry Production, Faculty of Animal Production, University of Khartoum, Sudan, Khartoum North

Summary A structured questionnaire was administered to 40 households in each of the three localities (Bahri, Dilling and Abu-neama). Morphological characteristics of 900 chickens were physically examined. Data analysis showed that the average flock size per household was 16.7 and that it varied significantly among localities. The flock structure study showed that 46.1 percent were chicks and growers, 34.6 percent hens and 19.35 percent cocks. Regarding the ecotypes, the flock was composed of 75.7, 8.8 and 15.5 percent large Beladi (LB), bare neck (BN) and Betwil (BT) dwarf, respectively. The Sudanese indigenous chicken are characterized by wide phenotypic variation. The plumage colour frequencies were mixed colour (37.8 percent), light brown (16.5 percent), black (13.1 percent), grey (3.5 percent) and white (6.2 percent). Feathering did not show a well-defined pattern; however, the few birds whose feathering pattern could be identified were crown feather (7.0 percent), partridge (3.5 percent), barred (3.5 percent), laced (3.1 percent) and frizzled (0.01 percent). The average adult body weight was 1 650.5 ± 125 g in males and 1 187 ± 70 g in females. Hens reached sexual maturity at 6.5 months and layed 13.2 eggs/clutch in 3–3.5 clutches per year. The average egg weight was 40.8 ± 1.6 g and the hatchability rate under natural incubation was 71.8 percent. The BN ecotype had significantly (P < 0.01) poorer maternal care (57.7 percent hatchability) compared with the other types (LB and BT).

Keywords: Sudan, indigenous chickens, morphological variation, phenotypic characteristics

Résumé Se facilitó un cuestionario estructurado a 40 hogares en cada uno de los siguientes tres municipios: Bahri, Dilling y Abu-neama. Se examinaron las características morfológicas de 900 aves. El análisis de los datos mostró que el número medio de aves por hogar fue de 16.7, aunque varió de manera significativa entre localidades. El estudio de la estructura de los gallineros reveló que el 46.1 por ciento eran pollitos y pollos en crecimiento, el 34.6 por ciento gallinas y el 19.35 por ciento gallos. En lo que concierne a los ecotipos, los gallineros se componían en un 75.7 por ciento, en un 8.8 por ciento y en un 15.5 por ciento de Gran Beladi, Cuello Pelado y Betwil Enana, respectivamente. Las gallinas autóctonas sudanesas se caracterizan por una amplia variación fenotípica. En el plumaje, se dieron los siguientes colores: color mixto (37.8 por ciento), marrón claro (16.5 por ciento), negro (13.1 por ciento), gris (3.5 por ciento) y blanco (6.2 por ciento). El plumaje no presentó un patrón bien definido; no obstante, en las pocas aves en que se pudo identificar un patrón para el plumaje, éste era con copete en un 7.0 por ciento, perdiz en un 3.5 por ciento, barrado en un 3.5 por ciento, ribeteado en un 3.1 por ciento y rizado en un 0.01 por ciento. El peso corporal medio de los adultos fue de 1650.5 ± 125 g en los machos y de 1187 ± 70 g en las hembras. Las gallinas alcanzaron la madurez sexual a los 6.5 meses y pusieron 13.2 huevos por nidada, con 3 a 3.5 nidadas por año. El peso medio de los huevos fue de 40.8 ± 1.6 g y la tasa de incubabilidad en incubación natural fue de un 71.8 por ciento. El ecotipo Cuello Pelado mostró un instinto materno significativamente menor (57.7 por ciento de incubabilidad, P < 0.01) que el de los otros ecotipos (Gran Beladi y Betwil).

Mots-clés: Soudan, poules autochtones, variation morphologique, caractéristiques phénotypiques

Resumen Un questionnaire structuré a été fourni à 40 foyers dans chacune des trois localités suivantes: Bahri, Dilling et Abu-neama. Les caractéristiques morphologiques de 900 volailles y ont été examinées. D’après l’analyse des données, le nombre moyen d’oiseaux par foyer a été de 16.7, avec une variation signiﬁcative entre localités. L’étude de la structure des poulaillers a mis en évidence que le 46.1 pour cent des oiseaux étaient des poussins et des poulets en croissance, le 34.6 pour cent des poules et le 19.35 pour cent des coqs. Pour ce qui est des écotypes, les poulaillers se composaient à 75.7 pour cent, 8.8 pour cent et 15.5 pour cent de Grand Beladi, Cou Nu et Betwil Naine, respectivement. Les poules autochtones du Soudan se caractérisent par une grande variation phénotypique. Les suivantes couleurs ont été observées dans le plumage: mélange de couleurs (37.8 pour cent), marron clair (16.5 pour cent), noir (13.1 pour cent), gris (3.5 pour cent) et blanc (6.2 pour cent). Bien que le plumage n’ait pas présenté un motif bien déﬁni, les

Correspondence to: C.E. Wani, Department of Animal Production, Faculty of Veterinary Science, University of Bahr Elgazal, South Sudan. email: yousiﬁ[email protected]; tel.: + 249912390684

33 34 C.E. Wani et al.

motifs suivants ont été reconnus sur le plumage d’un petit nombre d’oiseaux: à huppe (7.0 pour cent), perdrix (3.5 pour cent), barré (3.5 pour cent), dentelé (3.1 pour cent) et frisé (0.01 pour cent). Le poids corporel moyen des adultes a été de 1650.5 ± 125 g chez les mâles et de 1187 ± 70 g chez les femelles. Les poules ont atteint la maturité sexuelle aux 6.5 mois et elles ont pondu 13.2 œufs par couvée, avec 3 à 3.5 couvées par an. Le poids moyen de l’œuf a été de 40.8 ± 1.6 g et le taux d’éclosion, sous conditions naturelles d’incubation, a été de 71.8 pour cent. Compte tenu d’un taux d’éclosion plus bas (57.7 pour cent, P < 0.01) par rapport aux autres écotypes (Grand Beladi et Betwil), les poules de l’écotype Cou Nu sembleraient avoir un plus faible instinct maternel.

Palabras clave: Sudán, gallinas autóctonas, variación morfológica, características fenotípicas

Submitted 5 September 2013; accepted 11 March 2014

Introduction home lands of indigenous chickens threatened their exist- Poultry meat and egg production accounted for more than ence. Therefore characterization of these valuable indi- 28 percent of the total animal protein produced worldwide genous animal resources using genetic and phenotypic in 1997 (Tadelle et al (2003)). It is estimated that by the methods for the purpose of conservation has become very year 2020, the proportional contribution of poultry to the crucial. However, limited research has been conducted to total animal protein will rise to 40 percent with the characterize the local fowl of Sudan. In the present study, major increase being in the developing world (Delgado morphological, productive and reproductive characteristics et al., 1999). About 1.3 billion chickens are found in of Sudanese indigenous chickens were investigated. Africa producing approximately 1.7 and 2.1 million metric tonnes of egg and meat respectively of which 80 percent came from indigenous stock (FAO, 2006). In Sudan, Materials and methods there are no reliable statistics regarding the contribution of the conventional poultry sector. However, it was esti- Study sites mated to constitute about 70 percent of the total production This study was conducted in three different localities in the (Suleiman, 1996). Conventional poultry plays an important Sudan, namely El Dilling, Abu-Neama and Khartoum role in the socio-economic life of the rural community in North (Bahri). El Dilling locality is situated in South Africa as well as in Sudan; generating incomes and con- Kordofan State an area which lies 800 km south west of tributing to food security (Gueye, 2002). Moreover, poult- Khartoum (Capital of Sudan) at latitude 12°20′N and lon- ry raising is preferred by small holders because of the low gitude 29°28′E with rainfall of 500–800 mm. The area of input costs, high adaptability and resistance to diseases South Kordofan State is covered by the Nuba Mountains (Wimmers et al., 2000). Information about the flock struc- that occur as a series of isolated ranges covering an area ture of the Sudanese indigenous chicken is lacking. The of about 48 000 km2. average flock size of local Malawi chickens was reported Abu-Naema locality in Sinnar State (Central Sudan) is to be 12.9 ± 5 birds per household (HH) (Gondwe and an area located 400 km South of Khartoum at latitude Wollny, 2004). In comparison, large flocks per HH 12°44′N, longitude 34°08′E and altitude 445 m above (26 and 22 birds/HH) were reported for Pakistan and sea level. Bahri locality is in Khartoum State and includes Senegal native chickens, respectively (Missohou, Dieye a group of villages situated east of the River Nile. The and Talaki, 2002; Javed et al., 2003). area is semi desert or poor Savannah. Most of the people The Sudanese indigenous chicken breed was classified into in these three localities are either farmers or livestock three types as described by Desai (1962). The large Beladi herders. (LB) type is the most common, available in Northern, Central, Western and Southern Sudan. The birds are of good size (adult weighing 2½–3.0 Ibs) with a small Questionnaire administration crushed comb, plenty of plumage of many colour varia- A structured questionnaire was administered to 120 HHs tions. The bare neck (BN) type is smaller than the LB, from the three localities, 40 for each locality. Data captured characterized by its featherless neck and many colours in in the questionnaires were on chicken management prac- plumage. The home land of BN type is Southern Sudan tices such as flock structure, housing, feeding and market- and it was reported to be more resistant to endemic dis- ing. Production data included age at start of laying, clutch eases compared with other types (Wimmers et al., 2000). size, hatchability as well as live weights and egg weight. A The Betwil (BT) type is a small dwarf bird, mainly total of 900 chickens were characterized through physical found in the Nuba Mountains region of South Kordofan examinations that covered plumage colour and pattern, State, Sudan. It has a small compact body (adult bird skin colour, shank length, comb colour and shape. The weighing 1½–2.0 Ibs) with tiny black legs. Recently in questionnaire and phenotypic characterization were sup- Sudan, the expansion of modern poultry farming in the ported by group discussions among villagers. Characteristics of Sudanese native chicken 35

Table 1. Flock size in different localities.

Locality/district Number of HHs Mean ± S.E. Range (birds)

El-Dilling 40 15.74b ± 1.3 2–36 Abu-Neama 40 14.61b ± 1.24 3–30 Bahri 40 19.87a ± 1.29 6–36 Total 120 16.74 2–36

Means with the same letter are not signiﬁcantly different (P < 0.05). HH, hens per household.

Table 2. Flock size and structure of Sudanese indigenous chicken ecotypes.

Flock by age LB BN BT Total (n = 681) (n = 80) (n = 139) (n = 900) Plate 1. Mix colour feather LB chicken. Percentage kept by HH 75.7 8.9 15.4 100 Cock to hen ratio 1:2 1:1.2 1:1.2 1:1.8 Chicks and growers (%) 45.1 47.9 50 46.1 showed that 46.1 percent were chicks and growers, 34.6 Hens (%) 36.7 29.2 27.4 34.6 percent hens and 19.3 percent cocks. On the basis of eco- Cocks (%) 18.2 22.9 22.6 19.3 type, the flock was composed of 75.7, 8.8 and 15.5 percent LB, large Beladi; BN, bare neck; BT, Betwil; HH, hens per household. LB, BN and BT, respectively, and the average cock–hen ratio was 1:1.8. Most of the HHs tend to keep more Data analysis females than males and in all ecotypes surveyed, the chicks and growers formed the majority of the flocks (Table 2). Descriptive statistics were used to analyse phenotypic characterization data. The frequencies of phenotypic charac- Flock management practice teristics among the chicken ecotypes were calculated by dividing the number of birds having the trait by the total num- The results revealed that 70 percent of the HHs provide fl ber of the birds examined. Percentages were used to calculate night shelters to the chicken ocks they are keeping and the prevalence of a trait within each ecotype. The degree of it was observed that housing structures were made from association between prevalence of the trait and chicken eco- wood or bamboo poles. The non-housed chicken mostly ’ types was tested using Chi-square at the 95 percent confi- perched around farmer s houses (2 percent) or on the ’ dence level. SPSS Version 10 statistical package (SPSS trees near the farmer s homesteads (3 percent) and laid Inc., Chicago, and III, USA) was used for data analysis. eggs in nests placed around the vicinity or in the kitchen (20 percent) or store (5 percent) (Figure 1). Chicken were found to be scavengers for a variety of feed stuffs Results that included cereals, weeds, seeds, insects, worms and various herbs. Some farmers usually supplement their Flock size and structure chicken with whole cereals once every morning. In all The average flock size per HH was 16.74 ranging from 2 the areas surveyed, HHs do not keep records on feed sup- to 36 chickens/HH, the average varied significantly (P < plement, survival and mortality of chickens. They also do 0.05) according to locality. Bahri locality scored the high- not vaccinate their chickens nor use veterinary drugs for est number of chickens per HH with an average of 19.87 ± treatment or prophylaxis. The majority of the interviewed 1.29, followed by El Dilling (15.74 ± 1.3) then Abu- HHs (98.2 percent) declared that women were the owners Neama locality (14.61 ± 1.24) (Table 1). The data collected of chicken with their children assisting them for manage- on the Sudanese indigenous chicken flock structure ment (cleaning chicken house, providing supplementary

Figure 1. Chicken house provision. 36 C.E. Wani et al.

Plate 4. Mix colour feather Bare Neck chicken. (Source: Faculty of Animal Plate 2. Crown feather Large beladi chicken. (Source: Faculty of Animal Production, U of K, Sudan, 2011) Production, U of K, Sudan, 2011)

(40.3 percent) in LB (Plate 1) and the lowest (33.5 percent) food and water and selling eggs and chickens). However, in BN (Plate 4), followed by 16.5 percent light brown men were responsible for decision-making (95 percent), (Plate 3) and 13.1 percent black (Plate 5) plumage. giving permission for selling of chicken or eggs as well Feathering did not have a well-deﬁned pattern in the as home consumption or slaughter for special ceremonies. majority of chicken sampled (89.9 percent). Of the patterns that could be described was partridge 3.5 percent, barred 3.5 percent, laced 3.1 percent and frizzle feather (Plate Morphological characteristics 9) 0.01 percent (Table 3). The majority of chickens (91.1 Approximately, 37.8 percent of chickens had mixed colour percent) that include most of the BT (Plates 6, 7 and 8) plumage (Plates 1, 4 and 7), with the highest percentage and LB (Plates 1 and 2) were well covered with feather over all the body, while 8.9 percent were BN chickens without feather in the neck region (Plates 3, 4 and 5). There were four shank colours including black 38.5 percent, white 29.6 percent, yellow 20.4 percent and greenish 11.4 percent. On the other hand, four types of combs were observed and the most frequent was single comb (48.6 percent), crush or rudimentary comb (42.3 percent), strawberry comb (5.9 percent) and rose comb (3.2 percent).

Plate 3. Brown colour feather Bare Neck chicken. (Source: Faculty of Animal Plate 5. Black colour feather Bare Neck chicken. (Source: Faculty of Animal Production, U of K, Sudan, 2011) Production, U of K, Sudan, 2011) Characteristics of Sudanese native chicken 37

Plate 6. Red colour feather Betwill cock. (Source: Faculty of Animal Production, U of K, Sudan, 2011)

Plate 9. Frizzle Sudanese chicken. (Source: Faculty of Animal Production, U of K, Sudan, 2011)

The majority of the chickens possessed combs with red colour (88.0 percent) and the rest had pale red combs (12.0 percent). The results revealed that Sudanese chicken had a high frequency of white skin colour (91.4 percent) while yellow skin colour constituted a small percentage (8.6 percent). On the other hand, a well grown wattle was common in the majority of chickens (89.9 percent), whereas few lacked wattles or had rudimentary wattles (10.1 percent). However, it was observed that mainly males had projected wattles. Similarly, 70.7 percent of the sampled chickens had rudimentary spur while 29.3 percent (mainly males) had large spurs. The ear lobes were

Plate 7. Mix colour feather Betwill cock. (Source: Faculty of Animal Production, U of K, Sudan, 2011) Table 3. Plumage pattern and colour characteristics in percent for different Sudanese indigenous chicken ecotypes.

Morphology Trait LB BN BT Overall (n = 681) (n = 80) (n = 139) (n = 900)

Plumage Mix 40.3 33.5 39.5 37.8 colour colour Black 18.5 15.5 5.2 13.1 Red 12.1 12 7.3 10.5 Brown 13.4 13 10.9 12.4 Light 6.3 14 29.3 16.5 brown White 4.1 12 2.6 6.2 Grey 5.3 0.0 5.2 3.5 Total 100 100 100 100 Plumage Partridge 3.5 0.0 7.1 3.5 pattern Barred 2.6 8.0 0.0 3.5 Laced 1.3 4.0 5.0 3.1 Frizzle 0.4 0.0 0.0 0.01 feather Undeﬁned 92.2 88.0 87.9 89.9 Total 100 100 100 100 Plate 8. Light brown feather betwill. (Source: Faculty of Animal Production, U of K, Sudan, 2011) LB, large Beladi; BN, bare neck; BT, Betwil. 38 C.E. Wani et al.

Table 4. Morphological characteristics in percent for different Table 5. Reproductive performances of the Sudanese indigenous Sudanese indigenous chicken ecotypes. chicken ecotypes.

Morphology Trait LB BN BT Overall Ecotypes LB BN BT Total (n = 681) (n = 80) (n = 139) (n = 900) Age at maturity (month) 6b 6b 7a 6.3 Comb Crush 32.6 30.8 63.4 42.3 No. of eggs laid/clutch 13.2b 14.0a 13.0b 13.42 pattern No. of clutches/year 3.0 3.5 3.0 3.2 Single 55.8 60.6 29.3 48.6 Inter-clutch interval (month) 2–3.5 2–3.5 2–3.5 2–3.5 Strawberry 7.0 5.8 4.9 5.9 No. of eggs incubated 9.87a 9.9a 9.27a 9.68 Rose 4.6 2.8 2.4 3.2 No. of chicks hatched 7.7a 5.2b 7.27a 6.72 Comb colour Red 91.3 87.5 85.3 88.0 No. of eggs wasted 2.17b 4.7a 2.0b 2.96 Pale 8.7 12.5 14.7 12.0 Hatchability (%) 78.34a 57.7b 79.2a 71.8 Skin colour White 95.3 86.1 92.7 91.4 No. of chicks weaned 5b 2c 6a 4.3 Yellow 4.7 13.9 7.3 8.6 Hatch to wean period (months) 2.5 2.0 3.0 2.5 Shank colour White 44.2 27.8 17.1 29.7 Black 18.6 36.1 61.0 38.5 Means with the same letter are not significantly different (P < 0.05). Yellow 23.3 30.5 07.3 20.4 LB, large Beladi; BN, bare neck; BT, Betwil. Greenish 13.9 5.6 14.6 11.4 Earlobe White 34.6 30.6 41.5 35.6 colour hatchability was 78.34, 57.7 and 79.2 percent in LB, BN Red 65.4 69.4 58.5 64.4 fi Head feather Crown 12.3 2.5 05.8 6.9 and BT, respectively, with BN being signi cantly lower feather (P < 0.01) than the other two types. Hatch to rearing period Normal 87.7 97.5 94.2 93.1 averaged 2.5 months. These results indicated that there Wattle Presence 91.5 94.9 83.2 89.9 were highly significant differences (P < 0.01) among the Rudimentary 8.5 5.1 16.8 10.1 chicks reared by the respective ecotype hens up to wean- Spur Large 29.3 Rudimentary 70.7 ing, with the LB exhibiting excellent maternal care and the BN being the poorest in that respect (Table 5). LB, large Beladi; BN, bare neck; BT, Betwil.

Body weight mostly not prominent and they were mainly red (64.4 percent) and white (35.6 percent) in colour. Crown head fea- The average male body weights of LB, BN and BT chick- ther chicken is known in Sudanese colloquial language as en were 1 720.7 ± 312.0, 1 670.4 ± 175.9 and 1 560.3 ± (Abu-Guja). These had a relatively low frequency among 155.3 g, respectively, with an overall average of 1 650.3 ’ the Sudanese indigenous chickens sampled (6.9 percent), ± 125 g. While females corresponding body weights while the rest had normal feather on the head (Table 4). were 1 350.1 ± 53.5, 1 305.1 ± 98.7 and 906.5 ± 113.6 g The results revealed that the BT chicken ecotype possessed with an overall average of 1 187.6 ± 45 g. Sexual dimorph- the shortest shank length (4.5 ± 0.47 cm), showing a sign- ism was evident among Sudanese chicken types with fi ificant difference (P < 0.05) when compared with LB (7.9 males being signi cantly (P < 0.05) heavier than the ± 1.0 cm) and BN (7.7 ± 0.93 cm). females in the three ecotypes. However, comparisons among ecotypes revealed that the BT ecotype was significantly lighter (P < 0.05) than the other two ecotypes which Reproductive performance were almost similar in their body weight (Table 6). Results revealed that LB and BN pullets reached sexual maturity at 6 months of age, whereas BT pullets reached sexual maturity at 7 months of age (Table 5). Moreover, Egg characteristics most of the hens laid on average 13.42 eggs/hen/clutch, Average egg weights of LB, BN and BT chickens were with BN being (14.0 eggs/hen/clutch) significantly differ- 41.3 ± 3.1, 42.4 ± 1.5 and 38.8 ± 5.1 g, respectively. BT ent (P < 0.01) from LB and BT. The majority of hens pro- eggs were significantly (P < 0.05) smaller than the eggs duced 3–3.2 clutches of eggs per year with the inter-clutch of the other two ecotypes, LB and BN. The majority (95 interval ranging from 2.0 to 3.5 months. On the other hand percent) of the Sudanese chicken lay white eggs and the

Table 6. Average body weights (g) for the mature Sudanese indigenous chicken ecotypes.

Types LB BN BT Total

No. 374 42 69 485 Male body weight 1 720.7a ± 312.0 1 670.4a ± 175.9 1 560.3b ± 155.3 1 650.5 ± 125 Female body weight 1 350.1a ± 53.5 1 305.1a ± 98.7 906.5b ± 113.6 1 187.6 ± 45

Means with the same letter are not signiﬁcantly different (P > 0.05). LB, large Beladi; BN, bare neck; BT, Betwil. Characteristics of Sudanese native chicken 39

Table 7. Egg characteristics of the Sudanese indigenous chicken ecotypes.

Trait LB (n = 60) BN (n = 60) BT (n = 60) Total (n = 180)

Egg weight (g) 41.3a ± 3.1 42.4a ± 1.5 38.8b ± 5.1 40.8 ± 1.6 Egg shell colour (%) White 95.8 93.1 96.3 95 Light brown 4.6 6.9 4.7 5 Yolk colour (%) Yellow 100 100 100 100 White 0.0 0.0 0.0 0.0

LB, BN, BT, large Beladi, bare neck and Betwil, respectively. Means with the same letter are not significantly difference (P > 0.05). rest 5 percent lay light brown eggs. Yolk was 100 percent and comb type, is controlled by a single/few pairs of major yellow in all the ecotypes (Table 7). genes. Such traits usually influence the preferences of consumers. In the present study, we found a high variability in plumage colour within and between the three identified Sudanese indigenous chicken ecotypes; these variable col- Discussion ourations provide aid for camouflage against predators particularly in the rural areas. However, with the modern The flock structure of Sudanese indigenous chicken ob- trends of poultry production the importance of plumage served in this study was 46.1 percent chicks and growers, colour goes beyond the need for camouflage. White feath- 34.8 percent hens and 19.3 percent cocks. Obviously, the ering has become a desirable characteristic in poultry HHs tend to keep females more than males. Males are breeding particularly in broilers because it eliminates usually sold and are slaughtered in social and traditional dark feather pigment deposits in the skin and is conse- functions. The average flock size of Sudanese indigenous quently preferred for the clean appearance of carcass and chickens was 16.74 birds/HH which was higher than that cut-up parts. Some important alleles such as frizzle and of Malawian indigenous chicken flock size of 12.9 birds/ BN genes which are genetically conserved for their special HH (Gondwe and Wollny, 2004), but lower than that of utility in tropical environments were found to exist among Pakistan indigenous chickens which amounted to 26 the Sudanese indigenous chickens even though they are at birds/HH (Javed et al., 2003) and Senegal indigenous low frequencies. There was a high diversity in colour and chickens, 22 birds/HHs (Missohou, Dieye and Talaki, type of combs and earlobes observed between and within 2002). About 91 percent of village chicken in Sudan are the Sudanese indigenous ecotypes. The most frequent kept by women suggesting gender bias in raising chicken comb type was single comb, it was more common at the village level and that they are a source of women among LB and BN chickens, followed by crush comb empowerment. This finding is in agreement with that of type which was more common among BT chickens. Kitalyi (1998) and McAinsh et al. (2004). In the surveyed According to Nesheim, Austic and Card (1979), the size areas, chicken are mainly used for HH consumption, as and colours of combs and wattles are associated with well as for the purpose of generating income. This obser- gonad development and secretion of sex hormones. vation is in agreement with the conclusion drawn by Large wattle and long legs are important morphological Gueye (2002). traits that allow better heat dissipation in the hot tropical The observation that all chickens scavenged for a variety environment. On the other hand, the short tiny legs and of feeds is similar to the results reported by Maphosa the crush combs of BT chickens may be important for et al. (2004) and Muchadeyi et al. (2004). However, this adaptation to the mountainous nature of their home land method of feeding does not optimize the utilization of in the Nuba region of South Kordofan State. the limited feed resources since younger and weaker birds have to compete with mature birds and other scaven- Generally, the Sudanese indigenous chickens are charac- ging farm animals. This could be one of the reasons for the terized by small body size. However, the sub normal observed reduced number of weaned chicks coupled with body size of the BT chicken is an indication to the fi other factors such as predators and the lack of proper presence of dwar sm, which is due to a sex-linked reces- housing structures for the birds. Seventy percent of the sive gene that tends to reduce body and egg size. This HHs construct night shelters only and the non-housed trait is promising in the sense that feed requirements for fi birds are either perched on trees or elsewhere around maintenance are lowered and thereby the ef ciency of the farmer’s homestead. No records were kept for pro- feed utilization is increased (Bordas and Merat, 1984; duction activities and disease management, which makes Merat, 1990). the village chicken more vulnerable to diseases and The age at the start of laying for Sudanese indigenous predators. chicken varied from 180 to 210 days, it is higher than The inheritance of qualitative traits, such as plumage col- that of Egyptian indigenous chicken (Hanafi and Labban, our, plumage pattern, skin colour, egg colour, shank length 1984) and higher than that of the modern strains, such as 40 C.E. Wani et al.

Rhode Island Red and Rhode Island White (Simeonovova systems, particularly feeding and health care in rural et al., 1989). This late age of sexual maturity may be areas. Several qualitative traits of economic value were due to variation in management practices, environmental reported for the Sudanese indigenous chicken populations conditions and genetic potential (El-Zubeir, Salih and which can be helpful for preservation and future commer- El-Amin, 1991; Aganga et al., 2000). The number of cial exploitation. The field survey in this study revealed eggs per clutch in the present study is very close to that that the local chickens of Sudan are available in abundance reported for native chickens from other areas such as in their homelands with the exception of the BT ecotype in South Darfur region (Beladi chicken) (9–13 eggs/clutch), the Nuba Mountains which seemed to be endangered as a Morocco (12 eggs/clutch) and Bangladesh (10–16 eggs/ result of conflicts in the region and the ensuing displace- clutch) (Wilson, 1979; Sazzad, 1986; Benabdeljelil and ment of communities. This ecotype requires special con- Arfaoui, 2001). However, the low rate of egg production sideration for propagation and conservation. in the indigenous chickens may be associated in part with the behaviour of broodiness. Hatchability in BN chicken was found to be lower than that of the other ecotypes, and this might possibly be due to the presence of the Acknowledgements Na allele in the genotype of the embryos. Horst (1980) mentioned that there was a 13 percent reduction of embry- Authors would like to acknowledge the financial support onic survival associated with the Na allele. In addition, it from the University of Bahr Elgazal, South Sudan and might also be due to poor hatchability. BN hens also the Ministry of Higher Education and Scientific seem to have low numbers of weaned chicks, although it Research, Sudan. Thanks are also extended to the staff, was proved by Horst (1980) that there was no difference technicians and postgraduate students of the Faculty of in post embryonic mortality of naked neck chicken com- Animal Production, University of Khartoum and Faculty pared with normally feathered birds at different ages. All of Veterinary Science, University of Bahr Elgazal who these aspects of poor performance of BN chickens classify offered us technical assistance to conduct this research. them as poor mothers. The hatchability of the indigenous Sudanese chicken under natural incubation was moderately low when compared with that of White and Brown fi Hisex eggs hatched under arti cial incubation in Sudan References (El-Agraa, 1988). In general, a rural-oriented breeding strategy to produce a Aganga, A.A., Omphile, U.J., Malope, P., Chabanga, C.H. & dual purpose chicken seems to be vital in improving the Motsamai, L.G. 2000. Traditional poultry production and commer- productive and reproductive potential of the indigenous cial broiler alternative for small-holder farmer in Botswana. Livest. Res. Rural Dev., 12(4): 1–8. chickens for the benefit of rural communities. This could be achieved through implementing a selective breeding Benabdeljelil, K. & Arfaoui, T. 2001. Characterization of Beladi chicken and turkey in rural poultry flocks of Morocco: current state and programme for several generations, to be followed by future outlook. Anim. Genet. Resour. Inf., 31: 87–95. limited cross-breeding involving improved tropical or exotic breeds that proved to be adaptive to tropical Bordas, A. & Merat, P. 1984. Correlated responses in a selection experiment on residual feed intake of adult Rhode-Island Red cocks and environmental conditions. The newly improved genotypes hens. Ann. Agric. Fenn., 23: 233–237. can be tested, propagated and disseminated to rural HHs Delgado, C., Rosegrant, M., Steinfeld, H., Ehui, S. & Courbois, C. accompanied with concomitant withdrawal of the existing 1999. Livestock to 2020: the next revolution. Food Agriculture and unimproved flocks. This strategy must be supported by the Environment Discussion Paper 28. Rome, FAO. conducting an active extension programme to increase Desai, D.K. 1962. The status importance and development of poultry the awareness of HH members of the importance of raising keeping in the Sudan. Sudan J. Vet. Sci. Anim. Husb., 3: 140–143. these genetically improved flocks under relatively im- El-Agraa, S.M. 1988. Fertility and hatchability of Hisex brown parent proved management. layers under control environment in Sudan. In Symposium of Exotic White and Brown Breeds Under Sudan Condition, League of Arab States, Arab Organization for Agricultural Development (AOAD), Khartoum, 16–17 October 1988, pp. 72–77. Conclusions El-Zubeir, E.A., Salih, M.E. & El-Amin, H. 1991. Early reproductive characteristics for Leghorn pullets reared on diet containing sorghum The Sudanese indigenous chicken ecotypes exhibited wide gluten feed. Sudan J. Anim. Prod.,4:45–51. phenotypic variation. Characterization information is – important for implementing conservation and improve- FAO. 2006. Statistical year book 2005 2006. Vol. 2. Rome, FAO. ment programmes for sustainable utilization. Estimates of Gondwe, T.N. & Wollny, C.B.A. 2004. With Globalization will low production performance obtained from the extensive input production system prevail in Malawi? The case study of rural fi chicken. In Book of Abstract of the 55th Annual Meeting of the rural systems were signi cantly lower than those obtained European Association for Animal Production. 5–9 September 2004, under relatively improved managerial conditions. This Bled, Slovenia. Wageningen, Wageningen Academic Publishers, necessitates exerting efforts to improve management The Netherlands. Characteristics of Sudanese native chicken 41

Gueye, F.H.E. 2002. Employment and income generation through family Missohou, M., Dieye, P.N. & Talaki, E. 2002. Rural poultry production poultry in low income food-deﬁcit countries. World Poult. Sci. J., 58: and productivity southern Senegal. Livest. Res. Rural Dev.,14(2):1–8. 541–555. Muchadeyi, F.C., Sibanda, S., Kusina, N.T., Kusina, J. & Makuza, Hanaﬁ, M.S. & El-Labban, A.F.M. 1984. On estimating genetic para- S.M. 2004. The village chicken production system in Rushinga meters of partial egg production records and other related traits in pull- District of Zimbabwe. Livest. Res. Rural Dev., 16(6): 1–12. ets of Dokki-4 chicken produced from triallel mating. Egypt. J. Anim. Nesheim, C.M., Austic, E.R. & Card, E.L. 1979. Poultry production. Prod., 24: 51–67. 12th edn, Philadelphia, PA, Lea & Febiger, pp. 58–92. Horst, P. 1980. Genetically perspectives for poultry breeding on improved productive ability to tropical conditions. In 2nd World Sazzad, M.H. 1986. Reproductive performance of Desi (indigenous) hens Congress of Genetics Applied to Animal Production, Vol. 8, under scavenging and intensive system of rearing. In Proceedings of pp. 887–892. the First Annual Livestock Research Workshop. Savar, Dhaka, Bangladesh, Bangladesh Livestock Research Institute, pp. 63–67. Javed, K., Farooq, M., Mian, M.A., Durrani, F.R. and Mussawar, S. 2003. Flock size and egg production performance of backyard chicken Simeonovova, J., Ingr, I., Jerabek, S., Wnterova, J. & Dvorakova, V. reared by rural women in Peshawar, Pakistan. Livest. Res. Dev., 1989. An evaluation of eggs laid by Rhode Island Red (RIR) and 15(11). Rhode Island White (RIW) hens. Poult. Abstr., 16(7): 203. Kitalyi, A.J. 1998. Village chicken production system in rural Africa: Suleiman, M.E.F. 1996. Egg characteristics, genetic and phenotypic Household security and gender issues. Animal Production and relationships of body weight at various ages in indigenous chicken. Health Paper 142. Rome, FAO. University of Khartoum (MSc thesis). McAinsh, C.V., Kusina, J., Madsen, J. & Nyoni, O. 2004. Tadelle, D., Millon, T., Alemu, Y. & Peters, K. J. 2003. Village Traditional chicken production in Zimbabwe. World Poult. Sci. J., Chicken Production System in Ethiopia: 2. Use Patterns and 60: 233–246. Performance valuation and Chicken Products and socio-economic Functions of Chickens. Livestock Research for rural. Maphosa, T., Kusina, J., Makuz, N.T.S. & Sibanda, S. 2004. A monitoring study comparing production of village chickens between com- Wilson, R.T. 1979. Studies on the livestock of Southern Darfur, Sudan munal (Nharira) and small-scale commercial (Lancashire) farming VII. Production of poultry under simulated traditional conditions. areas in Zimbabwe. Livest. Res. Rural Dev., 16(7). Trop. Anim. Health Prod., 11: 143–140. Merat, P. 1990. Pleotropic and associated effects of major genes. In Wimmers, K., Ponsuksili, S., Hardge, T., Valle-Zarate, A., Mathur, P. R.D. Crawford, ed. Poultry breeding and genetics, pp. 429–467. K. & Horst, P. 2000. Genetic distinctness of Africa, Asian and South Amsterdam, Elsevier. American Local chickens. Anim. Genet., 31: 159–165. Animal Genetic Resources, 2014, 54, 43–51. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000113

Phenotypic characterization of indigenous chicken ecotypes in the north Gondar zone, Ethiopia

Addis Getu1, Kefyalew Alemayehu2 and Zewdu Wuletaw3 1Department of Animal Production and Technology, Faculty of Agriculture, Woldia University; 2Department of Animal Production and Technology, College of Agriculture and Environmental Sciences, Bahir Dar University; 3Sustainable Land Management Organization (SLM), Bahir Dar, Ethiopia

Summary An exploratory field survey was conducted in north Gondar zone, Ethiopia to identify and characterize the local chicken ecotypes. Seven qualitative and 12 quantitative traits from 450 chickens were considered. Chicken ecotypes such as Naked neck, Gasgie and Gugut from Quara, Alefa and Tache Armacheho districts were identified, respectively. Morphometric measurements indicated that the body weight and body length of the Naked neck and the Gasgie ecotypes were significantly (P < 0.01) higher than the Gugut ecotypes except in shank circumstances (circumference). Sex and ecotypes were significant (P < 0.01) sources of variation for both body weights and linear body measurements. The relationship of body weight with other body measurements for all ecotypes in both sexes was highly significant (r = 0.67, P < 0.01). Some traits like the spur length (r = 0.64, P < 0.01) for males and (r = 0.59, P < 0.01) for females of the Naked neck chickens are significantly correlated with body weight. Therefore, highly correlated traits are the basic indicators for the estimation of the continuous prediction of body weight of chicken. Identification and characterization of new genetic resources should be employed routinely to validate and investigate the resources in the country.

Keywords: chicken ecotypes, (Naked Neck), Gasgie, Gugut, north Gondar zone

Résumé En la zona Norte del distrito de Gondar (Etiopía), se llevó a cabo un estudio exploratorio de campo con el que se pretendía identificar y caracterizar los ecotipos locales de gallina. Se consideraron siete parámetros cualitativos y doce cuantitativos en una muestra de 450 gallinas. Se identificaron ecotipos de gallina tales como Cuello Pelado, Gasgie y Gugut en los distritos de Quara, Alefa y Tache Armacheho, respectivamente. Las medidas morfométricas indicaron que el peso y la longitud corporal de los ecotipos Cuello Pelado y Gasgie eran significativamente (P < 0.01) mayores que los del ecotipo Gugut, con la excepción de la circunferencia de los tarsos. El sexo y el ecotipo resultaron ser fuentes significativas (P < 0.01) de variación, tanto para el peso corporal como para las medidas lineales del cuerpo. Para todos los ecotipos y en ambos sexos, el peso corporal estuvo relacionado, de manera muy significativa (r = 0.67, P < 0.01), con el resto de medidas corporales. En las gallinas de Cuello Pelado, algunos rasgos, como la longitud de los espolones (r = 0.64, P < 0.01 en los machos y r = 0.59, P < 0.01 en las hembras), presentaron una correlación significativa con el peso corporal. Por tanto, los parámetros altamente correlacionados son indicadores básicos para estimar la evolución del peso corporal de las gallinas. La identificación y caracterización de nuevos recursos genéticos debería realizarse de manera rutinaria para validar y conocer los recursos presentes en el país.

Mots-clés: ecotipos de gallina, Cuello Pelado, Gasgie, Gugut, Norte del distrito de Gondar

Resumen Dans la zone Nord du district de Gondar (Éthiopie), une étude d’exploration a été menée sur le terrain pour identifier et caractériser les écotypes locaux de poule. Sept traits qualitatifs et douze quantitatifs ont été évalués sur un total de 450 poules. Des écotypes tels que Cou Nu, Gasgie et Gugut ont été identifiés dans les districts de Quara, Alefa et Tache Armacheho, respectivement. Les mesures morphométriques ont indiqué que le poids corporel et la longueur du corps étaient significativement (P < 0.01) plus élevés chez les écotypes Cou Nu et Gasgie que chez l’écotype Gugut, exception faite du tour des tarses. Le sexe et l’écotype ont été des sources de variation significatives (P < 0.01) aussi bien pour le poids corporel que pour les mesures linéaires du corps. Pour tous les écotypes et chez les deux sexes, des corrélations hautement significatives (r = 0.67, P < 0.01) ont été mises en évidence entre le poids du corps et les autres mesures corporelles. Chez l’écotype Cou Nu, certains traits, tels que la longueur des ergots, ont été significativement corrélés avec le poids corporel, tant pour les mâles (r = 0.64, P < 0.01) que pour les femelles (r = 0.59, P < 0.01). Ainsi, les traits hautement corrélés sont des indicateurs de base pour estimer l’évolution du poids corporel des poules. L’identification et la caractérisation de nou- velles ressources génétiques devraient se faire systématiquement pour valider et connaître les ressources présentes dans le pays.

Palabras clave: écotypes de poule, Cou Nu, Gasgie, Gugut, Nord du district de Gondar

Submitted 6 September 2013; accepted 27 February 2014

Correspondence to: Addis Getu, Department of Animal Production and Technology, Faculty of Agriculture, Woldia University. email: [email protected]

43 44 A. Getu et al.

Introduction Gondar town and 909 km from Addis Ababa with the temperature of 25–30 °C and annual rainfall of 900–1400 mm. Ethiopia takes the lead in livestock population and is a gate- Armacheho district is located 814 km northwest of Addis way of domestic animal migration from Asia to Africa, thus Ababa and 65 km northwest of Gondar town at an altitude rolled for widespread distribution and huge population size of 600–2000 m above MSL with the temperature of 25–42 in the country (Halima, 2007). Poultry contribute socio- °C and annual rainfall of 800–1800 mm (CSA, 2011). economic roles in food (security), generation of additional cash and income, etc. (Kondombo, 2005; Salam, 2005). Therefore, almost all rural and many peri-urban families Data collection methods fl keep a small ock of scavenging chickens (Jens et al., In addition to exploratory field survey, semi-structured ques- 2004). In Ethiopia, the population of chickens is estimated tionnaires and participatory rural appraisal, focus group at about 49.3 million of which 97.3, 0.38 and 2.32 percent discussion, field observation, trait characterization and are indigenous, hybrid and exotic breeds, respectively body measurements were employed to derive the required (CSA, 2011). Indigenous chickens have a good potential information. For the morphological and biometrical mea- to adapt in different agro-ecological conditions throughout surements, all matured chicken ecotypes n = 450, 150 their habitual management system (Tadelle and Alemu, males and 300 females, were measured. Qualitative traits 1997). Local chickens are non-descriptive type and show such as plumage size, body shape, comb type, shank colour, a large variation, which might be attributed to their wide- skin colour, head shape and eye colour were documented spread distribution (Tadelle, Alemu and Peters, 2003; through direct visualization. Whereas measurable traits Halima, 2007; Fisseha, Abera and Tadelle, 2010). like body weight (kg), body length, wing span, shank length Indigenous chickens are underestimated because of their and circumference, wattle length and width, keel length, poor performance. To this effect they have been neglected spur length, beak length, comb length and width were mea- and little attention has been given to them by researchers, sured using a spring balance and a measuring tape in cm, development workers and policy-makers (Tadelle, 2003). measuring to the nearest two digits (FAO, 2011). Some researchers (Tadelle, 2003; Halima, 2007; Dana et al., 2009) have made phenotypic and genetic characterization of indigenous chicken in some parts of Ethiopia. Data management and statistical technique Poultry production and market system were studied in Data from personal observation and focus group discussions Southern Ethiopia by Mekonnen (2007); characterization were summarized and synthesized by researchers, whereas, of poultry productivity and market system by Bogale the other quantitative and qualitative data were analysed (2008) and genetic parameters on Horro chickens for using SAS software version 9, 2002. Particularly, general weights and egg production traits was conducted by linear model was used to analyse quantitative traits (SAS, Dana, van der Waaji and Johan (2010). However, compre- 2002). Tukey’s comparison test was used to compare the fi hensive genetic resources identi cation in the remote sub factor brought significant difference. The model was fi districts of Northern Gondar zone in general and identi caused for body weight and linear body measurement of chick- tion, and characterization of new local chicken ecotypes in ens’ ecotypes by considering the fixed effects of sex and particular were not studied. Therefore, the objective of this ecotype. study was to identify and characterize the new local chick-

en ecotypes in the north Gondar zone of Ethiopia. Yijk = m + Ai + Dj + ADij + eijk, (1)

where Yijk is the observed body weight and linear body μ Materials and methods measurement of chickens; is the overall mean; Ai is the ﬁxed effect of ith eco-type (i = 1, 2 and 3); Dj is the effect ﬁ Description of the study area of kth sex ( j = male and i = female); ADji is the xed effect interaction of ith eco-type with jth sex; and E is the random The study was conducted in three districts of the north ijk residual error. Gondar zone (Quara, Alefa and Tache Armacheho) of Ethiopia (Figure 1). The altitude of the north Gondar Multiple liner regression analysis was performed to predict zone ranges from 528 to 4620 m mean sea level (MSL) the body weight of matured cocks and hens using 11 linear with annual rainfall of 880–1772 mm and temperature ran- body measurements (independent variables) in each eco- ging from 44.5 to –10 °C. Quara district is located in the type western part of the north Gondar zone between 11°47′ = b + b + b + b + b + b + b and 12°21N latitude and between 35°16′ and 35°47′E lon- Yj 0 1X1 2X2 3X3 4X4 5X5 6X6 + b + b + b + b + b + gitude. It is 1123 km from Addis Ababa and 324 km from 7X7 8X8 9X9 10X10 11X11 eijk , Gondar town and at an altitude between 528 and 654 m (2) above MSL. The annual temperature ranges from 25 to 44 °C with annual rainfall range of 600–1000 mm (CSA, where Yj is the dependent variable or predicted mean body 2011). Alefa district is located in the southwest of weight of chickens; β0 is the intercept; X1, X2, X3, X4, X5, Chicken ecotypes in Ethiopia 45

Figure 1. Map of the study area (Amhara region), and the three districts where the ecotypes identiﬁed are indicated in rectangular shape.

X6, X7, X8, X9, X10 and X11 are the independent variables tips (13.3 percent). About 53 percent of the birds have for wing span, body length, shank length, shank circumfer- white skin colour, 66 percent have single combs, 34 per- ence, keel length, spur length, beak length, wattle length, cent have rose combs and 70 percent have plain headed wattle width, comb length and comb width, respectively. facial appearance (Table 1 and Figure 2). The other pecu- β , β , β , β , β , β , β , β , β , β and β are partial liar features of this ecotype include aggressive behaviour, 1 2 3 4 5 6 7 8 9 10 11 higher feed intake, good productive and reproductive regression coefﬁcients of the variables X1, X2, X3, X4, X5, X , X , X , X , X and X and e is the residual error. performance, tolerance to common diseases and higher 6 7 8 9 10 11 ijk dressing percentage.

Results Gasgie chicken ecotype Gasgie chicken ecotype is distributed in the Alefa district Naked neck chicken and (Figure 3) most households keep this chicken sheltered Naked neck chickens are found in a very hot ecological in the family house during the night, while they spend the zone of the Quara district (Figure 2) and are maintained day scavenging in the backyards supplemented with grains under scavenging system with small feed supplementation and food leftovers. The chicken have a predominantly red and sheltered outside the family house (perch). These (32 percent) body plumage colour though have other chickens having predominantly white (28 percent) and diverse plumage colours such as white (9.3 percent), red- red (20 percent) body plumage colours. However, they brownish (9 percent), white with red tips (9 percent), have heterogeneity and diverse additional plumage colour black with white tips (9 percent), black (5.3 percent), like red-brownish (0.7 percent), white with red tips (5.3 multicolour (4 percent) and white black red tips (1.3 per- percent), black with white tips (10.7 percent), black (7.3 cent). About 60.7 percent of the ecotypes are rose combed percent), multicolour (5.3 percent) and white black red (Figure 3 and Table 1). Long neck (especially males), short 46 A. Getu et al.

Figure 2. Typical Naked neck male (right) and female (left) chicken ecotypes.

Table 1. Description of dominant body plumage colours, head shape, comb type and skin colour of newly reported indigenous chicken ecotypes (N = 450).

Character Attributes Ecotypes by proportions and their associations P-value

Naked neck = Gasgie Gugut Overall Cramer’s V 150 N = 150 N = 150 N = 450 Overall (%)

Plumage White black and redns tips (Kiy Tikur (0.67) 1.3 (1.34) (2.67) 1.56 0.13 0.323 colours Teterma) 2 0.22 Black with white tips (Tikur Teterma)ns (10.70) (8.67) (6.00) 0.22 8.44 0.13 0.286 White with red tips (Kiy Teterma)ns (5.33) (8.67) (8.00) 7.33 0.11 0.465 Grayish mixture (Gebsema)*** (10.70)b (22.00)a (10.00)b 14.20 0.33 0.005 Red-brownish (Kokima)** (13.30)a (8.67)a (2.00)b 8.00 0.34 0.003 Multicolour (Ambesa)* (5.33)a,b (4.00)b (10.00)a 6.44 0.22 0.037 Black (Tikur)*** (7.33)b (5.33)b (22.00)a 11.50 0.52 0.001 White (Nech)*** (28.00)a (9.33)b (9.33)b 15.60 0.66 0.001 Red (Kiy)*** (18.70)b (32.00)a (30.00)a 26.90 0.60 0.001 Head shape Plain (Ebaberas)ns (70.00)a (65.33)a (76.00)a 70.40 0.10 0.19 Crest (Gutya)ns (30.00) (34.67) (24.00) 29.60 0.08 0.23 Comb type Doublex (V-shape)* NA (9.33)b (24.00)a 32.70 0.21 0.024 Single*** (66.00)a (27.33)b (30.00)b 38.90 0.42 0.001 Rose*** (34.00)b (60.70)a (38.70)b 44.40 0.33 0.001 Peans NA (2.67)a (7.33)a 3.33 0.03 0.045 Skin colour Yellow*** (66.00)a (24.70)b (67.30)a 53.10 0.55 0.001 White*** (23.30)b (66.00)a (20.70)b 42.90 0.75 0.001 Black* (4.00)a,b (0.67)a (8.00)b 4.22 0.20 0.001 Green* (6.00)a NA (1.33)b 1.56 0.18 0.004 Red* (0.67)a (8.67)b (3.33)a 4.22 0.09 0.05 Eye colour Orangens (0.67)a (4.67)a (4.67)a 3.33 0.10 0.088 Blackns (6.00)a NA (3.33)a 3.11 0.16 0.058 Purl*** (1.33)b NA (3.33)a 1.56 0.13 0.021 Red*** (92)b (95.30)a (88.70)b 92 0.42 0.001 Body shape Triangular*** (40.7)a (15.30)b (30.70)a 29.00 0.32 0.001 Blocky*** (58.00)a (31.30)b (49.30)a 46.20 0.36 0.001 Wedge*** (1.30)b (5.33)b (20.00)a 24.90 0.32 0.001 Shank colour Yellow*** (68.00)b (66.00)a,b (51.30)a 62.00 0.19 0.002 White*** (16.70)a,b (14.00)b (26.00)a 19.10 0.20 0.001 Blackns (7.33)a (9.30)a (8.00)a 8.00 0.07 0.285 Greenns (7.33) (10.00) (8.00) 8.22 0.03 0.814 Red* (0.67)b (0.67)b (5.33)a 2.67 0.17 0.012

NA, not available and different superscripts within a row indicate signiﬁcantly different means (P<0.05). Chicken ecotypes in Ethiopia 47

Figure 3. Typical Gasgie male (left) and female (right) chicken types. weaning time, docile and good productive and reproduct- low performance, short beak length, ability to resist endemic ive performances are the unique features of this ecotype. disease, small body size, passive and easily exposed to predators are the unique behaviours of the ecotype. These chicken are dominated by white (22 percent) body plumage Gugut chicken ecotype colour and other diverse plumage colour such as multicolour Gugut chickens (Figure 4) are distributed in the Debresina (10 percent), red (9.3 percent), black (9.3 percent), white area of Tache Armacheho district. Most of the households with red tips (8 percent), red-brownish (2 percent), black keeping these chickens provide separate hanging shelters with white tips (2 percent) and white black red tips (2 per- during night time to protect them from hyperthermia. cent). The investigated chicken ecotype showed additional Dense feathers from the neck, absence of wattle in female, heterogeneity in quantitative traits (Table 1).

Figure 4. Typical Gugut male and female chicken types. 48 A. Getu et al.

The results indicated that the dominant average plumage (WW), beak length (bl), spur length (sl), keel length (KL) colour of newly identified average three local chicken eco- (cm) and body weight (Wt) (kg) for different sexes were types were 26.90 percent red followed by 15.60 percent considered. The least-squares mean of body weight and white and 14.20 percent greyish mixture. About 44.4 per- body measurements of Naked neck, Gasgie and Gugut cent are rose comb type, 42.90 percent have white skin col- chickens with Honestead significant difference comparison our, 46.20 percent have blocky body shape and 70.40 tests are presented in Table 2. The overall least-squares percent are plain headed and are the most dominant mean of wing span, shank length, shank circumference, observable traits of Gugut chickens ecotypes. About 34, body length, comb length, comb width, wattle length, 60.7 and 38.7 percent of chicken from Naked neck, wattle width, beak length, spur length, keel length (in cm) Gasgie and Gugut ecotype, respectively, were character- and body weight (in kg) were (37.04 ± 0.13), (7.79 ± 0.15), ized by rose comb type. The proportion of plain head (3.78 ± 0.07), (35.79 ± 0.09), (2.76 ± 0.09), (1.68 ± 0.04), shape in chicken populations of Naked neck, Gasgie and (1.76 ± 0.06), (1.51 ± 0.06), (2.03 ± 0.02), (0.18 ± 0.02), Gugut was comparable with 70, 65.3 and 76 percent, (8.24 ± 0.09) and (1.46 ± 0.01), respectively. respectively (Table 1 and Figure 3). This variation could be adaptation fitness to their environment (Dana, 2011). Overall sex effect body weight mean squares of male and female chickens were 1.63 ± 0.03 and 1.37 ± 0.02 kg, respectively. Naked neck chicken male body weight (1.78 ± 0.31 kg) is significantly (P < 0.01) higher than the Quantitative traits of (Naked neck, Gasgie Gugut male chicken 1.40 ± 0.04 kg but not body weight and Gugut) chickens of Gasgie 1.71 ± 0.05 kg. Further, the Naked neck cocks and hens were found to have significantly taller shank A total of 450 adult hens and cocks with 12 measurable length of 9.61 ± 1.03 and 9.043 ± 1.10 (cm), respectively, parameters such as wing span (WS), shank length (SL), than the Gugut. However, shank circumference of Gugut shank circumference (SC), body length (BL), comb length cocks and hens are inversely superior in shank circumfer- (CL), comb width (CW), wattle length (WL), wattle width ences than Naked neck and Gasgie male and female

Table 2. Comparison (LSM ± SE) of body weight (kg) and linear body measurements (cm) and extraction effect of independent variables of the three indigenous chickens.

Parameters Sex Naked neck Gasgie Gugut type CV% P-value Overall mean Grand mean

Sample size M 50 50 50 150 450

F 100 100 100 300

Effects and levels LSM ± SE LSM ± SE LSM ± SE LSM ± SE LSM ± SE

WS M 38.70 ± 2.6a 39.61 ± 0.42a 35.97 ± 0.23b 6.51 0.0001** 38.09 ± 0.24a 37.04 ± .13 F 37.17 ± 2.36a 37.36 ± 0.26a 35.03 ± 0.18b 6.19 0.0001** 36.52 ± 0.14b SL M 9.61 ± 1.03a 7.25 ± 0.10b 7.37 ± 0.73b 10.05 0.0001** 8.08 ± 0.11a 7.79 ± 0.15 F 9.043 ± 1.10a 6.80 ± 0.06c 7.08 ± 0.05b 9.10 0.0001** 7.64 ± 0.07b BL M 38.12 ± 2.14a 36.10 ± 0.34a 35.2 ± 0.09b 9.49 0.0002** 36.77 ± 03a 35.79 ± 0.09 F 36.90 ± 2.61a 34.60 ± 0.26b 34.37 ± 0.21b 6.93 0.0001** 35.29 ± 0.16b CL M 3.25 ± 0.87a 3.16 ± 0.12a 3.08 ± 0.09a 26.24 0.594ns 3.16 ± 0.07a 2.76 ± 0.09 F 2.99 ± 3.68a 2.28 ± 0.07b 2.40 ± 0.06a,b 35.61 0.0482* 2.55 ± 0.13b CW M 2.11 ± 0.82a 1.93 ± 0.13a 2.19 ± 0.05a 38.36 0.255ns 2.08 ± 0.07a 1.68 ± 0.04 F 1.78 ± 0.85a 1.07 ± 0.06b 1.59 ± 0.06a 45.55 0.0001** 1.48 ± 0.04b WL M 2.76 ± 0.69a 2.70 ± 0.14a 1.83 ± 0.23b 32.19 0.0001** 2.43 ± 0.07a 1.76 ± 0.06 F 2.44 ± 0.80a 1.84 ± 0.04b NA 37.23 0.0001** 1.42 ± 0.07b WW M 2.76 ± 1.01a 2.32 ± 0.16b 1.45 ± 0.09c 44.53 0.0001** 2.17 ± 0.09a 1.51 ± 0 0.06 F 2.34 ± 1.03a 1.19 ± 0.05b NA 56.29 0.0001** 1.18 ± 0.07b bl M 2.42 ± 0.45a 2.00 ± 0.02b 1.85 ± 0.10c 14.12 0.0001** 2.09 ± 0.03a 2.03 ± 0.02 F 2.28 ± 0.60a 1.93 ± 0.0b 1.78 ± 0.02c 18.67 0.0001** 1.99 ± 0.02b sl M 0.66 ± 0.8a 0.49 ± 0.10a 0.17 ± 0.09b 46.35 0.0009** 0.44 ± 0.05a 0.18 ± 0.02 F 0.09 ± 0.32a 0.08 ± 0.02a NA 48.83 0.0172* 0.011 ± 0.18a SC M 3.58 ± 0.50b 3.25 ± 0.07b 3.85 ± 0.03a 20.78 0.0001** 4.81 ± 0.18a 3.78 ± 0.07 F 3.31 ± 0.59a 3.11 ± 0.03b 3.38 ± 0.07a 17.23 0.0027** 3.27 ± 0.03b KL M 9.11 ± 1.02a 9.55 ± 0.15a 7.62 ± 0.23b 16.81 0.0001** 7.51 ± 0.24b 8.24 ± 0.09 F 8.56 ± 0.87b 9.27 ± 0.08a 7.98 ± 0.07c 9.08 0.0001** 8.60 ± 0.05a Wt M 1.78 ± 0.31a 1.71 ± 0.05a 1.40 ± 0.04b 18.15 0.0001** 1.63 ± 0.03a 1.46 ± 0.01 F 1.52 ± 0.26a 1.36 ± 0.03b 1.23 ± 0.02c 17.50 0.0001** 1.37 ± 0.02b

WS, wing span; SL, shank length; BL, body length; CL, comb length; CW, comb width; WW, wattle width; WL, wattle length; KL, keel length; sl, spur length; bl, beak length; SC, shank circumference, in the measurement of cm; Wt, weight (kg); NA, not available; LSM, least-squares mean; SE, standard error and different superscripts within a row indicate signiﬁcantly different means (P < 0.05). Chicken ecotypes in Ethiopia 49

chickens (Table 3). A non-signiﬁcant comb length vari- variables by considering other traits like sl, SC, CL, bl ation between sexes of Naked neck chicken was obtained. and KL at a time in the three chicken ecotypes. In linear While Naked neck and Gasgie cocks had the longest beak regression result, the body weight prediction value of length of 2.42 ± 0.45 and 2.00 ± 0.02 cm, respectively, Naked neck cocks and hens and Gasgie cocks were 0.40, than Gugut cocks 1.85 ± 0.10 (cm). Beak length variation 0.31 and 0.45, respectively. In addition to liner regression, is recorded among ecotypes but not with in ecotype in multiple regression analysis was considered to determine respective sexes (Table 3). the effects of other body measurements on body weight prediction (Table 4). To increase meat and egg production it requires genetic improvement of body weight of chickens. But proper measurement of this variable is often Correlations of body weight and other linear hard in villages due to lack of weighing scales. Hence, eas- body measurements ily measurable linear body measurements are more relevant for chickens’ body weight prediction at farmers’ Live weight was positively correlated (r = 55.5, P < 0.01) level rather subjectively judging manually. In addition, with wing span. Body length and spur length in Naked the present farmers are active at early morning by provid- neck were positively correlated, males (r = 0.62, P < 0.01) ing supplementary feed to their chicken before bringing and females (r = 0.55, P < 0.01). Whereas WL is the high- them to the market to increase the temporary body weight est correlated trait (r = 0.67, P < 0.01) with body weight of of their chickens. Therefore, prediction equation was ﬁ Gasgie male chickens. The high correlation coef cients important. between body weight and other body measurements (P < 0.01) helped to predict body weight of chickens (Table 3). Discussions

Prediction equation models Analysing the research result evidenced that more than 70 percent of the population of chicken ecotypes in the study First WS, BL and WW traits were used as linear regression area were carrying the Naked neck chicken characteristics. to predict body weight of chickens. Whereas stepwise mul- This is new and signiﬁcantly (P < 0.001) higher in number tiple regression was considered to predict the dependant than reported result in other parts of Ethiopia (7.9 percent;

Table 3. Coefﬁcient of correlations between body weight and linear body measurements for female and male in all ecotypes in the study area (N = 450).

Traits Variables Sex and ecotype

Naked neck Gasgie Gugut type

MFMFMF

WS N 50 100 50 100 50 100 r 0.64** 0.56** 0.35* 0.41** 0.50** 0.39** SL N 50 100 50 100 50 100 r 0.18ns 0.20* 0.26* 0.29** 0.54** 0.08ns BL N 50 100 50 100 50 100 r 0.59** 0.54** 0.49** 0.59** 0.33* 0.50** CL N 50 100 50 100 50 100 r 0.31* 0.05ns 0.54** 0.41** 0.40** 0.35** CW N 50 100 50 100 50 100 r 0.15ns −0.07ns 0.64** 0.39** 0.37** 0.21ns WL N 50 100 50 100 50 100 r 0.05ns −0.01ns 0.67** 0.39** 0.39** NA WW N 50 100 50 100 50 100 r 0.05ns −0.13ns 0.52** 0.47** 0.49** NA bl N 50 100 50 100 50 100 r −0.22ns −0.01ns 0.20ns 0.30** 0.24* 0.22* sl N 50 100 50 100 50 100 r 0.48** 0.27** 0.52** 0.28** 0.21ns 0 SC N 50 100 50 100 50 100 r 0.31* 0.13ns 0.35** 0.18ns 0.04ns −0.02ns KL N 50 100 50 100 50 100 r 0.37** 0.28** 0.62** 0.33** 0.23* 0.21*

WS, wing span; SL, shank length; BL, body length; CL, comb length; CW, comb width; WL, wattle length; WW, wattle width; bl, beak length; sl, spur length; SC, shank circumference; KL, keel length; N, number of samples and r, correlation coefﬁcients. 50 A. Getu et al.

Table 4. Prediction equations in multiple regression analysis of body weight on other variables of female and male.

Ecotype Male R2 Female R2

Naked neck Y = −1.34 + 0.08WS 0.40 Y = −0.78 + 0.06WS 0.31 Y = −1.12 + 0.07WS + 0.17sl 0.60 Y = −1.48 + 0.04WS + 0.04BL 0.41 Y = −2.12 + 0.06WS + 0.04BL + 0.14sl 0.65 Y = −1.60 + 0.05WS + 0.04BL − 0.10bl 0.47 Gasgie Y = 1.11 + 0.22WS 0.45 Y = −0.89 + 0.06BL 0.35 Y = 0.49 + 0.21WS + 0.2SC 0.53 Y = −0.77 + 0.05BL + 0.19WW 0.46 Gugut Y = 1.1 + 0.21WW 0.30 Y = −0.04 + 0.04BL 0.25 Y = −0.97 + 0.16WW + 0.06WS 0.44 Y = −0.47 + 0.04BL + 0.22bl 0.31 Y = −1.05 + 0.05WS + 0.09CL + 0.13WW 0.50 Y = −0.47 + 0.03BL + 0.07CL + 0.24bl 0.37 Y = −0.22 + 0.05WS + 0.12CL + 0.14WW + 0.03KL 0.55 Y = −0.8 + 0.03BL + 0.06CL + 0.23bl + 0.04KL 0.40 Y = −1.15 + 0.03BL + 0.06CL + 0.2bl + 0.05SC + 0.07KL 0.42

WS, wing span; sl, spur length; BL, body length; SC, shank circumference; WW, wattle length; CL, comb length and KL, keel length, all variables left in the model are signiﬁcant at the 0.05 level. No other variable met the 0.05 signiﬁcance level for entry into the model on body weight.

Aberra and Tegene, 2011; <2 percent; Dana, 2011), qualitative traits were considered among the three chicken Nigeria (6 percent; Gueye, 1998) and Botswana (3.6 per- ecotypes. Heavier adult body weight and longer shank cent; Badubi, Rakereng and Marumo, 2006). The other length were measured from the Naked neck, followed by results on plumage colours of the identified chicken eco- the Gasgie chicken ecotypes. Qualitatively, the Gasgie types are different from the report result from northwest chicken eco-type had normal feather morphology and Ethiopian (Halima, 2007). Variations of rose comb types, others like the Naked neck chicken ecotype is easily distin- white skin colour, blocky body shape and plain head guished by the complete absence of feather at neck and types are the dominant visible traits of chicken ecotypes. chest. Whereas Gugut chicken ecotype is characterized by This result was not in lined with the reported result done complete absence of wattle in hens, it is the smallest of all at Bure and Fogera districts in the Amhara region and and has dunce feather at neck in both the sexes. All these Dale district in Southern Ethiopia (Fisseha, Abera and findings indicated that the investigated chicken ecotypes Tadelle, 2010). This variation could be a breed-specific show heterogeneity in most traits considered. Thus, trait, nutritional status, genotype and reflected adaptation in-depth molecular characterization using genetic markers fitness to their environment (Aberra and Tegene, 2011; should be undertaken to confirm the level of genetic varia- Dana, 2011). Complete absence of wattle from Gugut tions and relationships among newly identified and other females, long neck and early weaning of Gasgie chicken indigenous chicken ecotypes. ecotype is the unique character from the previous studies in Ethiopia and elsewhere in the tropics by Halima (2007), Aberra and Tegene (2011) and Dana (2011). Acknowledgements Overall body weight of male and female chickens are varied from Ethiopian chickens reported by Dana (2011) This work is funded by the German International (GIZ) which is 1.63 kg for males and 1.27 kg for females. project to which we are grateful. We also thank all inter- Body weight of Naked neck and Gasgie chickens were viewed village chicken owners, district agricultural office higher than chicken in central (Danna, 2011) and north- experts, key informants and other participants who helped west Ethiopia (Halima 2007) in the body weight of 1.26 us to carry out every activity on this research. and 0.87 kg for adult male and female, respectively. The frequency of chicken ecotypes carrying the Naked neck fi gene that we studied was signi cantly (P<0.001) higher References than those reported in other parts of Ethiopia (<2 percent; Dana, 2011), Nigeria (6 percent; Gueye, 1998) and Aberra, M. & Tegene, N. 2011. Phenotypic and morphological charac- Botswana (3.6 percent; Badubi, Rakereng and Marumo, terization of indigenous chicken populations in southern region of 2006). Ethiopia. Anim. Genet. Resour., 49: 19–31. Badubi, S.S., Rakereng, M. & Marumo, M. 2006. Morphological characteristics and feed resources available for indigenous chickens in Botswana. Columbia, CIPAV, Livestock Research for Rural Conclusions and recommendations Development.

Naked neck, Gasgie and Gugut chickens are newly iden- Bogale, K. 2008. In situ characterization of local chicken eco-type for fi functional traits and production system in Fogera district, Amhara ti ed ecotypes from Quara, Alefa and Tache Armacheho regional state. Submitted to the Department of Animal Science, district in the northern parts of Ethiopia, respectively. The Haromiya University (MSc thesis). fi fi identi ed chicken ecotypes had diversi ed variations in CSA. 2011. Agricultural sample survey 2010/11. Volume 2: Statistical both qualitative and quantitative characters. As an example, Bulletin 05. Report on Livestock and Livestock Characteristics, phenotypic characterizations like 12 quantitative and seven Addis Ababa, 21 February. Chicken ecotypes in Ethiopia 51

Dana, N. 2011. Breeding programs for indigenous chicken in Ethiopia Jens, C.R., Anders, P., Charlotte, V., Ainsh, M.C. & Lone, F. 2004. analysis of diversity in production systems and chicken populations. Keeping of village poultry. A technical manual for small-scale poultry Submitted in fulﬁllment of the requirements for the degree of doctor production. Denmark 34. at Wageningen University (PhD thesis). Kondombo, S.R. 2005. Improvement of village chicken production in a Dana, N., Tadelle, D., Elisabeth, H.V. & Johan, A.M. 2009. mixed farming system in Burkina Faso. Wageningen Institute of Morphological features of indigenous chicken populations of Animal Sciences, Animal Nutrition Group, Wageningen University, Ethiopia. Animal Breeding and Genomics Center, Wageningen The Netherlands (PhD thesis). University. Anim. Genet. Resour., 46: 11–23. Mekonnen, G. 2007. Characterization of smallholder poultry production Dana, N., van der Waaji, E. & Johan, A.M. 2010. Genetic and pheno- and marketing system of Dale, Wonsho and Loka Abaya Woredas typic parameter estimates for body weights and egg production in of southern Ethiopia. Awassa College of Agriculture, Hawassa Horro chicken of Ethiopia. Submitted to Tropical Animal Health University (MSc thesis). and Production, Animal Breeding and Genomics, Wageningen SAS. 2002. Statistical Analysis System (SAS), SAS users guide, version University, The Netherlands. 9.1. NC, SAS Institute Inc. FAO. 2012. Phenotypic characterization of animal genetic resources. Salam, K. 2005. Improvement of village chicken production in a mixed FAO Animal Production and Health Guidelines No. 11. Rome. (chicken ram) farming system in Burkina Faso. Wageningen (accessible at http://www.fao.org/docrep/015/i2686e/i2686e00.pdf). Institute of Animal Sciences, Animal Nutrition Group, Wageningen Fisseha, M., Abera, M. & Tadelle, D. 2010. Assessment of village University, The Netherlands (PhD thesis). chicken production system and evaluation of the productive and repro- Tadelle, D. 2003. Phenotypic and genetic characterization of local ductive performance local chicken ecotype in Bure district North West chicken ecotypes in Ethiopia. Submitted to Humboldt University of Ethiopia. Afr. J. Agric. Res., 5(13):739–1748. Germany (PhD thesis). Gueye, E.F. 1998. Village egg and fowl meat production in Africa. Tadelle, D. & Alemu, Y. 1997. Studies on village poultry production sys- World’s Poult. Sci. J., 54: 73–86. tems in the central highlands of Ethiopia. Submitted to Swedish University (MSc thesis). Halima, H. 2007. Phonotypic and genetic characterization of indigenous chicken populations in Northwest Ethiopia. Submitted to the Faculty Tadelle, D., Alemu, Y. & Peters, K. 2003. Village chicken production of National and Agricultural Sciences, Department of Animal, Wild systems in Ethiopia: use patterns and performance evaluation and Life and Grass Land Sciences, University of the Free State, chicken products and socio-economic functions of chicken. Livest. Bloemfontein and South Africa (PhD thesis). Res. Rural Dev., 15(1). Animal Genetic Resources, 2014, 54, 53–63. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000095

Studies on morphometrical traits of Boran bulls reared on two feedlots in Southern Ethiopia

Sandip Banerjee1, Mohamed Beyan Ahmed1 and Girma Tefere2 1School of Animal and Range Sciences, Hawassa University, P.O Box 05, Hawassa, SNNPRS, Ethiopia; 2Dilla A.T.V.E.T College, Dilla, SNNPRS, Ethiopia

Summary The study was conducted on Boran bulls from three age groups (2, 3 and 4 years) reared at two different locations in Southern Ethiopia. The study was conducted to assess the different morphometrical measurements in Boran bulls, to estimate live weight of the bulls using step-down regression equations and also to calculate structural indices for assessment of type for which the breed was developed and thus assist in the selection of the bulls. The feedlots were situated in Meki district and varied both in location and the type of management, with the second feedlot being better managed than the ﬁrst. The morphometric traits included in the study were height at withers (HW), height at rump (RH), body length (BL), heart girth (HG), head width (WFH), neck circumference (NC), neck length (NL), chest width (CW), rump length (RL), hip width (WH) and ﬂank width (FW) and the body weight (BW). The results indicated that while the skeletal measurements HW, RH and BL did not vary (P < 0.05) between bulls of a particular age across the two feedlots, the other parameters were higher among the bulls reared in the feed lot with wind breaks and situated away from lake Ziway. The results of the step-down regression equations indicated that BW could be assessed using various morphometrical measurements viz. NL, NC, RL, WH and HG. The parameters included varied both across the age groups and locations. The results related to the structural indices indicated that the Boran bulls had posterior alignment and had higher CW than HW.

Keywords: Boran bulls, characterization, morphometric

Résumé L’étude a été menée sur des taureaux Boran de trois groupes d’âge (2, 3 et 4 ans) élevés à deux endroits différents dans sud de l’Ethiopie. Les parcs d’engraissement étaient situées dans le district de Meki et varié tant dans l’espace et le type de gestion, avec le deuxième parc d’engraissement est mieux géré que la première. Les traits morphométriques inclus dans l’étude étaient la hauteur au garrot (HW), la hauteur à la croupe (RH), la longueur du corps (BL), la circonférence de coeur (HG), la largeur de la tête (FMH), circonférence du cou (NC), la longueur du cou (NL ), la largeur de la poitrine (CW), la longueur croupe (RL), la largeur des hanches (WH) et la largeur de ﬂanc (FW) et le poids corporel de (BW). Les résultats indiquent que, bien que la mesure squelet- tiques HW, RH et BL ne varient pas (P < 0.05) entre les taureaux d’un âge donné dans les deux parcs d’engraissement, les autres paramètres étaient plus élevés chez les taureaux élevés dans le parc d’engraissement avec des brise-vent et situés loin du lac Ziway. Les résultats de l’étape vers le bas équations de régression ont indiqué que BW peut être évaluée en utilisant différentes mesures morphométriques à savoir. NL, NC, RL, WH et HG. Les paramètres inclus variée à la fois à travers les groupes d’âge et les lieux. Les résultats liés à des indices structurels indiqué que les taureaux Boran eu alignement postérieur et avait CW supérieur à HW.

Mots-clés: taureaux Boran, paramètres morphométriques, caractérisation

Resumen Este estudio fue llevado a cabo con toros Boran, de tres grupos de edad (2, 3 y 4 años), criados en dos lugares distintos del Sur de Etiopía. El estudio se realizó con el fin de medir diferentes parámetros morfométricos en toros Boran, de estimar el peso vivo de los toros con ecuaciones obtenidas por el método de regresión por eliminación de variables (análisis step-down) y con el fin también de calcular índices estructurales para la determinación del tipo de animal pretendido cuando se desarrolló la raza y, así, con- tribuir a la selección de los toros. Los corrales de engorde estuvieron ubicados en el distrito de Meki y difirieron tanto en el emplazamiento como en el tipo de manejo, siendo el segundo cebadero mejor gestionado que el primero. Los parámetros morfométricos incluidos en el estudio fueron la altura a la cruz (AC), la altura a la grupa (AG), la longitud corporal (LC), la circunferencia torácica (CT), la anchura de la cabeza (ACz), la circunferencia del cuello (CCu), la longitud del cuello (LCu), la anchura del pecho (AP), la longitud de la grupa (LG), la anchura de la cadera (ACd), la anchura de los flancos (AF) y el peso corporal (PC). Los resultados indicaron que, si bien las medidas esqueléticas AC, AG y LC no variaron significativamente, para una misma edad, entre los toros de los dos cebaderos, los valores de los otros parámetros fueron más elevados en los toros criados en el corral de engorde con cortinas cortaviento, situado a gran distancia del lago Ziway. Los resultados obtenidos con las ecuaciones de regresión señalaron que el peso corporal podía ser estimado usando algunas medidas morfométricas, en concreto LCu, CCu, LG, ACd y CT. Los parámetros

Correspondence to: Sandip Banerjee, School of Animal and Range Sciences, Hawassa University, PO Box 05, Hawassa, SNNPRS, Ethiopia. email: san- [email protected] and [email protected]; tel.: +251916011747

53 54 S. Banerjee et al.

considerados se vieron afectados tanto por el grupo de edad como por la ubicación del cebadero. En cuanto a los índices estructurales, los resultados mostraron que los toros Boran presentaban alineamiento posterior y que tuvieron mayor AP que AC.

Palabras clave: toros Boran, parámetros morfométricos, caracterización

Submitted 24 August 2013; accepted 27 February 2014

Introduction quality estimation such as feeling the loin area or by visual Livestock contribute significantly to poverty alleviation of estimates alone. Weighing scales are rare and they are a major section of society, especially in the developing mostly inaccurate due to lack of maintenance. Therefore world. Studies by Hoffmann (2010) indicate that livestock under such constraints, the pastoralists are usually unable are an important component of livelihood of resource chal- to receive a fair price of their livestock. Linear measure- fi lenged societies. CSA (2012) indicate that Ethiopia has ments provide a scienti c basis to describe the biological about 52.13 million cattle and the livestock production sys- variations between breeds and also for animals within a tem can be classified into three types of production sys- breed and thus can serve as a basis for measuring the per- tems: the lowland pastoral and agro pastoral system; the formance, productivity and carcass characters (Kayastha highland crop livestock production system; and a very et al., 2011; Kugonza et al., 2011). Studies by Gatesy small proportion of urban and peri-urban livestock produc- and Arctander (2000) indicated that morphological tion system. As indicated by ESPSLMMP (2010) the live- descriptions are useful tools for distinguishing livestock stock sector contributes about 20 percent of the Gross breeds and strains. These traits are also used to evaluate Domestic Product, supporting the livelihood (directly or the breeding goals of livestock (Zechner et al., 2001). indirectly) of about 70 percent of the nation´s population Accordingly, Mwacharo et al. (2006) indicated that the and representing 11 percent of the annual foreign earnings. relationship between size and breed can also be fairly According to ESPSLMMP (2011), Ethiopia earned about well estimated using several morphometrical measure- 211.1 million USD from export of 16 877 tonnes of meat ments. Many of the morphometrical traits are highly corre- and 472 041 head of livestock, and the revenue was lated with the body weight (BW) of livestock and hence enhanced by 69 percent in 2011 over the previous fiscal can serve as an important predictor for assessment of year. The study also shows that the earnings were mostly live weight of animals (Goe, Alldredge and Light, 2001; due to the sale of live animals (70 percent), while the Mwacharo et al., 2006). Skeletal measurements such as rest was from the sale of chilled carcasses. Among the dif- height at withers (HW) and body length (BL) indicate ferent classes of livestock, the sale of cattle on the hoof measurements pertaining to size of the livestock, which accounted for 46 percent of the total livestock exported. are better indicators of BW when compared to measurements which are related to muscle development and fat Studies by NEPAD–CAADP (2005) indicated that most of deposition, which are affected by nutrition and manage- the cattle raised for export purpose are bred in the lowlands ment (Kamalzadeh, Koops and van Bruchem, 1998). situated below 1 500 m a.s.l which covers about 60 percent One of the most commonly used procedures for the estima- of the total land of the country. These lowlands are situated tion of live-weight of cattle is by developing regression in the Eastern, Southern and Western parts of the Central equations based on some simple linear measurements as highlands. The report also indicates that these regions predictors. Studies by Heinrichs, Rogers and Cooper house 20 percent of the nation’s cattle, 25 percent sheep (1992) and Gilbert, Bailey and Shannon (1993) indicated and 75 percent of the goat population. The majority of that body measurements in livestock can serve as predic- the residents in these areas directly or indirectly depend tors of body weight. Msangi et al. (1999), Slippers, on livestock or livestock products to meet their daily Letty and De Villerrs (2000), Fouire et al. (2002), needs. Cattle play a multifarious role in the life of the Willeke and Dürsch (2002) and Bozkurt (2006) reported Agrarian Society: their dung is used as manure, they are that emphasis has shifted from subjective methods like also used for most of the agriculture-related work, from body condition score to objective methods such as the ploughing to transportation of the agriculture produce to use of linear measurements of different body parts for the market, with many other intermediate activities in the estimation of body weight. As indicated by Alderson between (Felleke and Geda, 2001; Little et al., 2001). As (1999), the importance of BW alone is of limited import- indicated by Solomon (2001), Borena plateau of ance in the absence of associated traits of type and Southern Ethiopia is the home of the best-known cattle conformation. Studies by Gilbert and Gregory (1952) breed from this country. The Boran cattle are the most indicate that ratios of BW/ HW can serve as an indicator sought after breed for consumption within the country of beef type in Hereford cattle. According to Alderson and for the export purpose (LMA 2001). The pastoralists (1999), indices can serve as useful tools in selection of usually grade their livestock using informal methods of breeding animals and also for assessment of cattle at Morphometrical traits of Boran bulls reared in Southern Ethiopia 55

maturity. They also serve as an alternative option for the season in the region. The cattle were grouped in three age assessment of BW as it incorporates measures, which are groups, i.e. 2, 3 and 4 years, respectively in both the feedlots, correlated with desirable conformation. Calculation of where they were separated by their ages, the bulls were pur- indices can also help potential livestock dealers as the traits chased by estimating their age and body weight. The age of considered in the development of the indices are associated the animals was assessed using the dentition method as sug- with the production traits. As indicated by Salako (2006), gested by Hammond, Mason and Robinson (1971). the desirable body conformation for traits pertaining to meat The Boran bulls (Figure 2) were purchased from their production is quite complex and hence cannot be truly breeding in Southern Ethiopia tract 6 months prior to the assessed taking into account a single linear measurement commencement of the study. In the feedlot, the bulls and is usually assessed using body condition score, which were sprayed with Cypermethrin at 2 ml/l of water and itself is a very subjective measurement and hence prone to drenched with Albendazole based on their BW and as per errors, besides most of the subjective scores have very the manufacturer’s recommendation. The cattle were pro- low heritability, and hence are rarely accurate in selection vided with concentrate supplements, which comprised rice of breeding stock. Indices are objective body measurements and wheat bran besides hulls from different pulses (haricot that are relatively accurate to estimate and are relatively easy bean (Phaseolus vulgaris), lentil (Lens culinaris)andpeas to understand. The term “type” indicates the structure of an (Pisum sativum) cotton seed cake and common salt the animal besides its body form and is hence considered as roughage (consisted of teff (Eragrostis tef), wheat ideal for understanding the purpose for which the animal (Triticum sp.) and barley (Hordeum vulgare)straw)and has been developed. The aim of this study was to investigate water was provided adlib, the proximate analysis of the morphometrical traits, assessment of live weight using step- feed indicated that it contained 19 percent crude protein, down regression equations and calculation of structural indi- 2.5 percent crude fat, 16.5 percent crude ﬁbre on dry matter ces values of Boran bulls aged 2, 3 and 4 years of age and basis, which was in accordance with the recommendation of reared at two locations of Southern Ethiopia. Hutcheson (2006) for Boran bulls reared for fattening purpose under Ethiopian condition. The concentrate was provided twice a day and the animals were group fed Materials and methods according to their respective age categories.

The study was conducted at Maki district of Oromia region. The BW and linear measurements were taken early in the The feedlots are situated at a distance of 133 km south of morning (Figure 3) after overnight fasting to ensure mini- Addis Ababa. The study was conducted on two feedlots, mum gut content in the bulls. The BW of the bulls was which varied significantly as far as the climate was concerned. assessed using a platform balance, which was calibrated The first feedlot is situated near Lake Ziway (7°59′19″N38° prior to being used. The linear measurements were taken 50′30″E) in an area devoid of any vegetation or tree breaks; by one of the authors during the course of the study and especially during the night when the temperature was low were carried out according to the methods suggested by and uncomfortable for the bulls. The second feedlot is situated Macjowski and Zieba (1982) and Tolenkhomba et al. 25 km west of the first feedlot. The climate of the second feed (2012). The traits included in the study were: HW, height lot was favourable for the livestock as there are sufficient trees at rump (RH), BL, heart girth (HG), head width (WFH), to serve as wind break thereby preventing the chilly winds neck circumference (NC), neck length (NL), chest width fl from the lake thus making the cattle comfortable during the (CW), rump length (RL), hip width (WH), ank width night (Figure 1). The study was conducted between January (FW) and BW of the bulls. The methodologies for the and April months of 2012, which is considered as the dry

Figure 1. Boran bulls reared in the second feed lot. Figure 2. Boran bulls of different age groups. 56 S. Banerjee et al.

Table 2. Calculations for structural indices.

Type of index Calculation

Height index Height at withers/body length × 1001 Rump length index Length of rump/body length × 1001 Over increase index Height at rump/height at withersX1001 Height slope Rump height–withers height2 Length index Body length/height at withers2 Width slope Hip width–chest width2 Body weight index Body weight/height at withers × 1003

1Alderson (1999). 2Szabolcs et al. (2007). 3Rohrer (1921).

method suggested by Röhrer and cited by Szabolcs et al. Figure 3. Restraining chute for taking the morphometric measurements. (2007).

assessment of the measurements are presented in Table 1. The data were analysed statistically using SPPS V 12 for Windows (SPSS, 2003). The data were analysed using Results and discussions descriptive statistics, the effect of feedlots on morphometrical traits and BW of Boran cattle from various age Morphometrical traits of Boran bulls aged 2 groups were compared using one-way analysis of variance years and reared in the two feedlots fi and were considered signi cant both at P < 0.01 and also The values pertaining to the various morphometrical traits P < 0.05. The prediction of BW taking into account differ- of Boran bulls aged 2 years and reared in two feed lots and ent morphometrical traits was conducted using step down of different age categories are presented in Table 3. The linear regression analysis and the accuracy was assessed 2 results from the table indicate that the average HW for using Radj values. The formulae used to calculate the vari- the bulls aged 2 years did not vary significantly between ous body indexes are presented in Table 2, height, rump the cattle reared in the two feedlots. The HW values as length, over increase was calculated according to the meth- assessed in the study are lower than those reported by odology suggested by Szabolcs et al. (2007), whereas the Alberro and Haile-Mariam (1982) for Boran cattle, Fasil values for height slope, length; width slope were assessed (2006) for Fogera cattle, Kugonza et al. (2011) for according to the methods suggested by Alderson (1999). Ankole cattle and Mwambene et al. (2012) for Nkasi cat- The weight index was assessed according to the tle. The values are comparable with those reported by Mason and Maule (1960) for Arado, Takele (2005) for Sheko cattle, Getinet et al. (2009) for Ogaden cattle, Table 1. Measurement techniques for the morphometrical traits. Fasil (2006) for Gojjam Highland Zebu cattle and Traits Measurements Nakachew (2009) for Abigar cattle. The results pertaining to the RH too did not vary significantly across the two Height at withers Distance between highest point of the wither to the locations. The values for the trait are comparable with ground fi Rump height Distance between the highest point of the rump to the gures reported by Nakachew (2009) for Abigar cattle. the ground The results of the study pertaining to BL also did not vary Body length Distance between point of the shoulder to the pin significantly across the two locations. The values for BL bone are comparable with those reported by Shiferaw (2006) Heart girth The circumference of the chest at the widest point for Kereyu cattle and also Nakachew (2009) for Abigar Width of fore The point between the two polls head cattle, Zulu (2008) for Angoni and Tonga cattle and also Neck Circumference of the neck at its widest point Cyprian et al. (2012) for Bunaji cattle. HG showed varia- circumference tions across the two locations, with higher values being Neck length The length between the atlas vertebrae and the first recorded among the cattle reared in the second feedlot, thoracic vertebrae which may be attributed to favourable climate prevailing Chest width Width of the thorax at the widest point just behind the elbow in the area. The HG values observed are in accordance Rump length The distance between the last thoracic vertebrae with the findings of Alberro and Haile-Mariam (1982) and just anterior of the hips for Boran cattle, Fasil (2006) for Fogera cattle and Width of the hips The widest point between the two pelvic bones Gojjam Highland Zebu, Getinet, Ayalew and Hegde fl Flank width The widest point of the rear ank (2009) for Ogaden, Kugonza et al. (2011) for Ankole cat- Body weight The weight of the bulls on a calibrated weighing bridge with an error margin of ±25 kg tle, Mwambene et al. (2012) for Nkasi cattle. The HG values as assessed were however lower than those reported Morphometrical traits of Boran bulls reared in Southern Ethiopia 57

Table 3. Morphometrical measurements of Boran bulls of three age groups and across two feedlots.

Traits Age-2 Age-3 Age-4

Feedlot-1 (N = 49) Feedlot-2 (N = 10) Feedlot-1 (N = 181) Feedlot-2 (N = 85) Feedlot-1 (N = 127) Feedlot-2 (N = 253)

HW (cm) 117.3 ± 3.4 117.2 ± 2.3 118.5 ± 5.0 119.1 ± 5.1 118.9 ± 6.2 119.6 ± 5.9 RH (cm) 118.7 ± 4.2 118.4 ± 4.0 120.1 ± 4.0 121.2 ± 4.9 120.6 ± 4.8 122.4 ± 8.1 BL (cm) 125.8 ± 7.9 123.0 ± 6.7 130.3 ± 11.1 129.6 ± 9.2 128.7 ± 16.8 129.5 ± 8.0 HG (cm) 151.4 ± 9.1 160.6 ± 8.0** 156.8 ± 9.8 164.3 ± 9** 158.0 ± 8.3 167.6 ± 8.0** WFH (cm) 19.5 ± 1.2 19.5 ± 2.3 19.7 ± 1.1 21.1 ± 2.19* 19.7 ± 1.2 21.5 ± 2.2* NC (cm) 72.4 ± 8.4 75.9 ± 7.4** 75.5 ± 8.5 79.1 ± 8.6** 77.9 ± 8.4 83.7 ± 7.5** NL (cm) 36.6 ± 2.8 37.0 ± 6.8 38.1 ± 2.8 40.8 ± 6.1** 38.1 ± 3.5 41.9 ± 5.6** CW (cm) 54.2 ± 5.3 57.1 ± 3.2* 55.2 ± 3.5 60.3 ± 3.7** 56.0 ± 3.2 61.1 ± 4.1** RL (cm) 45.5 ± 2.3* 42.5 ± 3.4 45.9 ± 2.6 46.3 ± 3.6 46.5 ± 2.7 46.3 ± 3.8 WH (cm) 29.9 ± 3.0 35.3 ± 4.0** 30.2 ± 2.7 36.8 ± 3.7** 30.2 ± 3.9 37.8 ± 3.2** FW (cm) 56.9 ± 5.2 60.2 ± 5.1** 57.7 ± 5.9 62.2 ± 5.8** 58.3 ± 7.1 62.9 ± 4.7** BW (kg) 248.0 ± 19.6 278.1 ± 23.8** 298.0 ± 43.9 326.4 ± 52.1** 316.5 ± 43.4 344.4 ± 39.5**

*P < 0.05, **P < 0.01. HW, height at withers; RH, rump height; BL, body length; HG, heart girth; WFH, width of forehead; NC, neck circumference; NL, neck length; CW, chest width; RL, rump length; WH, width of hips; FW, flank width; BW, body weight. by Zulu (2008) for Barotse, Tonga and Baila cattle. The The results pertaining to the morphometrical traits of WFH for the cattle reared in the two feedlots did not Boran cattle (reared in the two feedlots) aged 3 years are vary significantly; the values are similar to those reported also presented in Table 2. The study indicated that the by Szabolcs et al. (2007) for Charolais and Red Angus HW values did not vary between the two feed lots and cattle. NC values too varied across the two feedlots, with are lower than the values reported by Kugonza et al. (P < 0.01) higher values among the bulls raised in the sec- ( 2011) for Ankole cattle, Mwambene et al. (2012) for ond location, the values being similar to those reported by Nkasi cattle, Mwambene et al. (2012) for Sumbawanga Phanchung and Roden (1996) for Siri cattle from Bhutan. cattle Urban cattle, Mwambene et al. (2012)for The average NL indicated no significant difference among Sumbawanga cattle rural, Aamir et al. (2010) for Kenena the bulls reared across the two feedlots. The results as cattle and Abdelhadi and Babiker (2009) for Baggara cat- obtained are similar to those reported by Abdelhadi and tle. The values, however, being higher than those reported Babiker (2009) for Baggara cattle but higher values for by Mason and Maule (1960) for Arado, Takele (2005) for the trait were reported by Mason and Maule (1960) for Sheko cattle, Getinet et al. (2005) for Ogaden cattle, Fasil Arado cattle, Nakachew (2009) for Abigar cattle and (2006) for Gojjam Highland Zebu cattle, Nakachew (2009) Aamir et al. (2010) for Kenena cattle. The results pertain- for Abigar cattle and Kayastha et al. (2011) for Assam cat- ing to CW varied (P < 0.05) among the cattle reared in the tle. The values as assessed in the study find consonance two feedlots. The results pertaining to the RL were higher with the assessment of Alberro and Haile-Mariam (1982) (P < 0.05) among the bulls reared in the first feedlot. The for Boran cattle and Fasil (2006) for Fogera cattle. The RL values as recorded in the study were similar to those average values of RH also did not vary significantly across observed by Szabolcs et al. (2007) for Aberdeen Angus the two locations; the figures are comparable with the and Red Angus bulls. The results pertaining to WH, FW values reported by Nakachew (2009) for Abigar cattle. and BW too differed significantly (P < 0.01) between the However, the values are lower than those reported by two feedlots with higher values observed in the cattle Szabolcs et al. (2007) for Hungarian Simmental cattle, reared in the second location. The WH values as recorded Hereford cattle and for Aberdeen Angus cattle. The similar for the bulls raised in the second location are in accordance trend was also observed for BL with no differences among with the reports of Gilbert, Bailey and Shannon (1993) for the cattle reared in the two locations. The findings are Angus cattle; Suriya, Boonsaen and Innuruk (2011) for in close agreement with the observations of Milla, Kamphaengsaen (grass-fed) cattle. The FL values as Mahagoub and Bushara (2012) for Nilotic cattle, recorded in the study are lower than the values reported Mwambene et al. (2012) for Sumbawanga cattle Urban, by Khalafalla et al. (2011) and Abdelhadi, Babiker and Zulu (2008) for Angoni cattle, Zulu (2008) for Tonga cat- Kijora (2011) for Baggara and Zebu cattle from Sudan, tle, Cyprian et al. (2012) for Friesian cattle X Bunajii the lower values for the trait may be attributed to both gen- cattle. etic and non-genetic factors affecting the trait. The results However, the HG varied (P < 0.01) among the bulls reared pertaining to the BW of the bulls reared in the second site in the two locations with higher values observed among are in close accordance with the reports of Nakachew those raised in the second location. The values as assessed (2009) for Abigar cattle Mwambene et al. (2012) for for the first feed lot are in agreement with the observations Sumbawanga rural cattle and also for Baggara cattle of of Kugonza et al. (2011) for Ankole cattle and Mwambene Sudan as observed by Khalafalla et al. (2011). et al. (2012) for Nkasi cattle, while the results among the 58 S. Banerjee et al.

cattle reared the second feed lot find consonance with (2011) for Sudan Baggara and also by Abdelhadi, the observations of Alberro and Haile-Mariam (1982) for Babiker and Kijora (2011) for Zebu cattle from Sudan. Boran cattle, Shiferaw (2006) for Kereyu cattle, Mason The results pertaining to the average BW indicate that it and Maule (1960) for Arado, Milla, Mahagoub and varied (P < 0.01) among the bulls reared in the two loca- Bushara (2012) for Nilotic cattle, Takele (2005) for tions, the values being higher among the bulls raised in Sheko cattle, Getinet et al. (2005) for Ogaden cattle, the second feedlot. The values as observed among the Fasil (2006) for Gojjam Highland Zebu cattle and Fasil bulls reared in the first location find similarity with the (2006) for Fogera cattle. The findings indicate that the findings of Shiferaw (2006) for Kereyu, Milla, WFH too varied (P < 0.05) among the bulls reared in the Mahagoub and Bushara (2012) for Nilotic cattle, two feedlots with higher values observed among the Mwambene et al. (2012) for Sumbawanga cattle reared bulls reared in the second location. The findings are in in the urban areas. While BW of the cattle reared in the close accordance with the observations of Szabolcs et al. second location are similar to those reported by (2007) for Hereford cattle, Aberdeen Angus cattle, Red Phanchung and Roden (1996) for bulls of Siri breed cattle Angus cattle, Hungarian Simmental cattle, Lincoln from Bhutan. Red cattle, Shaver cattle and Charolais cattle. However, The results pertaining to morphometrical traits of bulls of the assessment of Gilbert, Bailey and Shannon (1993) Boran breed and aged 4 years and reared in the two feed for Hereford cattle and Angus cattle, Tolenkhomba et al. lots are also presented in Table 3. The results indicate (2012) for Manipuri cattle, Szabolcs et al. (2007) for that the HW were similar among the bulls reared in the Limousin cattle, and Blonde d’Aquitaine cattle are higher fi fi two locations, the gures are similar to the observations than those observed among the bulls reared in rst feedlot of Alberro and Haile-Mariam (1982) for Boran cattle, but lower than those reared in second location. The Shiferaw (2006) for Kereyu cattle, Mason and Maule fi ndings also indicate differences (P < 0.01) for NC (1960) for Arado cattle, Getinet, Ayalew and Hegde among the bulls reared in the two locations with higher (2009) for Ogaden cattl and Fasil (2006) for Fogera cattle. values observed among the cattle reared in the second The HW values are however higher than those reported by location, the values as assessed in the study are lower Fasil (2006) for Gojjam Highland Zebu, Kayastha et al. than the observations of Phanchung and Roden (1996) (2011) for Assam cattle and also Suriya, Boonsaen and for bulls of Siri cattle from Bhutan. Innuruk (2011) for Kamphaengsaen (grass-fed) cattle. The findings also indicate (P < 0.01) differences for NL The values for HW as assessed in this study are however between the bulls reared in the two locations. The values lower than the figures reported by Suriya, Boonsaen and being higher among the cattle reared in the second loca- Innuruk ( 2011) for Kamphaengsaen (feedlot-fed) cattle, tion. The observations (irrespective of the feedlots) are in Abdulmojeed et al. (2010) for Sokoto Gudali cattle, accordance with the results of Mason and Maule (1960); Cyprian et al. (2012) for Bunaji and also Angoni, Baila, Arado cattle, Abdelhadi and Babiker (2009); Baggara cat- Barotse and Tonga cattle by Zulu (2008). tle, Nakachew (2009); Abigar cattle and Aamir et al. The results of the study also indicated that the values of (2010); Kenena cattle. Similarly CW values too varied HR also did not vary among the bulls reared in the two among the cattle reared in the two locations, with higher locations, the findings are in consonance with the observa- values observed among the cattle reared in the second fi tions of Abdulmojeed et al. (2010) for Sokoto Gudali cat- location the values as assessed nd similarity with the tle, Shiferaw (2006) for Kereyu cattle and Nakachew reports of Getinet et al. (2005) for Ogaden cattle. The fi (2009) for Abigar cattle; higher values for the trait were RL did not show signi cant differences among the cattle reported by Szabolcs et al. (2007) for Hereford, reared in the two feedlots; the values are similar to Aberdeen Angus, Red Angus, Charolais and Limousin the results observed by Szabolcs et al. ( 2007) for Shaver, fi ’ breeds. The study shows that BL too was non-signi cant Charolais, Limousin and Blonde d Aquitaine cattle breeds. between locations, the values are in close accordance While the WH too varied (P < 0.01) between the bulls with the findings of Zulu (2008) for Baila, Tonga and reared in the two locations with higher values observed fi Angoni cattle, the values are however higher than those among the bulls reared in the second location, the ndings reported by Tolenkhomba et al. (2012) for Manipuri cattle, are in close agreement with the results of Gilbert, Bailey Rajendran et al. (2008) for Umblachery cattle, and are and Shannon (1993) for Hereford and Angus cattle and lower than those reported by Abdulmojeed et al. (2010) also Suriya, Boonsaen and Innuruk (2011) for for Sokoto Gudali and Bunaji cattle. The HG values indi- Kamphaengsaen (grass-fed) cattle; however, higher values cated differences (P < 0.01) between the bulls reared in the for the trait were reported for dairy cattle breeds, Holstein two locations. The HG values of the bulls raised in the first Friesian and Brown Swiss by Ozkaya and Bozkurt (2009). feedlot find similarity with the observations of Fasil (2006) The results pertaining to the average values of FW too find for Fogera cattle, Alberro and Haile-Mariam (1982)for differences among the cattle reared in the two feed lots Boran cattle, Aamir et al. (2010) for Kenena cattle and with higher value being observed among those being Mwambene et al. (2012) for Sumbawanga Urban cattle, reared in the second feedlot. The values of this study whereas those raised in the second location find similarity being lower than those reported by Khalafalla et al. with the observations of Mwambene et al. (2012) for Morphometrical traits of Boran bulls reared in Southern Ethiopia 59

Nkasi cattle, Zulu (2008) for Angoni, Baila and Tonga cat- The BW of the bulls indicate that those reared in the sec- tle. The WFH figures indicate the difference (P < 0.05) ond feedlot had higher (P < 0.01) values for the trait, those between the bulls reared in the two locations the results, reared in the first feedlot was higher than those reported for those raised in the first location being similar to by Shiferaw (2006), Abdelhadi and Babiker (2009), those reported by Szabolcs et al. (2007) for Red Angus, Khalafalla et al. (2011), Milla, Mahagoub and Bushara Aberdeen Angus and Charolais bulls, whereas those reared (2012) and Nakachew (2009). The results of a study by in the second feedlot are similar to those reported by Mwambene et al. (2012) indicated that the BW of Szabolcs et al. (2007) for Shaver, Hungarian Simmental Sumbawanga cattle reared in the rural and urban areas and Hereford cattle. was lower than those observed for the bulls reared in the fi The NC too varied between the bulls reared in the two rst feedlot, whereas the BW of Nkasi cattle was higher feedlots, the values being higher among the bulls raised than even the cattle reared in the second feedlot. in the second location similar values were reported by Phanchung and Roden (1996) for bulls of Siri breed. Effect of non-genetic factors influencing the The results of the study pertaining to the NL too indicate morphometrical traits and Body weight that there is variation for the trait among the bulls reared “ ” in the two feedlots, with higher values among the bulls The term adaptability refers to both genetic and physio- ’ reared in the second feedlot, the values as observed logical traits associated with an animal s response to both among the bulls from the first feedlot are in close accord- external and internal stimuli (Hafez, 1968). According to ance with the observation of Abdelhadi and Babiker Hafez (1968) the term physiological adaptation refers to (2009) for Baggara cattle, whereas those from the second the capacity of an individual to adjust to the external envir- feedlot are similar to those reported by Nakachew (2009) onment. All animals are genetically adapted to the environ- and Aamir et al. (2010) for Abigar and Kenena cattle, ment where they have evolved (Pianka, 2000), genetic respectively. The NL values as assessed by Mason and adaptation is the set of heritable traits that favour the sur- Maule (1960), Takele (2005) and Tolenkhomba et al. vival of the population in a particular environment. This is (2012) are lower than those assessed for the bulls reared all the more true for those animals, which are raised on nat- in the first feedlot, while higher values for the trait have ural grazing as they are subjected to several stress factors also been reported by Shiferaw (2006) for Kereyu cattle. when compared to those raised in the farm premises. The CW too showed difference (P < 0.01) among the The present study indicated that the skeletal traits (HW, bulls reared in the two feedlots with higher value observed RL, BL and RH) did not vary significantly between fee- among the cattle reared in the second feedlot, the results as dlots for a particular age category, which may be because assessed for the bulls maintained in the first feedlot is in skeletal traits are highly heritable and hence are least agreement with the results of Getinet et al. (2005) for affected by non-genetic factors the observations are in Ogaden bulls. The differences may be attributed to differ- accordance with the findings of Mwambene et al. (2012) ences in muscling among the bulls reared in the two fee- and Yunusa, Salako and Oladejo (2013). However, the dlots with lower muscling among the bulls reared in the other non-skeletal traits varied significantly across feedlots first feedlot. The effect of several non-genetic factors for a particular age category. The differences among the including the temperature prevailing in the region, which bulls due to feedlots are observed in the study may be may influence the muscling as a part of the ration is uti- attributed to the fact that the increase or decrease in live lized in maintenance of the body temperature, the observa- weight is a gross expression of many factors viz. organs, tions are in accordance with the reports of Young Manzi viscera, gut fill and also tissue mass, which are influenced et al. (2012). The results as assessed for RL indicated no by both genetic and non-genetic factors besides sex of the significant differences due to two feedlots. animal. As reported by Oladimeji et al. (1996), the climate fl The WH too indicated differences among the two feedlots of a region affects feed and water intake and also can in u- with higher (P < 0.01) values observed among the bulls ence the utilization of components within the feed con- fl reared in the second feedlot. The WH for those reared in sumed. It also in uences the heat produced by the the first location are similar to those reported by Gilbert, animal and thereby the net available energy and ultimately Bailey and Shannon (1993) for Angus and Hereford the productivity and phenotypic expression of the growing bulls, while those reared in the second feedlot are in con- livestock. sonance with the observations of Suriya, Boonsaen and The results also indicated that among the cattle aged 2 Innuruk (2011) for Kamphaengsaen (grass-fed) bulls, years the values for HG, NC, CW, BD, WH, FW and while the value was lower than those of Kamphaengsaen BW were significantly higher in cattle raised in the second bulls reared on feedlot. The study also indicates that the location, while RL was significantly higher among cattle FW of the bulls reared in the two feedlots too varies sign- reared in the first feedlot. The study also indicated that ificantly, with lower values observed among the bulls traits such as HG, WFH, NC, NL, CW, BD and BW values reared in the second feedlot, the findings are lower than were higher among bulls of 3 and 4 years of age raised in those reported by Khalafalla et al. (2011), and feedlot 2, whereas WH and FW for both the categories Abdelhadi, Babiker and Kijora (2011). were higher among the cattle raised in feedlot 1. It was 60 S. Banerjee et al.

also observed that the reared in the second feedlot were bet- The results of the regression studies also indicate that BW for ter managed. This may be attributed to adaptability of the the bulls aged 4 years and reared in the first feedlot was best cattle to the respective location; the findings pertaining to assessed using their HG values, while those reared in the sec- the effect of adaptability on variation in morpohometrical ond feedlot was estimated using RL, WFH and FW values traits have also been reported by Hafez (1968)and taken together, the importance of RL as an indicator of BW Bonsma (1980). The effect of climate prevailing in the is in accordance with the findings of Alderson (1999). The first location may have influenced the physiological factors; importance of NC, NL and WFH for assessment of BW of hence, the apparently cold climate influenced the muscle beef cattle has also been reported by Hammond, Mason and development thus affecting the BW but did not influence Robinson (1971). The importance of HG as a reliable tool to the skeletal development of the bulls, the findings being assess BW of cattle has also been reported (Branton and in consonance with the observations of Otuma (2004). Salisbury, 1946;Mullick,1950;ElKhidir,1980; Khogali, The findings are in consonance with the observations of 1999; Bagui and Valdez., 2007; Abdelhadi and Babiker several earlier studies in the tropics (Willamson and 2009). According to Afolayan et al., (2006), the morphometri- Pyne, 1978; Pagot, 1992; McDonald et al., 2002; Brito cal traits are moderately heritable and have a strong relationship et al., 2004). The muscle development of the cattle is sign- with the growth traits of livestock; however, the relationship ificantly affected by several non-genetic factors. This may vary from breed and also are influenced by several factors be the reason why there were differences among bulls such as age, type size, condition and degree of fattening of (within an age group) reared in the two feedlots; the obser- the animals (Heinrichs, et al., 1992; Yanar et al., 1995; vations are in accordance with the findings of Rothouge Ozkaya and Bozkurt, 2009). (2000). The difference in muscle development may also be attributed to the nutritional quality of the feed ingredi- ents and the way they were cultivated, harvested and post- Structural indices harvest methods of processing and storing, the observations The results pertaining to the structural indices of Boran are in accordance with the findings of Hunters and Buck bulls of different age groups and reared in the two feed (1992) and Bosman (1999). lots are presented in Table 5. The results indicate that The results of step-down regression equation as presented the height index values for the bulls aged 2 years varied in Table 4 indicate that NC was considered as the best pre- across the two feedlots with higher values being observed dictor among the cattle aged 2 years and reared in the first among the bulls reared in the second location. While the feedlot, whereas the assessment of BW was best assessed height index remained more or less similar among the using NL values for the bulls reared in second feedlot, this bulls aged 3 and 4 years and reared in both the feed was followed by considering the NL and RL values lots. The observations are in accordance with the findings together. The assessment of BW of the bulls aged 3 of Shackelford, Koohmaraie and Wheeler (1995)who years indicates that it was best assessed using HG values also reported that the skeletal maturity score increased which was followed by taking both HG and NC values till 2 years of age in cattle and thereafter remained together, whereas the BW of the bulls reared in the second more or less constant. The study also indicated that the location were best assessed taking into account the values RL index did not vary between the bulls across both of WFL and RL together. age categories and feedlots. This can be attributed to

Table 4. Step-down regression equations for assessment of body weight using morphometric measurements of Boran bulls of different age groups and reared in two different feedlots.

2 Age Location Radjusted Equations

Age-2 Location-1 0.25 161.480 + 1.194(NC) 0.29 111.112 + 1.153(NC) + 0.938(FW) Location-2 0.64 172.098 + 2.865(LN) 0.82 50.447 + 2.617 (LN) + 3.078 (RL) Age-3 Location-1 0.48 −190.406 + 3.114(HG) 0.55 −193.520 + 2.350(HG) + 1.626 (NC) 0.57 −19.630 + 2.253(HG) + 1.510(NC) − 1.263(HW) 0.58 12.000 + 2.310(HG) + 1.708(NC) − 1.231(HW) − 1.967 (WH) 0.60 −29.626 + 2.174(HG) + 1.598(NC) − 1.188(HW) − 2.122 (WH) + 1.859 (LN) Location-2 0.27 59.097 + 12.643(WFH) 0.46 −205.767 + 11.128(WFH) + 6.413 (RL) 0.51 −349.661 + 8.332(WFH) + 5.834 (RL) + 1.398(HG) Age-4 Location-1 0.15 −8.920 + 2.059(HG) Location-2 0.15 161.266 + 3.954 (RL) 0.21 86.179 + 3.514 (RL) + 4.446 (WFH) 0.24 −6.752 + 3.759 (RL) + 3.953 (WFH) + 1.464 (FW) 0.25 −41.410 + 3.513(RL) + 3.516 (WFH) + 1.418(FW) + .697 (NC) Morphometrical traits of Boran bulls reared in Southern Ethiopia 61

Table 5. Morphometrical indices of Boran bulls of different age groups and across the two feedlots.

Age (years) Index

Height Rump length Over increase Height Length Width Weight3 index1 index1 index1 slope2 index2 slope2

2 Feedlot-1 93.24 36.16 101.19 1.4 107.2 −24.3 211.40 Feedlot-2 95.28 34.55 101.02 1.2 104.9 −21.8 237.28 3 Feedlot-1 90.90 35.22 101.35 1.6 109.9 −25.0 251.40 Feedlot-2 91.89 35.72 101.76 2.1 108.8 −23.5 274.05 4 Feedlot-1 92.38 36.13 101.42 1.7 108.2 −25.8 266.19 Feedlot-2 92.35 35.75 102.34 2.8 108.3 −23.3 287.95

1Szabolcs et al. (2007). 2Alderson (1999). 3According to Röhrer as cited by Szabolcs et al. (2007). the character of the breed, the RL index values as Acknowledgements obtained in the study are higher than those reported by Szabolcs et al. (2007) for different taurine breeds of The authors acknowledge the assistance received from beef cattle. The results pertaining to the over increase Federal A.T.V.E.T co-coordinator, for the financial support. index too did not vary between the bulls reared in the The Dean and staff members of Dilla A.T.V.E.T College, two feed lots and also across the ages, the study indicated Head and staff members of School of Animal and Range that the frame of the Boran bulls sloped towards the anter- Sciences, Hawassa University. ior part of the body, the RH being higher than the height at withers, the observations are in accordance with the findings of Szabolcs et al. (2007). The results pertaining to the height slope as assessed in the study too collabo- References rated with the findings of the over increase index. The length index indicates that the Boran bulls are proportion- Aamir, H.M., Babiker, S.A., Youssif, G.M. & Hassan, Y.A. 2010. Phenotypic characterization of Sudanese Kenana cattle. ately longer bodied in comparison to their height. The Res. J. Anim. Vet. Sci.,5:43–47. width slope index values as obtained in the study indicate Abdelhadi, O.M.A., Babiker, S.A. 2009. Prediction of zebu cattle live that the Boran bulls are narrower at the hind quarters. The weight using live animal measurements. Livestock Research for weight index values indicate that the values were lower Rural Development. Volume 21 Article 133. Retrived August 2, (across all age groups) among the bulls reared in the 2013 from http://www.lrrd.org/lrrd21/8abde21133.htm fi rst feedlot, which may be attributed to the fact that the Abdelhadi, O.M.A., Babiker, S.A. & Kijora, C. 2011. Estimation of place was colder than the second feedlot and a part of zebu cattle carcass weight using body measurements. Livestock Res. the energy from the ration was utilized for body mainten- Rural Dev., 23: Article #12. Retrieved August 23, 2013 (available at ance rather than for BW development. The similar results http://www.lrrd.org/lrrd23/1/abde23012.htm), Addis Ababa, Ethiopia. were also reported by Young (1981) who reported that Abdulmojeed, Y., Kingsley Omogiade, I., Hadiza Salihu, H., under the similar levels of nutrition the cattle subjected Matthew, W. & Samuel, A. 2010. Multivariate analysis of phenotyp- to cold stress have poor productivity than those reared ic differentiation in Bunaji and Sokoto Gudali cattle. Acta Argiculturae Slovenica. 96(2): 75–80. under comfortable environment. Afolayan, R.A., Adeyinka, I.A. & Lakpini, C.A.M. 2006. The estimation of live weight from body measurements in Yankasa sheep. Czech J. Anim. Sci., 51: 343–348. Alberro, M. & Haile-Mariam, S. 1982. The indigenous cattle of Conclusion Ethiopia. Part I. FAO World Anim. Rev., 41: 2–10. Hence, it can be concluded from the study that Boran bulls Alderson, G.L.H. 1999. The development of a system of linear measure- if properly managed can be reared profitably for beef pur- ments to provide an assessment of type and function of beef cattle. Anim. Genet. Resour. Inf. FAO, 25: 45–55. pose. Linear measurements can be effectively used to assess BW of Boran bulls at all ages. Hence, morphometri- Bagui, N.J.G. & Valdez, C.A. 2007. Live weight estimation of locally raised adult purebred Brahman cattle using external body measure- cal measurements such as NL, NC, RL and WFH of Boran ments. Philip. J. Vet. Med., 44: 36–42. bulls can serve as indicators of selection of sires for higher BW. The results of the indices indicate that the Boran bulls Bonsma, J. 1980. Livestock production: a global approach. Cape Town, RSA, Tafelberg Pub. Ltd., p. 26–59. have a higher CW when compared to the hips, which may fi be a breed character. It can be concluded that Boran bulls Bosman, D.J. 1999. Selecting cattle for functional ef ciency. In M. M. Scholtz, L. Berg & D.J. Bosman, eds. Beef breeding in South can be selected for BW using the regression equations. Africa. Commemorating 40 years of beef cattle production testing, However, similar studies should be conducted using a 1959–1999, pp. 13–24. Irene, Agriculture Research Council Animal large population to further confirm the results. Improvement Inst. 62 S. Banerjee et al.

Bozkurt, Y. 2006. Prediction of body weight from body size measure- Heinrichs, A.J., Rogers, G.W. & Cooper, J.B. 1992. Predicting body ments in Brown Swiss feedlot cattle fed under small-scale farming weight and wither height in Holstein heifers using body measure- conditions. J. Appl. Anim. Res., 29: 29–32. ments. J. Dairy Sci., 75: 3576–3581. Branton, C. & Salisbury, G.W. 1946. The estimation of the live weight Hoffmann, I. 2010. Livestock biodiversity. Rev. Sci. Tech. Off. Int. Epiz., of bulls from heart girth. J. Dairy Sci., 29: 141–143. 29(1): 73–86. Brito, A.F.C., Silva, A.D.E.F., Barbosa, R.T. & Kastilec, J.P. 2004. Hunters, R.A. & Buck, N. 1992. Nutritional and climatic limits of beef Testicular thermoregulation in Bos indicus bulls: relationship with production in the tropics. In R. Jarrige & C. Beranger, eds. World ani- scrotal, testicular vascular cone and testicular morphology and its mal science, C5. Beef cattle production, pp. 379–387. The effect on semen quality and sperm production. Theriogenology, 61: Netherlands, Elsevier Science. 511–528. Hutcheson, D. 2006. Feeding to produce export quality Ethiopian beef CSA. 2012. Central Statistics Agency of Ethiopia. Report on Livestock requirements and recommendations. Ethiopia Sanitary and Phyto sani- and Livestock Characteristics, Agricultural Sample Survey 2010–11 tary standards and livestock and meat marketing program (2003 E.C). Statistical Bulletin No.532, Vol. II. March 2012. Addis (SPS-LMM), p. 18. Ababa Ethiopia. Kamalzadeh, A., Koops, W.J. & van Bruchem, J. 1998. Feed quality Cyprian, A., Gerald Nwachi, A.K.P.A., Pano, B.P., Finangwai Hosea restriction and compensatory growth in growing sheep: modeling Istifanus & Danbab, A.B. 2012. Comparative evaluation of linear changes in body dimensions. Livestock Prod. Sci., 53: 57–67. Udder and body conformation traits of Bunaji and Friesian X Bunaji cows. World J. Life Sci. Med. Res., 2(4): 134–140. Kayastha, R.B., Zaman, G., Goswami, R.N. & Haque, A. 2011. Physical and morphometric characterization of indigenous cattle of El Khidir, O.A. 1980. A note on prediction of live weight of growing Assam. Open Vet. J.,1:7–9. Kenana heifers from linear body measurements. Sudan J. Vet. Sci. Anim. Husbandry, 21(2): 102–104. Khalafalla, I.E.E., Atta, M., Eltahir, I.E. & Mohammed, A.M. 2011. Effect of body weight on slaughtering performance and carcass mea- ’ ESPSLMMP. 2010. Focus on Ethiopia s Meat and Live Animal Export, surements of Sudan Baggara bulls. Livestock Res. Rural Dev., 23: Ethiopia Sanitary and Phytosanitary Standards and Livestock & Meat Article #47. Retrieved August 23, 2013 (available at http://www. Marketing Program. Trade Bulletin Issue 1, Addis Ababa. lrrd.org/lrrd23/3/khal23047.htm). ESPSLMMP. 2011. Focus on Ethiopia’s Meat and Live Animal Export, Khogali, A.M. 1999. The effect of different dietary levels on perform- Ethiopia Sanitary and Phytosanitary Standards and Livestock & Meat ance, carcass characteristics and meat quality of the Baggara cattle. Marketing Program. Trade Bulletin 5, July. (Ph.D. thesis), University of Khartoum, Sudan. Fasil, G. 2006. On-farm phenotypic characterization of cattle genetic Kugonza, D.R., Nabasirye, M., Mpairwe, D., Hanotte, O. & Okeyo, resources and their production systems in Awi, East and West A.M. 2011. Productivity and morphology of Ankole cattle in three Gojjam zones of Amhara region. M.Sc. thesis, submitted to School livestock production systems in Uganda. Anim. Genet. Resour. Inf., of Graduate Studies Haramaya University Ethiopia. 48: 13–22. Felleke, G. & Geda, G. 2001. The Ethiopian dairy development policy. Little, P.D., Smith, K., Cellarius, B.A., Coppock, C.B. & Barrett, C. In: A draft policy document, Addis Ababa, Ethiopia: Ministry of 2001. Avoiding disaster: diversification and risk management Agriculture/AFRDRD/AFRDT. Food and Agriculture Organization/ among East African herders. Dev. Change, 32(3): 401–433. SSF. Fouire, P.J., Neser, F.W.C., Oliver, J.J. & Van der Westhizen, C. Livestock Marketing Authority (LMA). 2001. Study on causes of 2002. Relationship between production performance, visual appraisal cross-border illegal trades in South, Southwest and Eastern and body measurements of young Dorper rams. South African Ethiopia. Addis Ababa, Ethiopia, Market Research and Promotion J. Anim. Sci., 32: 256–262. Department. Gatesy, J. & Arctander, P. 2000. Hidden morphological support for the Macjowski, J. & Zieba, J. 1982. Genetics and animal breeding part-A. phylogenetic placement of Pseudoryx ngetinhensis with bovine Biological and genetic foundations of animal breeding. Warszawa, fi fi – bovids: a combined analysis of gross anatomical evidence and DNA Elisever Scienti c PWN. Polish Scienti c Publishers, pp. 30 37. fi – sequences from ve genes. Syst. Biol., 49(3): 515 538. Manzi, M., Junga, J.O., Ebong, C., & Mosi, R. 2012. Factors affecting Getinet, M., Ayalew, W. & Hegde, B.P. 2009. Growth and reproductive pre and post-weaning growth of six cattle breed groups at Songa performance of Ogaden cattle at Haramaya University, Ethiopia. Research station in Rwanda. Livestock Research for Rural Ethiopian J. Anim. Prod., 9(1): 13–38. Development. Volume 24, Article #68. Retrieved March 22, 2014, from http://www.lrrd.org/lrrd24/4/manz24068.htm. Gilbert, H.R. & Gregory, P.W. 1952. Some features of growth and development of Hereford cattle. J. Anim. Sci., 11: 3–16. Mason, I.L. & Maule, J.P. 1960. The indigenous livestock of Eastern and Northern Africa technical communication 14. Farnham Royal, Gilbert, R.P., Bailey, D.R.C. & Shannon, N.H. 1993. Linear body UK, CAB (Commonwealth Agricultural Bureaux). measurements of cattle before and after 20 years of selection for postweaning gain when fed two different diets. J. Anim. Sci., 71: Mc Donald, P., Edwards, R.A., Greenhalgh, J.F.D. & Morgan, C.A. 1712–1720. 2002. Animal nutrition, 6th edition. Edinburg, UK, Pearson Edu Ltd, pp. 5–8, 352–370. Goe, M.R., Alldredge, J.R. & Light, D. 2001. Use of heart girth to predict body weight of working oxen in Ethiopian highlands. Livestock Milla, A.P., Mahagoub, M.M.M. & Bushara, I. 2012. Estimation of Prod. Sci., 69: 187–195. live body weight from heart girth, body length and condition score in Nilotic cattle – Southern Sudan. J. Anim. Sci. Adv., 2(5): 453–457. Hafez, E.S.E. 1968. Principles of animal adaptation. In E.S.E. Hafez, ed. Adaptation of domestic animals, pp. 3–15. Philadelphia, Lea and Msangi, B.S.J., Bryant, M.J., Kavana, Y., Msanga, N. & Kizima, J.B. Fabiger. 1999. Body measurements as a management tool for crossbred dairy cattle at a smallholder farm Ksition. Proc. Tanzania Soc. Anim. Sci., Hammond, J. Jr., Mason, I.L. & Robinson, T.J. 1971. Hammond’s 26: 168–175. farm animals, 4th edition. London, Edward Arnold, p. 283. Morphometrical traits of Boran bulls reared in Southern Ethiopia 63

Mullick, D.N. 1950. The estimation of the weight of cattle and buffalo Shackelford, S.D., Koohmaraie, M. & Wheeler, T.L. 1995. Effects of from heart girth measurements. Indian J. Anim. Nutr.,3:52–58. slaughter age on meat tenderness and USDA carcass maturity scores of beef females. J. Anim. Sci., 73: 3304–3309. Mwacharo, J.M., Okeyo, A.M., Kamande, G.K. & Rege, J.E.O. 2006. The small East African shorthorn zebu cows in Kenya. Linear body Shiferaw, G. 2006. In-situ phenotypic Characterization of Kereyu cattle measurements. Trop. Anim. Health Prod., 38: 65–74. type in Fentale district of Oromya region. M.Sc. thesis submitted to Mwambene, P.L., Katule, A.M., Chenyambuga, S.W. & School of Graduate Studies Haramaya University Ethiopia. Mwakilembe, P.A.A. 2012. Fipa cattle in the southwestern highlands Slippers, S.C., Letty, B.A. & De Villerrs, J.F. 2000. Prediction of body of Tanzania: morphometric and physical characteristics. Anim. Genet. weight of Nguni goats. S. Afr. J. Anim. Sci., 30 (Suppl. 1): 127–128. Resour. Inf., 51: 15–29. Solomon, D. 2001. Cattle population Dynamics in the Southern Ethiopian Nakachew, M. 2009. Characterization of Abigar (Nuer) cattle breed at Rangelands. Utah State University Pastoral Risk Management Project. its production environment in Gambella Regional State, Ethiopia.A Research brief 01-02-PARIMA. thesis Submitted to the School of Graduate Studies Hawassa University, p. 159. SPSS. 2003. Statistical package for social sciences. SPSS 12.0 for windows. Chicago, IL, SPSS Inc. NEPAD–CAADP (New Partnership for Africa’s Development – Comprehensive Africa Agriculture Development Programme). Suriya, S., Boonsaen, P. & Innuruk, P. 2011. Body measurements of 2005. Ethiopia: investment Project Proﬁle “Live Animal and Meat male Kamphaengsaen beef cattle as parameters for estimation of Export”–Preliminary Options Outline, p. 3. live weight. Kasetsart J. (Nat. Sci.), 45: 428–434. Oladimeji, B.S., Osinowo, O.A., Alawa, J.P. & Hambolu, J.O. 1996. Szabolcs, B., Nagy, B., Nagy, L. & Kiss, B. 2007. Comparison of body Estimation of average values for pulse rate, respiratory rate and rectal measurements of beef cows of different breeds Arch. Tierz., temperature and development of a heat stress index for adult Yankassa Dummerstorf, 50(4): 363–373. sheep. Bull. Anim. Health Prod., 44: 105–107. Takele, T. 2005. On farm phenotypic Characterization of Sheko breed of Otuma, M.O. 2004. Inﬂuence of breeding designs and seasonal changes cattle and their habitat in Bench Maji. M.Sc. thesis submitted to on growth of Nigerian goats and their crosses. Trop. J. Anim. Sci.,7 School of Graduate Studies Alemaya University, Ethiopia. (2): 87–93. Tolenkhomba, T.C., Konsam, D.S., Shyamsana Singh, N., Prava, M., Ozkaya, S. & Bozkurt, Y. 2009. The accuracy of prediction of body Damodor Singh, Y., Ayub Ali, M. & Motina, E. 2012. Factor ana- weight from body measurements in beef cattle. Arch. Tierz, 52: lysis of body measurements of local cows of Manipur, India. Int. – 371 377. Multidisciplin. Res. J., 2(2): 77–82. Pagot, J. 1992. Animal production in the tropics and sub-tropics. Willamson, G. & Pyne, W.J.A. 1978. An introduction of animal hus- London, UK, McMillan Press Ltd., pp. 1–86, 232–252. bandry in the tropics, 3rd edition. USA, Longman Inc., p. 2. Phanchung & Roden, J.A. 1996. Characterisation of the Siri breed and Willeke, H. & Dürsch, T. 2002. Prediction of the body weight of the Mithun cross Siri in Bhutan. Anim. Genet. Resour. Inf., 20: 27–34. Simmental heifers using heart girth measurements. Arch. Tierzucht, Pianka, E.R. 2000. Evolutionary ecology, 6th edition. San Francisco, 45(1): 23–28. Addison-Wesley-Longman. Yanar, M., Tüzemen, N., Özhan, M., Aydın, R. & Ug˘ur, F. 1995. Rajendran, R., Raja, T.V., Thiruvenkadan, A.K., Mahalinga Nainar, Prediction of body weights from body measurements in Brown A. & Thangaraju, P. 2008. Morphobiometrical characteristics and Swiss cattle. Turk. J. Vet. Anim. Sci., 19: 357–360. management of Umblachery cattle from coastal region of Tamilnadu, India. Livestock Res. Rural Dev., 20: Article #40. Young, B.A. 1981. Cold stress as it affects animal production. J. Anim. – Retrieved August 23, 2013 (available at http://www.lrrd.org/lrrd20/3/ Sci., 52: 154 163. raje20040.htm). Yunusa, A.J., Salako, A.E. & Oladejo, O.A. 2013. Principal component Rohrer, F. 1921. “Der Index der Körperfülle als Maß des analysis of the morphostructure of Uda and Balami sheep of Nigeria. Ernährungszustandes” [The index of corpulence as measure of nutri- Int. Res. J. Agric. Sci., 1(3): 45–51. tional state]. Münchener medizinische Wochenschrift (in German), Zechner, P., Zohman, F., Sölkner, J., Bodo, I., Habe, F. & Marti, E. – 68: 580 582. 2001. Morphological description of the Lipizzan horse population. Rothouge, A. 2000. New Ecological perceptions of arid rangelands. Livestock Prod. Sci., 69(2): 163–177. Agricola, 11: 49–56. Zulu, D.N. 2008. Genetic Characterization of Zambian native cattle Salako, A.E. 2006. Application of morphological indices in the assess- breeds. M.Sc. thesis, submitted to Virginia Polytechnic Institute and ment of type and function in sheep. Int. J. Morphol., 24(1): 13–18. State University, p. 68. Animal Genetic Resources, 2014, 54, 65–72. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000125

Analysis of the body structure of Djallonke sheep using a multideterminant approach

Peter T. Birteeb1, Sunday O. Peters2 and Michael O. Ozoje3 1Department of Animal Science, University for Development Studies, P. O. Box TL 1882, Tamale, Ghana; 2Department of Animal Science, Berry College, Mount Berry, GA 30149, USA; 3Department of Animal Breeding and Genetics, University of Agriculture, Abeokuta, Nigeria

Summary This study aimed at using a multivariate approach to describe the body structure of Djallonke sheep in northern Ghana and to determine which approach explains better the variation in body composition. Live weight (LW) and linear body measurements including heart girth (HG), neck girth (NG), chest depth (CD), height at withers (HW), rump height (RH), body length (BL) and pin-bone width (PBW) were obtained from 172 sheep aged between two and three years. The fixed effects of sex and age were tested using the general linear model (GLM) while the Nearest Neighbor method of Hierarchical Cluster Analysis was used to group body traits into clusters. Principal Component Factor Analysis was used to describe the variation in body traits where extracted factors were varimax rotated to enhance interpretability. The analysis of variance revealed significant (P < 0.01) differences in the morphological traits of the two sexes with higher values recorded for the male in all traits examined except in PBW, which was insignificant (P > 0.05). Age had no significant influence (P > 0.05) on the body traits. The sheep weighed 26.92 ± 0.89 kg averagely and had averages of other body measurements to be: 71.74 ± 1.23, 40.52 ± 0.79, 27.73 ± 0.52, 60.72 ± 0.86, 59.61 ± 0.87, 58.87 ± 1.06 and 12.81 ± 0.23 cm for HG, NG, CD, HW, RH, BL and PBW, respectively. The product moments of correlation were positive and significant (P < 0.05, 0.01; r = 0.18–0.99) for all pairs of traits. The body traits were categorized mainly into two clusters with the first cluster comprising the HG, HW, RH and BL while NG, CD and PBW formed the second cluster. The grouping of the traits was slightly different in Factor analysis where two underlying principal components (PC) were extracted to discern the variance structure of the Djallonke sheep. The first principal component which consisted of CD, HW, RH, BL and PBW explained 61.26 percent and the second, 12.92 percent thereby giving a maximum of 74.17 percent generalized variance in body measurements. The traits loaded on the first principal component are closely associated with bone growth hence describing the general body size conformation while the traits (HG and NG) on the second component seem to describe only the thoracic region. It can be concluded that both the Hierarchical Cluster analysis and the Factor analysis grouped body traits similarly but the later is to be recommended because of the additional ability of indicating the amount of variation explained by the developed factors.

Keywords: body measurements, cluster, generalized variance, genetic resources, multivariate, West African dwarf

Résumé Cette étude a adopté une approche multivariée pour décrire la structure corporelle des moutons Djallonke du Nord du Ghana et pour déterminer quelle méthode permet d’expliquer le mieux la variation observée dans la conformation corporelle. Le poids vif (PV) et une série de mesures corporelles linéaires, à savoir le périmètre thoracique (PT), le tour du cou (TC), la profondeur de la poitrine (PP), la hauteur au garrot (HG), la hauteur à la croupe (HC), la longueur du corps (LC) et la largeur de la tubérosité ischiatique (LTI), furent obtenus sur un total de 172 moutons âgés de 2 à 3 ans. Les effets fixes du sexe et de l’âge ont été analysés suivant un modèle GLM alors que la méthode du plus proche voisin de l’analyse hiérarchique a été utilisée pour regrouper les traits corporels dans des grappes. L’analyse factorielle en composantes principales, avec rotation selon la méthode Varimax des facteurs identifiés (ceci a été fait pour améliorer l’interprétabilité), a été employée pour décrire la variation observée dans les caractères corporels. L’analyse de la variance a décelé des différences significatives (P < 0.01), entre les deux sexes, dans les caractères morphologiques, avec les valeurs obtenues chez les mâles étant les plus élevées pour tous les traits sauf pour la LTI, pour laquelle pas de différences significatives ont été observées (P > 0.05). L’âge n’a pas eu d’effet significatif (P > 0.05) sur les caractères corporels. Les moutons ont pesé en moyenne 26.92 ± 0.89 kg et les valeurs moyennes des mesures corporelles ont été: 71.74 ± 1.23, 40.52 ± 0.79, 27.73 ± 0.52, 60.72 ± 0.86, 59.61 ± 0.87, 58.87 ± 1.06 et 12.81 ± 0.23 cm pour PT, TC, PP, HG, HC, LC et LTI, respectivement. Les coefficients de corrélation ont été positifs et sign- ificatifs (P < 0.05, 0.01; r = 0.18–0.99) pour toutes les paires de caractères. Les caractères corporels ont été regroupés principalement dans deux grappes. La première grappe a compris les mesures PT, HG, HC et LC alors que TC, PP et LTI ont constitué la deuxième grappe. Le regroupement des caractères a été légèrement différent avec l’analyse factorielle qui a extrait deux composantes principales sous-jacentes pour percevoir la structure de la variance des moutons Djallonke. La première composante principale, formée par PP, HG, HC, LC et LTI, a expliqué le 61.26 pour cent et la deuxième le 12.92 pour cent, donnant ainsi un maximum de 74.17 pour cent de variance généralisée dans les mesures corporelles. Les traits compris dans la première composante principale sont étroitement liés à la croissance des os, comme quoi ils peuvent être utilisés pour décrire la conformation générale du corps. Par contre, les traits de la deuxième composante principale (PT et TC) semblent décrire seulement la région thoracique. Il peut être conclu que l’analyse

Correspondence to: P.T. Birteeb, Department of Animal Science, University for Development Studies, Tamale, Ghana. email: [email protected]; tel.: ( + 233) 24 981 5083, ( + 233) 20 827 8578

65 66 P.T. Birteeb et al.

hiérarchique de regroupement et l’analyse factorielle ont classé les caractères corporels de manière semblable. Néanmoins, la deuxième méthode est retenue pour la capacité additionnelle à indiquer la quantité de variation expliquée par les facteurs révélés.

Mots-clés: regroupement, mesures corporelles, variance généralisée, ressources génétiques, multivariée, moutons Nains d’Afrique Occidentale

Resumen En este estudio se adoptó un enfoque multivariante para describir la estructura corporal de las ovejas Djallonke del Norte de Ghana y para determinar qué método permite explicar mejor la variación en la conformación corporal. Se tomaron el peso vivo (PV) y una serie de medidas corporales lineales (la circunferencia torácica CT, la circunferencia del cuello CC, la profundidad del pecho PP, la altura a la cruz AC, la altura a la grupa AG, la longitud del cuerpo LC y la anchura de la tuberosidad isquiática ATI) en una muestra de 172 cabezas de ganado ovino, de edades comprendidas entre los dos y los tres años. Se analizaron los efectos fijos del sexo y la edad usando un modelo GLM mientras que se empleó el método del vecino más cercano del análisis jerárquico para agrupar los parámetros corporales en conglomerados. Se recurrió al análisis factorial de componentes principales para describir la variación en los rasgos corporales, siendo rotados, según el método Varimax, los factores identificados, lo cual se hizo con el fin de mejorar la interpretabilidad. El análisis de la varianza sacó a la luz diferencias significativas (P < 0.01), entre los dos sexos, en los rasgos morfológicos, siendo mayores en los machos los valores de todos los parámetros examinados, con la excepción de la ATI, para la cual no se dieron diferencias significativas (P > 0.05). La edad no tuvo un efecto significativo (P > 0.05) sobre los parámetros corporales. Los animales pesaron de media 26.92 ± 0.89 kg y los valores medios de las medidas corporales fueron: 71.74 ± 1.23, 40.52 ± 0.79, 27.73 ± 0.52, 60.72 ± 0.86, 59.61 ± 0.87, 58.87 ± 1.06 y 12.81 ± 0.23 cm para CT, CC, PP, AC, AG, LC y ATI, respectivamente. Los coeficientes de correlación fueron positivos y significativos (P < 0.05, 0.01; r = 0.18–0.99) para todos los pares de caracteres. Los parámetros corporales fueron agrupados principalmente en dos conglomerados. El primer conglomerado comprendió las medidas CT, AC, AG y LC mientras que CC, PP y ATI formaron el segundo conglomerado. La agrupación de los parámetros corporales fue ligeramente diferente en el análisis factorial en que se extrajeron dos componentes principales subyacentes para discernir la estructura de la varianza del ganado ovino Djallonke. El primer componente principal, formado por PP, AC, AG, LC y ATI, explicó el 61.26 por ciento y el segundo el 12.92 por ciento, dando así un máximo del 74.17 por ciento de varianza generalizada en las medidas corporales. Los parámetros comprendidos en el primer componente principal están estrechamente relacionados con el crecimiento de los huesos, de ahí que sirvan para describir la conformación general del cuerpo, mientras que las medidas CT y CC, del segundo componente, aparentemente sólo describen la región torácica. Se puede concluir que el análisis jerárquico de conglomerados y el análisis factorial agruparon los parámetros corporales de un modo similar. No obstante, se recomienda el segundo análisis por la capacidad adicional de indicar la cantidad de variabilidad explicada por los factores identificados.

Palabras clave: conglomerado, medidas corporales, varianza generalizada, recursos genéticos, multivariante, oveja Enana de África Occidental

Submitted 2 December 2013; accepted 5 March 2014

Introduction (AnGRs) in developing countries are being eroded through Sheep have long been recognized as major contributor to the rapid transformation of the agricultural systems in protein supply and family income in most developing which the main cause of the loss of indigenous AnGRs is countries. The choice of breed in a locality is inﬂuenced the indiscriminate introduction of exotic genetic resources by the production objectives and the environmental condi- before proper characterization, utilization and conservation tions, with increased productivity as the underlying object- of indigenous genetic resources (Yakubu, Salako and ive. There is a considerable potential for increased sheep Abdullah, 2011). Nevertheless, AnGRs conservation is production, if proper management is employed (Birteeb receiving keen attention of late as livestock researchers et al., 2012), and this demands a look at their genetic explore classical multivariate methods to study and charac- resources. At present, many countries are losing their genet- terize livestock populations genetically and phenotypically ic resources which may have lasting effects on food security especially through the use of quantitative information, par- and sustainable development especially in the light of global ticularly body measurements. warming changes (Hoffman, 2010). This is why the Food Linear body measurements have become very useful in live- and Agriculture Organization of the UN deemed it neces- stock research, as alternative body measurements and indices sary to promote the conservation of farm animal diversity estimated from various combinations of conventional and with the view that humankind may need to keep this speciﬁc non-conventional body parameters not only provide superior genetic biodiversity to face future (unknown) challenges guide to weight but also served as indicators of type and such as changes in demand for livestock products, spread function in domestic animals (Salako, 2006a). Body weight of new diseases, reducing environmental impact and climate is a major characteristic associated with animal breed types. change (FAO, 2011). In general, Animal Genetic Resources Precise measurement of weight is by using weighing scales Djallonke sheep body structure analysis 67

or bridges. However, proper weighing scales are neither read- Livestock Breeding Station, were used for this study. ily available to nor affordable by smallholder livestock farm- Pong-Tamale is located at latitude 9°40″N and longitude ers, but even if they were, it would be inconvenient and a huge 0°49″W and is about 32.7 km north of Tamale, the task to carry and assemble those weighing scales each time to Northern Regional capital in Ghana (Encarta, 2009). The weigh animals especially during management practices. sheep were managed semi-intensively. They were housed According to Traoré et al. (2008), the definition of the in properly constructed pens throughout the night and breed as the operational unit for the assessment of live- sometimes during the day when there was the need to stock diversity all over the world, necessitates the charac- restrict their movement. Feed and water were provided terization of local domestic animal populations in for the animals ad libitum throughout the year. developing countries. The characterization of African Conventional disease and pest control regimes were small ruminant populations through the use of classical practised. multivariate methods in analyzing morphological traits to The Breeding Station practises open nucleus breeding pro- assess variations within and between populations will gramme, where occasionally, outstanding sires from differ- play a major role in the maintenance of autochthonous ent farms are selected and brought to the station for genetic resources (Birteeb et al., 2012). improvement. Among the progenies, very superior rams Factor analysis is a multivariate technique for examining are retained on the station for breeding while all yearling the interrelationships among a set of variables that are cor- rams are supplied to farmers. The breeding sires and/or related. It deals with the reduction of a set of observable rams (which were all >2 years) included in this study, variables in terms of a small number of latent factors were second and third generation progenies of sires that (Rencher, 2002). Principal component factor analysis has were selected from different smallholder farms for been effectively used to study cattle (Yakubu, Ogah and improvement at the station. Hence the animals under Idahor, 2009a; Pundir et al., 2011), sheep (Riva et al., study were quite varied in terms of pedigree. 2004; Salako, 2006b; Yakubu, Salako and Abdullah, 2011; Birteeb et al., 2012), and even poultry (Ogah, Alaga and Momoh, 2009; Yakubu, Kuje and Okpeku, Data collection 2009b; Ogah, 2011). Cluster analysis is another multivariate technique in which a search is made for patterns in a The sex and age of each sheep were noted. Ages of ani- fi data set by grouping the (multivariate) observations into mals were identi ed using their birth records and categor- ≥ clusters, with the goal of finding an optimal grouping for ized into two groups (2 years and 3 years). Eight which the observations or objects within each cluster are biometric traits which include live weight (LW), heart similar, but the clusters are dissimilar to each other girth (HG), neck girth (NG), chest depth (CD), height at (Rencher, 2002). In grouping the observations, the similar- withers (HW), rump height (RH), body length (BL) and ities between all pairs of observations are considered. pin-bone width (PBW) were measured on each animal. Cluster analysis is similar to factor analysis in the follow- LW was measured with a hanging scale while all linear ing ways: (1) neither the number of groups nor the groups body dimensions were measured using a measuring tape. fi themselves are known in advance; (2) when clustering The traits were de ned and measured in accordance to variables, the similarity is based on a correlation (matrix). Birteeb et al. (2012) as follows: Among the existing sheep breeds, the Djallonke (West HG: The circumference around the chest just behind the African dwarf) sheep is known to be trypano-tolerant front legs and withers. hence has become a dependable asset to the local farmers NG: This is the circumference around the neck near the in West Africa (Wafula et al., 2005). It is well adapted to withers. and produces under the local environmental conditions and CD: The distance from the backbone at the shoulder to the so contributes significantly to the African small ruminant brisket between the front legs genetic resources. Despite the availability of this breed HW: The distance from the surface of a platform on which of sheep, detail information on its characteristics is limited. an animal stands, to the withers of the animal. Hence this study was undertaken to provide a detailed RH: This is the distance from the surface of a platform to description of the Djallonke sheep breed as well as deter- the rump. mine the multivariate approach that explains better the BL: The distance from the base of the tail to the base of the variation in its body composition. neck. PBW: It is the distance between the outer edges of the major hip bones on the right and left sides. Materials and methods Management of study animals Statistical analysis A total of 172 (28 males and 144 females) Djallonke (West The fixed effects of sex and age on the biometric traits African Dwarf) sheep, reared at the Pong-Tamale were tested using general linear model (GLM). The 68 P.T. Birteeb et al.

model was defined as: higher at the withers by 1.15 cm (Table 3). Also, the sheep were significantly taller than longer. Yij = m + Si + Aj + (SA) + 1ij, ij The cluster analysis revealed the formation of two distinct fi where Yij is the individual biometric trait; μ the overall clusters from the linear body traits of which the rst cluster fi mean; Si fixed effect of the ith sex (i = ram, ewe); Aj the comprised four traits (Figure 1). The rst agglomeration fixed effect of the jth age ( j = 2 years old, ≥3 years old); occurred between HW and RH, which then combined fi fi (SA)ij the interaction effect of ith sex and jth age; εij the with BL and nally HG to form the rst cluster. The sec- random error associated with each biometric trait [ε N ond cluster resulted from the agglomeration of NG, CD (0, δ)]. and PBW. A paired sample t-test was used to compare the heights Pearson’s correlation coefficients between all pairs of the (wither and rump) and BL. The Nearest Neighbor method various body traits are shown in Table 4. Even though of Hierarchical Cluster Analysis was used to group the LW was moderately (0.52–0.70) correlated with all the various body traits into clusters. Pearson’s coefficients of body conformation traits, the highest coefficient (0.99) correlation (r), among the various traits, were estimated, was obtained between RH and HW. All correlations which were then used for the principal component factor between pairs of traits were highly significant (P < 0.01) analysis. Anti-image correlations, Kaiser–Meyer–Olkin except that between NG and PBW, which was only sign- (KMO) measures of sampling adequacy and Bartlett’s ificant at (P < 0.05). Test of Sphericity were computed to test the validity of The KMO measure of sampling adequacy, which was the factor analysis of the data set. The factor matrix was signi ficantly in the middle (0.791) supported the use of subjected to varimax rotation to enhance interpretability the correlation matrix of the data set for factor analysis. of the extracted factors. All analyses were done using the Bartlett’s test of sphericity (chi-square = 1118.89; P < appropriate programmes in SPSS version 17 (SPSS, 2008). 0.001), coupled with the high communalities (0.639– 0.848), which represent the explained variance, and the determinant (0.001) obtained from the correlation matrix, all provided support for the validity and reliability of the Results principal component factor analysis. The analysis of variance of LW and linear body measure- The estimated factor loadings extracted by factor analysis, ments of Djallonke sheep are presented in Tables 1 and 2. eigenvalues and variation explained by each factor, are The biometric traits were significantly (P < 0.01) presented in Table 5. After a varimax (orthogonal) rotation influenced by sex, with the rams (males) being generally of the component matrix, two principal component factors superior to the ewes (females) in all the biometric traits were extracted. The variance of a variable was partitioned under consideration except for the PBW where the females into a common portion ‘communality’ shared with some or were larger (P > 0.05) in absolute terms (Table 2). all of the other variables, which showed that 63.9–84.8 However, age did not have any significant influence on percent of the variation in conformation traits were brought the LW and linear body traits as all the body traits of the about by the underlying factors (PC’s). The first principal 2–year-old sheep were statistically similar (P > 0.05) to component which was associated with five linear body those of the older sheep, even though the later recorded traits (CD, HW, TH, BL and PBW), explained about slightly higher values. Sex by age interaction was signifi- 61.26 percent of the generalized variance in the body mea- cant (P < 0.05) only for RH (Table 1). The S.E. values surements with moderate (0.651) to high (0.838) emphasis revealed that HG was the most varied trait, whereas the given to these body measurements. This factor can be con- animals were more consistent by the PBW irrespective of sidered as a generalized conformation or size of the sheep. their sex or age. A t-test between the HW and RH revealed The second factor comprised the HG and NG and that the animal was significantly (t = 29.9; P < 0.001) accounted for 12.92 percent of the generalized variance,

Table 1. Analysis of variance for morphological traits in Djallonke sheep.

Source of variation Mean squares and level of signiﬁcance**

DF LW HG NG CD HW RH BL PBW

Sex 1 362.20** 572.94** 997.05** 57.00** 177.94** 179.70** 213.58** 1.09 Age 1 2.13 5.12 6.92 7.41 15.32 26.28 19.42 0.53 Sex × age 1 0.01 4.63 2.66 7.77 28.52 43.87* 11.26 0.55 Residual 168 10.64 20.27 8.22 3.59 9.86 10.05 15.04 0.70

**Highly significant (P < 0.01); *significant (P < 0.05); all other traits are not significant. DF, degrees of freedom; LW, live weight; HG, heart girth; NG, neck girth; CD, chest depth; HW, height at Withers; RH, rump height; BL, body length; and PBW, pin-bone width. Djallonke sheep body structure analysis 69

Table 2. Least-square means (±S.E.) of live weight (kg) and linear body measurements (cm) of Djallonke sheep as affected by sex and age.

Traits Sex Age Overall

Male Female 2 years ≥3 years

LW 32.13 ± 1.76a 21.71 ± 0.29b 26.52 ± 0.71 27.32 ± 1.64 26.92 ± 0.89 HG 78.29 ± 2.43a 65.19 ± 0.41b 71.12 ± 0.98 72.36 ± 2.26 71.74 ± 1.23 NG 49.17 ± 1.55a 31.88 ± 0.26b 41.24 ± 0.63 39.80 ± 1.44 40.52 ± 0.79 CD 29.79 ± 1.02a 25.66 ± 0.17b 26.98 ± 0.41 28.47 ± 0.95 27.73 ± 0.52 HW 64.37 ± 1.70a 57.06 ± 0.28b 59.64 ± 0.69 61.79 ± 1.58 60.72 ± 0.86 RH 63.28 ± 1.71a 55.94 ± 0.29b 58.21 ± 0.69 61.02 ± 1.59 59.61 ± 0.87 BL 62.88 ± 2.09a 54.87 ± 0.35b 57.67 ± 0.85 60.08 ± 1.95 58.87 ± 1.06 PBW 12.52 ± 0.45ns 13.09 ± 0.08ns 13.01 ± 0.18 12.61 ± 0.42 12.81 ± 0.23 a,bMeans within the same row having different superscripts differ signiﬁcantly (P < 0.05/0.01) between the two sexes. ns = not signiﬁcant S.E., standard error; LW, live weight; HG, heart girth; NG, neck girth; CD, chest depth; HW, height at Withers; RH, rump height; BL, body length; and PBW, pin-bone width.

Table 3. Comparison of selected linear traits (cm) of Djallonke population (Toro et al., 2011). The superiority of males sheep. over females in this study was in agreement with the Paired traits Mean difference t-statistic P-value works of Birteeb et al. (2012), Legaz et al. (2011), Carneiro et al. (2010) and Yakubu and Akinyemi (2010) HW–RH 1.147 29.900 <0.000 who attributed such differences to sexual dimorphism. HW–BL 1.822 7.270 <0.000 Sexual dimorphism is thought to develop post-weaning – RH BL 0.675 2.758 0.006 because of faster mass gain by males during the age of 1–2 years (Festa-Bianchet et al., 1996). These same researchers suggested that males might have a longer sea- thereby bringing the total variance accounted for to 74.17 son to amass (gain) body mass each year throughout their percent. lives, while females divert annual resources into reproduction, rather than body mass. Discussion The insignificant influence of age on any of the body traits in this study was analogous to the findings of Kunene, Production of the autochthonous Ghanaian sheep breeds is Nesamvuni and Fossey (2007) in Zulu sheep where all directly associated with the concept of sustainable agricul- matured (2 and 3 years) sheep recorded similar body traits, ture (Birteeb et al., 2012). For centuries, the Djallonke suggesting that at maturity sheep exhibited negligible zoo- sheep has been an integral component of African AnGRs metric trait variations as a result of age difference. Hence as it is kept and reared throughout the continent. This the growth and development of Djallonke sheep reach a breed has varied names and exhibits biometric variations peak after 2 years of age and then seem to remain almost across geographic regions, hence could be referred to as constant afterwards. The report by Traoré et al. (2008) a traditional population (FAO, 2011). Genetic variation that the Djallonke sheep is generally small in size was is fundamental for livestock populations to adapt to vary- confirmed in this study since the sheep herein was much ing environments and to respond to artificial selection; smaller when compared to Zulu sheep in South Africa therefore, any conservation and development scheme (Kunene, Nesamvuni and Fossey, 2007) and Yankasa, Uda should start from assessing the state of variation in the and Balami sheep in Nigeria (Yakubu and Ibrahim, 2011).

Figure 1. Dendrogram indicating the agglomeration of body traits into clusters. 70 P.T. Birteeb et al.

Table 4. Phenotypic correlations among body traits of matured Djallonke sheep.

Trait LW HG NG CD HW RH BL

HG 0.70** NG 0.65** 0.59** CD 0.61** 0.54** 0.45** HW 0.62** 0.61** 0.50** 0.71** RH 0.60** 0.58** 0.48** 0.73** 0.99** BL 0.61** 0.40** 0.44** 0.61** 0.62** 0.64** PBW 0.52** 0.35** 0.18* 0.43** 0.46** 0.46** 0.48**

**Highly signiﬁcant (P < 0.01); *signiﬁcant (P < 0.05). LW, live weight; HG, heart girth; NG, neck girth; CD, chest depth; HW, height at Withers; RH, rump height; BL, body length; and PBW, pin-bone width.

Nevertheless, they were comparable to Sudan-Sahel sheep of (Yakubu, Salako and Abdullah, 2011; Birteeb et al., 2012) Burkina Faso (Traoré et al., 2008) and Karayaka sheep of like Principal Component, Factor and Cluster analyses in Turkey (Cam, Olfaz and Soydan, 2010). The animal is a typ- quantifying trait variability and characterizing breeds. The ical tropical breed, being higher (taller) than long except that agglomeration of the traits into two clusters suggests that it slopes slightly from the withers towards the rump just like body conformation of Djallonke sheep could be explained the Nigerian indigenous sheep (Yakubu and Ibrahim, 2011). in two dimensions. Explicitly the traits on the first cluster Since the Djallonke sheep is a traditional population (FAO, are associated with bone growth hence give information 2011) in sub-Saharan Africa, it may be biologically favoured about the general body conformation of the sheep. The sec- in its ability to adapt to changes in management, climate, ond cluster seemed to indicate body thickness which may nutrition and marketing. result from muscle development and/or fat deposition. The significant phenotypic correlations observed herein were However, the amount of variation attributable to each cluster similarly reported by Legaz et al. (2011), Cam, Olfaz and could not be estimated. Soydan (2010) and Salako (2006a, 2006b). Slightly higher The measure of sampling adequacy, KMO was signifi- values were earlier reported by Afolayan, Adeyinka and cantly high but slightly lower than 0.85 reported for Uda Lakpini (2006). High correlations indicate the interrelation- sheep (Salako, 2006b). Recently Birteeb et al. (2012) ships between/among the traits and such knowledge is very and Yakubu, Salako and Abdullah (2011) reported useful in breeding and management practices of livestock, KMO’s of 0.91 and 0.92 for sheep in Northern Ghana as selection for a given trait directly favours other positively and Nigeria, respectively. High KMO values were equally associated traits but disfavours negatively associated traits. reported for cattle (Yakubu, Ogah and Idahor, 2009a; Nevertheless, selection for positively or negatively asso- Pundir et al., 2011). The KMO test shows whether the par- ciated traits is influenced by the production objectives, selec- tial correlations among variables are low. In this study, the tion goals and even the socio-cultural demands or conditions high KMO obtained implied that the proportion of the of the farmer. Correlations among body traits also serve as variance in the body traits/measurements caused by the basis for employment of further multivariate techniques the underlying factors is high hence true factors existed and the data are factorable.

Table 5. Varimax rotated factor loadings and communalities of The extraction of two PC factors in the Factor Analysis is body traits of Djallonke sheep. comparable to the results of the Cluster Analysis where traits were grouped into two clusters. Except the replace- Trait Principal Communalities components ment of HG by the PBW, the association of major linear measures (HW, RH and BL) with the ﬁrst PC was analo- PC1 PC2 gous to the trait clustering in Figure 1. The similarity of HG 0.294 0.774 0.685 the results of these two procedures suggests that variations NG 0.081 0.879 0.780 in body composition of Djallonke sheep are best explained CD 0.651 0.520 0.694 in two dimensions. The result of this study is similar to earl- HW 0.702 0.592 0.843 ier reports on Djallonke sheep (Birteeb et al., 2012)and RH 0.724 0.569 0.848 Uda sheep (Salako, 2006b; Yakubu, Salako and Abdullah, BL 0.713 0.360 0.639 PBW 0.838 −0.043 0.704 2011) in which two underlying factors each were extracted Eigenvalue 4.288 0.904 to explain the generalized variance in body traits. However, Explained variance (%) 61.255 12.918 the respective total explained variances of 87.19, 86.27 and Cumulative variance (%) 61.255 74.173 78.7 percent by those earlier reports were higher than that LW, live weight; HG, heart girth; NG, neck girth; CD, chest depth; HW, obtained in this study. In studying Bergamasca sheep, height at Withers; RH, rump height; BL, body length; and PBW, pin- Riva et al. (2004) extracted two and four factors for bone width. young and adult sheep, respectively, in the sedentary Djallonke sheep body structure analysis 71

flock, and three factors each for both age groups in the trans- References humant flock. In a similar study in cattle, Pundir et al. (2011) extracted three latent factors to explain a total vari- Afolayan, R.A., Adeyinka, I.A. & Lakpini, C.A.M. 2006. The estima- ance of 66.02 percent. Again in the White Fulani cattle of tion of live weight from body measurements in Yankasa sheep. Czech Nigeria, two factors were extracted in the age group of J. Anim. Sci., 51: 343–348. 1.5–2.4 years to account for 85.37 percent of total variation, Birteeb, P.T., Peters, S.O., Yakubu, A., Adeleke, M.A. & Ozoje, M.O. while four factors extracted in the age group of 2.5–3.6 2012. Multivariate characterisation of the phenotypic traits of explained 86.4 percent of the total variance (Yakubu, Djallonke and Sahel sheep in Northern Ghana. Tropic. Anim. – Ogah and Idahor, 2009a). Except the lower coefficient of Health Prod., 45: 267 274. HG, the high loadings of body height (RH and WH), BL Cam, M.A., Olfaz, M. & Soydan, E. 2010. Body measurements reflect and PBW on the first factor in the present study is very body weights and carcass yields in Karayaka sheep. Asian J. Anim. – much comparable to the report of Pundir et al. (2011)in Vet. Adv., 5: 120 127. Kankrej cows although the later study was on a different Carneiro, H., Louvandini, H., Paiva, S.R., Macedo, F., Mernies, B. & species. Most researchers (Riva et al., 2004; Salako, McManus, C. 2010. Morphological characterization of sheep breeds – 2006b; Yakubu, Ogah and Idahor, 2009a;Pundiret al., in Brazil, Uruguay and Colombia. Small Rum. Res., 94: 58 65. 2011; Yakubu, Salako and Abdullah, 2011) have described Encarta (2009). World Atlas, Geography. Microsoft Student with the first factor as a representation of the general body size. Encarta premium. In this Djallonke sheep, the first factor equally describes FAO. 2011. Draft guidelines on phenotypic characterization of animal the general body size, while the second seems to indicate genetic resources, Commission on Genetic Resources for Food and the size of the thoracic region. Both principal components Agriculture, 13th Regular Session, 18–22 July, 2011, Rome (available could play a role in the ranking of the animals, and thus pro- at http://www.fao.org/docrep/meeting/022/am651e.pdf). vide an opportunity to select the animals based on a group Festa-Bianchet, M., Jorgenson, J.T., King, W.J., Smith, K.G. & of variables rather than on isolated traits (Yakubu, Salako Wishart, W.D. 1996. The development of sexual dimorphism: sea- and Abdullah, 2011). Such knowledge will be valuable in sonal and lifetime mass changes in bighorn sheep. Can. J. Zool., 74: 330–342. obtaining animals with highly coordinated bodies from fewer traits amidst many body measurements, as a result Hoffman, I. 2010. Climate change and the characterization, breeding of exploiting underlying factors. In recent times, linear and conservation of animal genetic resources. Anim. Genet., 41: 32–46. body measurements have become very useful in livestock studies especially breed characterization. Analyses of mor- Kunene, N., Nesamvuni, E.A. & Fossey, A. 2007. Characterization of phometric variables that are easy to measure make it pos- Zulu (Nguni) sheep using linear body measurements and some environmental factors affecting these measurements. S. Afr. J. Anim. Sci., sible to explore areas such as the structure of breeds, the 37: 11–20. degree of variability between various populations, the har- mony of morphological models and the definition of mor- Legaz, E., Cervantes, I., Perez-Cabal, M.A., de la Fauente, L.F., Martinez, R., Goyache, F. & Gutierrez, J.P. 2011. Multivariate phological models for a given breed (Birteeb et al., 2012). characterisation of morphological traits in Assaf (Assaf.E) sheep. Small Rum. Res., 100: 122–130. Ogah, D.M. 2011. Assessing size and conformation of the body of Conclusion Nigerian indigenous turkey. Slovak J. Anim. Sci., 44: 21–27. Ogah, D.M., Alaga, A.A. & Momoh, M.O. 2009. Principal component It may be concluded that the Djallonke sheep is generally factor analysis of the morphostructural traits of Muscovy duck. small in body size. Both the Cluster and Factor analyses Int. J. Poult. Sci., 8: 1100–1103. yielded very similar results, thus categorizing linear body Pundir, R.K., Singh, P.K., Singh, K.P. & Dangi, P.S. 2011. Factor ana- measurements into two groups each (clusters and factors). lysis of biometric traits of Kankrej cows to explain body conform- This implies that the linear body measurements considered ation. Asian Aust. J. Anim. Sci., 24: 449–456. herein can best be grouped into two main dimensions, and Rencher, A.C. 2002. Methods of multivariate analysis, 2nd edition. factor analysis approach is preferable because of its ability Canada, Wiley-Interscience, John Wiley & Sons, Inc. Publication, to quantify the explained variance in body composition. pp. 408–447. Finally, an easy way to describe the Djallonke sheep is Riva, J., Rizzi, R., Marelli, S. & Cavalchini, L.G. 2004. Body measure- by considering the linear measurements on the first factor. ments in Bergamasca sheep. Small Rum. Res., 55: 221–227. Salako, A.E. 2006a. Application of morphological indices in the assessment of type and function in sheep. Int. J. Morphol., 24: 13–18. Acknowledgement Salako, A.E. 2006b. Principal component factor analysis of the morphostructure of immature Uda sheep. Int. J. Morphol., 24: The authors are grateful to the staff of the Pong-Tamale 571–574. Livestock Breeding Station especially the Manager, Mr SPSS. 2008. Statistical package for social sciences. Chicago, IL, SPSS Inc. Ibrahim Shahadu and Head of the Sheep Unit, Mr Toro, M.A., Meuwissen, T.H.E., Fernandez, J., Shaat, I. & Lawrence Dartay for providing the experimental animals Maki-Tanila, A. 2011. Assessing the genetic diversity in small and support during the data gathering process. farm animal populations. Animal, 5: 1669–1683. 72 P.T. Birteeb et al.

Traoré, A., Tamboura, H.H., Kabore, A., Royo, L.J., Fernandez, I., Yakubu, A. & Ibrahim, I.A. 2011. Multivariate analysis of morphos- Álvarez, I., Sangare, M., Bouchel, D., Poivey, J.P., Francois, D., tructural characteristics in Nigerian indigenous sheep. Ital. J. Anim. Toguyeni, A., Sawadogo, L. & Goyache, F. 2008. Multivariate char- Sci., 10: 83–86. acterization of morphological traits in Burkina Faso sheep. Small Yakubu, A., Ogah, D.M. & Idahor, K.O. 2009a. Principal component Rum. Res., 80: 62–67. analysis of the morphostructural indices of White Fulani cattle. Wafula, P.O., Jianlin, H., Sangare, N., Sowe, J.M., Coly, R., Trakian J. Sci.,7:67–73. Diallo, B. & Hanotte, O. 2005. Genetic characterization of West African Djallonke sheep using microsatellite markers. The Yakubu, A., Kuje, D. & Okpeku, M. 2009b. Principal components as role of biotechnology. Turin, Italy, Villa Gualino, 5–7 March, measures of size and shape in Nigerian indigenous chickens. Thai – pp. 177–178. J. Agric. Sci., 42: 167 176. Yakubu, A. & Akinyemi, M.O. 2010. An evaluation of sexual size Yakubu, A., Salako, A.E. & Abdullah, A-R. 2011. Varimax rotated dimorphism in Uda sheep using multifactorial discriminant analysis. principal component factor analysis of the zoometrical traits of Uda Acta Agric. Scand., Sect A – Anim. Sci., 60: 74–78. sheep. Arch. Zootec., 60: 813–816. Animal Genetic Resources, 2014, 54, 73–77. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000071

Variation in qualitative traits in Bhutanese indigenous chickens

T. Tashi and N. Dorji Department of Animal Science, College of Natural Resources, Royal University of Bhutan, Lobesa, Punakha, Bhutan

Summary The objective of the study was to determine the phenotypic variation among four Bhutan indigenous chickens (BIC) namely Seim (SM), Phulom (PL), Yuebjha Narp (YN) and Khuilay (KL). A total of 120 chickens (KL, 30; YN, 30; PL, 30; SM, 30) were included for the present study. Based on descriptive statistic analysis, the results illustrate that there are three types of comb and single type predominant. The Bhutanese chickens do not have feather on the shank and the head type is plain. The common colours of earlobe in four indigenous birds are white and red. White skin and shank colour rank high among BIC groups; however, in SM the yellowish skin colour is observed high. The predominant feather colour of four indigenous chickens is pure black in YN (48.28 percent), white in KL (30.00 percent), reddish brown in SM (33.33 percent) and PL (23.33 percent). In addition, breast, back and neck feather followed the similar pattern of colour. Among BIC populations, KL and YN strains present the greatest and lowest plumage colour diversity, respectively. For example, eight different colours of breast feather are surveyed in KL and PL birds. Therefore, considering the variation in feather colourings, KL and PL seems to have more diversity in morphology but, it requires further studies. On the other hand, YN appears to have less diversity requiring for appropriate development of conservation strategies in Bhutan. Moreover, the large variations present the possibility of making improvement among BIC.

Keywords: indigenous chickens, phenotypic variation, diversity, conservation

Résumé Le but de l’étude a été de déterminer la variation phénotypique entre quatre poules autochtones du Bhoutan, à savoir Seim (SM), Phulom (PL), Yuebjha Narp (YN) et Khuilay (KL). Cette étude a compris un total de 120 poules (KL, 30; YN, 30; PL, 30; SM, 30). Basés sur une analyse statistique descriptive, les résultats ont décelé qu’il existe trois types de crêtes, dont le type simple est le type prédominant. Les poules bhoutanaises n’ont pas de plumes sur les tarses et leur tête est du type simple. Les couleurs les plus courantes des oreillons sont, chez les quatre poules autochtones, le blanc et le rouge. La peau et les tarses sont blancs chez la plupart des poules bhoutanaises. Néanmoins, de nombreux spécimens SM présentent une peau de couleur jaunâtre. La couleur prédominante du plumage est le noir pur chez YN (48.28 pour cent), le blanc chez KL (30.00 pour cent) et le marron rougeâtre chez SM (33.33 pour cent) et chez PL (23.33 pour cent). En outre, le plumage de la poitrine, du dos et du cou suivent un schéma de couleur similaire. Parmi les différentes populations de poules bhoutanaises, les lignées KL et YN sont celles qui présentent, respectivement, la plus grande et la plus faible diversité chromatique dans leur plumage. Ainsi, par exemple, huit couleurs différentes ont été relevées pour le plumage de la poitrine des poules KL et PL. Par conséquent, d’après la variation chromatique du plumage, les lignées KL et PL semblent être celles ayant la plus grande diversité morphologique. Toutefois, des études plus approfondies s’avèrent nécessaires. D’un autre côté, la lignée YN semble être celle qui présente la plus faible diversité, comme quoi des stratégies devraient être mises en place au Bhoutan pour assurer la conservation de cette lignée. De plus, les grandes variations observées entre les différentes poules bhoutanaises ouvrent la possibilité de procéder à des projets d’amélioration génétique.

Mots-clés: poules autochtones, variation phénotypique, diversité, conservation

Resumen El objetivo del estudio fue determinar la variación fenotípica existente entre cuatro gallinas autóctonas de Bhután, concretamente entre Seim (SM), Phulom (PL), Yuebjha Narp (YN) y Khuilay (KL). En este estudio, se incluyeron un total de 120 gallinas (KL, 30; YN, 30; PL, 30; SM, 30). Con base en un análisis estadístico descriptivo, los resultados mostraron que se dan tres tipos de cresta, de los cuales predomina el tipo sencillo. Las gallinas butanesas no tienen plumas sobre los tarsos y la cabeza es de tipo sencillo. En las cuatro gallinas autóctonas, los colores de orejilla más comunes son el blanco y el rojo. Por lo general, el blanco es el color de la piel y de los tarsos en las gallinas butanesas. No obstante, muchos ejemplares SM presentan piel de color amarillento. El color predominante en las plumas es el negro puro en YN (48.28 por ciento), el blanco en KL (30.00 por ciento) y el marrón rojizo en SM (33.33 por ciento) y en PL (23.33 por ciento). Asimismo, las plumas del pecho, del dorso y del cuello siguen un patrón cromático similar. Entre las distintas poblaciones de gallinas butanesas, las variedades KL y YN son las que presentan, respectivamente, la mayor y la menor diversidad cromática en su plumaje. Así, por ejemplo, se han registrado ocho colores diferentes para el plumaje del pecho en las gallinas KL y PL. Por tanto, de acuerdo con la variación en el colorido de las plumas, KL y PL son probablemente las variedades que mayor

Correspondence to: N. Dorji, Department of Animal Science, College of Natural Resources, Royal University of Bhutan, Lobesa, Punakha, Bhutan. emails: [email protected]; [email protected]

73 74 T. Tashi and N. Dorji

diversidad morfológica presentan. Sin embargo, se precisan más estudios. Por otro lado, parece que la variedad YN es la que menor diversidad presenta, con lo que se hace necesario un adecuado desarrollo de estrategias para la conservación de esta variedad en Bhután. Por último, las grandes variaciones observadas entre las distintas gallinas butanesas abren la posibilidad de hacer mejora genética.

Palabras clave: gallinas autóctonas, variación fenotípica, diversidad, conservación

Submitted 2 July 2013; accepted 23 January 2014

Introduction (KL = 30), Yuebjha Narp (YN = 30), Phulom (PL = 30) In developing countries, the majority of poultry products and Seim (SM = 30). come from traditional production practices. Indigenous chickens continue to be most popular due to their adapt- Data collection and analysis ability to harsh conditions and their resistance against local diseases (Ajayi, 2010; Dorji and Gyeltshen, 2012). The data captured parameters which are visually judged for This indigenous bird manifests a great deal of variation variation in the phenotypic traits such as colour of feather, which is due to genetic and environmental factors (Olori shank, ear lobe and shank, comb type and head type (FAO, and Sonaiya, 1992). Thus, they are reservoirs of genetic 2012). Descriptive statistics of SPSS 16 was used to ana- materials for genetic studies, improvement, preservation lyze the data. and conservation (Dorji, Duangjinda and Phasuk, 2012). Poultry production has been playing crucial roles in sustaining and supplementing a cheap protein to the poor Results and discussion (Dorji and Gyeltshen, 2012). Importantly, indigenous chickens are kept for their socio-cultural importance Feather colour (Dorji and Gyeltshen, 2012; Dorji, Duangjinda and Phasuk, 2012). For example, society believes in consum- The plumage colour of local chickens available was red- ing eggs and meat during pregnancy and after delivery. dish brown constituting the maximum proportion of plumage colour in SM and PL varieties with 33.33 and The traditional sub-sector consists of poorly uncharacter- 23.33 percent, respectively. An utmost plumage colour ized indigenous chickens (Dorji and Gyeltshen, 2012). was white (30.00 percent) and black (48.28 percent) in Recently, Food and Agriculture Organization of the KL and YN, respectively. Similarly, a predominant colour- United Nation (FAO) Domestic Animal Diversity ing as reddish brown (20.00 percent), black (30.00 percent) Information System (DADIS) listed ten indigenous strains and reddish (30.00 percent) in neck feather was observed of chicken in Bhutan. Dorji, Duangjinda and Phasuk for PL, YN and SM, respectively. Neck feather was (2012) conducted genetic study on four strains of more varied in PL population. On contrary, YN contains Bhutanese local chicken based on the economic import- the least variation of neck feather colours. The naked ance to the rural society and their popularity. Dorji, neck gene is a dominant gene and is responsible for gen- Duangjinda and Phasuk (2012) documented for greater eral reduction of feather over body surface and total loss genetic diversity of Khuilay (naked neck) and lowest for of feather in neck region (Merat, 1986; Ige et al., 2012). Yuebjha Narp (black chicken). Moreover, they did men- tion that the phenotypic variation was likely to be highest In terms of back feather colouration, the highest percentage for Khuilay but they did not quantify. Therefore, the was surveyed in SM, KL, PL and YN was reddish brown objective of the study was to access diversity of (40.00 percent), white (23.33 percent), brownish (20.00 Bhutanese local chicken based on qualitative trait. percent) and black (46.67 percent), respectively. The same colour domination was examined for breast feather colourings in SM (reddish brown, 33.33 percent), KL (white, 36.67 percent), PL (brownish, 20.00 percent) and Materials and methods YN (black, 53.33 percent). Importantly, plumage, neck feather, back and breast feather colour were found to be Sample sizes more varied in KL and PL populations (Table 1). This Considering the agro-ecology, socioeconomic significance probably explains the greater diversity with respect to fea- of chicken production and population of indigenous chick- ther colourations (Bhuiyan et al., 2005; Faruque et al., ens, Tsirang was selected and this district has been recog- 2010). On the other hand, lowest feathers colour was nized as important genetic resources of poultry in the observed in YN indicating the least diversified among country. Thus, farmers are being trained related to conser- the four Bhutan indigenous chickens (BIC) populations vation activity of traditional birds. The study was based on agreeing with Dorji, Duangjinda and Phasuk (2012) four Bhutanese indigenous chicken comprising Khuilay findings. Their report was based on FAO recommended Variation in qualitative traits in Bhutanese indigenous chickens 75

Table 1. Feather colour in four Bhutanese indigenous chicken populations (percent).

Trait Seim Khuilay Phulom Yuebjha Nap

Characteristic Percent Characteristic Percent Characteristic Percent Characteristic Percent

Plumage Reddish brown 33.33 White 30.00 Reddish brown 23.33 Black 48.28 Reddish black 16.67 Brownish black 26.67 Barred 20.00 Black/dark green 20.69 Reddish dark green 13.33 Black 6.67 White 16.67 Black/white 13.79 Brownish 13.33 Reddish/brown 6.67 Black 13.33 Black/brown 10.34 Reddish brown, 13.33 Barred 6.67 Brownish black 13.33 Black/dark green/ 6.90 black lines brown Reddish black 10.00 Brownish 6.67 Brownish 10.00 brown Reddish black 6.67 White, yellowish 3.33 brown Reddish 6.67 Brownish white 3.33 Neck Reddish 30.00 Reddish brown 20.00 Black 30.00 feather Reddish black 16.67 Brownish 16.67 Black/dark green 23.33 Brownish 13.33 White 13.33 Brownish 16.67 Reddish brown 13.33 Reddish 13.33 Black/brownish 16.67 Brownish black 10.00 Whitish black 10.00 Black/white 6.67 Multiple 10.00 Black 10.00 White 3.33 Brownish yellow 3.33 White black lines 6.67 Reddish 3.33 Reddish/dark green 3.33 Brownish black 6.67 Yellowish black 3.33 Back Reddish brown 40.00 White 23.33 Brownish 20.00 Black 46.67 feather Reddish/black 20.00 Brownish black 16.67 Whitish black 20.00 Black/dark green 40.00 Multiple 20.00 Reddish brown 16.67 Reddish brown 16.67 Black/brownish 10.00 Brownish 6.67 Black/white 16.67 Brownish black 16.67 Black/white 3.33 Reddish/dark green 6.67 Black/dark green 10.00 White 13.33 Black 3.33 Reddish 6.67 Reddish 6.67 Reddish 3.33 Brownish 6.67 Yellowish white 3.33 Brownish/dark 3.33 Black 3.33 green Breast Reddish brown 33.33 White 36.67 Brownish 20.00 Black 53.33 feather Brownish 23.33 Brownish 23.33 White 20.00 Light black 16.67 Reddish 16.67 Black 10.00 Reddish brown 16.67 Black/dark green 13.33 Reddish black 10.00 Reddish 10.00 Whitish black 10.00 Brownish/black 6.67 Brownish white 6.67 Brownish/black 6.67 Reddish 10.00 Black/white 6.67 Multiple 6.67 Black/white 6.67 Brownish black 10.00 Brown spot 3.33 Brownish yellow 3.33 Yellowish/white 3.33 Black 6.67 Reddish/brown 3.33 Brownish white 6.67 molecular tools. Moreover, the frizzling and naked neck from native chickens of Botswana (Badubi et al., 2006) genes is considered for increased feed efficiency, growth and Bangladeshis (Bhuiyan et al., 2005; Faruque et al., rate, egg production and disease tolerance (Ajayi, 2010; 2010; Uddin et al., 2011). This indicates that the bird Egahi, Dim and Momoh, 2013). Moreover, genetic dis- have been favoured by hot climatic conditions whereby tance between broiler and KL was sufficiently close large comb, such as single comb allows for efficient heat (Dorji, Duangjinda and Phasuk, 2012). regulation (Apuno, Mbap and Ibrahim, 2011). On the other hand, pea type was the most frequently sampled Comb type and earlobe for Ethiopian traditional birds (Dana et al., 2010). Only three types of comb (rose, pea and single) were found Except for the SM, all the BIC birds have ear lobes in all four indigenous chickens of Bhutan (Table 2). The (Table 2). Earlobe colour was dominated by red (50.00 finding agrees with the reports of Daikwo, Okpe and percent) in SM, white (46.67 percent) in KL, white and Ocheja (2011) and Ige et al. (2012) on native chickens red (36.67 percent) and white (43.33 percent) in YN of Nigeria. Egahi et al. (2010), Dana et al. (2010) and flocks. The dominant white colour was also reported Badubi, Rakereng and Marumo (2006) reported more in Nigerian (Egahi et al., 2010; Ige et al., 2012), than three comb types contradicting with present study. Bangladeshis (Faruque et al., 2010) and Ethiopian (Dana Nevertheless, the common comb occurring was single et al., 2010) indigenous chicken. Furthermore, Faruque comb type as examined for Nigerian local birds (Egahi et al. (2010) documented that the red and white earlobe et al., 2010; Apuno, Mbap and Ibrahim, 2011; Daikwo, colour dominants in naked neck population and is in line Okpe and Ocheja, 2011). Parallel result was also obtained with KL populations. 76 T. Tashi and N. Dorji

Table 2. Shank, skin and earlobe colour, comb type, heat type and ptilopody among four indigenous chickens of Bhutan (percent).

Trait Seim Khuilay Phulom Yuebjha Nap

Characteristic Percent Characteristic Percent Characteristic Percent Characteristic Percent

Shank colour Yellowish 36.67 White 26.67 White 33.33 Black 36.67 White 26.67 Yellow 23.33 Reddish 23.33 Ash 23.33 Ash 16.67 Black 20.00 Yellowish white 20.00 Light red 20.00 Yellowish white 13.33 Ash 16.67 Yellow 10.00 White 16.67 Black 6.67 Light red 10.00 Black 6.67 Pale yellow 3.33 Yellowish white 3.33 Ash white 6.67 Skin colour Yellowish 46.67 Reddish 33.33 White 30.00 White 33.33 White 30.00 White 30.00 Red 26.67 Light red 33.33 Light red 10.00 Reddish white 23.33 Yellowish 23.33 Yellowish 16.67 Red 6.67 Yellowish 10.00 Yellowish white 20.00 Ash 6.67 Pale yellow 6.67 Reddish ash 3.33 Reddish white 6.67 Yellowish white 3.33 Ear lobe Present 93.33 Present 100.00 Present 100.00 Present 100.00 Absent 6.67 Absent 0.00 Absent 0.00 Absent 0.00 Earlobe colour Red 50.00 White 46.67 White 36.67 White 43.33 White 32.14 Red 43.33 Red 36.67 Red 33.33 Yellowish white 17.86 Reddish white 3.33 Reddish white 16.67 Ash 13.33 Ash 3.33 Ash 10.00 Ash black 6.67 Yellowish 3.33 Yellowish white 3.33 Comb type Single 40.00 Single 36.67 Single 40.00 Single 36.67 Rose 33.33 Rose 33.33 Pea 33.33 Pea 36.67 Pea 26.67 Pea 30.00 Rose 26.67 Rose 26.67 Head shape Plain 100.00 Plain 100.00 Plain 100.00 Plain 96.67 Crested 0.00 Crested 0.00 Crested 0.00 Crested 3.33 Ptilopody Absent 100.00 Absent 100.00 Absent 100.00 Absent 100.00 Present 0.00 Present 0.00 Present 0.00 Present 0.00

Ptilopody and head type in Bangladeshis chickens. This contradicts with Mancha (2004)whosefinding in Plateau chicken was pink, dark Throughout the study crested head and ptilopody were not ash, ash and light yellow. Genes that affect plumage colour found in all BIC populations (Table 2). The absence of also affect shank colour (Crawford, 1990).The presence of feather on shank aligns with the finding of Halima large variations further confirms that higher diversity in KL (2007) in Northwest Ethiopian local birds. Faruque et al. and PL as suggested by Daikwo, Okpe and Ocheja (2011). (2010) also observed that the shank feather phenomenon was absent in three genotypes of Bangladeshis native chicken. On contrary, lower percent of ptilopody among Nigerian local birds was examined (Daikwo, Okpe and Conclusion Ocheja, 2011). Bhutanese indigenous chicken has their place in contributing to the genetic collection because of their hardiness Skin and shank colour and aptitude to survive under low input conditions. Development of conservation strategies is very important The skin colour ranged from white to red among the stud- to make genetic resources available for meeting the future ied BIC populations. However, the prime skin colour in unpredicted breeding requirement. Among four studied SM, KL, PL and YN was yellowish (46.67 percent), red- chicken strains in Bhutan, single comb type predominates dish (33.33 percent), white (30 percent) and (33.33 per- and there was no feather on shank and head. Based on fea- cent), respectively (Table 2). Daikwo, Okpe and Ocheja ther colours, Kuilay seems to be more diversed while (2011) also reports multiple shank colour for Nigerian Yuebjha Narp reflect lower morphological variation. It is traditional birds. This was contradicting with the further suggested to conduct morphometric study on Ethiopian local bird’s shank colour with white and yellow Bhutanese chicken populations to confirm our claims. (Dana et al., 2010). Nevertheless, high percent colour of shank follow the same pattern. Shank with white colour accounted the most in KL (26.67 per- Acknowledgements cent) and PL (33.33 percent), yellowish (36.67 percent) in SM and black (36.67 percent) in YN. Faruque et al. (2010) We would like to thank individual farmers for their active also mentioned that the white shank colouring dominants participation during this study. Variation in qualitative traits in Bhutanese indigenous chickens 77

References Egahi, J.O., Dim, N.I., Momoh, O.M. & Gwaza, D.S. 2010. Variations in qualitative traits in the Nigerian local chicken. Int. J. Poultry Sci., 9: 978–979. Ajayi, F.O. 2010. Nigerian indigenous chicken: a valuable genetic resource for meat and egg production. Asian J. Poultry Sci., 4: Egahi, J.O., Dim, N.I. & Momoh, O.M. 2013. The effect of plumage 164–172. modiﬁer genes on egg quality indices of the Nigerian local chicken. J. Agric. Vet. Sci., 2(2): 04–06. Apuno, A.A., Mbap, S.T. & Ibrahim, T. 2011. Characterization of local chickens (Gallus gallus domesticus) in Shelleng and Song local gov- Faruque, S., Siddiqee, N.U., Afroz, M.A. & Islam, M.S. 2010. ernment areas of Adamawa state, Nigeria. Agric. Biol. J. N. Am., 2(1): Phenotypic characterization of native chicken reared under intensive 6–14. management system. J. Bangladesh Agric. University,8:79–82. Badubi, S.S., Rakereng, M. & Marumo, M. 2006. Morphological Food and Agriculture Organization. 2012. Phenotypic characterization characteristics and food resources available for indigenous chickens of animal genetic resources. FAO Animal Production and Health in Botswana. Livestock Res. Rural Dev. (available at http://www. Guidelines 11, Rome. lrrd.org/lrrd18/1/badu18003.htm). Halima, H.M. 2007. Phenotypic and genetic characterization of indigen- Bhuiyan, A.K.F.H., Bhuiyan, M.S.A. & Deb, G.K. 2005. Indigenous ous chicken populations in Northwest Ethiopia. University of Free chicken genetic resources in Bangladesh: current status and future out- State, Bloemfontein, South Africa. (Ph.D. thesis). look. Animal Genet. Res. Inf. (AGRI), 36: 73–84, FAO, Rome. Ige, A.O., Salako, A.E., Yakubu, A. & Adeyemi, S.A. 2012. Qualitative Crawford, R.D. 1990. Poultry breeding and genetics. New York, traits characterization of Yoruba and Fulani ecotype indigenous chick- Elsevier Press, pp. 109–169. ens in derived Savannah zone of Nigeria. Int. J. Poultry Sci., 11(10): 616–620. Daikwo, I.S., Okpe, A.A. & Ocheja, J.O. 2011. Phenotypic characterization of local chickens in Dekina. Int. J. Poultry Sci., 10(6): Mancha, Y.P. 2004. Characterization of local chickens in Northern part 444–447. of the Jos Plateau. Animal Production Programme, School of Agriculture, ATBU, Bauchi. (Ph.D. thesis). Dana, N., Dessie, T., van der Waaij, L.H. & van Arendonk, J.A.M. 2010. Morphological features of indigenous chicken populations of Merat, P. 1986. Potential use of Na (naked neck) gene in poultry produc- Ethiopia. Animal Genet. Res., 46: 11–23. tion. Int. J. Poultry Sci., 42: 124–142. Dorji, N. & Gyeltshen, T. 2012. Characterisation of family poultry pro- Olori, V.E. & Sonaiya, E.B. 1992. Effect of length of lay of Nigerian duction in Haa and Mongar districts of Bhutan. Livestock Res. Rural indigenous chicken on their egg composition and shell quality. Dev. (available at http://www.lrrd.org/lrrd24/9/dorj24155.htm). Nig. J. Anim. Prod., 19: 95–100. Dorji, N., Duangjinda, M. & Phasuk, Y. 2012. Genetic characterization Uddin, A.H., Ali, A., Aktar, Y. & Khatun, M.A. 2011. Geographical of Bhutanese native chickens based on an analysis of Red Junglefowl, distribution, classiﬁcation, characterization and conservation of differ- domestic Southeast Asian and commercial chicken lines. Genet. Mol. ent native chicken varieties of Bangladesh. Bangladesh Res. Publ. J., Biol., 35(3): 603–609. 5(3): 227–233. Animal Genetic Resources, 2014, 54, 79–83. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000490

Genetic polymorphism of αS1 casein in Guéra and Sahel goat

D. Traoré, Y. Sanogo, R. Fané, A. Touré, O. Cissé and A.H. Babana Laboratoire de Recherche en Microbiologie et biotechnologie Microbienne (LaboREM-biotech), Faculté des Sciences et Techniques (FST), Université des Sciences, des Techniques et des Technologies de Bamako (USTTB), Mali

Summary The αS1 casein gene is one of the lacto protein genes, which are involved in the milk synthesis. The objective of this study was to identify genotypes of B and E alleles at the locus αS1 casein in the Guéra and Sahel goat herds in Mali. These herds were located in the Regional Center of Agronomic Research and the Ségala village in Kayes, Mali. A total of 101 blood samples (53 for Guéra goat and 48 for Sahel goats) were collected and analysed. Alleles B and/or E, located on exon 19, of the αS1 casein gene were amplified using allele specific amplification polymerase chain reaction and analysed using 2 percent agarose gel electrophoresis. Results from this analysis identified three genotypes: BB, BE and EE with frequencies 0.77, 0.17 and 0.06 for the Guéra goats. However, only the BB genotype was identified in the Sahel goats. The frequencies of alleles B and E in the Guéra goats were 0.86 and 0.14, respectively, with an observed and expected heterozygotes of 0.08 and 0.12, respectively.

Keywords: allele, αS1 casein, Guéra goat, Sahel goat

Résumé Le gène αS1 caséine fait partie du groupe de gène des lactoprotéines intervenant dans la synthèse du lait. L’objectif de notre travail était d’identifier les génotypes liés aux allèles B et E au locus αS1 caséine chez deux races caprines, la chèvre Guéra et du Sahel au Mali. Les analyses ont porté sur le gène αS1 caséine d’un troupeau de chèvre Guéra du Centre Régional de Recherche Agronomique de Kayes au Mali et d’un autre troupeau, la chèvre du Sahel en milieu rural dans un village (Ségala) près de Kayes au Mali. Des amorces spécifiques ont été utilisées pour l’identification des allèles B et/ou E du gène αS1 caséine (exon 19). Au total, 101 échantillons de sang (53 de la chèvre Guéra et 48 de la chèvre du Sahel) ont été prélevés et le gène αS1 caséine a été amplifié par la méthode de l’allèle spécifique polymerase chain reaction (AS-PCR). La révélation des allèles a été faite après électrophorèse sur un gel d’agarose 2 percent en utilisant le bromure d’éthidium. Trois génotypes, (BB, BE et EE) avec des fréquences respectives de 0,77; 0,17 et 0,06 ont été observés chez la chèvre Guéra. En revanche, un seul génotype (BB) a été identifié chez la chèvre du Sahel. Les fréquences des allèles B et E au niveau du troupeau Guéra étaient respectivement 0,86 et 0,14. L’hétérozygotie observée et attendue était de 0,08 et 0,12.

Mots-clés: allèles, αS1 caséine, chèvre Guéra, chèvre du Sahel

Resumen El gen de la caseína αS1 es uno de los genes, codificantes para proteínas lácteas, implicados en la síntesis de la leche. El objetivo de este estudio era identificar los genotipos de los alelos B y E del locus de la caseína αS1 en los rebaños de cabras Guéra y Sahel de Malí. Estos rebaños se encontraban en el Centro Regional de Investigación Agronómica y en el pueblo de Ségala, en la región de Kayes de Malí. Se tomaron y se analizaron un total de 101 muestras de sangre (53 de cabras Guéra y 48 de cabras Sahel). Los alelos B y/o E, situados en el exón 19, del gen de la caseína αS1 fueron amplificados de forma específica usando la reacción en cadena de la polimerasa y analizados por electroforesis en gel de agarosa (2%). Los resultados de este análisis permitieron identificar tres genotipos: BB, BE y EE, con unas frecuencias de 0,77, 0,17 y 0,06 para las cabras Guéra. Sin embargo, tan sólo el genotipo BB fue identificado en las cabras Sahel. Las frecuencias de los alelos B y E en las cabras Guéra fueron de 0,86 y 0,14 respectivamente, con una heterocigosis observada y esperada de 0,08 y 0,12, respectivamente.

Palabras clave: alelo, caseína αS1, cabra Guéra, cabra Sah

Submitted 7 August 2013; accepted 12 November 2013

Introduction

Small ruminants and especially goats play an important Correspondence to: D. Traore, Laboratoire de Recherche en Microbiologie et bio- role in the economies of rural smallholder families in technologie Microbienne (LaboREM-biotech), Faculté des Sciences et Techniques (FST), Université des Sciences, des Techniques et des Technologies de Bamako Mali. Goat milk is an important source of protein for (USTTB), Mali. email: [email protected] rural populations in the region of Kayes and particularly

79 80 D. Traoré et al.

in the fight against child malnutrition. In recent years, the Vacca et al. (2009) investigated on genetic diversity at the Institute of Rural Economy of Mali through the four casein genes in the Arbi goat, a native goat breed from Programme “Small Ruminants” conducted research activ- Tunisia. The most frequent allele at the CSN1S1 locus was ities to introduce the Guéra goats in rural areas with a B, followed by A and 01. Analysis on αS1 gene were car- view to improve milk production of local goats. ried out in a dairy goat herd (Saanen and Alpine) in Southeastern region of Brazil (Soares et al., 2009). Three The Guéra goats widely found in Mauritania was intro- different regions of the α casein gene (CSN1S1) were duced at the Agricultural Research Station in Kayes Mali S1 investigated to determine the frequencies of major alleles in 1998. This goat breed is raised for its high milk perfor null, low, intermediate and high milk protein expres- formance (Traoré, Nantoumé and Diarra, 2005). sion. Allele E was the most frequent in both Saanen According to the Mauritanian farmers, the origin of this (0.35) and Alpine (0.48) breeds followed with F allele breed is the Canary Islands (Spain). Information about (0.30). its genetic potential in Mali is sparse. The present research will contribute to a better understanding of the genetic Information on casein genes by Guéra goat is rare. The characteristics of Guéra and Sahel goats, through analysis relatively high milk yield in Guéra goat, 1–3 litres, in 142 days of lactation length (Traoré, Nantoumé and of αS1 casein gene. Diarra, 2005), could be better understood through molecu- α In the past few years, S1 casein gene in ruminants and lar genetics analysis. The objective of the investigations especially in goats has been used as a tool for genetic char- was to analyse the genetic polymorphism of αS1 casein acterization (Ramunno et al., 2001; Angiolillo et al., 2002; in Guéra and Sahel goats and to determine the genetic Feligini et al., 2005). This gene is part of the lactoproteins diversity in the studied populations. The present study α α β κ α β ( S1, S2, , casein, lactalbumin and lactoglobulin), was focused on identifying the B and E alleles of the which has been used in several studies to identify geno- CSN1S1 gene using allele specifics primers. types of animals and their links to milk performances (Boulanger, Grosclaude and Mahe, 1984; Grosclaude et al., 1987). Today, several alleles of these genes have been identified and characterized in different species and Materials and methods α breeds. For the S1 casein gene in goats, seven alleles Materials (A, B, C, D, E, F and O) with four levels of casein synthesis: 3.6 g/l for strong alleles (A, B and C), 1.6 g/l for the Descriptive characteristics of the breeds allele E, 0.6 g/l for low alleles (D and F) and 0 in goats Guéra goat fi with the allele null (0), were identi ed (Grosclaude In this study, 53 Guéra goats were randomly sampled from et al., 1987). the experimental herd of the Regional Center for In Alpine goats, it has been found that the goats with geno- Agronomic Research in Kayes (Mali). The breed is charac- type AA produce significantly less milk than those with terized by a great diversity in coat colour (uniform: white, genotypes AE, AF, EE and FF, the highest milk yield brown, black, etc. or pied). It is distinguished through its being observed in goats with EE genotype (Barbieri long hair. The average weight of mature males and females et al., 1995). At birth, kids with genotype AA had an aver- are 44 and 36 kg, respectively (Traoré, Nantoumé and age weight significantly lower than those with AE and EF Diarra, 2007). genotypes. Sahel goat Alleles A, B and C are recognized to be the ancestral types During the study 48 Sahel goats were also randomly carried nucleotide substitutions, whereas allele E is from sampled in farmer’s herds in Ségala in the region of an insertion of 458 bp nucleotides at the last exon of the Kayes (Mali). This breed is raised in the Sahelian zone gene (Martin, 1993). F allele is characterized through epi- of West Africa. It is found from Mali to Chad. It is kept tasis abnormalities in form of deletions of 37 amino acid for its milk and meat. The Sahel goat appears in various residues (Barbieri et al., 1995). coat colours: uniform white, brown, black or pied. The average weight of mature males and females are 35 and Studies on genetic characterization of casein gene complex 25 kg, respectively (Nantoumé, Traoré and Diarra, 2005). in West African goats (Red Sokoto, West African Dwarf Nigeria, West African Dwarf Cameroon and Borno) were carried out by Caroli et al. (2007). The four casein genes Methods αS1, β, αS2 and κ were typed at the DNA level. A polymerase chain reaction (PCR) single-strand conformation poly- Sampling and DNA extraction morphism method was implemented for the identification Blood samples from 53 Guéra goats and 48 Sahel goats of CSN1S1*F allele simultaneously with A/0(1), B/E, N were collected using confetti paper type Whatman 903 and the new allele (CSN1S1*B’). The investigations for molecular genetic analysis. showed a high frequency of CSN1S1*B’ in all the DNA was extracted from blood on confetti paper using the population. Truett et al. (2000) protocol. For alkaline lysis, a solution Genetic polymorphism of αS1 casein in Guéra and Sahel goat 81 of 25 mM NaOH, 0.2 mM disodium EDTA, pH 12 was Results used. The precipitation of DNA molecules was obtained with absolute ethanol in the presence of potassium acetate Three genotypes (BB, BE and EE) were observed in Guéra (8 M, pH 7.4). A solution of neutralization (40 mM Tris– goats samples (Figure 1a and b), whereas only genotype HCl, pH 5) was also used to stabilize DNA. Final DNA (BB) was observed in Sahel goat samples (Figure 2). was obtained using 70 percent ethanol by washing. In the Guéra goat, homozygous BB (41 animals) was the DNA concentrations ranging between 18 and 58 ng/μl most frequent genotype (0.77) followed by the heterozy- measured with a spectrometer (Eppendorf Bio gous BE (nine animals) with 0.17 and the homozygous Photometer) at 260 and 280 nm, were used for PCR. EE (three animals) with 0.06. In Sahel goat, all animals (48) were homozygous BB. Allelic specific amplification by PCR for The genotypic and allelic frequencies in the two breeds are identification of B and E alleles shown in Tables 1a and 1b. For the PCR, 50 ng of genomic DNA was amplified in a The frequency of allele B in the two breeds and that of E is reaction volume of 30 μl using 150 nM primer, 50 mM given in Table 2. KCL, 10 mM Tris–HCl (pH 8.3), 2.5 mM MgCl2, 200 μM dNTPs and 1.25 U Taq polymerase (Invitrogen). Different heterozygosity rates were evaluated. The genetic diversity index (FST) in the two breeds was low (0.07). Primers were selected according to the study of Feligini et al. (2005) for the identification of B alleles (90 bp) and/or E (550 bp) at the locus CSN1S1 (exon 19). The accession number of gene at the Genbank is X72221. Discussion The thermocycler thermal-type PTC-200 (M) Research Previous studies (Missohou, Talaki and Maman Laminou, fi was used for ampli cation with the following programme: 2006) on the casein gene on African goat breeds including initial denaturation at 94 °C for 3 min, denaturation at 94 ° those of West Africa have shown that the allelic frequen- C for 45 s by 25 cycles, hybridizing at 59 °C for 45 s, cies were low and the most identified allele were A and fi extension at 72 °C for 1 min, nal extension at 72 °C for B (strong allele). Further investigations (Caroli et al., 5 min and the conservation at 4 °C. 2007) on the casein group genes using four breeds or var- The PCR products were migrated in 2 percent agarose gel ieties of goats in Africa, namely the Sokoto red goat (BDH electran), for 60 min at 100 V. Ethidium bromide (1 (Nigeria), the Borno goat (Nigeria), the Nigerian and μM) was used to reveal the bands. Cameroon dwarf goat, showed that the A and B alleles were more frequent with a predominance of B allele ’ fi Statistical analysis (0.37). A new allele, named CSN1S1*B , was identi ed with a high frequency in all populations, ranging from The frequencies of B and E alleles were calculated using 0.295 (West African Dwarf Cameroon) to 0.405 (Borno). the formula: The intermediate E allele was also observed with a very pB = [2(BB) + (BE)]/2N, low frequency of 0.01 in Sokoto red goats. In our study, the results obtained in the Sahel goats were = + / qE [2(EE) (BE)] 2N, expected. They indicate the non-introduction of exotic breeds in the studied herd in Ségala (Kayes). In contrast, where pB and qE are the frequencies of B and E alleles, the Guéra goats introduced in Africa for several decades, BB and EE are the genotypes and N is the total number. from Spain was subject of many questions relating its The heterozygous rates (observed and expected) were eval- wide range of milk performances observed at the uated assuming that the herds were in genetic equilibrium of Hardy–Weinberg.

The observed heterozygosity (Ho)=1/N Σ Hi, where N is the total number of loci,

Hi is the heterozygosity at the locus i. 2 2 The expected heterozygosity (He)=1− (f1 +f2 + ...+ fk2)=1− Σ f2, where f are frequencies f1, f2,...,fk. 2 2 2 Total heterozygosity (HT)=1− (f1T +f2T + ...fkT)=1− 2 Σ fT, where, fT is the total frequency.

The genetic diversity index was calculated using FST Figure 1. (a) Ampliﬁcation of the αS1 casein gene (E 19) in 2 percent agarose gel for Guéra goat; M: marker 14; lane 1: EE lanes 2–7: BB lane C: negatif (HT − He)/HT,whereFST is the genetic diversity indice, control. (b) Ampliﬁcation of the αS1 casein gene (E 19) in 2 percent agarose HT is the total heterozygosity and He is the expected gel for Guéra goat; M: marker 14; lanes 6 et 7: BE lanes 5; 8; 9; 10; 11: heterozygosity. BB line C: negatif control. 82 D. Traoré et al.

Table 2. Observed, expected and total heterozgosity at the locus αS1 casein (exon 19) in Guéra and Sahel goat Heterozygoty

observed (Ho) expected (He)

Guéra goat: 0.17 Guéra goat: 0.24 Sahel goat: 0.00 Sahel goat: 0.00 Both breeds: 0.08 Both breeds: 0.12

Total (HT) = 0.13

Figure 2. Amplification of the αS1 casein gene (E 19) in 2 percent agarose gel for Sahel goat; M: marker 14; lanes 1 à 6: BB; lanes 7; 8; 9; 10: negatifs lane rate. The absence of the E allele in Sahel goats is largely C: negatif control. responsible for these low values (0.07). It would be inter- esting to carry out further studies including other popula- Agricultural Research Center in Kayes (Mali). The study tions of Sahel goat in Mali to confirm the absence of E made on the Guéra goat breed by Traoré, Nantoumé and allele. Diarra (2005) showed an average daily milk quantity of Allele B was predominant in both breeds. The method- 1.9 litres ranging from 1 to 3 litres, with a total milk pro- ology used in the present study cannot identify the differ- duction of 262 litres in 142 days. The milk protein percent- ence between alleles B and B’ of CSN1S1 like by Caroli age was estimated to 3.9 percent (Traoré, Nantoumé and et al. (2007). More investigations using other methodology Diarra, 2005). (PCR–SSCP, PCR–RFLP and DNA sequencing) are necessary to better genetic characterization of Malian It appears that the Guéra goat carried the allele (E) with a goat breeds. frequency of 0.14. This allele is widely identified in some European dairy goats (Feligini et al., 2005) such as the Alpine goat and in the Red Sokoto at a low frequency (Caroli et al., 2007). The allelic frequency of E in our Conclusions study is higher than those reported by Missohou, Talaki and Maman Laminou (2006) with 0.11. Genetic diversity is very important for animal breeding. Our study revealed polymorphism in αS1 casein gene in Studies on Spanish dairy goats, especially those of the Guéra goat. Allele B and the intermediate allele E were Canary Islands breeds showed a high frequency of the found in this breed. For the Sahel goat, no polymorphism allele E (Jordana et al., 1996). A phylogeny relationship was identified, B was the unique allele. The heterozygosity between the Guéra goat and Spanish goat breed was not rates and the genetic diversity index (FST) in the two investigated in the present study. However, the relatively breeds were low. high milk production of Guéra could be explained by the presence of the allele E in this breed. The genetic diversity The presence of allele E in Guéra goats could explain the observed was low, due to relatively low heterozygosity relatively high milk production of this breed and its origin in the Canary Islands. The information obtained for Guéra goat regarding the αS1 casein gene might be used in genet- α Table 1a. Genotype frequencies at the locus S1 casein (exon 19) ic improvement schemes of local goats in Mali. in Guéra and Sahel goat

Guéra goat Sahel goat Genotypes Number Frequencies Number Frequencies Acknowledgements

BB 41 0.77 48 1 The authors express their gratitude to “Institut d’Economie BE 9 0.17 0 0 ” EE 3 0.06 0 0 Rurale du Mali for the help and support of this research.

α Table 1b. Allele frequencies at the locus S1 casein (exon 19) in References Guéra and Sahel goat

Allele Angiolillo, A., Yahyaoui, M.H., Sanchez, A., Pilla, F. & Folch, J.M. 2002. Characterization of a new genetic variant in the caprine Breed B E κ-casein gene. Journal of Dairy Science 85: 2679–2680.

Guéra goat 0.86 0.14 Barbieri, M.E., Manfredi, E., Elsen, J.M., Ricordeau, G., Bouillon, J., Sahel goat 1.00 0.00 Grosclaude, F., Mahe, M.F. & Bibe, B. 1995. Effet du locus de la both 0.93 0.07 caséine alpha S1 sur les performances laitières et les paramètres génétiques des chèvres Alpines. Genetic Selection Evolution 27: 437–450. Genetic polymorphism of αS1 casein in Guéra and Sahel goat 83

Boulanger, A., Grosclaude, F. & Mahe, M.F. 1984. Polymorphisme des Nantoumé, H., Traoré, D. & Diarra, C.H.T. 2005. Mise au point de

caséines αs1 et αs2 de la chèvre (Capra hircus). Genetic Selection techniques d’amélioration des productions de lait, de viande et de Evolution 16: 157–176. laine des petits ruminants. Rapport final de Recherche, 11e Commission Scientifique du Comité National de Recherche Caroli, A., Chiatti, F., Chessa, S., Rignanese, D., Ibeagha-awemu, Agricole, Novembre 2005, Bamako. E.M. & Erhardt, G. 2007. Characterization of the casein gene complex in West African goats and description of a new αs1-casein poly- Ramunno, L., Longobardi, E., Pappalardo, M., Rando, A., Di morphism. Journal of Dairy Science 90: 2989–2996. Gregorio, P., Cosenza, G., Mariani, P., Pastore, N. & Masina, P. Cisse, M., Fall, Y. & ly, I. 2005. Performances laitières et état nutrition- 2001. An allele associated with a non-detectable amount of alpha – nel des chèvres du Sahel conduites sur parcours naturels: relations S2 casein in goat milk. Animal Genetics 32: 19 26. avec la croissance des chevreaux. In Proceedings of the Third Soares, M.A.M., Rodrigues, M.T., Mognol, G.P., Ribeiro, L.F.C., Biennial Conference of the African Small Ruminant Research Silva, J.L.C. & Brancalhão, R.M.C. 2009. Polymorphism of Network – UICC, Kampala, Uganda, 5–9 December y1994. alphas1-casein gene in a dairy goat herd in the southeastern region Feligini, M., Frati, M., Cubrik, V.C., Brambilla, A., Parma, P., Curik, of Brazil. Revista Brasileira de Zootecnia 38(6). I., Greppi, G.F. & Enne, G. 2005. Caprine alpha S1 casein poly- Traoré, D., Nantoumé, H. & Diarra, C.H.T. 2005. Amélioration de la morphism, characterization of A, B, E and F variants by means of production laitière par l’introduction de la chèvre Guéra. Rapport various biochemical and molecular techniques. Food Technology final de Recherche, 11e Commission Scientifique du Comité – Biotechnology 43(2): 123 132. National de Recherche Agricole, Novembre 2005, Bamako. Grosclaude, F., Mahe, M.F., Brignon, G., Di stasio, I. & Jeunet, R. Traoré, D., Nantoumé, H. & Diarra, C.H.T. 2007. La chèvre Guéra. 1987. A Mendelian polymorphism underlying quantitative variations Institut du Sahel; les Monographies Sahéliennes ISBN: α – of goat s1 casein. Genetic Selection Evolution 19(4): 399 412. 2-912693-48-9. Jordana, J., Amills, M., Diaz, E., Angulo, C., Serradilla, J.M. & Truett, G.E., Heeger, P., Mynett, R.L., Truett, A.A., Walken, T.A. & α Sanchez, A. 1996. Gene frequencies of caprine S1-casein polymorph- Warmen, M.L. 2000. Preparation of PCR quality mouse ism in Spanish goat breeds. Small Ruminant Research 20: 215–221. genomic DNA with hot sodium and tris (Hotshot). Biotechnics 29 Martin, P. 1993. Polymorphisme génétique des lactoprotéines caprines. (1): 52–54. Lait 73: 511–532. Vacca, G.M., Ouled Ahmed Ben Ali, H., Carcangiu, V., Pazzola, M. Missohou, A., Talaki, E. & Maman Laminou, I. 2006. Diversity and & Dettori, M.L. 2009. Genetic structure of the casein gene cluster genetic relationships among seven West African goat breeds. in tne Tunisian native goat breed. Small Ruminant Research 87(1): Asian-Australasian Journal of Animal Sciences 19(9): 1245–1251. 33–38. Animal Genetic Resources, 2014, 54, 85–91. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000519

Estimación de heredabilidad de la curva de crecimiento en el borrego de raza Chiapas en México

A.C. Méndez-Gómez1, R. López-Ordaz2,4, M. Peralta-Lailson1, R. Ulloa-Arvizu2, P. Pedraza-Villagómez1, F.J. Ruiz-López3, J.M. Berruecos-Villalobos2, C.G. Vásquez-Peláez2 1Centro Universitario de Investigación y de Transferencia de Tecnología, Universidad Autónoma de Chiapas, Campus III, San Cristóbal de las Casas, México; 2Facultad de Medicina Veterinaria y Zootecnia, Universidad Nacional Autónoma de México, México D.F.; 3Centro Nacional de Fisiología y Mejoramiento Animal, Instituto Nacional de Investigaciones Forestales y Agropecuarias, SAGARPA, Ajuchitlán Qro, México; 4Dirección actual: Universidad Autónoma Metropolitana, Unidad Xochimilco, México D.F.

Resumen Se utilizaron los registros genealógicos y de pesos mensuales de 790 ovinos de la raza Chiapas nacidos de 1991 a 2004 con objeto de caracterizar la curva de crecimiento y estimar componentes de (co)varianza de los estimadores (A, k y b) de la función de Gompertz. La estructura del pedigrí consistió por 790 ovinos, incluyendo 45 sementales y 379 madres. La estimación de los componentes de (co) varianza de cada uno de los estimadores de crecimiento se realizó utilizando un modelo animal univariado con la metodología de máxima verosimilitud restringida (REML). Los valores de heredabilidad estimados fueron: A (peso asintótico) 0.21 ± 0.06; b (constante de integración) 0.25 ± 0.07; k (tasa de madurez) 0.16 ± 0.06; edad a la inﬂexión se estimó como (ln(b)/k) 0.24 ± 0.07 y velocidad de crecimiento absoluto estimada como [k*(0.368*A)* ln(A/0.368*A)] 0.22 ± 0.07. Los machos muestran una velocidad de crecimiento 24% más rápida que las hembras (P < 0.05), siendo el crecimiento absoluto de 59 ± 2 g y 44 ± 2 g por día para machos y hembras respectivamente. La raza Chiapas es un borrego de tamaño pequeño con lenta velocidad de crecimiento debido probablemente a su adaptación a las condiciones ambientales donde es utilizado. La curva de crecimiento puede ser modiﬁcada a través de selección.

Palabras clave: heredabilidad, función de Gompertz, curvas de crecimiento, crecimiento absoluto, borrego Chiapas

Summary Genealogical and live body weight monthly records from 790 sheep of the Chiapas breed born between 1991 and 2004 were used to characterize the growth curve and estimate (co)variance components for the parameters (A, k and b) of the Gompertz function. The pedigree structure consisted of 790 sheep, including 45 rams and 379 ewes. Estimation of (co)variance components for each growth parameter was achieved using a univariate animal model with the restricted maximum likelihood (REML) method. Estimated values of heritability were: A (adult weight) 0.21 ± 0.06; b (integration constant) 0.25 ± 0.07; k (maturity rate) 0.16 ± 0.06; age at the inﬂexion estimated as (ln(b)/k) 0.24 ± 0.07; and absolute growth rate as [k*(0.368*A)* ln(A/0.368*A)] 0.22 ± 0.07. The growth of males was 24% (P < 0.05) faster than that of females; adult weight was 25 ± 0.7 kg for males and 23 ± 0.7 kg for females; absolute growth was 59 ± 2 g and 44 ± 2 g per day for males and females respectively. The Chiapas breed of sheep is a small animal with a slow growth curve, probably due to its adaptation to the environmental conditions where it lives. Growth curve can be modiﬁed through selection.

Keywords: heritability, Gompertz function, growth curves, absolute growth, Chiapas sheep

Résumé Archives généalogiques ont été utilisés et 790 pesos mensuels course de moutons Chiapas nés de 1991 à 2004 afin de caractériser la courbe de croissance et composants estimation de (co)estimateurs de la variance (A, K et b) de la fonction Gompertz. La structure pedigree composé de 790 moutons, dont 45 étalons et 379 mères. Estimation des composantes de (co)variance de chacun des estimateurs de croissance a été réalisée en utilisant un modèle animal univariée méthodologie REML (MVR). Les valeurs d’héritabilité estimées sont les suivantes: A (poids asymptotique) 0,21 ± 0,06, b (constante d’intégration) 0,25 ± 0,07, k (taux de maturité) 0,16 ± 0,06; âge d’inflexion a été estimée à [(ln (b) / k] 0,24 ± 0,07 et le taux de croissance absolue estimé que [k*(0,368*A)* ln(A/ 0,368*A)] 0,22 ± 0,07. mâles montrent un taux de croissance de 24% plus vite que les femelles (P < 0,05), avec la croissance absolue de 59 ± 2 g à 44 ± 2 g par jour pour les mâles et les femelles. Chiapas est une race de moutons de faible taux de croissance faible sans doute en raison de leur adaptation aux conditions environnementales où est utilisé. courbe de croissance peut être modifié par sélection.

Mots-clés: héritabilité, la fonction de Gompertz, courbes de croissance, croissance absolue, moutons Chiapas

Correspondencia: Dr Carlos Vásquez, Facultad de Medicina Veterinaria y Zootecnia, Universidad Nacional Autónoma de México, México D.F. email: [email protected]

85 86 A.C. Méndez-Gómez et al.

Introducción 1984; Pedraza et al.,1992; Perezgrovas y Castro, 2000; López et al., 2012); sin embargo, se desconoce la curva En la región de los Altos de Chiapas, México, se estableció y velocidad de crecimiento. López et al .(2012) estimaron desde el siglo XVI en la población indígena un sistema de heredabilidades de 0.09 a 0.34 para peso al nacimiento y producción ovina que prevalece casi sin cambios hasta de 0.09 a 0.16 para peso ajustado al destete (90 días), nuestros días. Esta población de borregos tiene su origen con una correlación genética entre estos pesos de 0.81 ± de las razas Churra, Manchega y Lacha y es conocida 0.18, sin que exista información en edades y pesos como raza Chiapas, con tres biotipos: Blanca, Negra y posteriores. Café (Figura 1) (Pedraza et al., 1992;FAO,2009). La estimación de la curva de crecimiento del nacimiento Estos ovinos tienen importancia cultural, social y hasta la edad adulta permite identificar etapas en el desar- económica por la producción de lana en la población rollo de esta raza, pudiendo establecer programas de man- Tzotzil de origen maya (Perezgrovas y Castro, 2000). ejo y alimentación así como estimar la tasa de crecimiento Son animales rústicos con partos anuales generalmente absoluto, la tasa relativa de maduración y peso adulto sencillos; el empadre se realiza con monta natural teniendo (Fitzhugh, 1976), e identificar animales genéticamente presente al macho durante todo el año, a pesar de existir sobresalientes a las condiciones ambientales. una marcada estacionalidad con partos al término de la Diferentes funciones no lineales han sido utilizadas para – época de lluvias (octubre noviembre). La trasquila se rea- describir y ajustar la curva de crecimiento en diferentes – liza cada 6 meses (abril octubre). Los rebaños son especies, entre las cuales se encuentra la función de pequeños de entre 8 a 10 animales por comunero y son alo- Gompertz que tiende a un mejor ajuste con respecto a fi jados en corrales de madera pequeños jos o móviles. Los otros modelos (Özdemir y Dellal, 2009; Forni et al., ovinos son llevados a pastorear en zonas comunales y a pie 2009; Renne et al., 2003; Delgadillo et al., 2009, de carretera donde consumen pastos (Sporobolus sp., Martínez et al., 2010). Esta función presenta dos Stipa ichu, Avena fatua, Cynodon dactylon, Pennisetum parámetros con interpretación biológica: A supone que el clandestinum, Tripogandra spp., Trifolium amabile, peso tiende a un valor final o asintótico con el tiempo Melilotus alba) y plantas nativas (Eupatorium mairetianum, (peso adulto) y k es la tasa de crecimiento la cual dismi- Eupatorium ligustrinum, Salvia spp., Phytolaccaceae spp., nuye monotónicamente a medida que aumenta el peso Rulus trilobus, Sambucus mexicana). El agua de bebida es hacia la madurez, mientras que el parámetro b está relacio- proporcionada de forma individual a cada animal con una nado con las condiciones iniciales (Blasco, 1999). cubeta procurando evitar que los animales beban de los aguajes o cauces de ríos a fin de evitar parasitosis. El objetivo de este estudio fue caracterizar la curva de cre- Cuando los dueños tienen oportunidad suplementan con cimiento utilizando la función de Gompertz desde el naci- rastrojo de maíz. Los animales enfermos son tratados con miento hasta la edad madura y la estimación de las (co) herbolaria tradicional (Pedraza et al., 1992; Perezgrovas varianzas de los componentes de la función en ovinos de y Castro, 2000). Por su respuesta reproductiva a estas con- la raza Chiapas que permitan ayudar a establecer sistemas diciones ambientales se le considera a la raza de borregos de manejo, alimentación y genéticos en esta raza. Chiapas como un recurso zoogenético de la región. Actualmente, la Universidad Autónoma de Chiapas conserva un núcleo de esta raza con las tres variedades, con apareamientos controlados exclusivamente dentro de Materiales y métodos estos biotipos, con el objetivo de caracterizar, evaluar y Se utilizó la información de los corderos nacidos entre 1991 apoyar la producción ovina de la región. y 2004, del rebaño de borrego Chiapas localizado en el El borrego Chiapas es una raza pequeña con pesos al naci- Centro Universitario de Investigación y Transferencia de miento de 2.5 kg y de 26 a 30 kg en la edad adulta en hem- Tecnología de la Universidad Autónoma de Chiapas bras y machos respectivamente (Perezgrovas y Pedraza, (CUITT-UNACH) ubicado en el municipio de Teopisca,

Figura 1. Ovino de la raza Chiapas en sus variedades Blanca, Negra y Café. (a) Variedad blanca Borrego Chiapas (Icsat en lengua Tzotzil) Blanco con manchas negras alrededor de los ojos, el hocico y las orejas (b) Variedad negra o Chamula (Saciol en lengua Tzotzil) Negro de piel y lana, con manchas blancas en la región frontal de la cabeza y la punta de la cola. (c) Variedad Cafó (Mesha en lengua Tzotzil) El color de la lana es blanco cremoso con tres variedades marrón, negro y dorado. Fuente: Adaptado de Perezgrovas y Castro, 2000. Estimación de heredabilidad de la curva de crecimiento en el borrego Chiapas 87

16°32′24″ de latitud norte y 92°28′19″ de longitud oeste lineal (GLM, SAS, 1993) incluyendo los efectos fijos y y a una altitud de 1780 msnm. El clima es templado las interacciones dobles, quedando en el modelo final los subhúmedo C(w2) (w) (clasificación de Köeppen), la efectos de año de nacimiento, tamaño de la camada, temperatura media anual oscila entre los 14 y 18°C con edad de la madre y sexo (P < 0.05); mientras que las inter- heladas de noviembre a febrero y una precipitación pluvial acciones dobles se eliminaron al no presentar significancia de 1200 mm anual de abril a octubre (Nahed, 1999). El estadística (P > 0.15). La estimación de los componentes CUITT-UNACH, tiene una extensión de 50 ha de praderas; de (co)varianza de los parámetros de crecimiento predi- 10 ha son dedicadas para la siembra de maíz y casi una ha chos a partir de la función de Gompertz se realizó utili- para instalaciones donde se localizan los corrales rústicos zando un modelo animal univariado bajo la metodología construidos con madera, y techo de lámina de zinc, sala de máxima verosimilitud restringida (REML) con el pro- de ordeño, bodega y oficina. Los borregos se mantienen grama ASReml versión 2.0 (Gilmour et al., 2006) a partir en un sistema mixto con pastoreo diurno de 6 a 8 horas de un modelo mixto representado como en pradera compuesta principalmente de zacate kikuyu (Pennisetum clandestinum) con encierro nocturno comple- Y = Xb + Za + e mentada con mazorca de maíz, harina de soya y sales donde Y es el vector N × 1 observaciones de los estima- minerales. El manejo sanitario del rebaño consistió en dores de los parámetros A, b y k del modelo de una rotación de principios activos para el control de enfer- fl medades gastroentéricas y respiratorias. En la época de Gompertz, peso y edad a la in exión, o velocidad de crecimiento absoluto; β es el vector de efectos fijos ambien- empadre se formaron lotes de 10 hembras por cada semen- tal, con una duración de 60 días de inicios de noviembre a tales de número de parto, tamaño de la camada, sexo; X finales de diciembre de 1991 a 1997; a partir de 1998 hasta es la matriz de incidencia que relaciona los efectos ambientales con las características de crecimiento; a es el vector 2004, las hembras fueron inducidas al estro con el empadre a principios de marzo; los nacimientos fueron desde agosto del valor genético aleatorio de los animales; Z es la matriz a finales de septiembre. El destete se realizó cuando los cor- de incidencia que relaciona los efectos aleatorios de los animales con las observaciones; e es el vector de efectos deros tenían 3 meses de edad aproximadamente. aleatorios residuales. Los supuestos en el modelos fueron Para cada uno de los corderos se registró la identificación E[y]=Xβ yE[e]=0y del padre, de la madre, el número de parto del cual provenían, tamaño de la camada, año de parto, sexo, s2 a = A a 0 Var s2 fecha de nacimiento, edad y peso semanal del nacimiento e 0 IN e al destete, y mensual del destete hasta los 24 meses de edad. Eliminando del análisis aquellos registros que con- donde N es el número de registros de las características taron con menos de 5 pesajes posdestete quedando un estudiadas; A es la matriz de relaciones genéticas aditivas total de 790 animales y 15,067 mediciones. entre los animales; IN es la matriz de identidad de orden N.

Análisis estadístico Resultados

La caracterización de la curva de crecimiento se realizó uti- En el Cuadro 1 se presentan las medias de cuadrados lizando la función de Gompertz (Proc NLIN, SAS, 1993), mínimos de los efectos ambientales sobre los estimadores calculando para cada animal los estimadores de los de la curva de crecimiento. El tamaño de la camada y parámetros de la función número de parto del que provenían los animales, no influyeron en A (P > 0.05), mientras que el valor estimado −be−kt yt = Ae en los machos fue de 25.4 kg (7.9%) mayor (P < 0.05) que en las hembras (24.8 kg). El parámetro b fue afectado por donde yt representa el peso corporal a la edad t (en días); todos los efectos fijos incluidos en el modelo (P < 0.05). A es el valor estimado del peso asintótico cuando la edad tiende a infinito, interpretado como peso adulto o maduro; El parámetro k no se vio influenciado por tamaño de la b es la constante de integración; k es la relación y/A que camada (P > 0.05); los machos tuvieron una mayor tasa representa la tasa de crecimiento (tasa de madurez); y e de madurez (24%) que las hembras (P < 0.05), mientras es el número de Euler 2.7182818. que los corderos provenientes de madres primíparas fl tuvieron el valor más bajo (0.0056), lo que representa La relación (ln(b)/k) estima la edad a la in exión, y ky ln del 80 a 87% de lo alcanzado por los corderos prove- (A/y) la velocidad de crecimiento absoluto al punto de nientes de hembras multíparas (P < 0.05). El peso a la inflexión, donde el valor de y se tomó como el peso a la fl fl in exión fue alcanzado 30 días antes en los corderos de in exión que es 0.368. parto sencillo comparados con los de parto múltiple (P < Para el cálculo de los efectos ambientales sobre los estima- 0.05); los machos alcanzaron el peso a la inflexión 8% dores de la curva de crecimiento se utilizó un modelo más rápido que las hembras (P < 0.05); los corderos de 88 A.C. Méndez-Gómez et al.

Cuadro 1. Medias de cuadrados mínimos y errores estándar para las características de crecimiento de acuerdo al tamaño de la camada, sexo y número de parto en borregos de la raza Chiapas

Efectos N Peso adulto Constante de Tasa de Peso de Edad de Velocidad de crecimiento (kg) (A) integración (b) madurez (k)* inﬂexión (kg) inﬂexión (días)† absoluto (g/día)††

Tamaño de la camada Sencillo 736 25.0 ± 0.5a 1.89 ± 0.02a 65 ± 2a 9.2 ± 0.1a 101.8 ± 2a 55 ± 4a Múltiple 54 23.9 ± 1.2a 2.10 ± 0.06b 56 ± 6a 8.7 ± 0.4a 131.5 ± 7b 48 ± 4a Sexo Macho 391 25.4 ± 0.7a 2.09 ± 0.04a 67 ± 4a 9.3 ± 0.2a 111.5 ± 2a 59 ± 2a Hembra 399 23.5 ± 0.7b 1.90 ± 0.04b 54 ± 4b 8.6 ± 0.2b 119.5 ± 3.b 44 ± 2b Número de Parto Primíparas 296 24.7 ± 0.7a 2.03 ± 0.04a 56 ± 4a 9.1 ± 0.2a 121.6 ± 3a 47 ± 2a Segundo 206 23.7 ± 0.8a 1.91 ± 0.04b 64 ± 4b 8.7 ± 0.3a 106.9 ± 3b 53 ± 2b >2 288 24.8 ± 0.8a 2.08 ± 0.04a 62 ± 4b 9.1 ± 0.2a 118.0 ± 3b 55 ± 2b

*1×10−4 † ln(b)/k †† ky ln (A/y), tomando y = 0.368*A. 1 × 10−3 a, b, c Valores con diferente literal dentro del efecto son medias diferentes (P < 0.05)

segundo parto tuvieron la menor edad a la inflexión (107 La estimación de heredabilidad para peso a la inflexión fue días), que representa un 13.7% menos que los corderos 0.21 ± 0.066; para edad a la inflexión, 0.24 ± 0.072 y para de primíparas y 10 % menos de las de tres o más partos velocidad de crecimiento absoluto, 0.22 ± 0.065. (P < 0.05). La Figura 2 muestra la curva de crecimiento y la tasa de crecimiento absoluto por sexo; la mayor tasa de creci- Discusión miento de los machos fue de 59 g/día (punto de inflexión), mientras que en las hembras la ganancia fue Con respecto al peso adulto, los valores encontrados en de 44 g/día (P < 0.05). Por otro lado, las crías de hembras este estudio son inferiores a lo observado en razas de primer parto tuvieron una tasa de crecimiento absoluto españolas Churra, Manchega y Lacha (Pedraza et al., menor que los corderos de dos o más partos (P < 0.05). 1992), así como con otras razas locales en otras latitudes (Lewis et al., 2002; Topal et al., 2004; Lambe et al., En general, los factores tamaño de la camada, número de 2006; Malhado et al., 2009; Braga et al., 2006; Ozder fi parto de la hembra y sexo del cordero, fueron signi cativos et al., 2009; Kubuc y Eyduran, 2009; Darkiran et al., en la variación de la curva de crecimiento del ovino 2010; Solomon et al., 2010), pero semejantes a lo infor- Chiapas. mado por Perezgrovas y Pedraza (1984): 25 y 28 kg en En el Cuadro 2 se presentan los componentes de varianzas hembras y machos respectivamente en esta misma raza aditiva y fenotípica así como los valores de heredabilidad de ovino Chiapas, siendo esto indicativo de que la estimados de los parámetros de la función de Gompertz población de ovino de raza Chiapas no ha sido seleccio- siendo para A (peso adulto) 0.21 ± 0.066, b (Constante de nada para características de crecimiento, ya que es una integración) 0.25 ± 0.068 y k (tasa de madurez) 0.16 ± 0.05. raza que es destinada para la producción de lana (Perezgrovas y Castro, 2000) y que el sistema de producción permanece prácticamente sin cambios (López et al., 2012). El modelo de Gompertz para peso maduro (A), mostró una subestimación de 1.5 kg aproximadamente, la cual ha sido también observada por otros autores (Lambe et al., 2006; Malhado et al., 2009), posiblemente debido a la cantidad de tejido graso, que es muy variable a lo largo del estado adulto del animal (Blasco, 1999), y a la disponibilidad de forraje a través del tiempo. También hay referencias que mencionan que la subestimación del peso maduro se debe únicamente al uso de datos realizados en la etapa temprana (Lambe et al., 2006). Figura 2. Pesos estimados con la funcion de Gompertz y velocidad de crecimiento absoluto (V.A.) para hembras y machos en el borrego de la raza El peso adulto en los machos fue 7.9% superior al de las Chiapas. hembras, siendo un porcentaje menor a lo observado en Estimación de heredabilidad de la curva de crecimiento en el borrego Chiapas 89

Cuadro 2. Componentes de varianzas y heredabilidades estimadas (h2 ± e.e) para las características de crecimiento en borregos Chiapas

Característica Varianzas h2 ± e.e

Aditiva Residual Fenotípica

Constante de integración (b) 0.04986 0.143 0.1929 0.25 ± 0.068 Peso adulto (A) 13.57 50.91 64.49 0.21 ± 0.066 Tasa de madurez (k) 0.000026 0.000136 0.00016 0.16 ± 0.059 Peso inflexión 1.8432 6.8915 8.735 0.21 ± 0.066 Edad de inflexión 2751.36 8565.17 11316.53 0.24 ± 0.072 Velocidad de crecimiento absoluto 0.00993 0.0357 0.04568 0.22 ± 0.065 razas de carne como la Suffolk (Lewis et al., 2002). El en borregos Chiapas tienen valores de A menores que hecho de que no existió efecto del tamaño de la camada los encontrados en otras razas, consecuentemente el peso y número de parto (P > 0.05) es explicado debido a que alainflexión también resultó menor. Sin embargo, como el peso adulto del animal está muy alejada de los efectos el valor de k también fue menor, dando resultado que la maternos que existen en el ambiente uterino y en la lactan- edad a la inflexión fuera superior a lo que presentan las cia (López et al., 2012), los resultados de este trabajo coin- otras razas (Lewis et al., 2002; Topal et al., 2004; cide con algunos estudios (Malhado et al., 2009), aunque Lambe et al., 2006; Malhado et al., 2009; Braga et al., otros autores han registrado un efecto del número de 2006; Ozder et al., 2009; Kubuc y Eyduran, 2009; parto sobre el peso maduro de sus crías (Bathaei y Darkiran et al., 2010; Solomon et al., 2010). Leroy, 1996; Lewis et al., 2002; McManus et al., 2003). La velocidad absoluta de crecimiento en el borrego El parámetro b se considera como una constante de Chiapas se presenta por arriba de los 50 g/día. Los machos integración sin significado biológico particular y con alta tuvieron una velocidad absoluta de crecimiento 1.34 veces correlación con el parámetro A (Blasco, 1999); de tal mayor que las hembras (P < 0.05) y 8 días antes (P < 0.05); modo que la expresión y0 = A exp (-b), puede ser interpre- los corderos provenientes de parto simple mostraron una tada como un estimador del peso al nacimiento; esta función velocidad de crecimiento similar que aquellos proveniente estimó el peso al nacer de 3.14 kg para machos y de 3.51 kg de parto múltiple (P > 0.05). La mayor velocidad de creci- para hembras, mostrando una sobrestimación en corderos de miento se da en la fase predestete, donde el cordero es la raza Chiapas; Perezgrovas y Castro, (2000) y López et al. dependiente del ambiente materno dado básicamente por (2012) observaron pesos al nacer de 2.15 y 2.27 kg para la producción de leche. Las bajas tasas de crecimiento hembras y machos de esta misma raza. Este sesgo se observadas después del destete indican que el manejo puede deber a la inconsistencia de la función. nutricional de estos animales no es el óptimo. En la mayoría de los estudios no se presenta el valor La estimación de componentes de varianza para los encontrado de b, sin embargo, hay informes en razas del parámetros de la función de Gompertz en este estudio para medio oriente (Topal et al., 2004) y brasileñas (Malhado la raza Chiapas mostraron un rango entre 0.16 y 0.25. Con et al., 2009) que presentan valores alrededor de 2, seme- respecto al peso adulto o maduro (A), Stobart et al.(1986) jante a la encontrado en el borrego Chiapas por López estimaron valores de 0.57 con una edad a la madurez de et al.(2012) y en este estudio. cinco años en razas Rambouillet, Targhee y Columbia, superiores a lo observado en este estudio (0.21) a los dos En este estudio, los valores de la tasa de madurez son años de edad. Lewis et al.(2002), utilizando animales de menores hasta en un 50% de los registrados en otras razas la raza Suffolk, estimaron una heredabilidad de 0.37 en el ovinas locales (Topal et al., 2004; Kubuc y Eyduran, mismo parámetro, y Lambe et al.(2006) estimaron hereda- 2009;Malhadoet al., 2009) y solo ligeramente inferiores bilidades de 0.427 con la función de Richards y 0.871 con la a los encontrados en la West African Dwarf (Gbangboche función de Gompertz para la raza Texel. En estudios con et al., 2008), indicando que la raza de ovinos en la raza otras razas locales, como la Mehraban (Bathaei y Leroy, Chiapas es de lento crecimiento, lo que puede estar asociado 1996), Scottish Blackface (Lambe et al., 2006) y Horro a su adaptabilidad al medio ambiente de la zona. Como se (Solomon et al., 2010), los valores de heredabilidad estima- esperaba, los machos presentaron una tasa de madurez 1.24 dos variaron entre 0.29 a 0.52. veces mayor que la de las hembras (P < 0.05); resultados En cuanto a la tasa de madurez (k), el valor de heredabili- similares se han observado en razas locales brasileñas dad estimado en este estudio fue similar a lo reportado en (Malhado et al., 2009). razas Santa Inés y Horro (Braga et al., 2006; Solomon El peso y la edad a la inflexión son los puntos donde la et al., 2010), mientras que Lambe et al.(2006) estimaron ganancia de peso llega al máximo y posteriormente tiende una heredabilidad de 0.06 en la raza Scottish Blackface; en asintóticamente a cero (Figura 2). En la función de contraste, Bathaei and Leroy (1996) encontraron un valor Gompertz este punto es fijo; el peso a la inflexión de 0.45 en la raza Mehraban Irani y Stobart et al. 1986 representó el 36.78% del peso a la madurez (A). Como de 0.32 en razas Rambouillet, Targhee y Columbia. 90 A.C. Méndez-Gómez et al.

En cuanto al parámetro b, que es considerada como una Referencias constante de integración que está ligado a las condiciones de inicio de la curva (Blasco, 1999), en la raza Suffolk, Abegaz, S., Van Wyk, J.B. & Oliver, J.J. 2010. Estimation of genetic Lewis et al.(2002) estimaron un valor de 0.38, superior and phenotypic parameters of growth and productivity of Horro – a lo encontrado en este estudio. sheep. Archiv für Tierzucht (1): 85 94. La forma en el crecimiento en ovejas se modifica con la Bathaei, S.S. & Leroy, P.L. 1996. Growth and mature weight of Mehraban Iranian fat-tailed sheep. Small Ruminant Research22: edad y su forma depende de efectos de raza y condiciones 155–162. ambientales como es la disponibilidad de alimento y man- Blasco, A. 1999. La descripción del crecimiento. Informe Técnico ejo. La función de Gompertz se ha sugerido como ade- Ocasional No. 6. Universidad Politécnica de Valencia. Departamento cuada, cuando no existen limitantes en las condiciones de Ciencia Animal. ambientales (Lewis et al., 2002). Braga, L.R.N., Vasques, V.L.C., Oliveira, L.A.M., Sousa, P.J.R. & Los resultados indican que es posible modificar a través de Oliveira, A.A. 2006. Parâmetros genéticos de características estima- selección la forma de la curva Selección directa en los das da curva de crescimento de ovinos da raça Santa Inês. Revista – parámetros A o b, tendrían como resultado un incremento Brasileira de Zootecnia 3:1012 1019. en el peso adulto así como en los pesos corporales al nacer Darkiran, I., Koncagul, S. & Bingol, M. 2010. Growth characteristics of e indirectamente en la producción de lana y la resistencia a indigenous norduz female and male lambs. Journal of Agricultural – parásitos (Safari and Fogarty, 2003). Sin embargo, tiene el Science 16: 62 69. inconveniente de incrementar el intervalo generacional. La Delgadillo, C.A.C., López, O.R., Montaldo, V.H.H., Berruecos, V.J. selección sobre la tasa de madurez (k) o la velocidad de M., Luna, A.E., Shimada, M.A. & Vásquez, P.C.G. 2009. Caracterización de la curva de crecimiento del ciervo rojo (Cervus ela- crecimiento, podría mejorar peso adulto, aunque habría phus scoticus) en el centro de México. Técnica Pecuaria en México fi que evaluar su impacto en la e ciencia dentro de este sis- 47: 117–123. tema de producción de los Altos de Chiapas. FAO. 2009. Domestic Animal Diversity Information System, http://www. fao.org/dad-is/. Food and Agriculture Organization of the United Nations, Rome. Conclusiones Fitzhugh Jr., H.A. 1976. Analysis of growth curves and strategies for altering their shape. Journal of Animal Science 42: 1036–1051. El borrego de la raza Chiapas muestra una curva de creci- Forni, S., Piles, M., Blasco, A., Varona, L., Oliveira, H.N., Lôbo, R.B. miento lento, el cual puede atribuirse a la adaptación al & Albuquerque, L.G. 2009. Comparison of different nonlinear func- medio ambiente. tions to describe Nellore cattle growth. Journal of Animal Science 87 – La variabilidad genética aditiva encontrada a partir de la (2): 496 506. función de Gompertz en el crecimiento de esta raza de ovi- Gbangboche, A.B., Glele-Kakai, R., Salifou, S., Albuquerque, L.G. & nos Chiapas, indica que es posible modificar la curva de Leroy, P.L. 2008. Comparison of non-linear growth models to describe the growth curve in West African Dwarf sheep. Animal crecimiento del borrego Chiapas pudiendo ser utilizada (7): 1003–1012. la estimación de velocidad o crecimiento máximo absoluto como criterio de selección. Gilmour, A.R., Gogel, B.J., Cullis, B.R. & Thompson, R. 2006. ASReml User Guode Release 2.0 VSN International Ltd, Hemel Hempstead, HP1 1ES, UK. Kubuc, M. & Eyduran, E. 2009. The determination of the best growth Agradecimientos model for Akkaraman and German Blackheaded Mutton x Akkaraman B1 crossbreed lambs. Bulgarian Journal of Agricultural Science 15: Se agradece al Consejo Nacional de Ciencia y Tecnología 90–92. (CONACYT-México) por la beca otorgada a la primera Lambe, N.R., Navajas, E.A., Simm, G. & Bünger, L. 2006. A genetic autora para realizar estudios de Maestría en Ciencias investigation of various growth models to describe growth of lambs of (FMVZ-UNAM); Proyectos UNAM-PAPIIT IN207707-3, two contrasting breeds. Journal of Animal Science 84: 2642–2654. UNAM-PAPIIT IN205710-3 y SAGARPA-CONACYT Lewis, R.M., Emmans, G.C. & Dingwall, W.S. 2002. A description of 2004-CO1-111/A1. the growth of sheep and its genetic analysis. Animal Science 74: 51–62. Esta investigación fue posible gracias al convenio de López, O.R., Olivera, I.V., Berruecos, J.M., Peralta, L.M., colaboración entre la Facultad de Medicina Veterinaria y Ulloa-Arvizu, R. & Vásquez, C.G. 2012.Genetic parameters for Zootecnia de la Universidad Nacional Autónoma de México birth and weaning weights in the local Chiapas sheep breed from Mexico, Revista Mexicana de Ciencias Pecuarias 3(1): 113–123. y el Centro Universitario de Investigación y Transferencia de Tecnología de la Universidad Autónoma de Chiapas. Martínez, C.A., Rodríguez, A.P., Jiménez, A. & Manrique, C. 2010. Descripción matemática de la función de Gompertz aplicada al crecimiento de animales. Revista de la Facultad de Medicina Veterinaria y de Zootecnia 57:76–80. Declaración de interés Malhado, C.H.M., Carneiro, P.L.S., Affonso, P.R.A.M., Souza, A.A. O. & Sarmento, J.L.R. 2009. Growth curves in Dorper sheep crossed Los autores declaran que no existen conflictos de interés en with the local Brazilian breeds, Morada Nova, Rabo Largo and Santa la presentación de este artículo. Inés. Small Ruminant Research 84: 16–21. Estimación de heredabilidad de la curva de crecimiento en el borrego Chiapas 91

McManus, C., Evangelista, C. & Fernandes, L.A.C. 2003. Curvas de Perezgrovas, R. & Pedraza, P. 1984. Ovinocultura indígena crescimento de ovinos Bergamasia criados no Distrito Federal. I. Desarrollo corporal del borrego Chiapas. Cuadernos de Revista Brasileira de Zootecnia 32: 22–27. Investigación 1 (Universidad Autónoma de Chiapas, Tuxtla Gutiérrez, México), pp. 1–13. Nahed, T.J. 1999. Alternativas para el desarrollo de sistemas de producción ovina sostenibles en los altos de Chiapas. Tesis doctorado, Renne, U., Langhammer, M., Wytrwat, E., Dietl, G. & Bünger, L. Facultad de Medicina Veterinaria y Zootecnia, Universidad Nacional 2003. Genetic-statistical analysis of growth in selected and unselected Autónoma de México, México D.F. mouse lines. Journal of Experimental Animal Science 42: 218–232. Özdemir, H. & Dellal, G. 2009. Determination of growth curves in Safari, A. & Fogarty, N.M. 2003. Genetic parameters for sheep produc- young angora goats. Tarim Bilimleri Dergisi 15 (4): 358–362. tion traits: estimates from the literature. Technical Bulletin 49, NSW Agriculture, Orange, Australia. Ozder, M., Sezenler, T., Onal, A.R. & Ceyhan, A. 2009. Genetic and non-genetic parameter estimates for growth traits in Turkish SAS Institute. 1993. SAS/STAT User’s Guide. In: 6.03 ed., Series SAS/ Merino lambs. Journal of Animal and Veterinary Advances STAT User’s Guide. SAS Institute, Cary, NC. (9):1729–1734. Stobart, R.H., Bassett, J.W., Cartwright, T.C. & Blackwell, R.L. Pedraza, P., Peralta, M. & Pérezgrovas, R. 1992. El borrego Chiapas: 1986. An analysis of body weights and maturing patterns in western una raza local mexicana de origen español. Archivos de Zootecnia 41 range ewes. Journal of Animal Science 63: 729–740. (extra): 355–362 Topal, M., Ozdemir, M., Aksakal, V., Yildiz, N. & Dogru, U. 2004. Perezgrovas, R. & Castro, G.H. 2000. El borrego Chiapas y el sistema Determination of the best nonlinear function in order to estimate tradicional de manejo de ovinos entre las pastoras tzotziles. Archivos growth in Morkaraman and Awassi lambs. Small Ruminant de Zootecnia 49: 391–403. Research 55: 229–232. Animal Genetic Resources, 2014, 54, 93–101. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000313

Caracterización genética de la población bovina criolla de la Región Sur del Ecuador y su relación genética con otras razas bovinas

L. Aguirre Riofrio1, G. Apolo2, L. Chalco2 y A. Martínez3 1Universidad Nacional de Loja, Centro de Biotecnología Reproductiva Animal (CEBIREA), Ecuador y aluno Doutorado em Biociência Animal, Universidade de São Paulo, Brasil; 2Tesistas, MVZ-UNL, 2012; 3Facultad de Veterinaria, Universidad de Córdoba, España.

Resumen Al no contar a la fecha con ningún estudio sobre la variabilidad genética de la población bovina criolla de la Región Sur del Ecuador (RSE), se ha procedido a su análisis y caracterización con fines de disponer de información técnica que permita planificar y desarrollar medidas tendientes a su conservación y utilización sustentable en este duro ecosistema en donde habita. Para ello se analizaron genéticamente 46 animales adultos, de una población de bovinos criollos de 114, agrupados por sus características fanerópticas en cuatro grupos, conocidos en el medio como: Negro Lojano (15 animales), Encerado (15), Colorado (9) y Pintado o Cajamarca (7). En el análisis de la variabilidad genética intrapoblacional, se obtuvo un Na de 228, un número promedio de alelos por locus de 8,14, y un PIC de 0,70 ± 0,10. Todos los 28 marcadores analizados son polimórficos y altamente informativos. Las He y Ho fueron de 0,74 ± 0,0178 y 0,676 ± 0,013, lo que indica la existencia de una alta diversidad molecular en la población estudiada. El HWE determina que el 32 por ciento de los alelos no se encuentran en equilibrio. El Gst fue de 4,5 ± 2,8 por ciento, y el Fst de 0,08 por ciento, lo que permite corro- borar que no existe diferenciación genética entre los cuatro grupos de esta población. Las medias del Fit yFis son de 8,89 y 8,92 por ciento, respectivamente. También se determinó el grado de distanciamiento genético entre esta población criolla con 18 razas taurinas y cebuinas. Los distanciamientos con todas las poblaciones superan el 0,10, las distancias menores están entre Ec:BC (0,1031), Ec:PAJ (0,1132), Ec:NAN (0,1302), Ec:VCA (0,1326) y Ec:FRI (0,1362). El mayor distanciamiento se observa entre Ec:GUZ (0,4725) y Ec: NEL (0,4624). Estos valores indican que los bovinos criollos de la RSE tienen un gran distanciamiento con las razas cebuinas y más bien sus troncos ancestrales están en las poblaciones ibéricas. Los resultados de variabilidad y distancias genéticas permiten disponer de herramientas para trazar estrategias que lleven a un manejo y conservación de la raza Criolla de la RSE y sugieren un manejo inde- pendiente de los cuatro grupos faneropticamente diferenciados a ser manejados en los distintos pisos altitudinales de la Región Andina del Ecuador.

Palabras clave: Bovino criollo de Ecuador, variabilidad genética, grado de distanciamiento, microsatélites

Summary In the absence to date of studies on genetic variability in the Creole cattle of the Southern Region of Ecuador, the analysis and characterization of this population was carried out in order to provide the technical information needed for the planning and development of measures aimed at the population’s conservation and sustainable use in the harsh ecosystem these animals inhabit. For this purpose, 46 adult animals, out of a population of 114 Creole cattle, were genetically analysed; on the basis of phaneroptical traits, they were classiﬁed into four groups known locally as Negro Lojano (15 animals), Encerado (15), Colorado (9) and Pintado or Cajamarca (7). The analysis of the intra-population genetic variability brought out a Na of 228, an average number of alleles per locus of 8.14 and a PIC of 0.70 ± 0.10. All 28 markers analyzed were polymorphic and highly informative. The He and Ho amounted to 0.74 ± 0.0178 and 0.676 ± 0.013, which is an indication of a high molecular diversity in the population studied. The HWE determines that 32 per cent of the alleles are not in equilibrium. The Gst amounted to 4.5 ± 2.8 per cent and the Fst was 0.08 per cent, which corroborates that there is no genetic differentiation among the four groups of this population. The average values for Fit and Fis were 8.89 and 8.92 percent, respectively. The degrees of genetic distance between this Creole population and 18 cattle and zebu breeds were also determined. The distance exceeded 0.10 for all populations, with the lowest distances being those between Ec:BC (0.1031), Ec:PAJ (0.1132), Ec:NAN (0.1302), Ec:VCA (0.1326) and Ec:FRI (0.1362). The greatest distances were observed between Ec:GUZ (0.4725) and Ec:NEL (0.4624). These values show that the Creole cattle of the Southern Region of Ecuador are very distant from the zebu breeds, and therefore their ancestors would be more likely to have originated from Iberian populations. The results concerning variability and genetic distance will enable strategies to be devised for managing and conserving the Creole breed of the Southern Region of Ecuador. These results also suggest that the four groups with phaneroptical differences should be managed independently at the different altitudes of the Andean Region of Ecuador.

Keywords: Creole cattle of Ecuador, genetic variability, degree of distance, microsatellites

Correspondence to: L. Aguirre, Universidad Nacional de Loja, Centro de Biotecnología Reproductiva Animal (CEBIREA), Ecuador. email: [email protected]

93 94 L. Aguirre et al.

Résumé Compte tenu du manque d’études sur la variabilité génétique des populations bovines créoles de la Région Sud de l’Équateur, le but de ce travail a été d’analyser et de caractériser ces bovins afindedisposerdel’information technique nécessaire à la planification et développement de mesures pour leur conservation et utilisation durable dans le dur écosystème où ils sont élevés. Pour ce faire, 46 animaux adultes d’une population de 114 bovins créoles ont été analysés génétiquement. Les bovins créoles sont classés, d’après certaines caractéristiques des phanères, en quatre types qui reçoivent localement le nom de: Negro Lojano (15 animaux), Encerado (15), Colorado (9) et Pintado ou Cajamarca (7). Il résulte de l’analyse de la variabilité génétique intra-populationnelle que le Na est de 228, le nombre moyen d’allèles par locus de 8,14 et le PIC de 0,70 ± 0,10. Il a de même été observé que les 28 marqueurs analysés sont tous polymorphes et très informatifs. La He et la Ho ont été de 0,74 ± 0,0178 et de 0,676 ± 0,013, ce qui indique l’existence d’une grande diversité moléculaire dans la population étudiée. D’après le HWE, le 32 pour cent des allèles ne se trouvent pas en équilibre. Le Gst aétéde 4,5±2,8pourcentetleFst de 0,08 pour cent, ce qui permet de confirmer qu’il n’existe pas de différenciation génétique entre les quatre groupes de cette population. Le Fit et le Fis ont été en moyenne de 8,89 et de 8,92 pour cent, respectivement. Le degré de d’éloignement génétique entre cette population créole et 18 races taurines et de zébu a aussi été déterminé. L’éloignement dépasse le 0,10 pour toutes les populations, les plus faibles distances étant celles entre Ec:BC (0,1031), Ec:PAJ (0,1132), Ec:NAN (0,1302), Ec:VCA (0,1326) et Ec:FRI (0,1362). Le plus grand éloignement a été observé entre Ec:GUZ (0,4725) et Ec:NEL (0,4624). Ces valeurs indiquent que les bovins créoles de la Région Sud de l’Équateur sont très éloignés des races de zébu, comme quoi leurs ancêtres seraient plutôt parmi les populations ibériques. Les résultats de variabilité et d’éloignement génétique vont aider à élaborer des stratégies d’élevage et de conservation de la race Créole de la Région Sud de l’Équateur. Ces résultats suggèrent de même une conduite indépendante, sur les différents niveaux d’altitude de la Région Andine de l’Équateur, des quatre groupes à différences phanériennes.

Mots-clés: Bovins créoles de l’Équateur, variabilité génétique, degré d’éloignement, microsatellites

Presentado: 13 Marzo 2013; aceptado: 8 Julio 2013

Introducción han logrado sobrevivir y desarrollarse en condiciones com- plejas, adquiriendo características propias que les permi- Según Lucena (1987), fue Sebastián de Belalcázar, proce- tieron adaptarse y prosperar en este hábitat. Al respecto, dente de Panamá en 1534, quien arribó al Ecuador en la De Alba (2009) y Aguirre et al. (2011) manifiestan que conquista del Reino de Quito, el que trajo los primeros estos animales poseen características de adaptación a bovinos Bos taurus, los cuales eran animales con una apti- estos ecosistemas, como: tolerancia al calor y sequedad, tud a más de la producción de carne y leche, de tracción de con temperaturas sobre los 26°C; estatura pequeña y carretas, función tan importante para aquella época en las gran fortaleza, lo que les permite caminar grandes distan- zonas alto andinas con un relieve tan irregular. Estos bovi- cias por pedregales y terrenos irregulares en busca de nos criollos han tenido un proceso de adaptación largo de alimento y abrevaderos; resistencia a los parásitos y exce- cinco siglos a los diversos climas y ambientes, primero en lente fertilidad e instinto materno. un proceso de selección natural y luego con la intervención Los bovinos criollos cumplen un rol importante en la vida del hombre, en la búsqueda de animales funcionales para de las comunidades campesinas marginales de la RSE, pues estos ecosistemas de montaña. Desde esa época el terri- constituyen una fuente de sustento y trabajo, su leche, su torio ecuatoriano estuvo poblado exclusivamente por bovi- carne, la venta misma de animales y su utilización como nos criollos (ibéricos naturalizados). Es a partir de los medio de tracción animal benefician estas regiones primeros años del 1900 (Pinzón, 1984), en que se inició geográficamente apartadas en donde la mecanización la introducción de bovinos extranjeros para la producción agrícola no es posible y la crianza de otros bovinos sería de carne y de leche por la creencia de que estas razas insostenible. son la mejor alternativa para los ganaderos, pues los resultados relativamente favorables obtenidos al inicio, fueron La presente investigación se orientó a realizar un estudio atribuidos a las aptitudes de las razas introducidas y no genético con el uso de marcadores moleculares para deter- al vigor híbrido que produce el mestizaje. Con el trans- minar la variabilidad intrapoblacional y el grado de curso del tiempo y los cruzamientos desordenados, se influencia de las razas ibéricas y cebuinas en esta han producido procesos de absorción de los animales crio- población de bovinos y promover de esta forma estrategias llos, y como consecuencia el número de estas poblaciones para el manejo y conservación de este importantísimo ha disminuido drásticamente. material genético de esta parte de los Andes del Ecuador. Sin embargo, a pesar de ello aún se reporta la existencia de ganado bovino criollo en algunos lugares apartados del Ecuador, como es en la RSE limítrofe con el norte del Materiales y métodos Perú, donde aún se pueden encontrar poblaciones de bovinos criollos en núcleos pequeños y heterogéneos, los mis- El análisis genético se realizó en 46 bovinos adultos de una mos que por sus particularidades fenotípicas y genotípicas población de 114 animales criollos, agrupados por sus Caracterización genética de la población bovina criolla de la Región Sur del Ecuador 95

Figura 1. Bovinos criollos ‘Negro Lojano’. características fanerópticas en cuatro grupos: 15 Negros respectivamente. El panel de 28 microsatélites empleados Lojanos (Figura 1, véase las páginas 00–00), 15 son: BM1314, BM8125, BM1818, CSSM66, ETH10, Encerados (Figura 2), 9 Colorados (Figura 3) y 7 Pintados INRA32, MM12, TGLA122, BM2113, CRSM60, ETH185, (Figura 4), provenientes de los sectores de Purunuma y HAUT27, HEL9, HEL13, ILSTS6, INRA23, INRA35, Nambacola en el cantón Gonzanama, provincia de Loja, INRA37, INRA63, SPS115, TGLA227, BM1824, ETH225, Ecuador (S 4°20′ / S 4°10′; W 79°30′ / W 79°0′5; Figura ETH3, HAUT24, ILSTS011, TGLA53, TGLA126 (FAO, 5). Los animales seleccionados fueron de ambos sexos y 2011). se evitó considerar para el análisis más de un animal proveniente de la misma familia. Se tomaron muestras de pelo de La caracterización genética se realizó de la siguiente man- la región terminal de la cola, que fueron identificadas y guar- era: Para el cálculo del PIC de cada microsatélite se aplicó dadas herméticamente hasta su procesamiento en el labora- la fórmula de Botstein et al. (1980), empleándose para ello el complemento ‘The Excel Microsatellite Toolkit’ (Park, torio del Departamento de Genética Molecular Aplicada de ® la Universidad de Córdoba, España, en donde se realizó el 2001), y utilizando el programa MS Excel 2003. análisis del ADN por medio de marcadores moleculares También se computarizaron los estadísticos F de Wright amplificados mediante la técnica de reacción en cadena de (1969) y Weir & Cockerham (1984), que corresponden: ɵ la polimerasa (PCR). Para la separación por tamaños de F=FIT, =FST yf=FIS, realizando un análisis factorial los fragmentos obtenidos mediante la PCR se sometieron de correspondencias mediante el programa Genetix estos a una electroforesis en gel de poliacrilamida en un v. 4.05 (Belkhir et al., 2003). Se calcularon las matrices secuenciador automático ABI 377XL; el análisis de los frag- de distancias genéticas Da (Nei & Chesser., 1983); para mentos y la tipificación alélica se realizó mediante los progra- ello se emplearon los análisis genéticos de 789 individuos mas informáticos Genescan Analysis 3.1.2 y Genotyper 2.5 pertenecientes a 19 razas taurinas y cebuinas del Consorcio

Figura 2. Bovinos criollos ‘Encerado’. 96 L. Aguirre et al.

Figura 3. Bovinos criollos ‘Colorado’.

Biobovis de la Red CONBIAND – Criollos RSE, Retinta, por Villalobos (2010) para las criollas panameñas Guabala Avileña, Rubia Gallega, Berrenda Colorada, Berrenda y Guaymi, y algunas razas ibéricas tales como Berrenda Negra, Mostrenca, Pajuna, Negra Andaluza, Canaria, Colorada, Berrenda Negra, Pajuna, Mostrenca, Retinta, Palmera, Frisona, Hereford, Pardo Suiza, Gyr, Brahman, Negra Andaluza, siendo comparables los resultados de Sindy, Guzera y Nelore – utilizando el programa Na nuestros, con los obtenidos por Dalvit et al. (2008) Populations 1.2.28 (Olivier, 2002: see http://bioinformat en cuatro razas italianas (Na = 8,2). También se determi- ics.org/tryphon/populations/). Con los valores obtenidos naron los alelos únicos (Au) en los cuatro grupos poblacio- se trazaron árboles de distancias, graﬁcando los mismos nales (datos no mostrados), destacando que 96 alelos son con el programa Treeview (Page, 1996) y con la matriz compartidos en los cuatro grupos, siendo el grupo de los de distancias Da de Nei, se construyeron ‘splitsgraph’ Negro Lojanos el que presenta el mayor número de Au mediante el algoritmo ‘Neighbor-Net’ con el programa (17) y que se correlaciona con una mayor variabilidad SplitsTree v 4.10 (Bryant & Moulton, 2004; Huson & alélica, en tanto que los Pintados poseen el menor Bryant, 2006). número de Au (5) y también la menor variabilidad. En general, los marcadores que presentaron los valores más elevados son los más recomendables en caso de Resultados y discusión asignación de nuevos individuos a esta población. De los 28 microsatélites analizados ninguno mostró alelos Variabilidad genética poblacional con frecuencias superiores a 0,95, por lo cual todos pueden En la Tabla 1, se observa que la población bovina de la considerarse polimórﬁcos y altamente informativos, excep- RSE presenta un total de 228 alelos, con una media de tuando el HEL13 y ILSTS011, que presentan un PIC medi- 8,14 ± 2,05 y un rango comprendido entre 5 (ILSTS6, anamente informativo; por tanto el 100 por ciento de estos INRA63, BM1824,INRA35) y 12 (HEL9, TGLA53) ale- marcadores son útiles para valorar la variabilidad genética los por locus. Estos valores son superiores a los reportados del bovino criollo de la RSE, es de destacar que el PIC

Figura 4. Bovinos criollos ‘Pintado’. Caracterización genética de la población bovina criolla de la Región Sur del Ecuador 97

Figura 5. Cantón Gonzanama en la provincia de Loja, Ecuador, área de ubicación de los bovinos criollos. medio de esta población (0,7) es ligeramente superior al (Menorquina), 0,590 y 0,512 (Lidia), 0,617 y 0,588 obtenido en la raza Guaymi panameña y la Limonero vene- (Pirenaica), 0,681 y 0,666 (Asturiana); igualmente se zolana, con 0,69 y 0,65 respectivamente (Villalobos, 2010; tiene información reportada en razas autóctonas de Villasmil et al., 2008). Portugal como la Alentejana, valores de 0,677 y 0,638, Mertolenga 0,696 y 0,747, Arauquesa 0,717 y 0,730 Los valores medios de Ho y He en la población criolla de (Mateus et al., 2004). En razas de Hungría y Austria la RSE muestran valores altos de 0,676 y 0,74 respectiva- como la Carinthian Blond y la Waldviertel Blond, se mente, existiendo únicamente cuatro marcadores genéticos reportan valores de 0,674, 0,662 y 0,654, 0,634 respectiva- que presentan un menor grado de heterocigosis: HAUT27, mente (Baumung et al., 2006). HEL13, ILSTS011 e INRA35, determinando que hay una buena variabilidad genética, si comparamos los mismos De los microsatélites analizados, nueve no se han mostrado en con los valores de He y Ho que se han obtenido en estu- equilibrio Hardy-Weinberg (P < 0.1). Estos son: BM2113, dios de caracterización genética que oscilan entre 0,755 HAUT27, BM1824, ETH225, ETH3, HAUT24, ILSTS011, y 0,895 en razas mexicanas (Quiroz, 2007); 0,461 y INRA35 y TGLA53, lo que representa el 32 por ciento del 0,694 en razas Argentinas (Martínez, 2008); 0,750 y total de alelos y revela que en estos animales se están operando 0,756 en el criollo Casanareño de Colombia (Sastre fuerzas que están cambiando las frecuencias genotípicas ya et al., 2003). Por otra parte Villalobos (2010) cita los estu- sea por deriva genética, selección o migración, y que pone dios hechos en seis razas autóctonas españolas que mos- de maniﬁesto que hay una ligera inestabilidad genética traron valores ligeramente inferiores de 0,564 y 0,570 en esta población. Estas desviaciones del equilibrio 98 L. Aguirre et al.

Tabla 1. Índices de variabilidad genética obtenida en los 28 loci Tabla 2. Coeficiente de variación genética (GST) y Estadísticos F analizados de la población criolla de la Región Sur del Ecuador. de Wright: coeficiente de consanguinidad dentro de la población (F ), coeficiente de consanguinidad total (F ) y coeficiente de Locus Na PIC Ho He HWE is it variación genética entre las subpoblaciones (Fst), por marcador BM 8125 7 0,64 0,61 0,70 0,23 en la población bovina criolla de la RSE. BM1314 9 0,70 0,78 0,75 0,91 Locus G F (θ)F(F) F (f) BM1818 7 0,62 0,63 0,69 0,92 st st it is CSSM66 10 0,84 0,85 0,87 0,42 BM8125 0,03 −0,02 0,12 0,13 ETH10 8 0,56 0,56 0,59 0,41 BM1314 0,03 −0,01 −0,04 −0,03 INRA 32 9 0,74 0,67 0,78 0,16 BM1818 0,01 −0,04 0,09 0,13 MM12 8 0,63 0,67 0,67 0,98 CSSM66 0,07 0,03 0,02 −0,01 TGLA122 11 0,76 0,70 0,79 0,13 ETH10 0,04 0,02 0,06 0,05 BM2113 8 0,69 0,55 0,73 0,00* INRA32 0,11 0,06 0,14 0,08 CRSM60 8 0,65 0,76 0,71 0,58 MM12 0,08 0,04 −0,01 −0,05 ETH185 8 0,72 0,77 0,76 0,92 TGLA122 0,03 0,01 0,14 0,13 HAUT27 9 0,75 0,50* 0,78 0,00* BM2113 0,09 0,03 0,26 0,23 HEL13 7 0,51 0,50* 0,54 0,16 CRSM60 0,04 −0,00 −0,08 −0,08 HEL9 12 0,74 0,80 0,78 0,19 ETH185 0,07 0,03 −0,00 −0,03 ILSTS6 5 0,76 0,83 0,80 0,95 HAUT27 0,03 −0,01 0,37 0,38 INRA 23 10 0,76 0,70 0,80 0,38 HEL13 0,02 −0,03 0,07 0,09 INRA 37 8 0,71 0,63 0,75 0,10 HEL9 0,01 −0,02 −0,05 −0,03 INRA63 5 0,59 0,72 0,66 0,84 ILSTS6 0,02 −0,02 −0,04 −0,01 SPS115 8 0,74 0,69 0,78 0,36 INRA23 0,05 0,00 0,12 0,11 TGLA227 10 0,84 0,89 0,87 0,15 INRA37 0,07 0,02 0,16 0,14 BM1824 5 0,70 0,76 0,75 0,05* INRA63 0,01 −0,03 −0,10 −0,07 ETH 225 9 0,78 0,76 0,82 0,02* SPS115 0,03 −0,01 0,11 0,12 ETH3 11 0,79 0,72 0,82 0,01* TGLA227 0,02 −0,02 −0,03 −0,01 HAUT24 7 0,78 0,62 0,82 0,03* BM1824 0,02 −0,02 −0,02 0,00 ILSTS 011 6 0,49 0,42* 0,52 0,01* ETH225 0,02 −0,02 0,07 0,09 INRA35 5 0,53 0,35* 0,58 0,00* ETH3 0,03 −0,01 0,13 0,14 TGLA53 12 0,85 0,77 0,88 0,00* HAUT24 0,05 −0,01 0,24 0,24 TGLA126 6 0,69 0,70 0,74 0,88 ILSTS011 0,09 0,04 0,20 0,17 Total 228 INRA35 0,06 −0,00 0,40 0,41 Promedio ± SD 8,1 ± 2,05 0,7 ± 0,1 0,68 ± 0,01 0,74 ± 0,02 TGLA53 0,08 0,02 0,11 0,09 TGLA126 0,03 −0,02 0,06 0,08 Na: número de alelos por locus; PIC: contenido de información Promedio±SD 0,04±0,03 0,001±0,025 0,09±0,13 0,09±0,12 polimórfica; Ho: heterocigosidad observada; He: heterocigosidad esperada; HWE: equilíbrio Hardy-Weinberg. Nota: Se destacan con * los marcadores con mayor grado de Nota: Se destacan con * los marcadores con menor grado de heterocigosis consanguinidad. y en desequilibrio Hardy-Weinberg

Hardy-Weinberg (HWE) pueden estar produciéndose aceler- lo que nos indica que la diferenciación genética entre los adamente debido a factores como los siguientes: cuatro grupos poblacionales es muy pequeña. Esta escasa diferenciación se corrobora con los valores de Fst (grado • en la selección y en el apareamiento de los animales hay de diferenciación genética entre las subpoblaciones) encon- intervención del propietario; trados en los cuatro grupos fenotípicamente definidos, con • fi existe migración de animales por la cercanía geográ ca un valor medio de Fst = 0,0008, valor muy inferior al encon- entre las poblaciones; trado en el ganado criollo mexicano con 0,033 (Quiroz, • no existe el criterio de tener núcleos de animales 2007), en los bovinos criollos panameños con 0,068 fenotípicamente homogéneos. (Villalobos, 2010) y entre las razas bovinas europeas con Vale destacar que los microsatélites ETH3, INRA35 y 0,068 y 0,07 (Cañón et al., 2001;Jordanaet al., 2003). TGLA53 también se encontraron en desequilibrio en los Vale indicar que si bien no hay diferencia genética entre bovinos criollos Patagónicos (Martínez, 2008) y los estas cuatro subpoblaciones, si se los debe manejar como microsatélites BM1824, HAUT24 y HAUT27 en la núcleos homogéneamente diferenciados por el color del población Guabala (Villalobos, 2010). pelaje, pues ello favorece el bienestar animal y la adaptación que tienen los mismos a los diferentes pisos altitudinales de la Región Andina, en donde el Negro Lojano Análisis intrapoblacional (Figura 1) y el Encerado (Figura 2) han ido evolucionando En la Tabla 2, se presentan los índices de variabilidad para un clima moderado por la altura, absorbiendo y mante- genética intrapoblacional considerando las cuatro subpobla- niendo más calor de los rayos solares para estas zonas frías y ciones del bovino criollo de la RSE. El valor de Gst (coefi- ventosas y no siendo por el color obscuro los más apropia- ciente de diferenciación genética) es entre 0,0091 para el dos para las zonas andinas bajas por el stress del calor y la locus BM1818 y 0,115 para el INRA32, y el valor medio mayor predisposición al ataque de los ectoparásitos para todos los loci representa apenas el 4,5 ± 2,8 por ciento, (De Alba, 2009). Para estas zonas es más conveniente Tabla 3. Matriz de distancias genéticas Da (Nei & Chesser, 1983), sobre 19 poblaciones de bovinos.

RS-EC RET AVI RGA BC BN MOS PAJ NAN VCA PAL FRI HER SPA GYR BRH SIN GUZ NEL

RS-EC 0 0,1524 0,1580 0,1566 0,1031 0,2379 0,1929 0,1132 0,1302 0,1326 0,2243 0,1363 0,1730 0,1593 0,4289 0,3429 0,4216 0,4725 0,4624

RET 0,1524 0 0,1153 0,1014 0,1328 0,2456 0,2159 0,1223 0,1466 0,1735 0,2634 0,2143 0,2150 0,1862 0,4798 0,3656 0,4801 0,5165 0,4712 99 Ecuador del Sur Región la de criolla bovina población la de genética Caracterización AVI 0,1580 0,1153 0 0,1186 0,1249 0,2662 0,2221 0,1331 0,1169 0,1804 0,2645 0,1913 0,2441 0,1890 0,5400 0,4307 0,5161 0,5547 0,5246 RGA 0,1566 0,1014 0,1186 0 0,1343 0,2491 0,2182 0,1256 0,1509 0,1587 0,2646 0,1774 0,1984 0,1789 0,5304 0,3942 0,4881 0,5536 0,5233 BC 0,1031 0,1328 0,1249 0,1343 0 0,2007 0,1564 0,0673 0,1012 0,1110 0,2079 0,1402 0,1595 0,1363 0,4850 0,3681 0,4426 0,4953 0,4855 BN 0,2379 0,2456 0,2662 0,2491 0,2007 0 0,2667 0,2057 0,1964 0,2310 0,3029 0,2759 0,2563 0,2370 0,5597 0,4394 0,5490 0,5903 0,5553 MOS 0,1929 0,2159 0,2221 0,2182 0,1564 0,2667 0 0,1573 0,1814 0,1815 0,2794 0,2342 0,2003 0,2043 0,5389 0,4556 0,5243 0,5846 0,5247 PAJ 0,1132 0,1223 0,1331 0,1256 0,0673 0,2057 0,1573 0 0,1055 0,1215 0,2177 0,1362 0,1742 0,1466 0,4796 0,3760 0,4771 0,5223 0,4994 NAN 0,1302 0,1466 0,1169 0,1509 0,1012 0,1964 0,1814 0,1055 0 0,1548 0,2413 0,1807 0,2050 0,1568 0,5115 0,4132 0,5054 0,5347 0,5153 VCA 0,1326 0,1735 0,1804 0,1587 0,1110 0,2310 0,1815 0,1215 0,1548 0 0,1603 0,1533 0,1813 0,1559 0,5058 0,3921 0,4697 0,5252 0,5006 PAL 0,2243 0,2634 0,2645 0,2646 0,2079 0,3029 0,2794 0,2177 0,2413 0,1603 0 0,2648 0,2878 0,2693 0,5767 0,4928 0,5118 0,5701 0,5775 FRI 0,1363 0,2143 0,1913 0,1774 0,1402 0,2759 0,2342 0,1362 0,1807 0,1533 0,2648 0 0,1842 0,1807 0,5580 0,4324 0,5253 0,5856 0,5521 HER 0,1730 0,2150 0,2441 0,1984 0,1595 0,2563 0,2003 0,1742 0,2050 0,1813 0,2878 0,1842 0 0,1719 0,5394 0,4335 0,5111 0,5602 0,5398 SPA 0,1593 0,1862 0,1890 0,1789 0,1363 0,2370 0,2043 0,1466 0,1568 0,1559 0,2693 0,1807 0,1719 0 0,5003 0,4022 0,4741 0,5426 0,5033 GYR 0,4289 0,4798 0,5400 0,5304 0,4850 0,5597 0,5389 0,4796 0,5115 0,5058 0,5767 0,5580 0,5394 0,5003 0 0,1407 0,2330 0,2185 0,2404 BRH 0,3429 0,3656 0,4307 0,3942 0,3681 0,4394 0,4556 0,3760 0,4132 0,3921 0,4928 0,4324 0,4335 0,4022 0,1407 0 0,1653 0,2048 0,1870 SIN 0,4216 0,4801 0,5161 0,4881 0,4426 0,5490 0,5243 0,4771 0,5054 0,4697 0,5118 0,5253 0,5111 0,4741 0,2330 0,1653 0 0,2333 0,2664 GUZ 0,4725 0,5165 0,5547 0,5536 0,4953 0,5903 0,5846 0,5223 0,5347 0,5252 0,5701 0,5856 0,5602 0,5426 0,2185 0,2048 0,2333 0 0,2526 NEL 0,4624 0,4712 0,5246 0,5233 0,4855 0,5553 0,5247 0,4994 0,5153 0,5006 0,5775 0,5521 0,5398 0,5033 0,2404 0,1870 0,2664 0,2526 0

Abreviación de las razas: RS-EC: Región Sur del Ecuador; RET: Retinta; AVI: Avileña; RGA: Rubia Gallega; BC: Berrenda Colorada; BN: Berrenda Negra; MOS: Marismeña; PAJ: Pajuna o Serrana; NAN: Negra Andaluza; VCA: vaca Canaria; PAL: vaca Palmera; FRI: Frisona; HER: Hereford; SPA: Pardo Suizo; GYR: Gyr; BRH: Brahman; SIN: Sindi; GUZ: Guzera; NEL: Nelore. 100 L. Aguirre et al.

mantener el grupo de los Colorados (Figura 3) y Pintados (Figura 4).

El Fit (coeﬁciente consanguinidad total población) y el Fis (coeﬁciente consanguinidad subpoblaciones) indican que hay nueve marcadores que presentan valores negativos (exceso de heterocigosis); cabe destacar también que marcadores que estén con valores de Fit yFis elevados (superiores a 0,1) indican una consanguinidad elevada, sea a nivel de población total o subpoblación. En este caso, son once los marcadores con un Fit yFis superior a 0,10 (BM8125, TGLA122, BM2113, HAUT27, INRA23, INRA37, SPS115, ETH3, HAUT24, ILSTS011 e INRA35). De acuerdo a los resultados de estos estadísticos, la media del Fit yFis de la población es de 8,89 y 8,92 por ciento respectivamente; valor ligeramente superior al encontrado por Martínez (2008) en los bovinos Patagónicos (0,06579) y muy superior al descrito en los bovinos criollos del Noroeste de Argentina (–0,00913). Esto revela que los vacunos criollos de la RSE muestran un nivel mayor de homocigosis.

Relaciones genéticas Para el cálculo de distancias se trabajó con 19 poblaciones (razas): 13 taurinas europeas, 5 índicas y la población criolla de la RSE. El tamaño de la muestra analizada fue de 789 individuos y 24 locus estudiados; se empleó el cálculo de distancia Da (Nei & Chesser, 1983), y la Tabla 3 muestra los valores de distancia genética de los criollos de la RSE con Figura 6. Árbol de distancias genéticas entre 19 poblaciones (razas) bovinas: el resto de poblaciones. Es de destacar que los distanciamien- Da, NJ. Abreviación de las razas: NEL: Nelore; GYR: Gyr; GUZ: Guzera; tos con las 18 poblaciones de bovinos, superan el 0,10. Las BRH: Brahman; SIN: Sindi; EC:Ecuador; VCA: vaca Canaria; PAL: vaca distancias menores están entre RSE:BC (0,1031), RSE:PAJ Palmera; FRI: Frisona; NAN: Negra Andaluza; AVI: Avileña; RET: Retinta; RGA: Rubia Gallega; BC: Berrenda Colorada; PAJ: Pajuna ó Serrana; BN: (0,1132), RSE:NAN (0,1302), RSE:VCA (0,1326) y RSE: Berrenda Negra; SPA: Pardo Suizo; MOS: Marismeña; HER: Hereford. FRI (0,1362); el mayor distanciamiento se observa entre RSE:GUZ (0,4725) y RSE:NEL (0,4624). Estos valores indican que los bovinos criollos de la RSE tienen un gran distan- En general se puede apreciar que los bovinos criollos de la ciamiento con las razas cebuinas y más bien sus troncos RSE están muy distanciados genéticamente de las razas ancestrales están en las poblaciones ibéricas traídas en la con- cebuínas y están más relacionadas con las razas de origen quista de América, destacándose en ello, en su orden: la taurino, especialmente las razas autóctonas españolas; en Berrenda Colorada, Pajuna ó Serrana, la Negra Andaluza, este análisis también hay que destacar que a excepción la vaca Canaria y la Frisona (Holstein Friesian), población de la rama de las razas índicas, prácticamente todos los esta última importada recientemente a la región. nodos del árbol están próximos al origen del mismo, lo que indica que las poblaciones representadas son indepen- En la Figura 6, se puede observar el árbol de distancias dientes entre sí. Estos resultados ﬁlogenéticos son simi- genéticas construido mediante el algoritmo Neighbor-Net, lares a los encontrados por Villalobos (2010) y Martínez de las 19 razas bovinas, analizadas mediante la matriz de dis- (2008) en los bovinos criollos panameños y argentinos tancias Da. Con un valor de replicación del 49 por ciento, se respectivamente, lo que demuestran el origen común que nota la rama distante que separa al grupo (clúster) de razas tienen los vacunos criollos de América. cebuinas de las taurinas, incluida la criolla de la RSE; así mismo, con un valor de replicación bajo de 39 por ciento, hay una relación genética cerrada entre la Hereford, la Marismeña y la Pardo Suiza formando un clúster; otra Conclusiones rama la conforman la Rubia Gallega, la Retinta y la Avileña. Hay una población separada del resto que es la Los bovinos criollos de la RSE presentan una elevada variabi- Berrenda Negra; y con un nivel de remuestreo alto (99 y lidad genética, demostrada en el número de alelos por locus, en 68 por ciento), se forma una rama con nodos cortos que el PIC y en la heterocigosidad, vale destacar el 32 por ciento de dan lugar a una relación genética entre la población criolla alelos que se encuentran en desequilibrio H-W, lo que demues- de la RS-EC, la Pajuna, la Berrenda Colorada, la Negra tra una ligera inestabilidad genética que de no tomarse correc- Andaluza, la vaca Canaria y la Frisona. tivos conllevaría a una disminución de esta valiosa diversidad. Caracterización genética de la población bovina criolla de la Región Sur del Ecuador 101

Esto se corrobora con los valores de Fit yFis encontrados donde De Alba, M.J. 2009. El libro de los bovinos criollos de América (Ed. se evidencia que existe un exceso de homocigosis que puede Papiro Omega, S.A., México), pp. 19–36, 425. deberse a una elevada consanguinidad que supera el 8 por FAO. 2011. Molecular genetic characterization of animal genetic ciento; si bien los valores de Gst yFst, indican que no hay resources. FAO Animal Production and Health Guidelines. No. 9. una diferenciación genética entre los cuatro grupos de bovinos Rome. (available at http://www.fao.org/docrep/014/i2413e/i2413e00. criollos, faneróptica y fenotípicamente sí se diferencian por lo pdf). que convendría manejarlos como núcleos en forma separada Huson, D.H. & Bryant, D. 2006. Application of phylogenetic networks por el color del pelaje, pues esto ayuda a un mejor bienestar in evolutionary studies. Mol. Biol. Evol. 23: 254–267. y desenvolvimiento de los animales de acuerdo a los diferentes Jordana, J., Alexandrino, P., Beja-Pereira, A., Bessa, I., Cañon, J., pisos altitudinales existentes en la región. En la determinación Carretero, Y., Dunner, S., Laloe, D., Moazami, K., Sanchez, A. de las distancias genéticas se aprecia un mayor grado de acer- & Ferrand, N. 2003. Genetic structure of eighteen local south camiento con razas ibéricas, como la Berrenda Colorada, European beef cattle breeds by comparative F-statistics analysis. Journal of Animal Breeding and Genetics 120: 73–87. Pajuna, Negra Andaluza, Canaria y Frisona, y un mayor aleja- miento genético con las cebuinas; en todo caso los valores de Lucena Salmoral, M. 1987. Sebastián de Belalcázar, Historia 15 distanciamiento superan el 0,10 en todos los pares en que se (Quórum, Madrid), p. 155. comparó a la criolla de la RSE con el resto de razas, lo que per- Martínez, R. 2008. Caracterización genética y morfológica del bovino mite afirmarquelapoblaciónbovinacriolladelaRSEpuede criollo argentino de origen patagónico. Tesis de doctorado, – ser considerada genéticamente como una raza taurina ‘propia’ Universidad Politécnica de Valencia, España, pp. 157 175. a la que hay que conservar, pues se corre el riesgo a un corto Mateus,J.C.,Penedo,M.C.,Alves,V.C.,Ramos,M.&Rangel-Figueiredo, plazodesudesaparición. T. 2004. Genetic diversity and differentiation in Portuguese cattle breeds using microsatellites. Animal Genetics 35: 106–113. Nei, M., & Chesser, R.K. 1983. Estimation of fixation indices and gene Agradecimiento diversities. Ann. Hum. Genet. 47: 253–259. Olivier, L., 2002. Populations 1.2.28. Disponible en: http://bioinformat- Al proyecto Biobovis de la Red CONBIAND, que con su ics.org/tryphon/populations/. banco de material genético, contribuyó a cumplir con los Page, R.D.M. 1996. Treeview: an application to display phylogenetic objetivos propuestos en la presente investigación. trees on personal computers. Computer Applications in the Biosciences 12: 357–358. Park, S.D.E. 2001. Trypanotolerance in West African cattle and the Referencias population genetic effects of selection. Tesis de doctorado, Trinity College, Dublín. Irlanda. Aguirre, L., Bermeo, A., Maza, D. & Merino, L. 2011. Estudio Pinzón, M.E. 1984. Historia de la Ganadería en Colombia. Suplemento fenotípico y zoométrico del bovino criollo de la sierra media y alta Ganadero CEGA, Centro de Estudios Ganaderos y Agricolas. Bogota, de la región sur del Ecuador (RSE). Actas Iberoamericanas de Colombia. Vol. 4, p. 208. Conservación Animal 1: 392–397. Quiroz, J. 2007. Caracterización genética de los bovinos criollos Baumung, R., Cubric-Curik, V., Schwend, K., Achmann, R. & mexicanos y su relación con otras poblaciones bovinas. Tesis Soelkner, J. 2006. Genetic characterization and breed assignment in de doctorado, Universidad de Córdoba, España, Austrian sheep breeds using microsatellite marker information. p. 155. Journal of Animal Breeding and Genetics 123: 265–271. Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N. & Bonhomme, F. Sastre, H.J., Rodero, E., Rodero, A., Azor, P.J., Sepúlveda, N.G., 2003. Genetix: 4.05 Logiciel sous WindowsTM pour la génétique Herrera, M. & Molina, A. 2003. Caracterización genética de la des populations. Laboratoire Génome, Populations, Interactions raza bovina colombiana criolla Casanare mediante análisis de CNRS UMR 5000. Université de Montpellier II, Montpellier, Francia. microsatélites. Revista Colombiana de Ciencias Pecuarias 16 (supl.): 49. Botstein, D., White, R.I., Skolnich, M. & Davis, R.W. 1980. Construction of a genetic linkage map in man using restriction fragment length poly- Villalobos, A. 2010. Caracterización genética de las poblaciones bovinas morphism. American Journal of Human Genetics 32: 324–331. Guaymí y Guabala y su relación con otras poblaciones bovinas mediante microsatélites. Tesis de doctorado, Universidad de Córdoba, Bryant, D. & Moulton, V. 2004. Neighbor-Net: an agglomerative España, pp. 70–72, 78–87. method for the construction of phylogenetic networks. Mol. Biol. Evol. 21: 255–265. Villasmil, Y., Roman, R., Yañez, L., Contreras, G., Jordana, J. & Aranguren, J. 2008. Diversidad genética de la raza criollo Cañon, J., Alexandrino, P., Bessa, C., Carleos, Y., Carretero, S., Limonero utilizando marcadores de ADN microsatélites. Revista Dunner, N., Ferran, D., Garcia, J., Jordana, J., Laloe, A., Científica FCV-LUZ 18 (4): 415–423. Pereira, A. & Moazami, K. 2001. Genetic diversity measures of local European beef cattle breeds for conservation purposes. Weir, B.S. & Cockerham, C.C. 1984. Estimating F-statistics for the ana- Genetics Selection Evolution 33: 311–332. lysis of population structure. Evolution 38: 1358–1370. Dalvit, C., De Marchi, M., Dal Zotto, R., Gervaso, M., Meuwissen, T. Wright, S. 1969. Evolution and the genetics of populations, vol. 2: The & Cassandro, M. 2008. Breed assignment test in four Italian beef cat- theory of gene frequencies. University of Chicago Press, Chicago. tle breeds. Meat Science 80: 389–395. Animal Genetic Resources, 2014, 54, 103–113. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000058

Genealogical and population viability analysis of a conservation nucleus of Brazilian Bergamasca sheep

H. Carneiro1, S.R. Paiva2, H. Louvandini3, R.M. Miranda4 and C. McManus1,4 1Faculdade de Agronomia e Medicina Veterinária, Campus Universitário Darcy Ribeiro, Universidade de Brasília, Brasília, DF 70910-900, Brazil; 2EMBRAPA Recursos Genéticos e Biotecnologia, Parque Estação Biológica, Avenida W5 Norte (Final), Brasília, DF 70770-900, Brazil; 3Centro de Energia Nuclear na Agricultura, Universidade de São Paulo, Piracicaba, SP, CEP 13416-000, Brazil; 4Departamento de Zootecnia, Universidade Federal de Rio Grande do Sul, Av. Bento Gonçalves, 7712, Porto Alegre, RS, CEP 91540-000, Brazil

Summary A genealogical and a viability analysis was carried out on the 1559 available registers for the Conservation Nucleus of Brazilian Bergamasca sheep of the University of Brasilia farm in Brazil using the ENDOG and the Vortex programs. To run the ENDOG it was used the registered data and for the Vortex it was used information obtained by a questionnaire answered by the curators of the herds. Of the animals registered, 767 had known parents, with signiﬁcantly more dams known at each generation. The number of pedigrees known has increased over the generations, with higher registration of parents of sires than dams. The Computed Mean Inbreeding calculated by ENDOG was 0.29 percent and mean average relatedness was 1.52 percent. Mean Generation interval was 3.71 years with this values being lower for sires than dams. The population probability of extinction, calculated by Vortex was 17 percent and the average time to extinction was 59 years. Forty-two additional scenarios were created to determine which factors most threaten these populations which were frequency of catastrophes, lack of animal entrance and adult and lamb mortalities, especially adult female mortality. These results indicate that future breeding plans should include exchange of sires between farms to maintain low inbreeding levels and increase genetic variability and upgrade the management to control the mortality rates of animals.

Keywords: average relatedness, inbreeding, population modelling, Vortex, Wright statistics

Résumé Une analyse généalogique et de viabilité a été réalisée avec les 1559 enregistrements disponibles au Centre de Conservation de la race ovine Bergamasca Brésilienne de l’Université de Brasilia (Brésil), en utilisant les logiciels ENDOG et Vortex. Les données des enregistrements ont été utilisées pour travailler avec le logiciel ENDOG alors que l’information traitée par le logiciel Vortex avait été obte- nue au moyen d’un questionnaire auquel ont répondu les personnes prenant soin des troupeaux. Pour 767 des animaux enregistrés, les parents étaient connus, avec le nombre de femelles connues étant significativement plus élevé à chaque génération. Le nombre de généalogies connues a augmenté au fil des générations, grâce à un enregistrement plus fréquent des parents des mâles. La Consanguinité Moyenne a été estimée par ENDOG à 0,29 pour cent alors que la Parenté Moyenne a été de 1,52 pour cent. L’Intervalle Générationnel Moyen a été de 3,71 ans, avec l’intervalle étant plus court chez les mâles que chez les femelles. La probabilité d’extinction de la population a été de 17 pour cent d’après Vortex et le temps moyen d’extinction a été de 59 ans. Quarante-deux scénarios additionnels ont été créés pour déterminer quels sont les facteurs menaçant le plus ces populations. Ceux- ci ont été: la fréquence de catastrophes, le manque de renouvellement des animaux et la mortalité des adultes et des agneaux, notamment la mortalité des femelles adultes. Ces résultats indiquent que les futurs plans de sélection devraient envisager l’échange de mâles entre les fermes afin de maintenir bas les niveaux de consanguinité et d’accroître la variabilité génétique. Il faudrait, de même, améliorer la gestion des exploitations dans le but de contrôler le taux de mortalité des animaux.

Mots-clés: consanguinité, modélisation populationnelle, parenté moyenne, statistiques de Wright, Vortex

Resumen Se realizó un análisis genealógico y de viabilidad con los 1559 registros disponibles en el Núcleo de Conservación de la raza ovina Bergamasca Brasileña, de la Universidad de Brasilia (Brasil), usando los programas ENDOG y Vortex. En el programa ENDOG se trabajó con los datos de los registros mientras que en el programa Vortex se usó la información obtenida mediante un cuestionario al que respondieron los encargados del cuidado de los rebaños. De los animales inscritos, se conocían los progenitores de 767, con un número signiﬁcativamente mayor de hembras conocidas en cada generación. El número de genealogías conocidas ha aumentado a lo largo de las generaciones, gracias a un mayor registro de los progenitores de los machos. La Consanguinidad Media calculada por ENDOG fue de 0,29 por ciento mientras que el Parentesco Medio fue de 1,52 por ciento. El Intervalo Generacional Medio fue

Correspondence to: C. McManus, Faculdade de Agronomia e Medicina Veterinária, Campus Universitário Darcy Ribeiro, Brasília – DF 70919-900, Brazil. email: [email protected]

103 104 H. Carneiro et al.

de 3,71 años, siendo este dato menor para los machos que para las hembras. La probabilidad de extinción de la población, calculada por Vortex, fue del 17 por ciento y el tiempo medio hasta la extinción de 59 años. Se recrearon 42 escenarios adicionales para determinar cuáles son los factores que, en mayor medida, amenazan a estas poblaciones. Éstos resultaron ser: la frecuencia de catástrofes, la falta de reposición de los animales y la mortalidad de los adultos y de los corderos, en especial la mortalidad de las hembras adultas. Estos resultados indican que los futuros planes de selección deberían contemplar el intercambio de machos entre las granjas para mantener bajos los niveles de consanguinidad e incrementar la variabilidad genética. Asimismo, se debería mejorar el manejo de las explotaciones para controlar las tasas de mortalidad de los animales.

Palabras clave: consanguinidad, estadísticos de Wright, modelización poblacional, parentesco medio, Vortex

Submitted 11 June 2013; accepted 10 January 2014

Introduction Genetic diversity can be considered the cornerstone of con- Sheep have been present in Brazil for almost 500 years, but servation genetics, since it is the basis of the evolutionary their numbers have not increased to the same extent as potential of species to respond to environmental changes. other farm animals (Hermuche et al., 2012, 2013). This The most threatened populations are subdivided into differ- has been largely due to herd structure (smallholders), per- ent groups in different fragments of habitat, nature reserves ceived use of sheep as a secondary culture and lack of or in groups of different breeds or lines, in the case of interest from major research groups. Taking into account domestic animals (Toro and Caballero, 2005). rapid genetic erosion (Hall and Ruane, 1993; Scherf, This study aims to analyse genealogical information inte- 2000) and the importance of Animal Genetic Resources grated with simulation tools in the risk of extinction and for the future (Notter, 1999), it may not be possible to molecular markers of a Conservation Nucleus of this breed wait for specific and detailed information on available to determine the effectiveness of current conservation efforts. breeds and ecotypes. The Brazilian Bergamasca is an Italian breed from the Lombardy region, which is widespread mainly in Material and methods Northern Italy (Dalvit et al., 2009). In Italy, this breed spe- cializes in meat production and is widely used in crossing The flock of Brazilian Bergamasca sheep used in this study with other breeds. It is thought to have been introduced in is kept on the University of Brasilia (UnB) research farm, Brazil in the 1930s with Italian immigration (Miranda and is part of the Brazilian National Program for Conservation McManus, 2000; Paiva et al., 2005). Although once wide- of Animal Genetic Resources and this is the only one of spread has been substituted for other breeds, especially this breed to participate in this program. The population hair sheep. In Brazil, it is considered a triple purpose studied was of 1 559 animals, which included all breed raised for milk, meat and coarse wool (Miranda Brazilian Bergamasca sheep on the UnB farm born and McManus, 2000), and at present is threatened with between 1990 and 2010. There were 48 percent males extinction in the country. and 52 percent females. Data for the population viability More recently, there has been some interest in this breed analysis were taken from the questionnaire answered by for milk production to make cheese, which is not tradition- the curator of the flock, the data from which are in al in Brazil, but a few herds are left. The rare ARK haplo- Table 1. On the other hand, information for the pedigree type has also been described in this breed (Luhken et al., analysis were obtained in the animal register. 2007) meaning that it may be useful in studying scrapie Two analyses were carried out: one looking at the genetic resistance in sheep. Embrapa (Brazilian Agricultural structure of the herd over the 20-year period to determine Research Corporation) included domesticated and natura- population parameters (pedigree analysis) and the other lized breeds threatened of extinction in its Program of simulating future events to evaluate the future of the Research on Genetic Resources in 1983. Currently, the herd (viability analysis). conservation of animal genetic resources is carried out by diverse Research Centers of Embrapa, Universities, State Research Companies, as well as private farmers. Pedigree analysis Currently, much of the naturalized animals are in small and The population genetic parameters were computed using scattered populations. Small populations are unstable, the ENDOG program, version 4.8. The program details since they are more exposed to stochastic factors that can can be found in Gutiérrez and Goyache (2005). ENDOG lead to extinction, such as demographic stochasticity, is a computer program for population genetics that per- environmental variation, genetic drift and inbreeding forms various demographic and genetic analyses from depression (Armstrong, Postiglioni and Gonzáles, 2006). pedigree information. A number of population and genetic Genealogy and viability analyses for sheep conservation 105

parameters were computed for the reference populations, Table 1. Demographic parameters used to simulate the population which consist of animals with both parents known. of sheep Brazilian Bergamasca with Vortex. Parameters included: pedigree completeness level was Input data Value computed as the proportion of ancestors known per parental generation; the number of equivalent to discrete genera- Number of iterations 1 000 tions (t) for each individual in a pedigree; the inbreeding Number of years 100 fi fi Reproduction system Polygamous coef cient (F), de ned as the probability that two alleles Age of first offspring (F/M) 2/2 at a randomly chosen locus are identical by descent; The Inbreeding depression? Yes average relatedness coefficient (AR), defined as the prob- Maximum number of lambings per year 2 ability that an allele randomly chosen from the whole Maximum number of progeny per lambing 3 population belongs to a given animal; the probability of Sex ratio at birth 50% Maximum age of reproduction 8 gene origin was characterized by computing: effective Adult female breeding (%) 80 number of founders ( fe) was computed according to Probability of simple birth 74 Lacy (1989); effective number of ancestors ( fa); the found- Probability of twin birth 23 er genome equivalents was calculated from the inverse of Probability of triple birth 3 twice the average coancestry of the individuals within the Female mortality age between 0 and 1 (%) 30 Female mortality age between 1 and 2 (%) 5 population. Effective population size (Ne) was computed Female mortality after age 2 (%) 5 following Gutiérrez et al. (2003). Male mortality age between 0 and 1 (%) 30 Male mortality age between 1 and 2 (%) 5 Male mortality after age 2 5 Population viability analysis Catastrophe 1: animal attack Frequency (%) 6 Simulations of interactions between demographic, environ- Effect on reproduction (%) 80 mental and genetic factors were carried out using the pro- Effect on survival (%) 90 gram Vortex, version 9.99. A questionnaire was developed Catastrophe 2: feeding error to be answered by the curators of the herds linked to the Frequency (%) 12 Brazilian National Program for Conservation of Animal Effect on reproduction (%) 60 Effect on survival (%) 50 Genetic Resources with the information required for pro- Males in breeding pool (%) 5 gram input. According to Miller and Lacy (2005), the Initial population 25 Vortex program begins by creating individuals to form a Carrying capacity (K) 400 base population and then steps through life cycle events Animals harvested? Sim on an annual basis. Stochastic events such as breeding suc- 100 males per year First year of harvest 5 cess, progeny number, sex at birth and mortality rate were Last year of harvest 100 determined by a predefined probability density function Interval between harvests 1 given by the operator (Table 1). Consequently, each iter- Population supplemented? Sim ation of the model gave a different result and by running One male every 2 years the model hundreds of times, and a range of possibilities First year of supplementation 2 Interval between supplementation 2 and outcomes were examined. The scenario described in Table 1 was called baseline and all the other scenarios created changed one parameter of it to test the population (Tables 5 and 6). indexes for microsatellites were calculated using the Molkin package (Gutiérrez et al., 2005). The data including catastrophes described above were informed by the curator of the nucleus. The loss of genetic To determine which factors most threaten this population, variability is simulated as the transmission of alleles from new scenarios were tested to see the behaviour of the para- parents to offspring in a hypothetical neutral site. For each meters. Forty-two additional scenarios were created, as animal at the beginning of the simulation is assigned two described in Tables 5 and 6, by changing baseline unique alleles at this locus. Then for each individual cre- (Table 1) values, maintaining other variables stable. ated during the simulation is randomly assigned one of the alleles from each parent. The program monitors how many alleles remain in the ori- Results ginal population, by giving the mean number of alleles remaining within extant population, the average heterozy- Pedigree data on 1 559 animals were analysed, the herd gosity and genetic diversity relative to the initial level. was classified as commercial using according to Forty-six Brazilian Bergamasca animals were genotyped Gutiérrez and Goyache (2005) as care has been taken using 19 microsatellite markers to provide input allele fre- over the years to buy rams where possible to avoid quency information, obtained in the Animal Genetics inbreeding. Pedigree Information increased over the gen- Laboratory in Embrapa Genetic Resources and erations were analysed, with more information available Biotechnology, Brasilia – DF, Brazil (Table 2). Diversity on dams than sires, with almost 100 percent of dams 106 H. Carneiro et al.

Table 2. Microsatellite loci information in Brazilian Bergamasca sheep.

Locus Number of alleles/locus Allele frequency

Inra23A 10 0.05 0.05 0.12 0.05 0.05 0.01 0.1 0.19 0.12 0.26 Oar304A 6 0.06 0.15 0.01 0.48 0.06 0.24 MAF214A 5 0.04 0.43 0.39 0.06 0.08 INRA63A 7 0.18 0.23 0.04 0.13 0.37 0.03 0.02 OARHH35A 9 0.02 0.19 0.02 0.12 0.19 0.02 0.37 0.04 0.03 INRA35A 9 0.12 0.27 0.1 0.07 0.15 0.14 0.13 0.01 0.01 OMHC1A 5 0.14 0.71 0.003 0.14 0.007 ILSTS87A 9 0.04 0.04 0.04 0.34 0.15 0.07 0.24 0.08 0.26 ILSTS05A 5 0.14 0.27 0.38 0.06 0.15 ILSTS11A 5 0.47 0.2 0.02 0.18 0.13 MAF65A 7 0.22 0.34 0.09 0.11 0.02 0.21 0.01 BM827A 7 0.03 0.11 0.36 0.34 0.04 0.11 0.01 OARFCB20A 11 0.17 0.49 0.02 0.04 0.04 0.04 0.02 0.07 0.02 0.08 0.01 OARCP20A 7 0.28 0.17 0.22 0.01 0.17 0.02 0.13 OAR129A 3 0.02 0.76 0.22 INRA172A 7 0.03 0.03 0.38 0.15 0.02 0.2 0.19 HUJ616A 3 0.32 0.67 0.01 SRCRSP05A 4 0.04 0.35 0.48 0.13 BM6526A 5 0.03 0.19 0.01 0.02 0.75

known in the present generation (Figure 1). The reference Effective population size per year (Figure 2) has remained populations were taken as the animals with both parents relatively stable over the years, varying about 30 animals. known. This population will be smaller than that one The sharp decrease in 1998 was due to an attack by dogs, used to analyse Founders. The effective founder popula- which killed the whole lamb ﬂock and a large number of tion Size was 229 with 791 animals in the reference popu- ewes. lation and 215 ancestors contributing to the reference Generation intervals (Table 4) varied from 4.08 (father– population. The effective number of founders/ancestors son) to 2.98 years (Father–daughter). This shows a rela- for the reference population was 42/34 with 12 animals tively fast turnover of reproducing animals. In terms of supplying 50 percent of the ancestors. mean age of parents when offspring were born these The Computed Mean Inbreeding was 0.29 percent and were higher for dams than sires. Alderson (1990) used Mean AR was 1.52 percent. These numbers are compar- the genetic conservation index (GCI) to calculate an effect- able to those for the Galega (Adán et al., 2007) and ive number of founders in the pedigree of an animal. The Mallorquina (Goyache et al., 2010) sheep in Spain and higher the GCI value the higher the value of an animal for well below the dangerous level (10 percent). conservation. Mean GCI was 1.36, varying between 0.77 and 5.07. This has increased steadily over the years with As the generations increased, there was little change in an increment of approximately 0.05 per year (Figure 3). inbreeding and AR (Table 1). This was due to careful con- Using sires to analyse differentiation, F was found to be trol of reproduction in the nucleus. Although very low (<0.4 is −0.080; F = 0.069 and F = −0.005, all indices being percent) there is an increase over the generations, which st it close to zero. In the present study, when sires were consid- should be carefully monitored. There were 2 percent mat- ered subpopulations there is low genetic differentiation ings between sibs and 1 percent between parent-offspring. (7.7 percent of total genetic variation). Self-coancestry These values are summarized in Table 3, per generation.

Figure 1. Proportion of parents known per generation in a Conservation Nucleus of Brazilian Bergamasca sheep. P – parent, GP – grandparent, Figure 2. Effective population size per year of birth in a Conservation GGP – great grandparent. Nucleus of Brazilian Bergamasca sheep. Genealogy and viability analyses for sheep conservation 107

Table 3. Inbreeding (F), average relatedness (AR) and effective population size (Ne) per generation in a Conservation Nucleus of Brazilian Bergamasca sheep.

F % AR Ne

Maximum generations 1 – 0.01 2 0.00 0.01 0.01 336.00 3 0.00 0.02 0.02 1 369.00 4 0.00 0.07 0.02 760.60 5 0.01 0.08 0.02 152.30 6 0.01 0.13 0.02 Figure 3. Genetic conservation index for year of birth in a conservation 7 0.01 0.20 0.02 57.70 nucleus of Brazilian Bergamasca sheep. 8 – 0.02 Complete generations 0 0 0.01 difference in expected values between Stoch.r and Det.r 1 0.00 0.08 0.02 115.80 fi 2 0.03 0.40 0.04 19.70 is clearly seen. While the rst had high variation, the second remained stable throughout the period of the simulation, which supports the idea that the population is highly susceptible to environmental and genetic effects, was 0.50, inbreeding 0.003 and Nei distance between sires such as variations in the feed, precipitation levels, inbreed- 0.069, showing that, according to the pedigree analysis, no ing and genetic drift. immediate problems in maintenance of genetic diversity within the herd is seen. When considering the influence of environmental variation, the population did not reach K = 400 (Figure 5). The values of genetic diversity were relatively high, Population viability analysis always above 0.8 (Table 5). The overall results did not The estimated probability of extinction in population via- change significantly when the default program was used bility analysis was 17 percent and therefore, the possibility (each animal at the beginning of the simulation is assigned of success of 83 percent. From the scenarios that have two unique alleles and not using real population genetic become extinct, the average time to extinction was 59 data (Table 5)), as the genetic diversity values were very years (Table 5). Considering only the deterministic growth similar in both scenarios. However, when calculating the of the population, i.e. no inbreeding depression, mating within diversity observed directly from the original micro- with no restrictions and no entry and exit of animals, the satellite markers, the following values of diversity was population would grow until it reaches the carrying cap- obtained: (1) observed heterozygosity of 0.6880 (SD acity of the environment (K) and the deterministic growth 0.0160); (2) expected heterozygosity of 0.6864 (SD rate (Det.r) remained stable for the full simulation period. 0.0335); (3) mean number of alleles 6.53 (SD 2.29); and However, the stochastic prediction considering the factors fixation index (Fis) of −0.002 (P > 0.05). and the stochastic growth rate (Stoch.r) was different The mean observed heterozygosity reported by Vortex is (Figure 4). the mean inbreeding coefficient of the population (Miller In the herd analysed, population growth was much lower and Lacy, 2005), in this case, a high inbreeding value of than expected considering only the Det.r (Figure 4). The 0.9238. One thousand iterations were needed to ensure sta- bilization of the results (Harris, Maguire and Shaffer, 1987). This population appeared to be influenced by gen- Table 4. Generation intervals (L), their standard deviations (SD) and standard errors (SE) for a Conservation Nucleus of etic effects inherent to small populations. This can also be Brazilian Bergamasca sheep. noticed by observing the average number of alleles, which declined to less than half of the initial value for the Interval Standard deviation SE

Generation intervals Father–Son 4.08 1.48 0.66 Father–Daughter 2.98 1.20 0.10 Mother–Son 3.49 1.06 0.29 Mother–Daughter 4.02 2.04 0.12 Total 3.71 1.87 0.09 Mean age of parents when offspring born Father–Son 3.03 1.77 0.08 Father–Daughter 2.83 1.56 0.07 Mother–Son 4.21 2.25 0.08 Mother–Daughter 4.18 2.32 0.08 Figure 4. Difference between the deterministic (Det.r) and stochastic (Stoch.r) Total 3.74 2.17 0.05 growth rates for the period of Vortex simulations in Brazilian Bergamasca sheep. 108 H. Carneiro et al.

and the lamb mortality in this population is already high (30 percent).

When comparing the Effective Number of Founders and Ancestors for the reference population found by ENDOG (42/34) with the initial population size (25) answered by the curator of the herd, it is noticed that the value found by pedigree analysis was higher than the actual initial population. This may be due to error in animal records or the reduction of the herd size due to pressure to reduce Figure 5. Total number of animals achieved at the end of Vortex simulations in Brazilian Bergamasca sheep considering only the stochastic growth rate, i.e. the number of animals to increase the commercial herd. environmental and genetic factors. To include a value of the initial population obtained in the analysis with ENDOG in the simulation with Vortex, we chose the 12 animals supplying 50 percent of the ancestors to also check the impact of a smaller initial population. This increased the probability of extinction, but did not affect the other parameters (Table 5).

The lowest percentage of females breeding each year had a strong impact on the population with a probability of extinction of 100 percent. On the other hand, raising the percentage of males available for breeding increased Figure 6. Total number of alleles obtained at the end of Vortex simulations in only the genetic variability of the population and did not Brazilian Bergamasca sheep estimated for the entire period simulated. affect the other parameters (Table 5).

projected period (Figure 6), indicating a great loss of allelic The population seemed to be viable for the simulation per- diversity, which can endanger the long-term population iod of 100 years, but increasing the simulation time to 500 persistence. and 1 000 years there was a probability of extinction of 99 and 100 percent, respectively (Table 5). Fieberg and Ellner The results of the simulations (Tables 5 and 6) showed that (2000) found that extinction probabilities seem to be an removing inbreeding from the analysis (Table 5) only uncertain measure, sensitive to various parameters and reduced the probability of extinction, but the other para- data quality. According to these authors, attempting to meters remained unchanged. This was expected, since incorporate density dependence, genetic effects, interac- the program models inbreeding depression as a reduction tions with other species to the models will inevitability in the survival of offspring during the ﬁrst year of life, introduce even more uncertainty.

Table 5. Result summary obtained from Vortex simulations in Brazilian Bergamasca sheep considering baseline and the alternative scenarios.

Scenario r GI female GI male PE Years to ﬁrst extinction He Ho Alleles

Baseline 0.103 4.34 4.34 17 59 0.8601 0.9238 18.71 Without inbreeding 0.103 4.34 4.34 11 75 0.8256 0.8754 17.38 Using ENDOG output as initial 0.103 4.34 4.34 38 31 0.8565 0.9178 18.78 population 20% of females breeding −0.192 4.91 4.91 100 15 0 0 0 100% of males in the breeding pool 0.103 4.34 4.34 15 65 0.9206 0.9573 27.98 Without the allelic frequency data 0.103 4.34 4.34 17 59 0.8574 0.9195 18.64 Simulated for 500 years 0.103 4.34 4.34 99 71 0.4859 0.4968 2.00 Simulated for 1000 years 0.103 4.34 4.34 100 55 0 0 0 Extreme scenarios Scenario r L female L male PE Years to ﬁrst extinction He Ho Alleles 70% of lamb mortality −0.086 4.70 4.70 99 23 0.9139 1.00 17.67 30% of adult female mortality −0.127 3.69 4.78 100 17 0 0 0 Higher catastrophe severity 0.069 4.30 4.30 49 42 0.8961 0.9596 24.46 Higher catastrophe frequency −0.003 4.21 4.21 90 33 0.8969 0.9801 18.68 Smaller initial population 0.103 4.34 4.34 55 38 0.8875 0.9382 27.17 70% of female lamb mortality −0.086 4.70 4.70 100 19 0 0 0 70% of male lamb mortality 0.103 4.34 4.34 11 22 0.8994 0.9597 29.22

r, deterministic growth rate; L, generation interval; PE, probability of extinction; He, expected heterozygosity; Ho, observed heterozygosity; Alleles, estimated ﬁnal number of alleles. Genealogy and viability analyses for sheep conservation 109

Table 6. Result summary for the scenarios testing alternative management options in Vortex simulations in Brazilian Bergamasca sheep.

Scenario r L female L male PE Years to ﬁrst extinction He Ho Alleles

Ten males entering 0.103 4.34 4.34 10 30 0.953 0.9811 82.54 Ten females entering 0.103 4.34 4.34 0 0 0.8936 0.923 85.51 15% of catastrophe 1 frequency 0.089 4.33 4.33 26 47 0.8597 0.9277 18.26 20% of catastrophe 1 frequency 0.082 4.32 4.32 32 41 0.8591 0.9293 18.08 25% of catastrophe 1 frequency 0.074 4.32 4.32 35 39 0.8627 0.9356 18.11 25% of catastrophe 1 frequency 0.018 4.23 4.23 86 22 0.8554 0.9607 13.20 30% of catastrophe 2 frequency −0.017 4.18 4.18 97 18 0.8407 0.9682 10.34 35% of catastrophe 2 frequency −0.052 4.14 4.14 100 15 ––– 15% of female adult mortality 0.076 4.38 4.38 27 45 0.8659 0.9362 18.64 20% of female adult mortality 0.062 4.41 4.41 37 43 0.8697 0.9454 18.37 30% of female adult mortality 0.031 4.47 4.47 64 34 0.8771 0.9605 17.48 40% of female adult mortality −0.003 4.54 4.53 82 30 0.8794 0.9739 15.70 50% of female adult mortality −0.043 4.61 4.61 97 26 0.8781 0.9845 13.67 60% of female adult mortality −0.089 4.70 4.70 100 23 ––– 20% of male adult mortality 0.103 4.34 4.34 20 37 0.8536 0.9164 18.51 40% of male adult mortality 0.103 4.34 4.34 17 53 0.8564 0.9197 18.63 50% of male adult mortality 0.103 4.34 4.34 17 59 0.8573 0.9226 18.54 60% of male adult mortality 0.103 4.34 4.34 18 57 0.8574 0.9227 18.74 40% of female lamb mortality 0.066 4.40 4.40 36 44 0.866 0.9393 18.66 50% of female lamb mortality 0.024 4.48 4.48 64 35 0.8748 0.9622 16.43 60% of female lamb mortality −0.025 4.58 4.58 92 25 0.8747 0.9832 13.21 70% of female lamb mortality −0.09 4.70 4.70 99 21 0.8889 0.999 12.00 80% of female lamb mortality −0.17 4.86 4.86 100 15 ––– 40% of male lamb mortality 0.103 4.34 4.34 19 47 0.8611 0.9282 18.78 50% of male lamb mortality 0.103 4.34 4.34 16 53 0.8609 0.9293 19.09 60% of male lamb mortality 0.103 4.34 4.34 15 45 0.8651 0.9376 19.55 70% of male lamb mortality 0.103 4.34 4.34 19 35 0.8680 0.9416 19.72 80% of male lamb mortality 0.103 4.34 4.34 18 35 0.8717 0.9510 20.09 r, deterministic growth rate; L, generation interval; PE, probability of extinction; He, expected heterozygosity; Ho, observed heterozygosity; Alleles, estimated ﬁnal number of alleles.

Increasing the mortality of lambs had a strong impact on (Table 5), showing that it is better to have fewer severe cat- the population, with a negative growth rate, increase in astrophes than several mild disasters. the generation interval and increased the probability of From the conclusions of the extreme scenarios, we extinction (Table 5). Separating the mortality of male attempted to determine the extent to which the population and female lambs, it is noticed that the major impact is would be still viable by changing some parameters to due to the mortality of female lambs, probably due to determine critical points in the management of these ani- the role of females in reproduction and the characteristic mals and suggest measures to be adopted in the conserva- of the herd of having more females than males. Working tion of this herd (Table 6). In the first two scenarios, ten with Vortex, Brook et al. (1997) found that small changes male and ten female adults were added to the population in mortality, potentially due to inbreeding depression, every 2 years for 100 years. The entrance of this animals minor chronic disease, or new exotic predators, had a pro- increased genetic variability, greatly influenced the final found effect on the population projected future. This same number of alleles, reduced the probability of extinction, author observed that the mortality of the juveniles was but did not affect the growth rate and the generation inter- more important in determining the extinction probability val. The increase in the final number of alleles was than that of the adults. Any increase in mortality also great- expected, since the Vortex considers that all migrants are ly reduced the size of any extant populations. However, in not related to the initial population and have unique alleles. this study the mortality of adult females had a greater Another implication of this, according to Brook et al. impact on the population than the mortality of young (1997), as Vortex assumes all supplemented animals are females (Table 5). unrelated; inbreeding could not be realistically modelled. The extreme scenarios, higher mortality of lambs, higher Catastrophe 1, animal attack, being less severe (Table 6), mortality of female lambs (Table 5), affected the popula- even at higher frequencies did not affect the population tion growth rate, the general interval and the probability to a great extent, once the genetic variability was main- of extinction, but did not affect the genetic variability, tained high and did not affect either the growth rate and probably because the population was extinct in the simula- generation interval. Catastrophe 2, feeding error, being tions before losing the genetic variability. The frequency more severe, rapidly affected population. Up to 25 percent of the catastrophe seemed to be worse than the severity did not show much effect, but from 30 percent the 110 H. Carneiro et al.

population is already in a declining trend and at 35 percent Mallorquina sheep (Goyache et al., 2010) in Spain. the probability of extinction was 100 percent. These are on the lower limit for those calculated by The mortality of adult female (Tables 5 and 6) has a Huby et al. (2003) in six French meat sheep breeds in greater effect than the mortality of males. At 30 percent France but higher than those found by Goyache et al. mortality of adult females affects the population, but it is (2003) for the Xalda of Asturias breed, which is also threa- still viable. For example, 40 percent adult male mortality tened with extinction. High generation intervals are due to does not affect most of the population parameters, but at dams and sires remaining in reproduction for a long time. the same 40 percent of adult female mortality the trend This is due to high lamb mortality (~30 percent per year) is a decrease in the population growth rate. At 60 percent mainly caused by harsh dry winters (8 percent relative humid- mortality of adult females, the probability of extinction is ity and temperatures reaching 5 °C at night rising to 30 °C 100 percent. during the day) and infections caused by Haemonhcus contortus, which are resistant to antihelminthics. The population was more “resistant” to female lamb mortality than to adult female mortality (Table 6). Up to 50 Moureaux et al. (1996) reported that when inbreeding percent mortality of female lambs, the population was using pedigree studies on horse breeds are considered, affected, but was still viable. At the 60 percent mortality two groups of horse breeds may be distinguished: one the population growth rate was negative and only at 80 per- group of international breeds, showing values ranging cent of female lamb mortality that the probability of from 0.81to 2.89 percent, and the other group with small extinction was 100 percent. On the other hand, the mortal- population sizes with values ranging from 2.25 to 14.7 per- ity of male lambs did not affect much, since up to 80 percent. Although the population here was small, the inbreed- cent mortality and the population was still viable. ing is less than found in the international breeds. This may be due to two factors: careful choice of sires in the conservation herd or lack of pedigree information (Sabbioni Discussion et al., 2007). Low AR means that an animal shares a low percentage of genes with the rest of the population; Pedigree analysis these animals could therefore serve to disseminate the breed to other farms or regions in the country. This herd is part of the Brazilian National Program for Conservation of Animal Genetic Resources and the only The herd size has not increased and sires are becoming Brazilian Bergamasca herd in the program. Of the five nat- more difficult and expensive to find on the national market, uralized breeds of sheep analysed by Paiva et al. (2005), this reduced herd size may bring problems for the conser- this breed had showed the lowest gene diversity and per- vation nucleus in the coming years. The effective popula- centage of polymorphic loci. The lack of sire information tion size is an important parameter to monitor breeds and is due to multiple ram mating groups, errors in recording breeding programs since it is related to inbreeding depres- and frequent exchange of farm helpers. Nevertheless sion and loss of genetic variability. According to Gandini these numbers are significantly higher than those found et al. (2004) Ne is the measure of choice to determine the by Goyache et al. (2010) and Álvarez et al. (2007) for risk of extinction of a population. The determination of this Mallorquina and by Alfonso et al. (2006) for Latxa parameter allows one to infer the rate of inbreeding and Black-Faced sheep breed in Spain. These latter authors loss of genetic diversity within the population. This popu- note that the lack of genealogies leads to computation of lation kept Ne relatively stable and could recover from the high generation intervals. sharp decrease of 1998, but probably with the difficulty of introducing new animals, the Ne could not stay stable for The effective number of founders/ancestors for the refer- very long. ence population was 42/34 with 12 animals supplying 50 percent of the ancestors. The effective number of ancestors for the reference population of Conservation Nucleus of Galego sheep in Spain was 13 animals with five explaining Population viability analysis 50 percent of the ancestors (Adán et al., 2007). The The Det.r equals the average growth of the population if Brazilian Bergamasca Conservation Nucleus has therefore the flock is large enough to avoid being affected by ran- a higher genetic base and a low effective number of ances- dom or stochastic factors. The Det.r will correctly predict tors for the reference population. The low number of the population development if the birth and mortality ancestors, which explain half the genetic variation, indi- rates are constant over time and within the expected cates a super usage of a low number of sires. This is due values, if there is no inbreeding depression, no limitation fi fi to dif culty in nding rams of this breed, distances and or restriction on random mating and has not limiting to cost required to transport these animals when found (usu- the carrying capacity of the environment. However, usual- fi ally >1 000 km) as well as funding dif culties for a ly one or more of these assumptions are not respected, thus Federal University sheep herd. leading to a slightly lower population growth than Generation intervals are similar to those found for a expected by the deterministic growth rate (Ballou, Lacy Conservation Nucleus of Galega (Adán et al., 2007) and and Miller, 2005). Genealogy and viability analyses for sheep conservation 111

The Stoch.r is calculated for each year of the simulation. concurrently, the likelihood of extinction increases dramat- Usually, this parameter is lower than Det.r and only ically (Brook et al., 1997). shows values close to Stoch.r if the population growth is Both catastrophes analysed in this study, are easily steady and robust. The Stoch.r will be noticeably smaller avoided, improving the feeding control by making a fl than the Det.r if the population is under wide uctuations stock for the drought period, training the handlers and pro- due to environmental variation, catastrophes and demo- tecting the flock from attacks installing fences and arrest- graphic or genetic instability inherent to the small popula- ing the animals at night, period on which most attacks tions (Frankham, Ballou and Briscoe, 2008). As seen in have occurred. These procedures would reduce the fre- Figure 4, there is a difference between the Det.r and quency of the catastrophes, which seemed to be worse to stoch.r, with a wide variation of the second, supporting the population (Table 5). the idea of environmental and genetic effects on susceptibility. Working with Vortex, Lindenmayer, Lacy and Pope (2000) found that in the absence of animals entering, the According to Miller and Lacy (2005), the expected hetero- population rapidly collapsed. One migrant per generation zygosity remaining in a simulated population, in the case of is an appropriate lower limit to the amount of gene flow this study 0.8601, is a useful metric of genetic decay for that is desirable, but sometimes more than one migrant comparison across scenarios and populations. From this per generation will be necessary to achieve the desired result, it is found that the value of heterozygosity calculated genetic goals (Mills and Allendorf, 1996). by Vortex, using real allele frequency data or not, should be analysed considering only the decrease from the value of According to these same authors, one migrant per gener- fi 100 percent. We do not suggest using the final value of ation is suf cient to avoid the loss of alleles in subpopula- He to confirm whether the population is threatened or not. tions caused by genetic drift and will allow the allele frequencies within subpopulations to respond to local The instability present in the population is probably due to selective pressures. However, deviations from the ideal inbreeding depression and low allelic diversity (Figure 6), population structure will tend to compromise the effective- since the housing environment is controlled and does not ness of migration relative to expectations, including the present major challenges to the animals, unless the mortality social structure and taking into account the relative repro- of lambs that can be considered high (around 30 percent). ductive success of immigrants. This could mean that more The genetic instability is caused possibly by the small num- than one migrant per generation may be necessary. The ber of animals entering the population and the use of a few acquisition of new animals for this herd is difficult as sires. Population managers should concern with the variation there is a lack of pure commercial flocks for purchase or depleting effects of genetic drift, which can be countered by exchange, the distances between farms, priorities for the introduction of occasional immigrants to prevent deleteri- research and bureaucracy in trading animals with a herd ous inbreeding within population (Lacy, 1987). belonging to a public university in Brazil. Therefore, it fi The extinction probability should not be considered alone would be very dif cult to increase the entry of animals and should not be the only factor taken into account when beyond what has already occurred in the analysed period. determining conservation strategies. It can be used to indi- Based on these results (Table 5), we noticed that the mor- cate the population trend of extinction or persistence and tality rates of adults and lambs, males and females should warn if conservation measures must be urgent or not. always be monitored and controlled to ensure the viability The reliable predictions of extinction probabilities can be of the population. The lamb mortality rate could be made only for short-term time horizons, 10 to 20 percent decreased by improving the management, especially taking as long as the period over which the population has been better care of the new born, such as navel dipping, colos- monitored (Fieberg and Ellner, 2000). These authors stated trum, protecting from the cold and wind, dam mastitis con- that the given typical amounts of data available it will rare- trol, as well as vaccinating at the appropriate time. ly be meaningful to estimate a long-term extinction prob- However, the mortality of the adult females seemed to be ability, but it may sometimes be possible to reliably more critical to the population and should be closely mon- estimate a short-term extinction probability. Considering itored by the curators, even though it is already low, about the amount of data available in this study, neither it 5 percent. Changes to mortality and fecundity were a dan- would be appropriate nor reliable to estimate the probabil- gerous threat to the short-term persistence of the population ity of extinction for a period longer than 100 years. Brook et al. (1997) studied. Even relatively minor increases According to Table 5, it is better to have fewer severe cat- in mortality or decreases in fecundity translated into strong astrophes than having several mild disasters. Keith et al. effects on population size and extinction probability. (2008), working with plants found that species populations were more viable under the less frequent catastrophe simulations, in this case, fire. The frequency and severity of cat- Conclusion astrophes had major effects on both the probability of extinction and the final size of extant populations. When While no immediate problems are observed in the conser- a catastrophe’s frequency and severity were increased vation nucleus, reduced herd size and lack of other herds 112 H. Carneiro et al.

of this breed could cause problems in the near future. From Goyache, F., Gutiérrez, J.P., Álvarez, I., Fernández, I., Royo, L.J. & all of the scenarios, it is concluded that the number of Gomez, E. 2003. Genetic analysis of calf survival at different pre- – females is critical to sustain the population. The key para- weaning ages in beef cattle. Livestock Prod. Sci., 83: 13 20. meters to be observed in a conservation management were Goyache, F., Fernández, I., Espinosa, M.A., Payeras, L., Pérez-Pardal, providing the entrance of animals, control the frequency of L., Gutiérrez, J.P., Royo, L.J. & Álvarez, I. 2010. Análise fi – catastrophes, not so much the severity, and control the demográ co y genetic de la raza ovina Mallorquina. ITEA,106:3. 14 mortality of females, especially the adults ones (up to 30 Gutiérrez, J.P. & Goyache, F. 2005. A note on ENDOG: a computer percent). program for analysing pedigree information. J. Anim. Breeding Genet., 122: 172–176. Gutiérrez, J.P., Altarriba, J., Díaz, C., Quintanilla, R., Cañón, J. & Piedrafita, J. 2003. Pedigree analysis of eight Spanish beef cattle Acknowledgements breeds. Genet. Sel. Evol., 35: 43–63. Gutiérrez, J.P., Royo, L.J., Àlvarez, I. & Goyache, F. 2005. Molkin The authors acknowledge to the CNPq, INCT-Pecuária, v. 2.0: a computer program for genetic analysis of populations using FAPDF and Finatec for financial support. molecular coancestry information. J. Hered., 96: 718–721. Hall, S.J. & Ruane, J. 1993. Livestock breeds and their conservation: a global overview. Conserv. Biol., 7: 815–825. Statement of interest Harris, R.B., Maguire, L.A. & Shaffer, M.L. 1987. Samples sizes for minimum viable population estimation. Conserv. Biol., 1: 72–76. No conflict of interest was identified in this study. Hermuche, P., Silva, N.C., Guimarães, R.F., Carvalho Júnior, O.A., Paiva, S.R., Gomes, R.A.T. & McManus, C.M. 2012. Dynamics of sheep production in Brazil using principal components and maps of auto-organization characteristics. Rev. Bras. Cart., 64: 821–832. References Hermuche, P., Maranhão, R.L.A., Guimarães, R.F., Carvalho Júnior, O.A., Gomes, R.A.T., Paiva, S.R. & McManus, C.M. 2013. Adán, S., Fernández, M., Justo, J.R., Rivero, C.J., Rois, D. & Lama, Dynamics of Sheep Production in Brazil. ISPRS Int. J. Geo-Info.,2: J. 2007. Análisis de la información genealógica en la raza ovina ovella 665–679. galega. Arch Zootec., 56: 587–592. Huby, M., Griffon, L., Moureaux, S., De Rochambeau, H., Burge, C. Alderson, L. 1990. Genetic conservation of domestic livestock. Wallin, D. & Verrier, E. 2003. Genetic variability of six French meat sheep CAB International, US page. breeds in relation to their genetic management. Genet. Sel. Evol., 35: – Alfonso, L., Parada, A., Legarra, A., Ugarte, E. & Arana, A. 2006. 637 655. Effects on geneticvariability of selection against scrapie sensitivity Keith, D.A., Akçakaya, R., Thuiller, W., Midgley, G.F., Pearson, R. – in the Latxa black-faced sheep. Genet. Sel. Evol., 38: 495 511. G., Phillips, S.J., Regan, H.M., Araújo, M.B. & Rebelo, T.G. Álvarez, I., Fernández, I., Espinhosa, M.A., Payeras, L., Gutiérrez, J. 2008. Predicting extinction risks under climate change: coupling sto- P., Royo, L.J. & Goyache, F. 2007. Análisis del libro genealógico de chastic population models with dynamic bioclimatic habitat models. la raza ovina mallorquina. In Proceedings SEOC 2007. pp. 163–166. Biol. Lett., 4: 560–563. (available at http://www.exopol.com/seoc/docs/6t02xe4o.pdf) Lacy, R.C. 1987. Loss of genetic diversity from managed populations: Armstrong, E., Postiglioni, A. & Gonzáles, S. 2006. Population viabil- interacting effects of drift, mutation, immigration, selection, and popu- ity analysis of the Uruguayan Creole cattle genetic reserve. Anim. lation subdivision. Conserv. Biol., 1: 143–158. Genet. Resour. Inf., 38: 19–33. Lacy, R.C. 1989. Analysis of founder representation in pedigrees: founder Ballou, J., Lacy, B. & Miller, P. 2005. Population and Habitat Viability equivalents and founder genome equivalents. Zoo Biol., 8: 111–123. Assessment (PHVA): Briefing Book. I Maned Wolf International Lindenmayer, D.B., Lacy, R.C. & Pope, M.L. 2000. Testing a Workshop – CENAP/IBAMA, Atibaia, SP, Brazil. simulation model for population viability analysis. Ecol. Appl., 10: Brook, B.W., Lim, L., Harden, R. & Frankham, R. 1997. How secure 580–597. is the Lord Howe Island woodhen? A population viability analysis using Vortex. Pacific Conserv. Biol., 3: 125–133. Luhken, G., Buschmann, A., Brandt, H., Eiden, M., Groschup, M.H. & Erhardt, G. 2007. Epidemiological and genetical differences Dalvit, C., De Marchi, M., Zanetti, E. & Cassandro, M. 2009. Genetic between classical and atypical scrapie cases. Veter. Res., 38: 65–80. variation and population structure of Italian native sheep breeds undergoing in situ conservation. J. Anim. Sci., 87: 3837–3844. Miller, P.S. & Lacy, R.C. 2005. Vortex user´s manual. A stochastic simulation of the extinction process. Chicago Zoological Society, Chicago. Fieberg, J. & Ellner, S.P. 2000. When is it meaningful to estimate an extinction probability? Ecology, 81: 2040–2047. Mills, L.S. & Allendorf, F.W. 1996. The one-migrant-per-generation rule in conservation management. Conserv. Biol., 10: 1509–1518. Frankham, R., Ballou, J.D. & Briscoe, D.A. 2008. Genética e extinção; Conseqüências genéticas do tamanho populacional pequeno. In: Miranda, R.M. & McManus, C. 2000. Desempenho de ovinos Fundamentos de Genética da Conservação, pp. 77–101. Ribeirão Bergamácia na região de Brasília. Rev. Brasil. Zootec., 29: 1661– Preto, Sociedade Brasileira de Genética. 1666. Gandini, G.C., Ollivier, L., Danell, B., Distl, O., Georgoudis, A., Moureaux, S., Verrier, É., Ricard, A. & Mériaux, J.C. 1996. Genetic Groeneveld, A., Martyniuk, E., Van Arendonk, J.A.M. & variability within French race and riding horse breeds from genea- Woolliams, J.A. 2004. Criteria to assess the degree of endangerment logical data and blood marker polymorphisms. Genet. Sel. Evol., 28: of livestock breeds in Europe. Livestock Prod. Sci., 91: 173–182. 83–102. Genealogy and viability analyses for sheep conservation 113

Notter, D.R. 1999. The Importance of genetic diversity in livestock from pedigree analysis. Ann. Facoltà Med. Veterin. Parma., 27: populations of the future. J. Anim. Sci., 77: 61–69. 199–210. Paiva, S., Silvério, V.C., Egito, A.A., McManus, C., Faria, D.A., Scherf, B.D. 2000. World watch list for domestic animal diversity, Third Mariante, A.S., Castro, S.R., Albuquerque, M.S.M. & Dergam, edition. Rome, Food and Agricultural Organization of the United J.A. 2005. Genetic variability of the Brazilian hair sheep breeds. Nations. Pesquisa Agropecuária Brasil., 40: 887–893. Toro, M.A. & Caballero, A. 2005. Characterization and conservation of Sabbioni, A., Valentino, B., Francesca, T.M. & Paola, S. 2007. Genetic genetics diversity in subdivided populations. Phil. Trans. R. Soc. B, variability and population structure in the Italian Haﬂinger Horse 360: 1367–1378. Animal Genetic Resources, 2014, 54, 115–125. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000489

Population structure and genealogical analysis of the Brazilian Crioula Horse

F.C. Maciel1, C.D. Bertoli2, J. Braccini Neto2, J.A. Cobuci2, S.R. Paiva3 and C.M. McManus2,4 1Faculdade de Veterinária, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil; 2Departamento de Zootecnia, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil; 3EMBRAPA Recursos Genéticos e Biotecnologia, Brasília, DF 70770-970, Brazil; 4Universidade de Brasília, Brasília, DF 70910-900, Brazil

Summary A genealogical analysis of Crioula horses registered by the Brazilian Association of Crioula Horse Breeders was performed. The state of Rio Grande do Sul is the largest producer of animals with 89.85 percent of registered animals, of which 45.29 percent were males and 54.71 percent females. The inbreeding coefﬁcient was calculated at 0.88 percent and the average relatedness was 0.65 percent in the total population (animals born in Brazil and imported). Inbreeding increased over the generations. An increase was seen in the average genetic conservation index in animals born after the year 1927 (0) until 2010 (8.67 percent). The average generation interval was 10.3 years and the average age of parents when offspring born were 10.5 years, falling in recent years. The effective population size of founders was 95.19 animals, the number of ancestors contributing to this population was 5086 where 56 ancestors explained 50 percent of the genetic diversity of the breed. Inbreeding is under control in the Crioula horse. The increase in registrations reﬂects the increased interest from farmers in this breed.

Keywords: average relatedness, founding, ancestors, genetic conservation index, inbreeding, statistics of Wright

Résumé Une analyse généalogique des chevaux Créoles inscrits à l’Association Brésilienne d’Éleveurs de Chevaux Créoles a été réalisée. L’état du Rio Grande do Sul est le plus grand producteur d’animaux avec 89,85 pour cent des animaux inscrits, dont le 45,29 pour cent sont mâles et le 54,71 pour cent femelles. Le coefficient de consanguinité a été estimé à 0,88 pour cent, avec une parenté moyenne de 0,65 pour cent dans la population totale, y compris les animaux nés au Brésil et ceux importés. La consanguinité a augmenté au fil des générations. Un accroissement de l’indice moyen de conservation génétique a été observé pour les animaux nés entre 1927 (0) et 2010 (8,67 pour cent). L’intervalle générationnel moyen a été de 10,3 ans et l’âge auquel les parents ont en moyenne leur premier descendant a été de 10,5 ans, cet âge ayant diminué au cours des dernières années. La taille effective de la population des fondateurs a été de 95,19 animaux. Le nombre d’ancêtres ayant contribué à cette population a été de 5086, dont 56 sont responsables du 50 pour cent de la diversité génétique de la race. La consanguinité est sous contrôle dans la race équine Créole au Brésil. L’augmentation des inscriptions reflète l’intérêt croissant des éleveurs pour cette race.

Mots-clés: consanguinité, parenté moyenne, indice de conservation génétique, statistiques de Wright, ancêtres fondateurs, généalogie, cheval créole, Brésil

Resumen Se realizó un análisis genealógico de los caballos Criollos registrados en la Asociación Brasileña de Criadores de Caballo Criollo. El estado de Río Grande del Sur es el mayor productor de animales con el 89,85 por ciento de los animales registrados, de los cuales el 45,29 por ciento son machos y el 54,71 por ciento hembras. El coeﬁciente de endogamia se estimó en un 0,88 por ciento, con un parentesco medio de 0,65 por ciento en la población total, contando tanto con animales nacidos en Brasil como con animales importados. La endogamia aumentó de generación en generación. Se detectó un incremento en el índice medio de conservación genética en los animales nacidos entre 1927 (0) y 2010 (8,67 por ciento). El intervalo generacional medio fue de 10,3 años y la edad media de los progenitores al nacimiento del primer descendiente fue de 10,5 años, habiéndose reducido esta edad en los últimos años. El tamaño efectivo de la población de fundadores fue de 95,19 animales. El número de ancestros que han contribuido a esta población fue de 5086, de los cuales 56 eran responsables del 50 por ciento de la diversidad genética de la raza. La endogamia está bajo control en la raza equina Criolla en Brasil. El aumento de registros reﬂeja el interés creciente de los ganaderos por esta raza.

Palabras clave: endogamia, parentesco medio, índice de conservación genética, estadísticos de Wright, ancestros fundadores, genealogía, caballo criollo, Brasil

Submitted 19 April 2013; accepted 17 October 2013

Correspondence to: C.M. McManus, Universidade de Brasília, Brasília, DF 70910-900, Brazil. email: [email protected]

115 116 F.C. Maciel et al.

Introduction According Laat (2001), population aspects of horse herds, especially those related to inbreeding and effective size, Brazil has the third largest herd of horses in the world with present peculiarities since the horse herds are generally ~5.9 million animals, second only to China with 7.9 mil- small, selection does not have well-defined goals and lion and Mexico with 6.2 million. The equine generates economic characteristics to evaluate progress, and sports about 640 000 direct jobs and if we consider indirect prizes, morphology and associated functionality are not jobs about 3.2 million, according to ESALQ/USP (2006). objectively measured. Furthermore, the low uses of reproductive biotechnologies such as artificial insemin- The Crioula horse of southern Latin America is a direct – descendant of the horses brought to the New World by ation, embryo transfer and IVF which are prohibited the Spanish and Portuguese conquistadores during the six- by the Association of Crioula breeders, affect population parameters. teenth century (Rodero, Delgado and Rodero, 1992). Many horses escaped or were abandoned, and returned to a wild Despite its importance, no studies on the level of inbreed- state. These were Portugese, Barbe and Spanish (particu- ing and population structure of the Crioula breed and, larly Andaluse) horses. therefore, the objective of this study was to analyse the current situation of the breed using population genetics Studies with Uruguayan (Kelly et al., 2002) and parameters, focusing especially on some aspects of genetic Argentinian Crioulas (Mirol et al., 2002) using blood variability and population structure. group and protein polymorphism variants that are considered to be breed markers of Spanish Pure-bred and Barb horses were detected in the Crioula breed but some micro- Material and methods satellites and protein polymorphisms alleles were found uniquely in the Crioula horse. Argentinean Crioula horses The analysis was based on data from the studbook where shared two haplotypes with the Peruvian Paso from the population was 341 616 animals that included all regis- Argentina, and the commonest haplotype of the Crioula tered animals (final and interim) by the Brazilian horses is identical to one of the Andalusian horses. Even Association of Crioula Horse Breeders (ABCCC) since when there was substantial subdivision between breeds its foundation up to 8th April 2011 (154 663 males and with highly significant Wright’s Fixation Index (FST), 186 843 females). Data were analysed using ENDOG pro- the parsimony and distance-based phylogenetic analyses gramme to calculate population parameters for the breed. failed to show monophyletic groups and there was no The following parameters were calculated: the pedigree clear relationship in the trees between the South American completeness level was computed as the proportion of and any of the other horses analysed. Although this result ancestors known per parental generation (MacCluer could be interpreted as mixed ancestry of the South et al., 1983); The number of equivalent to discrete genera- American breeds with respect to the Spanish breeds, it is tions (t) for each individual in a pedigree (Boichard, probably indicating the retention of very ancient maternal Maignel and Verrier, 1997); The inbreeding coefficient lineages in the breeds analysed. (F) (Malécot, 1948); average relatedness coefficient (AR) (Goyache et al., 2003; Gutiérrez, Goyache and The South American horse breeds form two main groups, Cervantes, 2009). The probability of gene origin was char- the Crioula types (including Brazilian, Argentinean and acterized by computing the following parameters: effective Uruguayan) and the other Brazilian breeds (Cothran number of founders ( f ) (James, 1977) computed from the et al., 1998) and the Crioula is most closely related to e genetic contribution of founders to the descendant gene Iberian Sorraia, Andalusian, Paso Fino and Lusitano pool of the population (Lacy, 1989); effective number of breeds. The pairing of the Sorraia with the Argentine ancestors ( f ), defined as the minimum number of ances- Crioula was probably due to the combination of low vari- a tors, not necessarily founders, explaining the complete ability of the Sorraia, small sample size of the Argentine genetic diversity of a population (Boichard, Maignel and Crioula and the fact that an Argentine Crioula stallion Verrier, 1997); the founder genome equivalents (Ballou was used in the regeneration of the Sorraia. and Lacy, 1995) obtained by the inverse of twice the aver- According to the Brazilian Association of Crioula Horse age coancestry of the individuals within the population Breeders (ABCCC), from 2001 until the end of 2010 (Caballero and Toro, 2000); Effective population size there was an increase of 143.7 percent in applications for (Ne) was computed following Gutiérrez, Goyache and registration of temporary records, jumping from 10 271 Cervantes (2009); F-statistics where a population has a in 2001 to 25 020 in 2010. Based on this growth, the population structure of two levels; one from the individual Association estimates that this number could reach (I ) to the subpopulation (S) and one from the subpopula- 57 502 in the year 2020, an increase of approximately tion to the total (T ). F-statistics describe the amount 129.82 percent over 2010. Trade of animals increased inbreeding-like effects within subpopulations FIS, among 985 percent between 2001 and 2010, rising from $9.2 to subpopulations FST, and within the entire population FIT. $100 million, respectively, and this number could exceed In this case Wright’s F-statistics are obtained according $300 million in the year 2020, according to data from to Caballero and Toro (2000, 2002); The within-breed the ABCCC. coancestry ( fii) and the between-breeds coancestry matrix Population structure of the Brazilian Crioula Horse 117

( fij) were computed averaging all pairwise coancestry coefﬁcients of the individuals belonging, respectively, to a given breed i or to two different breeds i and j. Following Caballero and Toro (2000, 2002) the between- breeds Nei’s minimum distance (Dm) matrix was also computed as Dm =((fii + fjj)/2) − fij where fii and fjj are the average coancestry within two breeds i and j and fij the coancestry between two breeds i and j. In the present study, municipalities, farms and sires were considered separate populations; the effective number of founders in an animal’s pedigree was calculated using the genetic conservation index, genetic conservation index (GCI; Alderson, 2 1991): GCI =1/ΣPi where Pi is the proportion of genes of founder animal i in the pedigree. Subsets were also analysed using sires with more than 100 offspring (803) giving 163 817 foals, the fathers of these animals (196) giving 53 632 foals. Farms (106) and municipalities with more than 500 horses registered were also investigated. Two further reference populations were created – families of stallions with more than 500 offspring (184 005 animals – FAMILY) and animals registered since 1998 (184 669 animals – RECENT). In the latter two populations Figure 1. Geographical distribution of the population (Pop) of Crioula horses 135 449 animals appeared in both. Patterns of inbreeding in in Brazil. the three populations studied (ALL, RECENT, FAMILY). Levels of inbreeding were grouped (0.00, 0.001–0.05 (0.05), 0.051–0.10 (0.10), 0.101–0.15 (0.15), 0.151–0.20 (0.20), 0.201–0.25 (0.25) and >0.25 (0.30).

Results

The Crioula horse is present in 23 of the 27 states of the Federative Republic of Brazil (Figure 1). The Crioula horse does not have animals born in states of Alagoas, Maranhão, Rio Grande do Norte and Amapá. The largest producers are the three southern states of the country that together add up to 96.34 percent, and Rio Grande do Sul, the largest producer with 89.85 percent, 3.81 percent in Paraná and Santa Catarina with 2.67 percent of the Figure 2. Number of Crioula horses imported per year from Uruguay, population total. Horses were also imported from Argentina and Chile. Argentina (866), Chile (868) and Uruguay (852) of In this year, there was a decrease in the number of records which ~80 percent were females. While horses from of animals to 1996. From 1996 to 1998 the number of Argentina and Uruguay have been imported since the records was almost stable, and from 1999 there was a start of breed registration, those from Chile have been more recent becoming the most important source of importation (Figure 2). Of all registered animals, 97.98 percent have their pedigrees fully known. The number of known pedigrees has increased over the generations with emphasis on the growth of the ninth generation. In the following, six generations this increased from 4.95 percent of the known pedigrees to 97.98 percent in the current 15th generation (Figure 3). The increase in the number of records per year is evident (Figure 4). Since the 1970s there has been a steady growth in the breed until the mid-1990s, more precisely in 1992. Figure 3. Percentage of known generation pedigrees of the Crioula horse in Before 1980, 9.3 percent of all animals were registered. Brazil. 118 F.C. Maciel et al.

Figure 4. Percentage of records by year of birth for the Crioula horse in Figure 5. Percentage of records by month of birth for the Crioula horse in Brazil. Brazil.

higher growth than seen in the 1970s and 1980s. From the Cabanha Paineiras (2994), Cabanha St. Angelo (2370) and year 2005, growth has intensified, with an increase in 3042 Estância San Francisco (2252), and these five farms, from animals from 2006 to 2007. The numbers of records for the a total of 7949, were responsible for 4.5 percent of regis- years 2010 and 2011 should be disregarded, since the data tered animals. When you consider the 50 largest farms, were collected in April 2011 and most of the animals born they, together, represent 19.67 percent of the animals in the spring of 2010 onwards had not been through tech- reared in the country, reaching 63 977 for a total of nical inspection on the date on which the data were 325 301 animals. gathered. The inbreeding coefficient was calculated at 0.88 percent Of the animals recorded by the ABCCC, 54.71 percent and the AR was 0.65 percent in the total population are female and 45.29 percent are males, in a total of (animals born in Brazil and imported). With the increase 341 497 observations. As for the month of birth, 90.71 per- of generations, an increase in inbreeding was observed cent of registered animals are born between September and (Table 1). This increase is evident from the first generation, January, the largest concentration in the months of October and especially observed in the fourth generation full gener- fi and November where we nd, respectively, 25.85 and ation. 96.58 percent of the animals which have some 26.44 percent of births (Figure 5). degree of inbreeding and inbreeding coefficient in this gen- Farms with the largest number of animals were registered eration was 2.78 percent. There is an increasing trend up to for Estância Nazareth (3930), Cabanha Tupambaé (3114), 100 percent of animals in the 15th generation being inbred,

Table 1. Inbreeding (F), average relatedness (AR) and effective population size (Ne) per generation in Brazilian Crioula Horse.

Generation No. of animals F Mean (%) % Endogamic animals F Mean for endogamic animals % Mean AR Ne

Endogamy per generation1 0 5882 0.00 0.01 1 7157 0.00 0.16 2 7794 0.68 3.77 18.15 0.57 73.0 3 10 997 0.93 8.11 11.41 0.88 206.5 4 15 147 1.08 14.72 7.33 0.90 323.2 5 23 019 0.94 18.73 5.04 1.04 6 38 604 0.93 27.46 3.40 1.35 7 59 015 1.10 41.55 2.65 1.60 2117.7 8 67 583 1.30 57.17 2.28 1.79 243.7 9 56 469 1.44 68.72 2.10 1.96 353.2 10 34 183 1.72 78.87 2.18 2.14 180.6 11 13 532 1.85 84.85 2.18 2.24 370.4 12 2004 2.16 85.83 2.52 2.24 155.9 13 186 1.99 87.10 2.29 2.18 14 42 2.27 95.24 2.38 2.13 461.7 15 2 13.10 100.00 13.10 1.30 4.5 Endogamy per complete Generation2 0 7901 0.00 0.08 1 64 559 0.27 2.39 11.12 0.83 188.0 2 18 1704 1.25 48.33 2.59 1.77 50.4 3 84 471 1.89 80.99 2.33 1.89 77.7 4 2981 2.78 96.58 2.88 1.69 55.0

1Maximum number of generations traced. 2Complete generations traced as deﬁned by Gutiérrez, Goyache and Cervantes (2009). Population structure of the Brazilian Crioula Horse 119

Table 2. Average number of generations, inbreeding per generation and effective population size for the Brazilian Crioula horse.

Mean Increase in Effective generations endogamy (%) population size

All Maximum 12.09 0.05 994.64 Complete 1.85 0.50 99.18 Equivalent 4.62 0.23 214.46 Family Maximum 7,15 0.14 348.53 Complete 2.03 0.76 65.45 Equivalent 3.40 0.54 93.28

which is not complete since the animals can be registered Figure 7. Genetic conservation index (GCI) and average relatedness (AR) definitively up to 4 years after birth. The number of gene- average per year of birth for the Brazilian Crioula horse. rations, complete and equivalent, shows that the breed is expanding (Table 2). Large variations in effective herd After 1970 there was a reduction in the coefficient of aver- size per generation are an indication of animals being age individual inbreeding, and the animals born in the past imported into the herd (for example, from Uruguay, decade had rates of ~1.50 percent (Figure 7). An increase Chile or Argentina). in the average rate of genetic conservation was seen in ani- The highest level of inbreeding was 0.41 with 6755 animals mals born after the year 1927 until 2010, which had the with a value >0.1. These animals were on average in the highest average GCI, 8.67 percent (Figure 7). seventh generation, but there were animals in the second The interval between generations (Table 3) was higher for generation with high levels of inbreeding. Analysing the mares than stallions. This indicates that mares are kept full herd, 72 (0.02 percent) animals were from crosses longer for reproduction or start reproductive life later. between full-sibs, 3047 (0.89 percent) between half-sibs There was no significant genetic differentiation between and 1784 (0.52 percent) between parents and offspring. the states of the federation where animals are raised The mean coefficient of individual relatedness showed an when considered sub-populations (not shown). Average increase in population over the years, especially after coancestry within subpopulations was 0.0084 and average the1950s, reaching 2.09 percent in animals born in the coancestry in the metapopulation was 0.0079. Wright para- year 2010 (Figure 6). meters show that the inbreeding in the population is greater An increase in average individual relatedness was found than expected if matings were performed at random and from the 1940s until 1970, the largest observation in which occur between the mating subpopulations. 1955 was 3.11 percent. In recent years, there was an aver- The effective population size of founders was 95.19 ani- age increase of inbreeding about 0.50 percent (Figure 7). mals, expected inbreeding due to an imbalance in the con- An increase in the average individual coefficient of tribution of founders was calculated as 0.53 percent and inbreeding was found in animals born after the 1940s, the mean inbreeding coefficient computed was 1.21 and in 1955 an average rate of 2.95 percent was seen. Table 3. Generation interval and the average age of parents when their offspring born for the Brazilian Crioula horse.

Relation N L Se

Generation interval Sire–Son 10.580 9.986 0.051 Sire–daughter 74.037 9.737 0.018 Dam–son 10.647 10.611 0.045 Dam–daughter 74.006 10.781 0.018 Total 169.270 10.264 0.012 Mean age of parents when offspring born Sire–Son 153.294 9.880 0.013 Sire–daughter 180.964 9.983 0.012 Dam–son 153.901 11.042 0.012 Dam–daughter 181.173 11.097 0.011 Total 669.332 10.504 0.006

Figure 6. Mean individual inbreeding (F) and Average individual change in N is the number of animals, L is the generation interval and SE is the inbreeding (AF) by year of birth for the Brazilian Crioula horse. standard error. 120 F.C. Maciel et al.

Table 4. Measures of genetic variation and Wright’s statistics (FIS, Table 6. Structure of the Brazilian Crioula Horse herds by state. F and F ) for Brazilian Crioula Horse. ST IT State1 Region Type % parents % males Parameter All Recent Family born in born in state state/ which are Mean coancestry within 0.008414 0.009137 0.010384 number of sires/number Subpopulations births of sires Autocoancestry 0.506043 0.506043 0.506043 Endogamy 0.012087 0.012087 0.012087 Acre N Commercial 0.00 0.00 Nei’s Distance 0.000516 0.001238 0.002486 Bahia NE Multiplier 20.19 77.78 Mean Coancestry in Metapopulation 0.007898 0.007898 0.007898 Ceará NE Commercial 0.00 0.00

FIS 0.003704 0.002977 0.001721 Federal CO Multiplier 0.48 3.51 FST 0.000520 0.001249 0.002505 District FIT 0.004222 0.004222 0.004222 Espirito SE Commercial 0.83 100.00 Santo Goiás CO Multiplier 5.25 46.81 Minas Gerais SE Commercial 5.47 100.00 percent for the population of 341 616 animals and the base Mato Grosso CO Multiplier 8.25 39.76 population (animals with one or more parents unknown) do Sul was 7901 (Table 4). The ratio of founders to ancestors Mato Grosso CO Multiplier 0.74 15.79 was 1.21 (91/75) meaning that in general no speciﬁc Pará N Commercial 0.00 0.00 lineages were created within the Crioula population. Pernumbuco NE Commercial 0.00 0.00 Piauí NE Multiplier 32.79 9.13 The reference population (both parents known) was smal- Paraná S Multiplier 36.01 44.95 ler than that used for the analysis of founders. The number Rio de SE Commercial 6.82 100.00 Janeiro of ancestors contributing to this population was 5086 for a Rondônia N Commercial 8.06 100.00 reference population of 333 715 animals, where 56 ances- Roraima N Commercial 0.00 0.00 tors explained 50 percent of the genetic diversity of the Rio Grande S Multiplier 81.74 92.70 breed. Animals 9838, 5174, 18 182, 5178, 5532, 5497, do Sul 5208, 5418, 5439 and 5449 had the largest contribution Santa S Multiplier 12.83 70.37 Catarina to the herd, reaching ~25 percent of the total genetic vari- Sergipe NE Commercial 0.00 0.00 ation. The largest number of offspring per stallion was São Paulo SE Multiplier 12.29 37.97 1428 (sire 63 592) with four animals (9838, 97 059 and Tocantins N Multiplier 16.36 60.27 5208) with more than 1000 offspring. Animals 63 592 1Brazilian States; N, North; NE, Northeast; S, South; SE, southeast; CO, and 97 059 are offspring of 9838, and eight stallions had Midwest. more than 500 offspring (forming the second reference population – FAMILY). Animal 9838 had the highest percent (7/11) did not use their own sires for reproduction, average relatedness. Mare 49 893 had the most offspring while 36.36 percent (4/11) using both purchased and own (20), but more than 12 292 mares had more than ten males for reproduction. Rio Grande do Sul state (RS) had offspring, showing the proliﬁcacy and longevity of the the highest percentage of sires born in the State, which is breed. not surprising as this is the home state for this breed. The total population of animals was divided by their state of birth and these subpopulations were analysed for type Even when subpopulations are studied, genetic parameters (Table 5) and structure (Table 6). (F, AR, AF) vary little from the mean (Table 7), showing

Only Commercial and Multiplier herds were seen, with no Table 7. Genetic variation by subpopulation (sires and grandsires nuclei or disconnected herds. Most multiplier states (14/ with largest number of offspring, farms and municipalities) for the 15) use their own breeding males, buy in animals and Brazilian Crioula horse. sell sires to other states. As for commercial herds, 63.64 Grandsire Sire Farms Municipalities

Number 196 803 106 49 Offspring 53 632 163 817 935.68 4242.22 Table 5. Number of herds by type and characteristics for the F 0.010 0.011 0.013 0.011 Brazilian Crioula horse. AR 0.018 0.018 0.016 0.015 Type Buy-in Use own Sell No. of % Sires Gen Max 6.043 6.952 6.878 7.308 sires sires sires states bought Gen Com 1.552 1.852 1.915 2.085 Gen Eq 2.737 3.196 3.236 3.509 Nucleus No Yes Yes 0 0.00 AF 0.005 0.005 0.005 0.004 Multiplier Yes Yes Yes 14 24.22 GCI 4.722 6.014 2.818 1.793 Multiplier Yes No Yes 1 100.00 Commercial Yes Yes No 4 93.69 Genetic conservation index (GCI), average relatedness (AR), mean indi- Commercial Yes No No 7 100.00 vidual inbreeding (F) and average individual inbreeding (AF), GenMax, Isolated No Yes No 0 0.00 maximum number of generations; Gen_Com, common number of generations; Gen_Eq, equivalent number of generations. Population structure of the Brazilian Crioula Horse 121

Table 8. Population data depending on recently born (from 1998) or family of sires with largest offspring numbers (family) in Crioula horses from Brazil.

All Recent Family

No. of reference animals 333 715 184 646 184 005 No. of ancestors 5086 4230 3334 Effective number of founders 91 62 49 Effective number of ancestors 75 (1.21) 45 (1.38) 35 (1.40) No. of ancestors explaining 50% 56 41 31 Base population (one or more unknown parents) 7901 Actual Base Population (one unknown parent = half founder) 6891.5 Effective Population Size of Founders 95.19 Expected F by unbalancing of founders contribution 0.53% Computed mean F 1.21% Mean Average Relatedness 1.58% Age first becoming parent 10.21 6.92 10.19 Regression coefficient of age becoming parent on inbreeding coefficient −13.18 ± 8.89 −1.11 ± 0.51 −21.35 ± 1.00 Regression coefficient of age becoming parent on increase in inbreeding coefficient −15.75 ± 16.71 0.32 ± 1.22 −34.77 ± 2.25 Ne from regression birth date 274.48 215.59 Ne from log regression birth date 270.80 230.19 Ne via individual increase in inbreeding 98.94 ± 20.01 97.23 ± 19.83 Ne from Log regression on equivalent generations 68.88 58.06 Ne via regression on equivalent generations 70.14 58.90

Ne, effective population size.

Figure 8. Percentage of animals registered per year depending on inbreeding level on all (ALL) animals registered, those from the major families (FAMILY) and the present generation (RECENT)Levels of inbreeding (0.00 (0.0), 0.001–0.05 (0.05), 0.051–0.10 (0.10), 0.101–0.15 (0.15), 0.151–0.20 (0.20), 0.201–0.25 (0.25) and greater than 0.25 (0.30).

Figure 9. Mean inbreeding (F) of Crioula Horse depending on the date the farm started to register animals, the period of time the farm registered animals and the number of animals registered per farm. that sires are imported into farms as a general rule. GCI that the major sire families produce a large proportion of was significantly higher than herd average for those stal- the recently born animals, reflected by the large number lions that left more offspring, as expected. of animals in common in both groups. More recently animals are becoming parents at a relatively It can be seen that the more recently farms started to register younger age. This may be a reflection of the rapid expan- animals the higher mean F (Figure 8, Figure 9). This is also sion of the breed and the need for breeding females to reflected in the longer the farm has been registering animals reproduce (Table 8). Effective population sizes were simi- and the higher the number of animals registered the lower lar for different subgroups. This is a reflection of the fact the mean F. 122 F.C. Maciel et al.

Discussion In domestic animals some individuals assume importance in the origin of the breed and its development (Fletcher, Due to their high commercial value, the horses, especially 1945, 1946; Rhoad and Kleberg, 1946; Gazder, 1954)as those used for sports, need have their pedigree known, evidenced in this population. This is due to several factors, both for registration and purchase. Confirmation of geneal- such as propaganda, success in auction rings and competi- ogy is extremely important, not only for ensuring the tions as well as functional and morphological traits. ascendancy of the animals, but also because a pedigree The concentration of births in the months from September can reliably enable the buyer to identify the origin of gen- to January is mainly due to the occurrence of marked etic problems in the herd and reduce or eliminate them photoperiod in southern Brazil (Winter, 2007). Most (Coelho and Oliveira, 2008). The study of genetic variabil- mares show anestrus and transitional phases in summer ity and population structure in some horse breeds using and autumn and, under natural conditions, the transition pedigree analysis, alone or in combination with informa- ends near or after the equinox, in September and October. tion such as genetic markers, has grown in recent years (Chiofalo et al., 2003; Valera et al., 2005). The mean coefficient of inbreeding found in Crioula was lower compared to other local breeds. Schurink, Arts and The results on the distribution of this breed in Brazil were Ducro (2012) found average inbreeding coefficient of in accordance with expected, with a high concentration of ani- 0.053 while the Crioula horse was found 0.88 percent mals in the southern region, since the breed was developed in for the total population (including imported) and 1.21 per- Southern Latin America, a region known as South American cent for those born in Brazil. Wolc and Balin´ska (2010) Pampa (Argentina, Southern Brazil and Uruguay). The also found much higher values (5.9, 5.1 and 5.6 percent) increase in the use of the Crioula Horse in southeastern states on three farms breed Polish Konik. Abrahão et al. such as Sao Paulo, as well as in Northern and Northeastern (2002) found similar levels of inbreeding (0.8 percent) to fl Brazil is also seen. This re ects the expansion and distri- those found in Crioula when studying inbreeding in bution of the breed, where the largest breeder (Estância Thoroughbred mares reared in Brazil. Low levels of Nazareth) has only 1.2 percent of the animals, unlike other inbreeding found in Crioula in Brazil can be explained local breeds, such as the Pantaneiro Horse, where the largest by the large population base, large number of founders breeder has 7.25 percent of animals recorded. and ancestors, and the introduction of animals from other The amount of information on the Crioula has increased countries (Argentina, Uruguay and Chile), introducing dif- with the number of generations. These data are consistent ferent racial strains. with those found in studies with other breeds where the The number of inbred animals found in the full third gen- most distant generations have less genealogical informa- eration (80.99 percent of the population) was similar to tion, as this may have been lost over the years or were that found by Mota et al. (2006) found that 88.0 percent not recorded (Valera et al., 2005). Vicente, Carolino and of inbred animals in Mangalarga horses in the period Gama (2009), as here, found an increase in complete 1936–2003. The results in Crioula show a clear increase Lusitano pedigrees over ten generations. in the number of inbred animals, reaching 100 percent at the 15th generation, although there are only two animals The increase in the number of records per year is evident in this generation. Inbreeding of 10 percent was estimated (Figure 2), mainly due to two factors: (a) partnership for ethnic breeds such as Spanish Purebred (Valera et al., between the Breeder’s Association and the Rural TV 2005) and Lipizzano (Zechner et al., 2002), and after the Channel, where the most important competition involving closing of the studbook this parameter has aggravated the Crioula, Freio de Ouro (which involves several tests, (Vicente et al., 2009). For Mangalarga inbred animals, such as breed standard, gait, barrel racing, separating cat- the mean coefficient of inbreeding was 5.7 percent, with tle, lassoing, horse control and obedience), is televised a maximum of 46.9 percent (Costa et al., 2005), similar across the country and worldwide over the internet. This to those found by Procópio et al. (2003) in the Campolina. channel also televises most of the auctions with these horses, so farmers from distant regions can acquire animals The increase of inbreeding (Boichard, Maignel and without leaving home; (b) projects developed by the Verrier, 1997) can be used to derive the effective size of ABCCC, such as Vaquejada Project (A northeastern trad- a population. However, this method reflects mainly the ition where horses are used to drive cattle in a confined long-term effects of selection and, furthermore, is very space.), created to insert Crioula horse in this activity in sensitive to incomplete information from a pedigree. An the northeast. The number of registers in the last two alternative would be additional parameters based on the years (2010 and 2011) should be disregarded as the ani- probability of gene origin, and the effective number of mals are subjected to technical inspection until they are founders and effective number of remaining genomes are nine months old and most of the animals born in this per- commonly used in wild populations, but less common in iod had not passed inspection at the time of data collection. studies of domestic animals. However, there are several The greater number of females compared to males can be studies in domestic species and different breeds, such as explained by fact that some breeders do not register males in cattle (Faria et al., 2002; Vercesi Filho et al., 2002; which do not meet their specifications. Pereira et al., 2005; Vozzi et al., 2006, 2007; Hammami Population structure of the Brazilian Crioula Horse 123

et al., 2007; Martínez, García, and Gallego, 2008), in In Andalusian horses, Valera et al. (2005) found 39.6 horses (Costa et al., 2005), sheep (Goyache et al., 2003), effective founders and 27 ancestors, with only six animals pigs (Toro et al., 2000) and even donkeys (Gutiérrez responsible for 50 percent of genetic variability in the et al., 2005), based on some sort of pedigree analysis. breed. The situation of Crioula is better, with nearly Ideal levels of inbreeding range from 0.05 (Nicholas, 95 founders, but this is lower than for Pantaneira 1989) to 0.01 (FAO, 2000) per generation, and the (McManus et al. 2013), perhaps due to rapid expansion Crioula are below these levels. of the breed in recent years. The small number of founder animals highlights the need for monitoring of inbreeding of Moureaux et al. (1996) reported that when studying the herd studied, control of breeding and introduction of inbreeding in horse breeds, two groups can be distin- animals that have no direct relationship with the principal guished: a group of international breeds, with values ran- ancestors identified in this study. Alderson (1991) used ging from 0.81 to 2.89 percent, and another group with the GCI to calculate an effective number of founders in small population sizes with values ranging from 2.25 to the pedigree of an animal. The higher the value GCI better 14.7 percent. The high variation in mean individual levels it is for use in conservation of a breed. The GCI can be of inbreeding and inbreeding rate seen here between 1940 used both by individual breeders as an aid in selecting and 1970 may have been a reflection of the lack of breed- breeding stock, or within a breed to formulate a compre- ing animals, the difficulty in finding stallions of the breed hensive improvement program. However, the index has as well as in moving between farms, since the increase in limitations, as it is not representative of any concentration effective herd size came after 1970 (Figure 3). of breeding animals for non-founders in subsequent genera- The mean coefficient of relatedness (kinship) increased tions and cannot be used without pedigree records from the start of breed registration until about the year (Alderson, 1991). 2003. These results can be explained by selection of animals which showed excellent results in morphological The interval between generations found in Crioula was and functional competitions, and therefore many breeders higher when compared with other breeds such as Costa sought to acquire its offspring. Mean coefficient of related- (2002) and Gonçalves (2010), who evaluated animals ness provides additional information to explain relations Mangalarga Marchador found generation interval of 8.9 between relatives (Gutiérrez, Goyache and Cervantes, and 8.98 years, respectively. This may be due to the pro- 2009). In conservation genetics, knowledge of relatedness hibition of use of reproductive biotechnologies until is required to optimize conservation strategies. Coancestry 2011, as well as the large number of morphological and relationship is expressed relative to the base population in functional competitions in which these horses compete, which all alleles are not defined as being identical by des- means that these horses begin reproductive life at a later cent, so that the population coancestry base is zero by age. As natural breeding is still used, many mares exposed definition (Falconer and Mackay, 1996; and Lynch to the stallion do not impregnate. Vicente, Carolino and Walsh, 1998). Kinship average here was low compared to Gama (2009) also found a high generation interval for other breeds of horse (Zechner et al., 2002; Valera et al., the Lusitano breed (10.4 years: males 11.2 and females 2005). Valera et al. (2005) found average values of 9.6 years), also possibly due to use after the end of their inbreeding and AR of the Andalusian horse population sporting career. McManus (2013) found lower intervals were, respectively 8.48 and 12.25 percent, while between generations in the Pantaneiro horse and consid- Cervantes et al. (2008) found inbreeding average of 7 per- ered this was because this breed is used on-farm animals cent to Spanish Arabian. instead of sports and therefore enter reproduction earlier. The average individual inbreeding and variation in inbreed- The use of artificial insemination and embryo transfer are ing increased from the mid-1930s to mid-1970s, and beginning to become more widely used in horse popula- remained relatively constant thereafter. This can be explained tions (2011 in Crioula) but less frequently than in other by an intense exchange of animals between Brazil, Argentina, farm animals (Laat, 2001). The population aspects of Uruguay, and especially Chile, where different strains were horses, especially those related to inbreeding and effective incorporated into the local population after the 1970s. size therefore have an individual character (Costa et al., 2005) and continued monitoring is necessary to avoid When comparing inbreeding in the three populations stud- future problems. ied (ALL, FAMILY and RECENT), a clear increase in the percentage of animals with relatively low levels of inbreed- Vinocur et al. (2003) determined the allele frequencies of ing (0.05). The percentage of animals with F > 0.05 has seven blood group systems and eight protein systems in not grown since the late 1990s and although an increase six herds of Crioula horses raised in Rio Grande do Sul in inbreeding (F > 0.1) was seen in the ALL population of State. The herds presented a significant component in the 1940s to 1980, this has largely disappeared. due to isolation (FST =0.0866, p < 0.01). They found high Higher levels of inbreeding are not seen in RECENT or values for average heterozygosity (0.4631). When all FAMILY. The percentage of animals with no inbreeding herds were considered in the analysis, the inbreeding level has steadily decreased in all populations and should (FIS) was zero, in line with seen here where values near increase with the closure of the herd book. zero indicate that levels of genetic variability of the flock 124 F.C. Maciel et al.

are high, since these values indicate an excess of heterozy- small population conservation. New York: Columbia University gous animals, both within and paternal lineages (subpopula- Press, p.76–111. tions) and for the population as a whole (metapopulation). Boichard, D., Maignel, L. & Verrier, É. 1997. The value of using prob- For the FST index, the value calculated was 0.000, demon- abilities of gene origin to measure genetic variability in a population. strating that there is no differentiation between the parental Genet. Sel. Evol., 29: 5–23. strains with the formation of subpopulations. The Crioula is Caballero, A. & Toro, M.A. 2000. Interrelations between effective not at risk status (FAO 1998; Database http://dad.fao.org population size and other pedigree tools for the management of con- DAD-IS) as opposed to other breeds of naturalized horses served populations. Genet. Res., 75: 331–343. (Luís et al., 2005). The integration of herdbooks from Caballero, A. & Toro, M.A. 2002. Analysis of genetic diversity for the other countries with Crioula horses used for crossing with management of conserved subdivided populations. Conserv. Genet., – the Brazilian Crioula (such as Uruguay, Argentina and 3: 289 292. Chile) could help to identify some genetic bottlenecks that Cervantes, I., Molina, A., Goyache, F., Gutiérrez, J. P. & Valera, M. may exist but were not seen here. 2008. Population history and genetic variability in the Spanish Arab Horse assessed via pedigree analysis. Livestock Sci., 113: 24–33. The fact that the smaller, younger farms have higher levels of inbreeding may reflect a policy of the larger farmers to Chiofalo, L., Portolano, B., Liottal, L., Rundo Soteral, A. & Finocchiaro, R. 2003. Demographic characterization, inbreeding import stallions or sell off inbred animals to those starting and genetic variability within Sanfratellano population horse from up in the business. It may also reflect a lack of breeding genealogical data. Ital. J. Anim. Sci., 2(Suppl. 1): 592–594. policy by the smaller farmers who may have a single stal- Coelho, E.G.A. & Oliveira, D.A.A. 2008. Testes genéticos na lion. This is also from the case seen with some pig breeds eqüideocultura. Revista Brasileira de Zootecnia, 37: 202–205. in the USA (Welsh et al., 2010) but to a lesser extent. Special Number. Costa, M.D. 2002. Caracterização demográfica e estrutura genética da raça Mangalarga Marchador. Universidade Federal de Minas Gerais, Conclusion Belo Horizonte, Brazil (Thesis). Costa, M.D., Bergmann, J.A.G., Resende, A.S.C. & Fonseca, C.G. The indices obtained from the genetic data of genealogical 2005. Análise temporal da endogamia e do tamanho efetivo da record of the Crioula Breed show that in general inbreed- população de equinos da raça Mangalarga Marchador. Arquivo – ing is under control in the total population and the effective Brasileiro de Medicina Veterinária e Zootecnia, 57: 112 119. population size is not at a critical level. Recent decrease in Cothran, E.G., Santos, S.A., Mazza, M.C.M., Lear, T.L. & Sereno, J. generation intervals may be a reflection of the rapid expan- R.B. 1998. Genetics of the Pantaneiro horse of the Pantanal region of Brazil. Genet. Mol. Biol. (available at http://dx.doi.org/10.1590/ sion of the breed and the need for mares for reproduction. S1415-47571998000300009). Escola Superior de Agricultura Luiz de Queiroz – ESALQ. Centro de Estudos Avançados em Economia Aplicada da ESALQ. 2006. Acknowledgements Estudo do Complexo do Agronegócio do Cavalo no Brasil. Brasília, CAN, MAPA. Thanks are due to the Breeder’s Association for the Falconer, D.S. & Mackay, T.F.C. 1996. Introduction to Quantitative Crioula Horse, INCT – Pecuária (CNPq, FAPEMIG, Genetics. Harlow, Longman. MCT) and CAPES for scholarships. Faria, F.J.C., Vercesi Filho, A.E., Madalena, F.E. & Josahkian, L.A. 2002. Estrutura populacional da raça Nelore Mocho. Arquivos Brasileiro de Medicina Veterinária e Zootecnia, 54: 501–509. Statement of interest Fletcher, J.L. 1945. A genetic analysis of the American Quarter Horse. J. Hered., 36: 346–352. No conflicting interests were identified. Fletcher, J.L. 1946. A study of the first fifty years of Tennessee Walking horse breeding. J. Hered., 37: 369–373. Gazder, P.J. 1954. The genetic history of the Arabian Horse. J. Hered., – References 45: 95 98. Gonçalves, R.W. 2010. Efeito da endogamia nas características Abrahão, A.R., Mota, M.D.S., Oliveira, H.N. & Madureira, A.P. morfométricas e reprodutivas dos eqüinos da raça Mangalarga 2002. Endogamia em éguas da raça Puro-Sangue Inglês. In Marchador. Universidade Estadual de Montes Claros, Montes Proceedings IV Simpósio da Sociedade Brasileira de Melhoramento Claros, Brazil. (Dissertation) (available at http://www.dominiopu- Animal, Campo Grande, Brazil, pp. 224–225. blico.gov.br/pesquisa/DetalheObraForm.do?select_action=&co_obra= 193645). Alderson, G.L.H. 1991. A system to maximize the maintenance of genetic variability in small populations. In L. Alderson, ed. Genetic conservation Goyache, F., Gutiérrez, J.P., Fernández, I., Gomez, E., Alvarez, I., of domestic livestock,pp.18–29. Wallingford, UK, CAB International. Díez, J. & Royo, L.J. 2003. Using pedigree information to monitor genetic variability of endangered populations: the Xalda sheep breed Ballou, J.D. & Lacy, R.C. 1995. Identifying genetically important indi- of Asturias as an example. J. Anim. Breed. Genet., 120: 95–105. viduals for management of genetic diversity in pedigreed populations. In: Ballou, J.D., Gilpin, M. & Foose, T.J. (Eds) Population manage- Gutiérrez, J.P., Marmi, J., Goyache, F. & Jordana, J. 2005. Pedigree ment for survival & recovery. Analytical methods and strategies in information reveals moderate to high levels of inbreeding and a Population structure of the Brazilian Crioula Horse 125

weak population structure in the endangered Catalonian donkey breed. (Eds) Evolution and animal breeding. Wallingford: CAB J. Anim. Breed. Genet., 122: 378–386. International, p.201–209. Gutiérrez, J.P., Goyache, F. & Cervantes, I. 2009. User´s Guide, ENDOG Pereira, M.C., Mercadante, M.E.Z., Albuquerque, L.G. & Razook, A. v4.6, A Computer Program for Monitoring Genetic Variability of G. 2005. Estimativa de ganho genético a partir de diferenciais de Populations Using Pedigree Information, (available at http://www. seleção e parâmetros populacionais em um rebanho Caracu. Revista ucm.es/info/prodanim/html/JP_Web_archivos/EN_Us_G_.pdf). Brasileira de Zootecnia, 34: 2245–2252. Hammami, H., Croquet, C., Stoll, J., Rekik, B. & Gengler, N. 2007. Procópio, A.M., Bergmann, J.A.G., Costa, M.D. 2003. Formação e Genetic diversity and joint-pedigree analysis of two importing demografia da raça Campolina. Arq. Bras. Med.Vet. Zoot. 55:361– Holstein populations. J. Dairy Sci., 90: 3530–3541. 365. James, J.W. 1977. A note on selection differentials and generation length Rhoad, A.O. & Kleberg, J.R. 1946. The developments of a superior – when generations overlap. Anim. Prod., 24: 109 112. family in the modern Quarter Horse. J. Hered., 37: 227–238. Kelly, L., Postiglioni, A., De Andrés, D.F., Vega-Plá, J.L., Gagliardi, Rodero, A., Delgado, J.V. & Rodero, E. 1992. Primitive Andalusian R., Biagetti, R. & Franco, J. 2002. Genetic characterisation of the livestock and their implication in the Discovery of America. Arch. Uruguayan Crioula horse and analysis of relationships among horse Zootec., 41: 383–400. breeds. Res. Vet. Sci., 72(1): 69–73. Schurink, A., Arts, D.J.G. & Ducro, B.J. 2012. Genetic diversity in the Laat, D.M. 2001. Contribuição genética de fundadores e ancestrais na Dutch harness horse population using pedigree analysis. Livestock raça Campolina. Universidade Federal de Minas Gerais, Belo Sci., 143: 270–277. Horizonte, Brazil. (Dissertation) (available at http://www.ebookcult. com.br/produto/Contribuicao_genetica_de_fundadores_e_ancestrais_ Toro, M.A., Rodrigañez, J., Silio, L. & Rodriguez, C. 2000. na_raca_campolina-10151). Genealogical analysis of a closed herd of Black Hairless Iberian pigs. Conserv. Biol., 14: 1843–1851. Lacy, R.C. 1989. Analysis of founder representation in pedigrees: founder equivalents and founder genome equivalents. Zoo Biol.,8: Valera, M., Molina, A., Gutiérrez, J.P., Gómez, J. & Goyache, F. 111–123. 2005. Pedigree analysis in the Andalusian horse: population structure, fl Luís, C., Cothran, E.G., Oom, M.M. & Bailey, E. 2005. Major histo- genetic variability and in uence of the Carthusian strain. Livestock – compatibility complex locus DRA polymorphism in the endangered Prod. Sci., 95: 57 66. Sorraia horse and related breeds. J. Anim. Breed. Genet., 122: 69–72. Vercesi Filho, A.E., Faria, F.J.C., Madalena, F.E. & Josahkian, L.A. Lynch, M. & Walsh, J.B. 1998. Genetics and analysis of quantitative 2002. Estrutura populacional do rebanho Tabapuã registrado no traits. Sunderland: Sinauer Assocs. Inc., 980p. Brasil. Arquivos Brasileiro de Medicina Veterinária e Zootecnia, 54: 609–617. Maccluer, J.W., Boyce, A.J., Dyke, B., Weitkamp, L.R., Pfenning, D.W. & Parsons, C.J. 1983. Inbreeding and pedigree structure in Vicente, A., Carolino, N. & Gama, L.T. 2009. Indicadores Standardbred. J. Hered., 74: 394–399. Demográficos no Cavalo Lusitano. Arch. Zootec., 58(Supl. 1): 501–504. Malécot, G. 1948. Les Mathématiques de l’Hérédité. Masson et Cie, Vinocur, M.E., Brass, K.E., Rubin, M.I.B. & Silva, C.A.M. 2003. Paris Genetic variability in the Brazilian Crioula horse breed. Cienc. Rural, 33: 137–142. Martínez, R.A., García, D., Gallego, J.L. 2008. Genetic variability in Colombian Crioula cattle populations estimated by pedigree informa- Vozzi, P.A., Marcondes, C.R., Magnabosco, C.U., Bezerra, L.A.F. & tion. J. Anim. Sci., 86: 545–552. Lôbo, R.B. 2006. Structure and genetic variability in Nellore (Bos indicus) cattle by pedigree analysis. Genet. Mol. Biol., 29: 482–485. McManus, C. 2013. Evaluation of Conservation Program for the Pantaneiro Horse in Brazil, Revista Brasileira de Zootecnia, in press. Vozzi, P.A., Marcondes, C.R., Bezerra, L.A.F. & Lôbo, R.B. 2007. Pedigree analyses in the Breeding Program for Nellore Cattle. McManus, C., Santos, S.A., Dallago, B.S.L., Paiva, S.R., Saraiva, R. – F., Braccini Neto, J., Marques, P.R. & Abreu, U.G. 2013. Genet. Mol. Res., 6: 1044 1050. Evaluation of conservation program for the Pantaneiro horse in Welsh, C.S., Stewart, T.S. & Schwab, C. 2010. Pedigree analysis for 5 Brazil. Rev. Bras. Zoot. 42:404–413. swine breeds in the United States and the implications for genetic con- – Mirol, P.M., Peral García, P., Vega-Pla, J.L. & Dulout, F.N. 2002. servation. J. Anim. Sci., 88:1610 1618. Phylogenetic relationships of Argentinean Crioula horses and other Winter, G.H.Z. 2007. Características reprodutivas sazonais da égua South American and Spanish breeds inferred from mitochondrial Crioula em um propriedade latitude 29°38′S no Rio Grande do Sul. – DNA sequences. Anim. Genet., 33: 356 363. Universidade Federal de Santa Maria, Brazil. (Dissertation) (available Mota, M.D.S., Prado, R.S.A. & Sobreiro, J. 2006. Characterization of at http://www.openthesis.org/documents/Reprodutivas-sazonais-da- the Mangalarga horse population in Brazil. Arch. Zoot., 55:31–37. crioula-em-473051.html). Moureaux, S., Verrier, É., Ricard, A. & Mériaux, J.C. 1996. Genetic Wolc, A., Balin´ska, K. 2010. Inbreeding effects on exterior traits in variability within French race and riding horse breeds from genea- Polish konik horses. Arch. Anim. Breed., 53: 1–8. logical data and blood marker polymorphisms. Genet. Sel. Evol. Zechner, P., Sölkner, J., Bodo, I., Druml, T., Baumung, R., Achmann, 28:83–102. R., Marti, E. Habe, F. & Brem, G. 2002. Analysis of diversity and Nicholas, F.W. 1989. Incorporation of new reproductive technology in population structure in the Lipizzan horse breed based on pedigree genetic improvement programs. In: Hill, W.G., Mackay, T.F.C. information. Livestock Prod. Sci., 77: 137–146. Animal Genetic Resources, 2014, 54, 127–134. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633613000362

Pedigree and population viability analyses of a conservation herd of Moura pig

H. Carneiro1, S.R. Paiva2, M. Ledur3, E.A.P. Figueiredo3, V.H. Grings3, F.C.P. Silva1 and C. McManus1 1Faculdade de Agronomia e Medicina Veterinária, Brasília, DF, Brazil; 2EMBRAPA Recursos Genéticos e Biotecnologia, Brasília, DF, Brazil; 3EMBRAPA Suínos e Aves, Concordia, SC

Summary Pedigree and population viability analyses of a conservation herd of Moura breed from Embrapa Swine and Poultry were carried out. ENDOG software was used for pedigree analysis and VORTEX for population viability analysis. The data were obtained with the curator of the herd and a total of 411 animals born between the years of 1999 and 2011 for the analysis with ENDOG. The input data in VORTEX were obtained from a questionnaire answered by the curator of the herd. The Inbreeding Coefﬁcient (F) was 0.06 percent, the average effective population size (Ne) was 30 animals with an average generation interval of 3.48 years. The extinction probability was 0 percent for a simulation period of 100 years and 3 percent for a period of 500 years. The determinant parameters for the viability of this herd were the entry of animals and proliﬁcacy of the breed. Although the study did not indicate immediate risk of extinction of this herd, it should be constantly monitored with controlled population parameters to prevent the extinction of this last herd of Moura pig breed.

Keywords: inbreeding, conservation population modelling, Moura pig

Résumé Les analyses de la viabilité de la population et pedigree ont été effectuées sur un harde de la race Moura de l’Embrapa Porcins et Oiseaux. ENDOG logiciel a été utilisé pour l’analyse de pedigree et VORTEX pour l’analyse de la viabilité de la population. Les données ont été obtenues avec le conservateur du harde et un total de 411 animaux nés entre les années 1999 et 2011 ont été analysés avec ENDOG. Les données d’entrée dans VORTEX ont été obtenus a partir d’un questionnaire auquel ont répondu par le conservateur du harde. Le coefﬁcient de consanguinité (F) était 0,06%, la taille effective de la population moyenne (Ne) était 30 animaux, avec un intervalle de 3,48 ans en moyenne de la production. La probabilité d’extinction était de 0% pour une période de simulation de 100 ans, et 3% pour une période de 500 ans. Les paramètres déterminant pour la viabilité de ce harde a été l’introduction des animaux et la proliﬁcité de la race. Bien que l’étude ne mentionne pas le immédiat extinction, la dernière race de porcs Moura doit être contrôlée en permanence avec les paramètres de population pour éviter l’extinction.

Mots-clés: endogamie, modèles de conservation de la population, porcs Moura

Resumen Se realizó el análisis de pedigrí y de viabilidad de la población de un rebaño de la raza Moura Embrapa Suínos y Aves. Se utilizó el Software ENDOG para el análisis de pedigree y VORTEX para el análisis de viabilidad de la población. Los datos se obtuvieron con el curador de lo rebaño y un total de 411 animales nacidos entre los años 1999 y 2011 se analizaron con ENDOG. Los datos de entrada en VORTEX se obtuvieron de un cuestionario contestado por el curador de lo rebaño. El coeﬁciente de endogamia (F) fue de 0.06%, el tamaño medio efectivo de la población (Ne) fue de 30 animales, con un intervalo medio de generación de 3,48 años. La probabilidad de extinción fue de 0% para un período de simulación de 100 años y 3% durante un período de 500 años. Los parámetros más importantes para la determinación de la viabilidad de este rebaño fuerón el ingreso de animales y proliﬁcidad de la raza. Aunque el estudio no indica peligro inmediato de extinción, el último rebaño de la raza Moura debe ser monitoreado constantemente con los parámetros de población para evitar la extinción.

Palabras clave: Endogamia, modelos de conservación de la población, cerdos Moura

Submitted 23 July 2013; accepted 16 October 2013

Introduction

Pigs were introduced in Brazil since the sixtieth century during the colonization period. The animals probably ori- Correspondence to: C. McManus, Faculdade de Agronomia e Medicina Veterinária, Campus Universitário Darcy Ribeiro, Brasília, DF 70919-900, Brazil. ginated from the Iberian Peninsula (Cavalcanti, 2000). email: [email protected] The variety of environment conditions found in farms in

127 128 H. Carneiro et al.

different regions of Brazil resulted in a wide variety of do Paraná in 1985, and later by the Company Café do breeds that for centuries sustained livestock production in Paraná. The first genealogical records of the breed were the country (Mariante et al., 2009). issued by the Brazilian Association of Swine Breeders (ABCS) in 1990, the year that marked the opening of Naturalized swine produce a higher fat to meat ratio than Brazilian Pig Book (PBB) of the Moura breed. Two commercial breeds, so there was little interest in using other breeders, also in Paraná, started the records of ani- them for improvement programmes. However, in rural mals since the year 1992. The records issued by ABCS areas these were the preferred animals, not only due to stopped in 1995. But returned in 2002, when Embrapa the large production of lard but also because of the taste Swine and Poultry created a pure herd using animals of their meat. They are adapted to different ecosystems from the Universidade Federal do Paraná, and this persists in different regions of Brazil and have a high capacity until today. One thousand, five hundred and twenty eight for survival in unfavourable conditions. The naturalized (1528) births were registered as pure animals in the pig breeds of Brazil are characterized by their toughness, ABCS from 2003 to 2011 (ABCS, 2011). Currently, resistance to diseases, low management requirements and there are records on animals in the municipalities of feeding as well as a high adaptability. These traits make Teutônia, Roca Sales, Cruzeiro do Sul, Venâncio Aires them an important genetic resource to be used in breeding and Carlos Barbosa. In the state of Santa Catarina, there programmes and research, thus justifying efforts for their are reports of the occurrence of the breed in the region conservation (Castro, Albuquerque and Germano, 2002). of Lages. In Paraná state, the records of occurrence are Eight naturalized Brazilian swine breeds have already in the municipalities of Bituruna and Roncador. In addition been described in the literature: Canastra, Canastrão, to these states, boars were sent to the Southern, UNESP Nilo, Piau, Caruncho, Pirapetinga, Moura and Tatu. (Ilha Solteira-SP), and Noreasthern, Universidade Federal Currently, there is no standardization in the phenotypic do Piauí (Teresina-PI), regions to compose pure herd. characterization of pig breeds as well as their denomina- The aim of this study was to monitor the conservation pro- tion leading to a possible loss of genetic diversity. On gramme of the herd of Moura pigs maintained by Embrapa the other hand, these breeds can also be considered as a Swine and Poultry, located in the state of Santa Catarina potential source of new allelic combinations of extreme (Brazil) through pedigree and population viability analysis. importance for the future of the Brazilian pig industry (Sollero, Paiva and Faria, 2009). However, the lack of data recording and specifying traits of zootechnical interest coupled with the large distances between farms and the Materials and methods small size of the breeding population makes it difficult to effectively work with naturalized swine herds (Garcia Experimental design and Barbosa, 2005). Crossbreeding and selection can rapidly modify characteristics of a breed, giving rise to Animals different genetic groups and complicating breed identifi- Pig data from the conservation herd of Moura pig breed cation. Although molecular techniques have aided in the maintained by Embrapa Swine and Poultry, located in identification and characterization of individuals and Concordia, Santa Catarina, Brazil were analysed. breeds, phenotypic characterization is also necessary. A The herd comprised a pure breeding herd of about 20 few studies have been carried out on the diversity of natur- females and ten males of different families. These families alized swine breeds in South American countries and sub- are crossed with each other to expand the genetic base for jective data are frequently missing for the most naturalized sire and dam replacement. When there is no need to breeds of farm animals (McManus et al., 2010). replace animals, females are crossed with boars of The Moura breed is a primary source of edible fat or lard. MS115 strain, a third-generation light pig developed by fi In addition, its relatively efficient conversion of food into Embrapa, and the piglets sold for nishing. weight gain makes it a useful cross with modern breeds. The data from all purebred pigs for pedigree and popu- These features make this breed a good alternative for sub- lation viability analyses were obtained directly from the sistence farming in the various regions of Brazil (Castro curators of the Moura herd. et al., 2001). The Moura is usually blue roan in colouration, occasionally red roan (Mason, 1996). The breed emerged in southern Brazil (Machado, 1967), although Analysis of pedigree the precise period remains unknown. This breed was The genetic population parameters were analysed using the widely used in the farms of the region before the introduc- ENDOG program, version 4.8 (Gutiérrez and Goyache, tion of the breeds such as Duroc, Landrace, Large White 2005). The information included pedigrees of animals and others (Fávero et al., 2007). born between 1999 and 2011, totalling 411 animals: 255 females and 156 males. The first initiatives aiming the formation of breeding herds for selection programmes of Moura pigs occurred in the Two analyses were performed, one with all the animals of state of Paraná, initially led by the Universidade Federal the pedigree and another considering only the animals that Moura Pig genetic resources’ conservation 129

Table 1. Allele frequencies in the conservation herd of Moura pig from Embrapa Swine and Poultry.

Locus Allele frequency

Number of alleles per locus 1 2 3 4 5 6 7 8 9 10

S0026 8 0.01 0.05 0.01 0.03 0.42 0.01 0.29 0.18 S0155 9 0.01 0.18 0.01 0.09 0.01 0.23 0.01 0.43 0.03 SW857 6 0.05 0.01 0.22 0.17 0.09 0.46 S0101 5 0.32 0.38 0.11 0.18 0.01 SW830 3 0.80 0.19 0.01 S0097 7 0.27 0.30 0.01 0.01 0.13 0.08 0.20 SW911 7 0.01 0.04 0.10 0.50 0.20 0.10 0.05 SW2406 5 0.73 0.01 0.02 0.03 0.21 S0068 10 0.10 0.17 0.22 0.01 0.28 0.01 0.01 0.10 0.02 0.08 OPN 6 0.34 0.06 0.01 0.51 0.05 0.03 SW936 9 0.07 0.27 0.03 0.01 00.03 0.01 0.31 0.17 0.10 SW1517 8 0.09 0.09 0.18 0.07 0.02 0.17 0.14 0.24 S0002 5 0.01 0.01 0.6 0.36 0.02 S0178 6 0.03 0.09 0.01 0.07 0.01 0.79 SW72 8 0.28 0.01 0.01 0.01 0.54 0.01 0.01 0.13 S0228 4 0.84 0.14 0.01 0.01 SW445 6 0.34 0.11 0.06 0.03 0.18 0.28

are currently part of the herd. For this purpose, animals Table 2. Demographic parameters input in VORTEX used to born after 2006 were used, totalling 364 records. simulate the population of Moura pigs.

Input data in VORTEX Value

Analysis of population viability Number of interactions 500 The known distribution of allele frequencies was obtained Number of years 100 by microsatellite marker analysed in Animal Genetics Reproduction system Polygamous Laboratory – LGA, Embrapa Cenargen, coming from the Age of first reproduction (F/M) 1/1 genotypic analysis of 35 animals (Table 1). Simulations Inbreeding depression? Yes Maximum number of farrowing per year 2 of interactions between demographic, environmental and Maximum size of the litter 11/farrowing genetic factors were made in the program VORTEX Mean sexual rate at birth 50% (males) Population Viability Analysis Software, version 9.99 Maximum age of reproduction 5 years (Miller and Lacy, 2005). Table 2 shows the demographic Adult females in reproduction (%) 80 parameters used to simulate the Moura pig herd, obtained Female probability having zero farrowing 2 Female probability having one farrowing 40 mainly from the answers of the curators of the conservation Female probability having two farrowing 58 herds linked to the National Programme for Conservation of Mortality of females aged between 0 and 1 (%) 8 Animal Genetic Resources. The base scenario was used in Mortality of adult females >1 (%) 5 the VORTEX model. VORTEX models population Mortality of males aged between 0 and 1 (%) 8 dynamics as discrete sequential events that occur according Mortality of adult males >1 5 Catastrophe 1: not determined by curator to the distribution of the random variables informed. A Frequency (%) 1 population was simulated as undergoing a series of events Effect on reproduction (%) 5% that describe the annual cycle of a sexually reproducing Effect of the survival (%) 10% organism: mate selection, reproduction, mortality, age Male reproduction 10 increase each year, migration, among others. At the end Initial population 12 Support capacity of the environment (K) 120 of the simulations, the programme provides: (1) the prob- Animals are removed? Yes ability of extinction of the population; (2) the mean time 80 males and 50 females per year to extinction if the population is extinguished in at least Animals entering the population? Yes 50 percent of the simulations; and (3) the average size and 20 females and three males per year genetic variation within the remaining population. Loss of genetic diversity is estimated by simulating the transmission of alleles from parents to offspring in a hypothetical neutral Results locus. The programme monitors how original alleles remain in the population, the average heterozygosity and the gen- The reference population was formed by parents of known etic diversity relative to the initial level. animals. The pedigree information with the identification To determine which factors mostly affect the population, of parents was 85 percent. The values found for the effec- 26 new scenarios were created by changing just one of tive number of founders (fe), effective number of ancestors the parameters of the base scenario (Table 3). (fa), average relatedness (AR), coefficient of inbreeding 130 H. Carneiro et al.

Table 3. Population viability analysis of the Moura herd, considering the base and additional scenarios.

r L females L males PE Final population He Ho N alleles

Base scenario 1.71 1.77 1.77 0 118.58 84 86 34.46 No entry 1.71 1.77 1.77 100 0.00 0 0 0.00 Two catastrophes 1.54 1.71 1.71 0 115.67 86 88 42.83 100% males reproduction 1.71 1.77 1.77 0 93 94 46.00 Lower proliﬁcacy 1.34 1.91 1.91 0 116.13 87 89 47.37 100% females reproduction 1.90 1.71 1.71 0 119.26 80 81 28.19 20% mortality of piglets 1.60 1.81 1.81 0 120.67 85 87 39.04 50% mortality of piglets 1.24 1.96 1.96 0 114.02 88 91 54.09 70% mortality of piglets 0.89 2.15 2.15 0 95.32 93 96 69.39 60% mortality of piglets females 1.08 2.04 2.04 0 104.94 95 98 72.00 More severe catastrophe 1.59 1.74 1.74 0 106.38 87 90 42.63 Fewer animals entering 1.71 1.77 1.77 0 96.91 73 76 14.16 Greater frequency catastrophe 1.64 1.75 1.75 0 118.34 84 86 36.70 70% mortality of piglets females 0.89 2.15 2.15 0 96.97 97 98 83.00 Greater adult mortality 1.65 1.43 1.43 0 118.70 84 86 36.02 No inbreeding depression 1.71 1.77 1.77 0 118.91 77 79 28.73 No genetic data 1.71 1.77 1.77 0 119.37 84 85 34.16 500 years 1.71 1.77 1.77 3 118.48 .01 .01 1.00 Initial population ENDOG 1.71 1.77 1.77 0 119.43 83 85 34.52 Breed lower 1.70 1.51 1.51 0 119.09 83 85 35.35 More animals coming out 1.71 1.77 1.77 0 115.31 86 88 41.26 Most one farrowing per year 1.36 1.91 1.91 0 117.33 87 89 47.49 Lower reproduction. more output and less input 1.34 1.58 1.58 100 0.00 0 0 0.00 More output and less entry of animals 1.35 1.91 1.91 100 0.00 0 0 0.00 Lower reproduction and more output of animals 1.34 1.58 1.58 2 2.91 81 86 24.67 Lower reproduction and less input of animals 1.34 1.58 1.58 100 0.00 0 0 0.00

r, rate of population growth; L, interval between generations; PE, probability of extinction; He, expected heterozygosity; Ho, observed heterozygosity; N alleles, The average number of alleles at the end of the simulation period.

(F) and mean index of genetic conservation (ICG) for the their genealogy. The average number of complete gener- reference population with pedigree are listed in Table 4. ations observed was 1.32, the average maximum number The coefﬁcient of inbreeding was 0.06 percent, which is of generations observed was 2.64 and the average equival- considered low for the herd. ent complete generations observed was 1.78.

The reference population (animals with both parents known) The interval between generations of the conservation herd was smaller than that used for analysing the founders. From from the Moura pigs and average ages at birth are shown in a total of 411 animals in the pedigree, 353 were included in Table 5. In the pedigree analysis, there were 27 records of the reference population. The base population, with at least animals born in 1999. The next birth was registered only in one known parent, had 58 animals. When considering 2004, indicating gaps in the registry of pedigree. only the current herd, the base population fell to 52 animals. The number of ancestors that contributed to the reference The generation interval in Moura pigs was obtained in the population was 31. Analysing only the current herd, the simulation with VORTEX. The interval between gener- reference population was 364 animals and 35 ancestors. ations refers to the average value at the end of the 100

There are a few generations available for evaluating the Moura herd. This is common in conservation programmes, Table 5. Interval between generations of the conservation herd since the interest in the conservation of these animals is from the Moura pigs (in years) and average ages at birth. recent, and insufﬁcient to obtain detailed information on Generation interval Full population Present population

Sire – offspring (male) 4.21 4.43 – Table 4. Population parameters from ENDOG pedigree analysis Sire offspring (female) 5.41 5.44 – in Moura pigs. Dam offspring (male) 3.32 2.89 Dam – offspring (female) 3.48 3.48 Parameter Full dataset Present herd Overall mean 4.31 4.21 Average ages at birth Effective number of founders (fe) 14 14 Boar – offspring (male) 2.94 3.74 Effective number of ancestral (fa) 14 14 Boar – offspring (female) 3.89 3.84 Average relatedness (AR) 8.38% 8.38% Dam – offspring (male) 2.45 2.40 Coefﬁcient of inbreeding (F) 0.06% 0.06% Dam – offspring (female) 2.39 2.39 Mean index of genetic conservation (ICG) 1.36 1.36 Overall mean 3.13 3.10 Moura Pig genetic resources’ conservation 131

years simulation and was close to the reality of the species. The Det.r assumes no inbreeding depression, no mating The generation interval in male was 1.77, in female 1.77 restrictions and no entry or exit to the animal population. and the average size of the final population was 118.58. Therefore, it simulates how the herd would react without any fluctuation due to chance or environment. The The fixation index (F-statistics) describes the level of Stoch.r considers the occurrence of environmental events heterozygosity in the population. The F statistics shows at random. Figures 2 and 3 show the comparison of the how inbreeding influences within the subpopulation evolution of inbreeding and the final average number of (F ), among subpopulations (F ) and in the whole popu- IS ST alleles, at different scenarios, respectively. Figure 4 lation (F ). The values found for the Moura herd were: IT shows the comparison of the growth rate, Stoch r. between F = −0.043073, F = 0, and F = −0.043073. The IS ST IT different scenarios of mortality in piglets and Figure 5 zero value indicates no population subdivision. shows the comparison of the evolution of genetic diversity The Ne averaged by year of birth of the animals was 30 and inbreeding in scenario 500 years. and decreased in the period analysed. The herd size did not increase, probably due to lack of market for the products, the absence of replacements and to the structural difficulties in maintaining a larger herd. This population Discussion size may cause problems in the viability of this herd. The parameters that support population viability are the Factors threatening the herd and possible management strat- high prolificacy of the breed and the entry of new animals. egies for conservation were tested in the new scenarios in The number of animals added to the herd was considered the PVA to determine the major threats. The average final high. Evaluating the number of births in each year added to herd size was very close to the environmental carrying the entry of animals and subtracting the animals from the capacity (K), indicating the growth capacity of the herd, population, intentionally or by mortality, it was noticed considering the rearing conditions. The probability of an important increase of animals every year in this herd. extinction was 0 based on 500 iterations performed in the Scenarios were created by changing only the factors: pro- VORTEX, indicating no extinction of the herd. The herd lificacy and fecundity; entry of animals and animals leav- increased rapidly to environmental carrying capacity (K) ing the population. In scenarios with “fewer offspring”, with an average deterministic growth rate (Det.r) of 1.71. “more animals produced” and “more than one farrowing Figure 1 shows the population growth comparison between per year”, the values of the maximum number of offspring, deterministic (Det.r) and stochastic (Stoch.r) simulation. number of animals leaving the herd and number of

Figure 1. Comparison between growth rates Det r. and Stoch r.

Figure 2. Comparison of the evolution of inbreeding in different scenarios. 132 H. Carneiro et al.

Figure 3. Comparison of the evolution of the ﬁnal average number of alleles at different scenarios.

Figure 4. Comparison of the growth rate, Stoch r. between different scenarios of mortality in piglets.

Figure 5. Comparison of the evolution of genetic diversity and inbreeding in scenario 500 years.

farrowings per year changed. The scenario that most per generation were sufficient to maintain the inbreeding impacted the herd was of the maximum number of farrow- within the acceptable level. ings per year. The other parameters were not affected in relation to the base scenario. Other scenarios were that The coefficient of inbreeding observed was 0.06 percent there was a decrease in the entry of animals caused the (Table 4). Estimating the inbreeding coefficient is complex extinction of the herd in all simulations. It was noted, because it usually requires knowledge of the pedigree of therefore, that maintaining the entry of animals in the individuals. In addition, the calculation of the coefficient herd is important for its persistence. This result corrobo- from the pedigree information requires the specification of rated with Fimland (2007), who also noted that the most the degree of relationship between the founders (Hedrick effective way to maintain genetic variability is the immi- and Kalinowski, 2000). According to Toro Rodriganez gration of individuals carrying alleles from different popu- and Rodriguez (2000), for the calculation of the inbreeding lations. Hamann and Distl (2008) found the 19 immigrants coefficient, it is preferable to use pedigree information Moura Pig genetic resources’ conservation 133

instead of using the molecular marker information. prolificacy. This may explain the low difference found Molecular marker can be useful to verify, correct or even between the base scenario and that without inbreeding complete pedigree data. The pedigree analysis always depression, as the mortality rate of piglets reported was begins with a founder population, in other words, individual low (8 percent). Figure 3 shows that there was no signifi- ancestry unknown and presumably unrelated, from which cant difference in the mean number of alleles between the the population is descended from pedigree (Lacy, 1989). base scenario without inbreeding depression, and that with When the founder population information is not available the analysis of true genetic information from population. (as here), it is assumed that the founders are not related The exception was for the scenario of “no entry” of ani- and are not endogamous (Hedrick and Kalinowski, 2000). mals, whose population rapidly went into extinction and, Sollero Paiva and Faria (2009) observed that the Moura therefore, the final average number of alleles was 0. and Monteiro had the lowest genetic diversity among native When analysed the impact of increased probability of repro- Brazilian pig breeds in terms of the effective number of duction in males and females, was observed that increasing alleles and observed heterozygosity. Pedigree analysis is the number of males in reproduction had more impact on the always recommended to allow planning to maximize the final average number of alleles than the increase in the num- effective size of the reproduction in relation to the total ber of females (data not shown). This can be explained by population size, and consequently, decrease the variability two factors. The information provided by the herd curator in the effects of genetic drift (Lacy, 1989). who reported that 80 percent of females and 10 percent A few generations were available for the evaluation of the males were available for reproduction. The second is the Moura herd, and time was insufficient to obtain detailed method of calculating these parameters by VORTEX. It information of their genealogy. The intervals between gen- was considered that every female of reproductive age has erations (Table 5) were considered high for the species com- 80 percent chance of reproducing each year. As for the pared with studies of Toro Rodriganez and Rodriguez males, this was only 10 percent. Therefore, any increase (2000), who found that a generation interval of 2.44 years in male number will have a greater impact than female. for Iberian pigs. This is probably because the animals Analysing the mortality rates of the base scenario it was from the Moura herd were not used for commercial pur- noted that a higher mortality rate of piglets affected the poses and the substitution in the herd only took place rate of population growth, the interval between the gener- when they became ill or died. The herd size has not ations and the final number of animals. These parameters increased during the time period studied, probably due to do not alter the probability of extinction, heterozygosity the absence of available replacements for the parents and or number of alleles. The mortality of piglets had a more fi structural dif culties in maintaining a larger herd. severe effect on the population than adult mortality, and There was a small difference between Det.r and Stoch.r, especially mortality of female piglets (Figure 4). which could indicate low environmental influences on However, mortality of animals up to 1 year informed by population (Figure 1). However, the answers given by the curator was low, indicating that the conservation herd the herd curators were considered very optimistic, which has good management of the offspring. could explain the positive results when the base scenario It was observed that increasing the number of catastrophes was analysed alone. Stoch.r will be smaller than Det.r if reduced the rate of growth and the final number of animals the population suffers wide fluctuations in environment, in the population, but did not affect the number of alleles catastrophes or genetic instabilities, as well as due to the and heterozygosity. Changing other scenarios such as the demographics of small populations (Miller and Lacy, frequency and severity of the catastrophe, had greater 2005), as seen here. effects on the population. The higher frequency of catastrophes had greater impact, with more effect on genetic To demonstrate the changing pattern of inbreeding coeffi- diversity. The highest frequency of one catastrophe cient of the herd, the scenario without animal entry was affected the population more than the lower frequency observed. The maximum value was attained in the first and severity of two catastrophes (data not shown). Thus, 10 years of the simulation (data not shown). The best per- it was inferred that minimizing the consequences of fre- formance of this parameter (Figure 2) was when 100 per- quent catastrophes in the herd conservation is more impor- cent of males were available for reproduction. No tant than specific effects of more severe catastrophes. significant difference was seen between the base scenario and that without the true population genetic data. The When the viability of the population was tested for a period inbreeding coefficient is calculated in each year of the of 500 years, it was noted that the probability of extinction VORTEX simulation and, if no true allele frequency is was small (3 percent), but the population was seriously provided, the programme determines a unique allele for compromised at the end of this period, when looking at each animal in the initial population. The data from the the final number of alleles and heterozygosity. The other pedigree analysis with ENDOG, even though not very parameters were not affected. Figure 5 shows the evolution deep, are more accurate. VORTEX considers inbreeding of genetic diversity and inbreeding coefficient in this depression solely as increased mortality of animals up to period, and it is seen that in the first 100 years these par- 1 year of age without considering decline in fertility and ameters remain relatively stable. From the year 100, these 134 H. Carneiro et al.

indices are seriously affected. This result demonstrates the Castro, A.M.M.G., Alvares, E.P., Zocoller-Seno, M.C., Neves, M.F. & ability of the pigs to adapt to the negative effects of Passipiéri, M. 2001. Morphological study of the small intestine of – small population size for a period, but after a certain Moura pigs (Sus scrofa Lineaus, 1758) during fetal development. Braz. J. Morphol. Sci. 8: 95–101. point, this control is no longer possible and the population is seriously compromised. According to Rangel (2012), a Castro, S.T.R., Albuquerque, M.S.M. & Germano, J.L. 2002. Census of Brazilian naturalizes swine breeds. Arch. Zootec. 51: 235–239. species does not go extinct immediately after the reduction of habitat or climate change beyond the tolerance limit of Cavalcanti, S.S. 2000. Suinocultura Dinâmica. Rome, Fep – Mvz Ed, the animals. Extinction can occur several generations Contagem. 494 pp. after the initial impact on the population. This initial period, Fávero, J.A., Figueiredo, E.P., Fedalto, L.M. & Woloszyn, N. 2007. A until effective extinction, creates a window of opportunity raça de suínos Moura como alternativa para a produção agroecológica de carne. Rev. Bras. Agroecol. 2: 1. for conservation, in which it is still possible to ensure the viability of the population (Wearn Reuman and Ewers Fimland, E. 2007. Genetic diversity and sustainable management of ani- – 2012). Simulations performed for a shorter period of time mal genetic resources, globally. Anim. Genetic Resources Inf. 41: 45 51. may not be sufficient to predict what conservation strategy is needed to ensure persistence of the population. A popu- Garcia, S.K. & Barbosa, A.S. 2005. Conservação e Estudo de raças – lation may seem feasible when the simulation considers suínas brasileiras na UFMG 12 anos. I Simpósio Brasileiro de Melhoramento Animal, 295. only 100 years, but the risk of extinction may increase considerably if a longer period is considered, as observed in Gutiérrez, J.P. & Goyache, F. 2005. A note on ENDOG: a computer program for analysing pedigree information. J. Anim. Breed Genet. Figure 5. Simulations should therefore be over a long 122(3): 172–176. period of time (over 100 years) or performed for two different periods and check if the results found in the short-term Hamann, H. & Distl, O. 2008. Genetic variability in Hanoverian Warmblood horses using pedigree analysis. J. Anim. Sci. 86: 1503– simulation remain in the long term (Nilson, 2003). 1513. Hedrick, P.W. & Kalinowski, S.T. 2000. Inbreeding depression in conservation biology. Annu. Rev. Ecol. Syst. 31: 139–162. Conclusion Lacy, R.C. 1989. Analysis of founder representation in pedigrees: founder equivalents and founder genome equivalents. Zoo Biol. 8: 111– The characterized Moura herd has low population size. Only 123. a few breeders or producers are interested in this breed, a Machado, L.C.P. 1967. Os suínos. Porto Alegre-RS, Editora A granja usual situation in production of Animal Genetic Resources Ltda. 622p. in Brazil. In the present study, this small effective population Mariante, A.S., Albuquerque, M.S.M., Egito, A.A., McManus, C., size may cause problems for the population in the near Lopes, M.A. & Paiva, S.R. 2009. Present status of the conservation future, despite the inbreeding coefficient being considered of livestock genetic resources in Brazil. Livestock Sci. 120: 204–212. low. The herd curators should continue to import animals Mason, I.L. 1996. A World dictionary of livestock breeds, types and var- from other breeders as this is essential for the viability of ieties, 4th edition. C.A.B International, Wallingford, UK 273. this conservation herd, as well as care should be taken to McManus, C., Paiva, S.R.P., Silva, A.V.R., Murata, L.S., Louvandini, maintain the reproductive rate of the animals. In conserva- H., Cubillos, G.P.B., Castro, G., Martinez, R.A., Dellacasa, M.S.L. tion programmes with small herds of domestic breeds, it is and Perez, J.E. 2010. Phenotypic characterization of naturalized important to evaluate the risk of extinction and which man- swine breeds in Brazil, Uruguay and Colombia. Braz. Arch. Biol. agement activities should be prioritized to reduce this risk. Technol. 53: 583–591. However, no specific risk factor is identified in this study. Miller, P.S. & Lacy, R.C. 2005. Vortex User’s Manual. A stochastic simulation of the extinction process. Version 9.50. Brookfield, IL: Chicago Zoological Society. Acknowledgments Nilson, T. 2003. Integrating effects of hunting policy, catastrophic events, and inbreeding depression, in PVA simulation: the Scandinavian wolf The authors acknowledge CNPq for scholarships and population as an example. Biol. Conserv. 115: 227–239. financial aid and INCT-Pecuária – Informação Genético Rangel, T.F. 2012. Amazonian extinction debts. Science 337: 162–164. Sanitária da Pecuária Brasileira (CNPq/MCT/FAPEMIG), as well as the farmers and researchers who contributed Sollero, B.P., Paiva, S.R. & Faria, D.A. 2009. Genetic diversity of Brazilian pig breeds evidenced by microsatellite markers. Livestock with data and access to the animals. Sci. 13: 1–8. Toro, M.A., Rodriganez, L.S. & Rodriguez, C. 2000. Genealogical References analysis of a closed herd of black hairless Iberian pigs. Conserv. Biol. 14: 1843–1851. ABCS – Associação Brasileira de Criadores de Suínos. 2011. Relatório Wearn, O.R., Reuman, D.C. & Ewers, R.M. 2012. Extinction debt and anual do registro genealógico e provas zootécnicas de suínos. windows of conservation opportunity in the Brazilian Amazon. Available at the website www.abcs.com.br Science 337: 228–232. Animal Genetic Resources, 2014, 54, 135–140. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000101

Evaluación del semen criopreservado de toros Curraleiro Pé Duro, perteneciente al Banco Brasilero de Germoplasma Animal

I.S. Montesinos1, J.O. Carvalho2, J.V. Malaquias3, E. Arnhold1, G.E. Freneau1, M.A.N. Dode2, M.C.S. Fioravanti1 y J.R.B. Sereno3 1Universidad Federal de Goiás, CEP 74001-970. Goiânia-GO, Brasil; 2Embrapa Cenargen, CEP 70770-900, Brasilia-DF, Brasil; 3Embrapa Cerrados, CEP 73310-970, Planaltina-DF, Brasil

Resumen Esta investigación tuvo el objetivo de evaluar la calidad del semen criopreservado de 21 toros Curraleiro Pé Duro en banco de germoplasma, categorizando las dosis de los reproductores según parámetros relacionados a la viabilidad espermática, y veriﬁcando la capacidad fecundante in vitro de los espermatozoides. Todas las muestras fueron descongeladas y analizadas en cuanto a morfología, cinética, integridad de membrana plasmática e integridad de acrosoma. Para evaluar las tasas de fecundación, muestras de sólo cuatro toros fueron utilizadas. Los resultados indicaron que las dosis de semen Curraleiro Pé Duro, almacenadas en el Banco Brasilero de Germoplasma Animal, mostraron calidad espermática variable y alta capacidad fecundante in vitro, posibilitando delinear estrategias adecuadas de reproducción asistida para la conservación de este recurso genético brasilero en peligro de extinción.

Palabras clave: conservación ex situ, espermatozoide, razas criollas

Summary This investigation aimed to evaluate the quality of cryopreserved semen from 21 Curraleiro Pé Duro bulls in a germplasm bank, categorizing the doses of the sires according to parameters related to sperm viability, and verifying the fertilizing capacity in vitro of the spermatozoids. All the samples were thawed and analyzed as to morphology, kinetics, plasma membrane integrity and acrosome integrity. Samples of only four bulls were used for evaluating the fertilization rates. The results indicated that the Curraleiro Pé Duro semen doses, stocked in the Brazilian Animal Germplasm Bank, showed variable sperm quality and high in vitro fertilizing capacity, making it possible to design appropriate strategies for assisted reproduction in order to conserve this Brazilian genetic resource, which is in danger of extinction.

Keywords: creole breeds, ex situ conservation, spermatozoid

Résumé Cette recherche visait à évaluer la qualité du sperme cryoconservés de 21 taureaux Curraleiro Pé Duro dans une banques de gènes, categorisant les doses des reproduteurs selon les paramètres liés à la viabilité du sperme, et de vériﬁer la capacité fecondante in vitro des spermatozoïdes. Tous les échantillons ont été décongelés et analysés pour la morphologie, la cinétique et la membrane plasmique et intégrité de l’acrosome. Pour évaluer le taux de fecondation, des échantillons de seulement quatre taureaux ont été utilisés. Les résultats ont indiqués que les doses de sperme Curraleiro Pé Duro, stokée dans la Banque Bresilienne de Gènes Animal, ont montré de haute qualité du sperme variable capacité de fécondation in vitro, permettant délimiter les stratégies appropriés de reproduction assistée pour la préservation de cette ressource génétique bresilienne en voie de disparition.

Mots-clés: créoles races, la conservation ex situ, spermatozoïde

Presentado: 25 Agosto 2013; aceptado: 27 Febrero 2014

Introducción selección natural y adaptación por más de 400 años. Este taurino tropical representa un tesoro genético, ya que a El bovino Curraleiro Pé Duro es una raza criolla brasilera, pesar de no tener niveles superiores de producción, es originaria del Bos taurus ibericus, traído por los coloniza- rústico, resistente al calor, enfermedades y parásitos, y dores europeos en el siglo XVI, y sometido a constante además aprovecha eﬁcientemente pastos de baja calidad. Debido a la importación de razas especializadas con

Correspondencia: I.S. Montesinos, Universidad Federal de Goiás, Brasil. e-mail: mayor desempeño zootécnico, el Curraleiro Pé Duro fue [email protected] descartado de los rebaños, lo que casi provocó su

135 136 I.S. Montesinos et al.

desaparecimiento (Mariante et al., 2009). Gracias a progra- Evaluación de morfología espermática mas de conservación aún existen ejemplares, en el Centro La morfología espermática fue evaluada de acuerdo con Oeste y Noreste brasilero, registrando 3.692 animales en el Barth y Oko (1989), utilizando microscopio de contraste año 2008, distribuidos entre los estados de Pará, Piauí, de fase (Nikon Eclipse E200) con inmersión y aumento Tocantins y Goiás (Fioravanti et al., 2011). La de 1000x. Se contaron 200 espermatozoides por muestra, conservación consta de dos estrategias: una es la in situ, clasificándolos en normales y con defectos (cabeza o donde los animales permanecen en ambientes naturales o cola), siendo expresado en porcentajes. artificiales, y otra la ex situ, criopreservando en bancos de germoplasma los espermatozoides, ovocitos, embriones y otras células, como fuente de multiplicación, Evaluación de cinética espermática regeneración y distribución de una población animal en La cinética espermática fue evaluada por el Computer extinción (Mariante y Egito, 2002). Según Dode y Assisted Sperm Analysis (CASA), modelo IVOS Rumpf (2002) técnicas de reproducción asistida son utili- Ultimate 12 de la Hamilton Thorne Biosciences, ajustado zadas en conservación animal, destacando la (set up) para semen bovino. Se colocó 10 μL de semen inseminación artificial (IA), transferencia de embriones en la cámara de Makler, con selección de tres campos de (TE), producción in vitro de embriones (PIV) e lectura, evaluando: motilidad total (MT, %) y progresiva inyección intra-citoplasmática de espermatozoides (ICSI). (MP, %); velocidad de trayecto (VAP, μm/s), rectilínea Desde 1983, Embrapa Recursos Genéticos y Biotecnología (VSL, μm/s) y curvilínea (VCL, μm/s); amplitud lateral (CENARGEN) realiza la conservación ex situ de razas de cabeza (ALH, μm); frecuencia de batimientos de cola criollas brasileras, criopreservando diferentes células en (BCF, Hz); rectilinearidad (STR, %); linearidad (LIN, %) el Banco Brasileiro de Germoplasma Animal (BBGA) y células rápidas (RPD, %). (Mariante y Egito, 2002). Para asegurar el futuro del bovino Curraleiro Pé Duro (Figura 1) se necesitan más estudios enfocando parámetros reproductivos, por lo que Evaluación de integridad de la membrana objetivamos evaluar la calidad del semen criopreservado plasmática y acrosoma de 21 toros de esta raza, perteneciente al BBGA. La integridad de membrana plasmática fue evaluada utilizando diacetato de 6 carboxifluoresceína (C-FDA) y yoduro de propidio (PI) (Harrison y Vickers, 1990). Una Material y métodos muestra de semen (10 μL) fue diluida en solución color- ante (40 μL), compuesta por 480 μL de citrato de sodio Esta investigación fue realizada en el año 2011, en las a 3%; 5 μL de formol salina a 1,6%; 5 μL de IP 0,75 instalaciones del laboratorio de reproducción animal mM; y 10 μL de C-FDA 0,46 mg/mL en DMSO. Los CENARGEN (DF). Se utilizaron pajillas de semen de 21 espermatozoides teñidos de rojo se consideraron con mem- toros Curraleiro Pé Duro, donadas por el BBGA en brana plasmática dañada y los de verde con membrana número de dos unidades por reproductor, excepto cinco íntegra. Para evaluar la integridad acrosomal se utilizó la toros, con diez dosis cada uno. Este material genético se conjugación isotiocianato de fluoresceína (FITC) con lec- criopreservó entre 1994 y 2008, mediante protocolos con- tina Peanut aglutinin (PNA) y PI (Klinc y Rath, 2007). vencionales desarrollados por el CENARGEN. Las dosis Una muestra de semen (10 μL) fue diluida en solución col- fueron descongeladas por 30 segundos en baño María orante (30 mL), con los mismos componentes usados en la (37°C), y el semen depositado en un tubo para evaluación de membrana plasmática, excepto C-FDA que evaluación in vitro. fue substituido por 10 μL de FITC-PNA (1 mg/mL en PBS). Las células teñidas de rojo se contabilizaron como muertas. En los espermatozoides vivos el acrosoma teñido de verde fue considerado como reaccionado, y el acrosoma sin fluorescencia, íntegro. Ambas evaluaciones se realizaron en microscopio de epifluorescencia (Axiophot Zeiss: filtro de longitud de onda 395/420 nm y 494/517 nm excitación/emisión). Aproximadamente 200 células fueron examinadas en cada muestra y los resultados se expresaron en porcentajes.

Evaluación de tasas de fecundación Para evaluar las tasas de fecundación fueron utilizadas dosis de sólo cuatro toros Curraleiro Pé Duro, debido a la escasez de pajillas disponibles. Se utilizó la metodología propuesta por Carvalho et al.(2010), con Figura 1. Toro Curraleiro Pé Duro en el bioma Cerrado. tres replicas independientes, donde a 18 horas post Evaluación del semen criopreservado de toros Curraleiro Pé Duro 137

inseminación in vitro los posibles cigotos fueron retirados nutricionales, interferencia climática y escasa selección de las gotas de fecundación y transferidos para una gota de de este grupo genético nativo. Además las altas tasas de 200 μL de tampón fosfato (PBS). Estos fueron desnudados gota proximal y cola fuertemente doblada con gota y fijados en solución de etanol y ácido acético glacial (3:1), indicarían problemas de madurez sexual, epidídimo y fer- por 24 horas. Después se lavaron y tiñieron con solución tilidad (Rao, Bane y Gustafsson, 1980). Saacke (2008) car- lacmoide a 1% en ácido acético glacial 45%. La lectura acteriza cabeza piriforme, cráter y pouch formation como de las láminas fue en microscopio de contraste de fase defectos no compensables, ya que suelen poseer aberra- bajo inmersión, con aumento de 1000x (Nikon Eclipse ciones en la cromatina que afectan la fecundación y/o E200). Se utilizó 336 ovocitos, siendo clasificados en no desarrollo embrionario, causando sub-fertilidad fecundados (presencia de cromatina femenina y ausencia espermática. de masculina) y fecundados (presencia de cromatina femenina y espermatozoide en el citoplasma, presencia de Cinética espermática cabeza descondensada, pronúcleo, clivaje y polispermia). Según Kathiravan et al.(2011), altos porcentajes de MP y RPD revelan resistencia a criopreservación. Además la MP Análisis estadístico está correlacionada con MT, y contribuye 62,6% en varia- Los datos obtenidos en los exámenes espermáticos de las ciones de las tasas de fecundación, teniendo como prome- 21 dosis fueron analizados a través del programa R (R dio grupal 27,3% ± 9,7% y RPD 35,4% ± 11,4%. En el Development Core Team, 2011). Primero se realizó promedio individual destacaron las dosis DT5 (MP 45%, estadística descriptiva (promedio y desvío estándar). RPD 55%), DT9 (MP 40%, RPD 48%), DT16 (MP Después se escogieron parámetros para la construcción 39%, RPD 47%) y DT19 (MP 31%, RPD 51%); contrario de un dendrograma, mediante el método UPGMA a DT10 (MP 15%, RPD 17%), DT18 (MP 9%, RPD 11%) (Unweighted Pair Group Method with Arithmetic y DT20 (MP 13%, RPD 17%) que fueron inferiores. Las Means), utilizando la distancia Euclidiana padronizada. velocidades VAP, VSL y VCL (μm/s) fueron altas en El Cluster formó tres grupos, los cuales fueron comparados DT5 (105,6; 88,4; 179,2), DT11 (88,7; 72,1; 156,9), en cada examen espermático a través de análisis de var- DT19 (87,2; 60,9; 168,4) y DT21 (91,3; 75,0; 156,5). ianza y test Duncan, a nivel de significancia del 5%. Las Para ALH (μm) obtuvieron tasas superiores DT8 (7,3), tasas de fecundación y estados de cromatina 18 horas DT14 (7,6) y DT19 (7,7), lo que concuerda con sus post FIV se compararon por el test Chi-cuadrado. bajos valores en BCF (DT8, 20,4 Hz; DT14, 22,4 Hz; DT19, 23,6 Hz), STR y LIN (DT8, 67,0%, 39,0%; DT14, 75,0%, 41,0%; DT19, 70,0%, 38,0%). La DT19 Resultados y discusión mostró hiperactividad espermática (HE), por sus altos valores de velocidad; ALH superior en el grupo, así Evaluación de la calidad espermática como BCF, STR y LIN inferiores. La HE conduce a movi- Morfología espermática mientos vigorosos y desordenados, provocando trayectoria Considerando 70% como mínimo de espermatozoides nor- irregulary y curva, situación que sólo ocurre in vivo en el males para semen descongelado bovino (CBRA, 1998), las local de fecundación (Mortimer, 1997; Verstegen, 21 dosis Curraleiro Pé Duro presentaron buen promedio Iguer-Ouada y Onclin, 2002); mas in vitro podría favorecer (81,1% ± 13,3%). Individualmente tres dosis fueron infer- en la penetración del ovocito. La criopreservación origina mudanzas en membrana plasmática, similares a las de iores: DT1 (60,2%); DT10 (37,5%) y DT17 (68,2%). “ ” Los defectos mayores (DMA) en una muestra seminal no capacitación. Esa criocapacitación post descongela- deben superar el 20% (CBRA, 1998), a pesar de haber miento talvez sería responsable por la HE detectada encontrado cinco dosis con alto porcentaje (DT1, 34,7%; (Januskauskas et al., 1999). DT7, 24,3%; DT10, 39,5%; DT16, 22,5% y DT17, 28,1%), el promedio general de DMA era 14,7% ± 9,9% Integridad de membrana plasmática y acrosoma y los defectos menores 4,2% ± 7,2%. Este promedio fue Se consideraron la membrana plasmática íntegra (MPI) y mejor al encontrado por Sousa et al.(2004) en ocho los espermatozoides vivos con acrosoma íntegro (EVAI), toros Curraleiro Pé Duro (21,1% ± 8,4%). Ciertos DMA porque reflejan buena conformación estructural y funcional superaron el límite de 5% que relata Freneau (2011) para de la célula (Garner et al., 1986; Graham, 2001). En el pro- anormalidades individuales: cráter (DT1, 21,8%; DT16, medio general MPI obtuvo 36,7% ± 14,2% y EVAI 34,0% 9,0%); contorno anormal (DT1, 5,6%); gota proximal ± 14,4%. Los valores más bajos fueron: MPI (DT10, (DT10, 19,0%); cola fuertemente doblada (DT11, 7,2%); 15,0%; DT14, 21,3% y DT18, 10,5%) y EVAI (DT2, cola fuertemente doblada con gota (DT7, 7,1%; DT17, 18,0%; DT10, 12,5%; DT14, 21,5%; DT18, 9,5% y 15,1%; DT10, 6,5%) y cabeza piriforme (DT20, 6,5%). DT20, 22,0%). Para Januskauskas, Johannisson y Britto y Mello (1988) obtuvieron individualmente DMA Rodriguez-Martinez (2003), la integridad de membrana exacerbados, en el semen de cinco toros Curraleiro Pé plasmática es la variable con mayor impacto en la fertili- Duro, destacando pouch formation (17,1%). Las altera- dad, caso contrario a Carvalho et al.(2010) que no identifi- ciones encontradas pueden tener origen en deficiencias caron alta correlación con la fecundación. Cabe resaltar 138 I.S. Montesinos et al.

comparación al G2. El G2 obtuvo la mayor tasa de defectos morfológicos; no en tanto cinética, viabilidad y células vivas con acrosoma íntegro aceptables, características que son propias de espermatozoides con defectos no compensables (Saacke, 2008). G3 obtuvo altos porcentajes de morfología espermática como G1, pero la peor cinética, integridad de membrana plasmática y acrosoma, indicando mayor presencia de espermatozoides muertos, ya que según Garner et al.(1986) células con membrana plasmática dañada no poseen movimiento.

Evaluación de las tasas de fecundación Se utilizaron dosis de sólo cuatro toros Curraleiro Pé Duro (DT18, DT19, DT20 y DT21), debido al amplio estoque de pajillas donadas por el BBGA para esos reproductores. El semen presentó valores NRM y DMA, dentro de los padrones CBRA (1998). Sólo la DT18 y DT20 fueron Figura 2. Dendrograma de las 21 dosis de semen. inferiores en cinética y exámenes de fluorescencia. Previo al uso de las dosis para FIV, estas se seleccionaron que la reacción acrosómica prematura disminuye el poten- mediante gradiente Percoll (45:90%), eliminando los cial reproductivo del espermatozoide (Silva y Gadella, espermatozoides con defectos. Después fue evaluada la 2006). Casi todas las dosis obtuvieron un porcentaje bajo MP en el microscopio de contraste de fase, tanto pre (≤ 2,0%), para espermatozoides vivos con acrosoma reac- como post pasaje por el gradiente, siendo que cionado, excepto DT17 con 14,5%, que puede estar rela- incrementó considerablemente. Rodriguez-Martinez cionado a su alta tasa de DMA (28,1%). (2005) relata que esa prueba mide la capacidad de migración espermática, seleccionando las células con Parámetros escogidos para construcción de un mayor potencial para ultrapasar in vitro la barrera del dendrograma Percoll, que simularía el trayecto in vivo hasta el local de Fueron analizados parámetros que según Amann (1989)son fecundación. Una vez tratadas las dosis seminales para parte de los atributos espermáticos para fecundación (NRM, FIV, se seleccionó e inseminó in vitro 336 ovocitos, DMA, MP, RPD, MPI y EVAI). Se construyó un dendro- evaluando post 18 horas las tasas de fecundación grama (Figura 2) utilizando la distancia Euclidiana, para for- (Cuadro 2). mar tres grupos (G1, G2 y G3). La DT19 y DT10 resultaron Las tasas de fecundación fueron altas y próximas, sin difer- excluidas, por ser consideradas outliers, al presentar valores encia estadística significativa, revelando resultados no muy distantes en relación a las demás dosis (Sartorio, 2008). esperados antes del Percoll (45:90%). Como fue el caso Se sabe que la DT19 mostró HE y DT10 obtuvo los valores de DT18 y DT20, que tenían bajos porcentajes para MP más bajos en los exámenes in vitro. (33%; 30%), en el microscopio de contraste de fase, y exámenes de fluorescencia (MPI 10,5%; 28,3% y EVAI Caracterización de los grupos y comparación 9,5%; 22%), pero post Percoll mostraron buena MP estadística (70%; 53%) y fecundación ovocitaria (74,4%; 68,8%). En los tres grupos formados (Cuadro 1), el G1 alcanzó Según Tangue et al.(2002), las tasas de fecundación obte- altos valores en morfología; la mejor cinética; así como nidas son aceptables (> 60%), revelando la capacidad viabilidad y tasas de espermatozoides vivos con acrosoma espermática in vitro del semen Curraleiro Pé Duro, para íntegro razonables, mostrando cierta superioridad en ligar y penetrar la ZP del ovocito, así como promover la

Cuadro 1. Parámetros espermáticos analizados en los tres grupos formados por el dendrograma.

Parámetros espermáticos (%) Grupo 1 (G1) Grupo 2 (G2) Grupo 3 (G3)

Promedio DS Promedio DS Promedio DS

Espermatozoides normales (NRM) 84,7a 4,5 66,3b 5,5 88,9a 6,8 Defectos mayores (DMA) 12,0b 5,3 29,0a 5,2 8,5b 4,4 Motilidad progresiva (MP) 34,2a 5,3 25,0b 4,5 15,2c 5,2 Células rápidas (RPD) 41,5a 6,4 34,0a 7,2 23,2b 9,1 Membrana plasmática íntegra (MPI) 39,1a 8,1 41,6a 9,5 24,9b 9,5 Espermatozoide vivo acrosoma íntegro (EVAI) 36,9a 9,0 35,5a 4,9 22,4b 8,5

DS: desvío estándar. Promedios con letra diferente (a,b,c) no son iguales, por el Test Duncan al 5%. Evaluación del semen criopreservado de toros Curraleiro Pé Duro 139

Cuadro 2. Tasas de fecundación ovocitaria 18 horas post industrial con razas comerciales, para otorgar mayor rusti- inseminación in vitro. cidad y adaptación al trópico brasilero. Dosis Toro NO Fecundado No P valor X2 fecundado N%N% Referencias DT18 90 67 74,4 23 25,6 DT19 87 67 77,0 20 22,9 0,4 Amann, R.P. 1989. Can the fertility potential of a seminal sample be pre- DT20 77 53 68,8 24 31,2 dicted accurately? Journal of Andrology 10 (2): 89–98. DT21 82 66 80,5 16 19,5 Barbas, J.P. & Mascarenhas, R.D. 2009. Cryopreservation of domestic animal sperm cells. Cell and Tissue Banking 10: 49–62. NO: número de ovocitos; N: número; X2: Chi-cuadrado. Barth, A.D. & Oko, R.J. 1989. Abnormal morphology of bovine spermatozoa. Iowa State University Press, Ames (IA), USA. fertilización. Este resultado fue algo similar al obtenido por ≥ Britto, C.M.C. & Mello, M.L.S. 1988. Anomalias de cabeça em Obando, Luque y Martínez (2003), en las tasas de FIV ( espermatozóides de gado “Pé Duro”. Revista Brasileira de Genética 61,9%) de cinco toros de la raza criolla San Martinero en el 11 (1): 63–71. trópico colombiano. Carvalho, J.O., Sartori, R., Machado, G.M., Mourão, G.B. & Dode, Se constató en los cigotos, que los estados de cromatina mas- M.A.N. 2010. Quality assessment of bovine cryopreserved sperm culina 18 horas post inseminación in vitro presentaron valores after sexing by flow cytometry and their use in in vitro embryo pro- – estadísticamente diferentes. Con tasas superiores para duction. Theriogenology 74: 1521 1530. pronúcleo (> 60%), siendo que la DT19 obtuvo el mayor por- CBRA. 1998. Manual para exame andrológico e avaliação de sêmen ani- centual, quizás debido a su HE que contribuyó en la FIV, en mal. 2da edición, Colégio Brasileiro de Reprodução Animal, Belo Horizonte, Brasil. cuanto la DT20 el menor, y además un valor elevado para cabeza descondensada. Según Eid, Lorton y Parrish (1994), Dode, M.A.N. & Rumpf, R. 2002. Produção in vitro de embriões na – esto revelaría en DT20 una fecundación tardía, posiblemente espécie bovina. Biotecnologia Ciência & Desenvolvimento 26: 32 37. debido a sus porcentajes de cabeza piriforme (6,5%), defecto Eid, L.N., Lorton, S.P. & Parrish, J.J. 1994. Paternal influence on fi espermático no compensable, que podría haber ultrapasado el S-Phase in the rst cell cycle of the bovine embryo. Biology of Reproduction 51: 1232–1237. gradiente de Percoll, y afectaría una futura embriogénesis si fuese desarrollada. Finalmente, los resultados obtenidos Fioravanti, M.C.S., Juliano, R.S., Costa, G.L., Abud, L.J., Cardoso, sugieren que el G1 sea utilizado en la IA a campo, por V.S., Carpio, M.G. & Costa, M.F.O. 2011. Conservación del bovino Curraleiro: cuantificación del censo y caracterización de los criadores. tener valores favorables para fecundación in vivo.Diferente Animal Genetic Resources 48: 109–116. de la DT19, DT10, G2 y G3, que deben utilizarse con prior- Freneau, G.E. 2011. Aspectos da morfologia espermática em touros. idad, para evitar mayor deterioro de estas dosis y en estricta Revista Brasileira de Reprodução Animal 35 (2): 160–170. asociación a biotecnologías reproductivas, como el pasaje por gradiente Percoll y PIV de embriones, que optimizarán Garner, D.L., Pinkel, D., Johnson, L.A. & Pace, M.M. 1986. Assessment of spermatozoal function using dual fluorescent los requisitos espermáticos para fecundación, como ya fue staining and flow cytometric analyses. Biology of Reproduction 34: verificado, sabiendo que sólo espermatozoides competentes 127–138. tienen contacto con el ovocito (Barbas y Mascarenhas, Graham, J.K. 2001. Assessment of sperm quality. En: Proceedings of 2009). De esa forma se evitará la pérdida de la valiosa varia- the 47th Annual Convention of the AAEP, 24–28 November 2001, bilidad genética de estas dosis de semen del BBGA, garanti- San Diego (CA), USA. zando la máxima producción de becerros y embriones Harrison, R.A.P. & Vickers, S.E. 1990. Use of fluorescent probes to Curraleiro Pé Duro. assess membrane integrity in mammalian spermatozoa. Journal of Reproduction and Fertility 88: 343–352. Januskauskas, A., Gil, J., Soderquist, L., Haard, M.G.M., Haard, M. Conclusiones C., Johannisson, A. & Rodriguez-Martinez, H. 1999. Effect of cooling rates on post-thaw sperm motility, membrane integrity, cap- La calidad seminal encontrada en las dosis del BBGA acitation status and fertility of dairy bull semen used for artificial – presentó variabilidad, mostrando características favorables insemination in Sweden. Theriogenology 52: 641 658. y desfavorables para fecundación. Esto indica la importan- Januskauskas, A., Johannisson, A. & Rodriguez-Martinez, H. 2003. cia de la evaluación espermática para determinar el uso Subtle membrane changes in cryopreserved bull semen in relation with sperm viability, chromatin structure, and field fertility. adecuado de las dosis seminales, tanto in vivo como in Theriogenology 60: 743–758. vitro, en asociación a biotecnologías reproductivas que busquen resultados positivos. Permite así la conservación Kathiravan, P., Kalatharan, J., Karthikeya, G., Rengarajan, K. & Kadirvel, G. 2011. Objective sperm motion analysis to asses dairy del bovino Curraleiro Pé Duro, a través de programas de bull fertility using computer-aided system – a review. Reproduction recuperación del efectivo poblacional y criopreservación in Domestic Animals 46: 165–172. en bancos de germoplasma, tal como viene siendo ejecu- Klinc, P. & Rath, D. 2007. Reduction of oxidative stress in bovine tado por instituciones de investigación y criadores de spermatozoa during flow cytometric sorting. Reproduction in esta raza. También surge su posible uso en el cruzamiento Domestic Animals 42: 63–67. 140 I.S. Montesinos et al.

Mariante, A.S. & Egito, A.A. 2002. Animal genetic resources in Brazil: Saacke, R.G. 2008. Sperm morphology: Its relevance to compensable Result of ﬁve centuries of natural selection. Theriogenology 57: 223–235. and uncompensable traits in semen. Theriogenology 70: 473–478. Mariante, A.S., Albuquerque, M.S.M., Egito, A.A., McManus, C., Sartorio, S.D. 2008. Aplicações de técnicas de análise multivariada em Lopes, M.A. & Paiva, S.R. 2009. Present status of the conservation experimentos agropecuários usando o software R. Dissertação em of livestock genetic resources in Brazil. Livestock Science 120: 204–212. Estatística e Experimentação agronômica, Escola Superior de Mortimer, S.T. 1997. A critical review of the physiological importance Agricultura Luiz de Queiroz, Universidade de São Paulo, and analysis of sperm movement in mammals. Human Reproduction Piracicaba, Brasil. Update 3 (5): 403–439. Silva, P.F.N. & Gadella, B.M. 2006. Detection of damage in mammalian – Obando, H., Luque, O. & Martínez, R. 2003. Evaluación de la capaci- sperm cells. Theriogenology 65: 958 978. dad fertilizante in vitro del semen de toros San Martinero y Brahman. Sousa, A.P.F., Freneau, G.E., Fioravanti, M.C. & Juliano, R.S. 2004. – ICA Informa 30 (2): 56 62. Estudo da espermatogênese pelo sêmen e a morfologia testicular do R Development Core Team. 2011. R: A language and environment for Gado Pé Duro. Anais do 1° Congresso de Pesquisa, Ensino e statistical computing. R Foundation for Statistical Computing, Vienna, Extensão da Universidade Federal de Goiás, 18–24 Outubro 2004, Austria. Disponible en http://www.r-project.org/. Goiânia (GO), Brasil. Rao, A.R., Bane, A. & Gustafsson, B.K. 1980. Changes in the morph- Tangue, S., Van Soom, A., Sterckx, V., Maes, D. & De Kruif, A. ology of spermatozoa during their passage through the genital tract in 2002. Assessment of different sperm quality parameters to predict dairy bulls with normal and impaired spermatogenesis. in vitro fertility of bulls. Reproduction in Domestic Animals 37: Theriogenology 14: 1–12. 127–132. Rodriguez-Martinez, H. 2005. Methods for semen evaluation and their Verstegen, J., Iguer-Ouada, M. & Onclin, K. 2002. Computer assisted relationship to fertility. En: Anais do 16° Congresso Brasileiro de semen analyzers in andrology research and veterinary practice. Reprodução Animal, 1–5 Agosto 2005, Goiânia (GO), Brasil. Theriogenology 57: 149–179. Animal Genetic Resources, 2014, 54, 141–152. © Food and Agriculture Organization of the United Nations, 2013 doi:10.1017/S2078633613000349

Unique cultural values of Madura cattle: is cross-breeding a threat?

T.S.M. Widi1,2, H.M.J. Udo1, K. Oldenbroek2, I.G.S. Budisatria3, E. Baliarti3 and A.J. van der Zijpp1 1Animal Production Systems Group, Department of Animal Sciences, Wageningen University, PO Box 338, 6700 AH Wageningen, The Netherlands; 2Centre for Genetic Resources, PO Box 16, 6700 AA Wageningen, The Netherlands; 3Department of Animal Production, Universitas Gadjah Mada, Jl. Fauna no 3, Kampus Bulaksumur UGM, Yogyakarta, Indonesia

Summary In Indonesia, cross-breeding local cattle with European beef breeds is widely promoted to stimulate beef production. This cross-breeding is threatening local breeds that have often different functions, including cultural roles. This study analysed the cultural values of Madura cattle and the effects of cross-breeding on local traditions in Madura Island. Bull racing (karapan) and cow conformation contests (sonok) are traditional cultural events on Madura. Since 2001, cross-breeding with Limousin is allowed. The local government also promotes a conformation contest for cross-bred (madrasin) cattle. Quantitative and qualitative information were collected through participatory approaches involving farmers (n = 97), government officials, community groups, key informants and through direct observation of sonok, karapan and madrasin events. Phenotypic characteristics were collected from 184 cows. The Madura cattle population and production systems are not homogeneous. Four cattle types could be distinguished: karapan, sonok, madrasin and conventional Madura cattle. These cattle were found in three discrete areas, differing in land sizes, cropping and cattle keeping in terms of management practices and importance of specific cultural practices. Sonok and madrasin cows were significantly bigger and had higher body condition scores than karapan cows and the conventional Madura cows in the madrasin area. Madura cattle participating in cultural events were valued at prices that were 2–3.5 times higher than Madura cattle not participating in cultural events. Cross-breeding will not directly influence the cultural events or the management practices of Madura cattle in the karapan and sonok areas; however, outside the karapan and sonok areas, crossbreds are rapidly replacing conventional Madura cattle. The present top-down approach towards conservation and breeding strategies has to be turned into bottom-up approaches that consider the needs of the sonok and karapan Madura cattle sub-populations. Monitoring and characterization studies have to collect information at different aggregation levels and have to be aware of the sub-populations. The Madura example shows that the cultural values of livestock can be a main driver for maintaining relatively small populations of local breeds.

Keywords: Madura, bull racing, cattle contests, conservation, cultural values

Résumé En Indonésie, le croisement des bovins locaux avec des races bovines à viande européennes est largement promu afind’augmenter la production de viande. Cette pratique menace les races locales qui ont souvent différentes fonctions, y compris des rôles culturels. Cette étude a analysé les valeurs culturelles des bovins Madura et l’impact des croisements sur les traditions locales sur l’île de Madura. Les courses de taureaux (karapan) et les concours de conformation des vaches (sonok) sont des évènements culturels traditionnels en Madura. Depuis 2001, le croisement avec la race Limousine est permis. Le gouvernement local promeut également un concours de conformation pour les bovins croisés (madrasin). Des informations quantitatives et qualitatives ont été recueillies suivant une approche par- ticipative impliquant des éleveurs (n = 97), des fonctionnaires des gouvernements, des groupements communautaires et des informateurs clés, et à travers l’observation directe des évènements sonok, karapan et madrasin. Les caractéristiques phénotypiques ont été mesurées sur un total de 184 vaches. La population bovine Madura et les systèmes de production ne sont pas homogènes. Quatre types de bovins ont pu être distingués: karapan, sonok, madrasin et les bovins Madura conventionnels. Ces bovins se trouvent dans trois zones distinctes, qui diffèrent par la taille des terres, les cultures et la production animale en ce qui concerne les pratiques d’élevage et l’importance de pratiques culturelles spécifiques. Les vaches sonok et madrasin ont été significativement plus grandes et ont obtenu des notations de conformation corporelle plus élevées que les vaches karapan et les vaches Madura conventionnelles dans la zone madrasin. Les bovins Madura participant aux évènements culturels atteignent des prix qui sont de 2 à 3,5 fois plus élevés que ceux des bovins Madura qui ne participent pas aux manifestations culturelles. Les croisements n’influenceront pas directement les évènements culturels ou les pratiques d’ élevage des bovins Madura dans les zones karapan et sonok. Cependant, en dehors des zones karapan et sonok, les animaux croisés sont rapidement remplacés para des bovins Madura conventionnels. L’actuelle approche descendante dans les stratégies de conservation et de sélection doit être tournée vers une approche ascendante qui tienne compte des besoins des sous-populations sonok et karapan de bovins Madura. Les études de surveillance et de caractérisation doivent recueillir l’information à différents niveaux d’agrégation et doivent prendre en considération les sous-populations. L’exemple Madura montre que les valeurs culturelles du bétail peuvent être un levier important pour le maintien de populations relativement petites de races locales.

Mots-clés: Madura, courses de taureaux, concours bovins, conservation, valeurs culturelles

Correspondence to: T.S.M. Widi, Animal Production Systems Group, Wageningen University, PO Box 338, 6700 AH Wageningen, The Netherlands. email: [email protected]

141 142 T.S.M. Widi et al.

Resumen En Indonesia, el cruzamiento del ganado bovino local con razas europeas de aptitud cárnica es fomentado ampliamente con el finde incrementar la producción de carne de bovino. Esta práctica supone una amenaza para las razas locales, que tienen a menudo diversas funciones, incluidos roles culturales. Este estudio analizó los valores culturales del ganado bovino Madura y el impacto del cruzamiento sobre las tradiciones locales en la isla de Madura. Las carreras de toros (karapan) y los concursos de conformación de las vacas (sonok) son acontecimientos culturales tradicionales en Madura. Desde 2001, se permite el cruzamiento con ganado Limousin. El gobierno local promueve también un concurso de conformación para ganado bovino cruzado (madrasin). Se recogió información cuantitativa y cualitativa siguiendo un enfoque participativo que implicaba a ganaderos (n = 97), funcionarios públicos, grupos comunitarios y personas de alto rango, y mediante la observación directa de los actos sonok, karapan y madrasin. Las características fenotípicas fueron evaluadas en 184 vacas. La población de ganado bovino Madura y los sistemas de producción no son homogéneos. Ha sido posible diferenciar cuatro tipos de ganado bovino: karapan, sonok, madrasin y ganado bovino Madura convencional. Estos tipos de ganado se encuentran en tres zonas distintas, que difieren en el tamaño de los terrenos, en los cultivos y en la actividad ganadera en relación con el manejo de los animales y la importancia de prácticas culturales específicas. Las vacas sonok y madrasin son significativamente más grandes y alcanzan puntuaciones de condición corporal mayores que las vacas karapan y las vacas Madura convencionales en la zona madrasin. El ganado Madura que participa en actos culturales alcanza precios entre 2 y 3,5 veces mayores que el ganado Madura que no interviene en estos acontecimientos. El cruzamiento no afectará directamente a los actos culturales o a las prácticas de manejo del ganado bovino Madura en las zonas karapan y sonok. Sin embargo, fuera de las zonas karapan y sonok, los animales cruzados están siendo rápidamente sustituidos por ganado bovino Madura convencional. El actual enfoque verticalista descendente en las estrategias de conservación y mejora debe ser redirigido hacia un planteamiento ascendente que considere las necesidades de las subpoblaciones sonok y karapan de ganado bovino Madura. Los estudios de seguimiento y caracterización deben recoger información de los diferentes niveles de agregación y tomar en consideración las subpoblaciones. El ejemplo de Madura muestra que los valores culturales del ganado pueden ser un factor principal para el mantenimiento de poblaciones relativamente pequeñas de razas locales.

Palabras clave: Madura, carreras de toros, concursos de ganado bovino, conservación, valores culturales

Submitted 3 May 2013; accepted 5 August 2013

Introduction is also preferred by Madura farmers) because of its red col- In Indonesia, population increase is a major driving force our, similar to the Madura cattle. As a result, the cross- for the rising demand for animal products. To satisfy the breeding through AI is threatening the survival of the demand for red meat, the government has implemented a Madura breed (Barwegen, 2004). policy to import meat and live exotic livestock. From the Currently, the number of cattle in Madura Island is about 1980s, the government has also promoted an Artificial 600 000. It is not possible to obtain accurate data about Insemination (AI) programme using beef breeds from tem- the population of pure Madura cattle as not all regional perate regions, such as Simmental, Limousin and Angus, government offices separate the data in terms of genetic to improve the beef performances of local cattle. At pre- group. sent, the AI organisation uses almost exclusively exotic semen. Thus, cross-breeding is rapidly progressing. Understanding and assessing the cultural values of livestock are important in the implementation of interventions Many publications state that cross-breeding leads to a loss aiming at conservation and utilization of animal genetic of indigenous breeds and loss of adaptation of livestock to resources (FAO, 2012). Madurese people have strong cul- local environments (Anderson, 2003;FAO,2007). On tural ties to their animals, caring for their cattle as if they Indonesia’s Madura Island, until 2001, cross-breeding are their family members (De Jonge, 1990). The signifi- was not allowed to protect the local Madura cattle breed. cance of cattle within the culture is also evident from the Madura cattle developed approximately 1 500 years ago appearance of bulls in local folk-tales, proverbs and from crossing wild banteng or Bali cattle and zebu wood carvings depicting bulls, and from the horns on tra- (Payne and Rollinson, 1976). Continuous selection by ditional houses. Bull racing (karapan) and cow confor- the farmers in Madura created a uniform breed population mation contests (sonok), are important cultural events on (Anonymous, 2003). Its coat is reddish-brown with a non- the island (De Jonge, 1990). specific white pattern on the rump and legs (Anonymous, Cross-bred cattle are kept primarily for meat production. 2003); see Figure 1. Madura cattle are extremely well Therefore, cross-breeding may be a threat not only to the adapted to the local conditions and traditional manage- existence of purebred Madura cattle, but also for the cul- ment. They are reported to be one of the best draught ani- tural activities involving cattle in Madura. mals in the world relative to their size (Barwegen, 2004). Since the removal of import restrictions, the local govern- In Indonesia, Madura cattle offer a unique opportunity to ment has allowed crossing with the Limousin breed (which study cultural practices involving cattle and the effect of Unique cultural values of Madura cattle 143

Figure 1. Colour pattern of the rump and legs of Bali versus Madura cattle (source: Widi, T.SM). cross-breeding. This paper aims to analyse the cultural are expected to have larger body sizes than Madura cattle values of Madura cattle and the effect of cross-breeding and, consequently, require more feed. Cross-bred cattle are on these local traditions. not used for draught purposes. All the crossbreds resulted from Madura cows being inseminated with semen of Limousin bulls. The farmers named the cross-bred cattle madrasin. This study area will be described as “madrasin” Materials and methods area. Since 2007 the local government has promoted a conformation contest for madrasin cattle. Madura cattle in the Study areas madrasin area will be described as conventional Madura Madura (Figure 2) is a small, densely populated cattle. Indonesian island located off the northeast coast of Java The three study areas do not only differ in cattle keeping Island. It is administered as part of East Java Province practices, but also differ in cropping practices. In the kar- and consists of four districts: Bangkalan, Sampang, apan area, cropping conditions are difficult because of Pamekasan and Sumenep. Besides the mainland, Madura small farm sizes and poor fertility of the soil. In the has some small outer islands. The climate is tropical, exhi- fi sonok area, tobacco is the main crop, which results in a biting de ned wet (September to February) and dry relatively good income (Smith, 2011). In the madrasin (March to August) seasons, with an average annual rainfall – area, more crop by-products are available mainly because of 1 500 2 000 mm. of the large paddy growing areas in this lowland area. Two districts were used in this study (Figure 2). One was Sumenep, within which is Sapudi, an isolated island that is demarcated by the national government as a conservation Data collection area for Madura cattle, particularly as a source of karapan Secondary data collection cattle. For the study, Nong Gunong, one of two sub- Secondary data were collected from the Agricultural districts, was used. This study area will be described as Department of Indonesia, the Animal Husbandry Office “ ” the karapan area. The second district was Pamekasan; of East Java Province, Pamekasan and Sumenep District Waru, Pasean, Larangan and Pegantenan sub-districts offices, government websites, publications and insti- were used in this study. Waru and Pasean sub-districts tutional reports. These provided insights into the local situ- are famous as the source of sonok cattle and are well- ation and policies. known as areas where the sonok contest started. This study area will be described as the “sonok” area. Larangan and Pegantenan sub-districts were selected by Participatory observations local government as cross-breeding areas owing to Quantitative and qualitative information were collected the abundance of crop by-products. Cross-bred cattle through participatory approaches involving farmers (cattle- 144 T.S.M. Widi et al.

Figure 2. Map of Indonesia and Madura Island with the three study areas (source: http://geography.about.com/library/cia/ncindonesia.htm).

keepers), government ofﬁcials, community groups and of Maduranese people, and the special characteristics of other key informants and through direct observation of cattle that are important in maintaining local traditions. sonok, karapan and madrasin events. A historic overview of the cultural events was based on the secondary data and Phenotypic characteristics of Madura and cross-bred open interviews with key informants, comprising of per- cattle sons well-versed in cattle keeping and cultural events, Phenotypic characteristics of cows in the study farms were such as heads of cultural event groups, experienced farm- determined by taking height at withers, and estimation of ers or village elders (three persons per study area). body weight by measuring chest girth, using a measure- Semi-structured interviews were conducted with three cat- ment tape (FHK Ogawa Seki Co. Ltd, Tokyo, Japan). egories of cattle-keepers: 30 karapan farmers, 37 sonok The age of each animal (in years) was determined by farmers and 30 madrasin farmers. The farmers were inter- inspecting its teeth. viewed about their background and motivation, technical aspects such as cattle management and the economics of cattle keeping. The input costs and the selling and buying Data analysis prices of individual cattle in a farm over a period of one The qualitative data were analysed descriptively. The year (February 2010–January 2011) were estimated by quantitative data were analysed using one-way ANOVA farmers. The input costs consisted of purchases of forage to compare the different areas (Ott and Longnecker, and supplements, such as maize and rice bran, and 2001). Ranking was used to determine the level of agree- additional feeds such as herbs, eggs and palm sugar for ment of farmers’ motivations for keeping cattle. General animals participating in the cultural events, veterinary ser- Linear Model (GLM) (Ott and Longnecker, 2001) was vices, AI or natural mating, training for and participating in used to analyse ranking preferences. The smallest score the cultural events and hired labour costs. The interviews of rank meant the most important motivation and the high- were used to evaluate the role of cattle in the livelihoods est score of rank meant the least important motivation. Unique cultural values of Madura cattle 145

Results physical and age criteria for karapan. In 2011, in the big- gest racing event, no more than 30 pairs of bulls competed. The history of cultural cattle events in Madura The total number of bulls actually participating in karapan Island is about 400 pairs on Sapudi Island and 200 pairs in Madura Karapan mainland (Kuswadi, 2010, personal communication). Karapan has a very long history. Noer and Maduratna (1975) noted that on Sapudi Island, during the twelfth to Sonok thirteenth century, Prince Katandur wanted to make the The key events in the sonok have resulted from the common soils more fertile. He introduced ploughing using a pair of practice in Madura that, from about 9 am until 1 pm, cattle bulls. To encourage people to keep cattle, he created a bull- are tethered between two pillars, locally named tacek,in racing game, karapan, in which pairs of bulls raced against view of the front veranda of the house. The cattle’s forefeet each other in a field. The equipment used in the race was stand on a platform, a piece of wood about 15 cm high, similar to equipment used in ploughing (Figure 3). while they are tethered. This tethering practice gives farmers the opportunity to show their great pride in caring for their Over time, karapan became a favourite game for cattle. In the meantime, farmers clean the cattle and the Maduranese people, organized after the harvest season. barn, and undertake other activities, such as foot care and These are now held in nearly every village in the eastern horn shaping. In 1963, a village head became particularly sub-districts. In these karapan races, two or three pairs interested in farmers’ behaviour in caring for their cattle. of bulls are raced over a course of 100–140 m pulling a He showed cattle of good conformation tethered in the sledge with a jockey. The winners and losers of the first front of farmers’ houses to high officials who were visiting round are assigned to separate categories, after which the his village. This idea to show cattle daily became a routine races continue according to the “cup system”: the winners activity in the area (Anonymous, 2007) go on, while the losers drop out. A major competition is held each year in which all the island’s districts are rep- In 1927, Sommerfeld (1927) described pajengan in which resented. Since 1982, races for tourists are organized in farmers bring their cattle, especially female cattle, to be western Madura Island. tethered together in a field. In 1967, a local government official organized pajengan activities twice a week during In the 1970s, local government set some rules for karapan the tobacco harvest season. Pajengan is currently still (Noer and Maduratna, 1975): taking place every month. It gives farmers an opportunity (1) Karapan bulls have to be purebred Madura bulls. to meet, showing their good cattle and sharing information (2) The minimum height at the withers of the bulls is 120 with the extension officers. Year after year, this event is cm and the maximum 130 cm. well supported by farmers (Rudi Haryanto, 2009, personal (3) The minimum age of the bulls is 2 years. communication). (4) Mistreating bulls during racing is not allowed. Key informants mentioned that in the 1980s, to make (5) The minimum number of bull pairs having to partici- pajengan more attractive, the competing farmers chose pate in a sub-district is 40. At that time, there were pairs of good females, harnessed to make them walk in 20 sub-districts in Madura Island, so 1 600 bulls pairs and dressed with beautiful adornments. The pairs had to be used every year. of cattle, guided by a jockey, have to walk 25 m to reach fi Nowadays, some of the rules are not respected because of a nishing line, designed like a gate, while their forefeet the limited availability of Madura bulls that fulfil the step in a harmonious manner. Saronen, traditional music,

Figure 3. Karapan race (source: Schultinga, M). Figure 4. Sonok contest (source: Widi, T.S.M). 146 T.S.M. Widi et al.

is played to accompany the cattle when they walk to the Livestock farming systems stage. This event was named sonok; see Figure 4. The Table 1 gives selected household and farm characteristics total number of cattle participating in different sonok for the livestock farming systems in the three study areas. events in Madura is about 600 pairs (Rudi Haryanto, Karapan farmers were relatively older than farmers in the 2009, personal communication). other two areas. Most of the farmers have crop pro- Sonok cattle are judged by conformation traits, such as duction as their main source of farm income. Family height at withers, colour, body conformation, body con- size was significantly larger in the madrasin area. More dition, health and harmonious walking in a pair. Unlike than half of the farmers in the three study areas only karapan, there are no winners and losers in this event. completed elementary school. Farmers had a long experi- Every contestant farmer receives a gift from the officials ence in cattle keeping, about 20 years or more; karapan without exception, and there is dancing accompanied by farmers had significantly more experience than madrasin traditional music. There are juries who provide the audi- farmers. ence with information about the farmer, scores of perform- Cattle are the main livestock (Table 1). Goats, sheep and ance during walking and their opinion about the matching chickens are of secondary importance. Even though the appearance of the pairs. Often the jury consists of influen- land sizes of farmers in Sapudi Island were quite differ- tial people, for example, the head of the sonok community ent from those of farmers in the sonok area, the number group, the inseminator and the head of the district. Since of cattle was not much different. Farmers kept about there are no winners, there is no gambling. Although three head of cattle. Only in the madrasin area were there are no prizes, the contest is very important because the numbers of cattle slightly smaller than in the two during the contest day many transactions among farmers other areas. Farmers in the sonok area will participate take place; farmers can show their cattle and trade them in the sonok contest only if they have pairs of animals for good prices. The cows that perform well are very pop- that meet the criteria for participation and if they have ular for breeding. sufficient cash available to pay for the fee to join the contest. If farmers in the karapan area have a pair of Madrasin potential racing bulls, they will first enter this pair in a In the area where cross-breeding is practiced, farmers, village race. If this pair runs well, they will enter the encouraged by the government, have organized the madra- pair in a regional contest. Again, a farmer will need sin contest since 2007. The criteria are body size and body sufficient cash to join the race and to cover the additional weight, and, for cows, the conformation of their calves. costs. When the farmers have insufficient cash to pur- The cattle are tied up next to each other in the traditional chase cattle, the farmers join with other farmers. way for Madura cattle: rope goes through the nose of the Around 12 percent of the farmers were sharing cattle cattle and is tied between two pillars, and cattle stand (Table 1). The main occupation of the interviewees with their front legs on a platform. In the madrasin contest, was farming, although in the madrasin area, 23 percent the cattle do not parade and are not harnessed or dressed were traders or businessmen. with adornments; see Figure 5. Responses from the interviews showed that the number of participants and the number of participating cattle are growing every year, but it is Motivations for keeping cattle less attractive to spectators than the karapan and sonok As in other areas in Indonesia, in Madura, cattle-keeping events. serves various objectives. Table 2 shows the ranking of the motivations for cattle-keeping in the three study areas. Farmers mentioned the following motivations: financial security (saving), income, providing manure, utilization of crop by-products, raising the social status of their owner, cultural events, draught purposes and hobby. Most farmers mentioned “saving” as the most important motivation. Farmers consider saving in terms of being able to sell cattle to meet unexpected or large expenditures, such as sending children to school, paying hospital bills for a family member, financing a wedding party or a pilgrim- age to Mekka. Sonok and madrasin farmers gave income as the second motive. Farmers consider income as the cash they receive regularly from the annual sale of progeny. Karapan farmers ranked manure as the second most important motive for keeping cattle. Farmers in the kar- Figure 5. Picture of a winner in madrasin contest (source: Widi, T.S.M). apan area depend on their cattle for ploughing. Since Unique cultural values of Madura cattle 147

Table 1. Characteristics of households and farms in the three study areas.

Area

Karapan (N = 30) Sonok (N = 37) Madrasin (N = 30)

Age of household head (year) 51.6a ± 13.5 45.4b ± 10.0 47.1a,b ± 10.7 Land size (ha) 0.31a ± 0.19 0.56b ± 0.76 0.73b ± 0.61 Family size 3.7a ± 0.3 4.5a ± 0.3 5.7b ± 0.4 Experience in cattle keeping (year) 26. 4a ± 2.7 22. 3a,b ± 1.7 20.1b ± 2.1 Number of livestock per farm (n)1 - Cattle 3 (1–13) 3 (1–6) 2 (1–6) -Cows 2(0–6) 2 (0–4) 2 (1–6) - Heifers 0 (0–2) 1 (1–2) 0 (0–2) - Males 0 (0–6) 0 (0–1) 0 (0–4) - Calves 0 (0–2) 0 (0–2) 0 (0–2) - Small ruminants (goat/sheep) 0 (0–6) 0 (0–4) 0 (0–2) - Chickens 0 (0–6) 0 (0–6) 0 (0–10) Cattle ownership (%) - Private 87 89 87 - Sharing 13 11 13 Main occupation (%) - Farming 100 73 70 - Government official 0 16 3 - Trader/private business 0 8 23 - Labourer 0 3 3 Crops Maize, cassava and tobacco, maize, groundnut and Rice, maize, tobacco and groundnut cassava cassava a,b,cDifferent superscripts indicate significant differences between areas (P < 0.05). nsNon-significant. 1Number of livestock is presented in mode and range.

field sizes are small, and land is stony, a bit hilly and are used extensively in the madrasin area, which is a dry, cattle are the best option for ploughing. They use lowland paddy-growing area. both pairs of cows or bulls for ploughing. If needed, they will borrow an animal from a neighbour to match their own animal. Sonok cattle are rarely used for Time allocation for caring for cattle ploughing. Sonok farmers said that they are afraid that Table 3 gives the time spent caring for cattle in the three differ- ploughing will negatively affect the beauty of their ent areas. Sonok farmers and their family members spent sig- cattle. They only use cattle that are not participating in nificantly more time per day managing their cattle than sonok events for ploughing. Madrasin cattle are never karapan and madrasin farmers. They spent time on cleaning used for ploughing, because of the absence of a hump cattle, sunbathing, foot care and horn shaping. In the sonok (needed for the ploughing harness used) and their lim- area, significantly more time was spent on these activities ited endurance. Another reason is that hand tractors than in the other two areas. Sonok respondents emphasized

Table 2. Farmers’ ranking of motivations for keeping cattle in the three study areas.

Role of cattle Area

Karapan (N = 30) Sonok (N = 37) Madrasin (N = 30)

Average score1 Rank2 Average score1 Rank2 Average score1 Rank2

1. Saving 2.2a ± 2.1 1 1.7a ± 0.9 1 2.1a ± 2.1 1 2. Income 5.1b ± 3.2 3 2.4b ± 1.7 2 3.2b ± 2.9 3 3. Manure 2.6a ± 0.9 2 4.5c ± 1.8 4 3.4b ± 2.0 2 4. Social status 5.9bc ± 3.0 5 4.4c ± 2.1 3 6.3c ± 2.4 4 5. Cultural values 6.4c ± 2.4 6 4.9c ± 2.3 5 8.0d ± 0.0 7 6. Draught power 5.7bc ± 2.8 4 7.5d ± 1.1 7 8.0d ± 0.0 7 7. Utilization of crop by- 7.5d ± 1.3 7 7.3de ± 1.7 6 7.5de ± 1.3 5 product or backyard land 8. Hobby 8.4d ± 1.7 8 7.3e ± 2.4 6 7.8e ± 2.2 6

1Different superscripts indicate signiﬁcant differences between reasons (P < 0.05). 2Rank = the smallest score of rank (1) means the most important motivation and the highest score of rank (8) means the least important motivation. 148 T.S.M. Widi et al.

Table 3. Time allocation per farm in caring for cattle in the three study areas.

Activity (h/day) Area

Karapan Sonok Madrasin (N = 30) (N = 37) (N = 30)

1. Cleaning barn 0.48a ± 0.40 0.74b ± 0.47 0.39a ± 0.27 2. Cleaning the cattle, sunbathing, foot caring, horn shape, etc. 3.40a ± 0.88 3.94b ± 1.35 2.12c ± 0.34 3. Gathering feed 2.17a ± 0.48 2.42ab ± 0.85 2.50b ± 0.81 4. Offering feedns 1.08 ± 1.69 1.36 ± 0.43 1.08 ± 1.98 5. Training 0.30a ± 0.57 0.38a ± 0.59 0.00b ± 0.00 Total time spent (h/day) 7.42a ± 1.69 8.84b ± 1.98 6.08c ± 1.12

a,b,cDifferent superscripts indicate signiﬁcant differences between areas (P < 0.05). nsNon-signiﬁcant.

the cleanliness of their cattle and the surrounding environ- comparable with madrasin cows in these measurements. ment, since the daily appearance of sonok cattle, in particular Both sonok and madrasin cows had a significantly higher during the contests and pajengan meetings, is very important. body condition score compared with karapan cows and conventional Madura cows in the madrasin area. Madura cattle are very docile and easy to handle. About once a month, farmers pay attention to horn shape and foot caring. Farmers have a special preference for horn Selling prices, costs and profit estimates shape. They shape the horns when the cattle are around 4 months of age, by punching holes in the two horns The time to sell cattle varies with the purpose of cattle- and binding them using a wire: this results in small to med- keeping. In general, farmers sell animals when they need “ ” ium size horns, curving inwards perfectly symmetrically. cash, but they also sell surplus animals because they fi Most farmers shape cattle’s horns, even when the cattle keep a speci c number of animals in relation to the will not be used for cultural events. They check the feeds available and the time they have available to care hoofs of cattle participants in the karapan or sonok contest for their animals. Farmers sell weaned cattle for fattening – very regularly. Farmers use simple equipment for foot or breeding, usually at 4 6 months after weaning, mainly caring. to traders, whereas cattle for cultural events are sold directly to other farmers who are looking for cattle that Madrasin farmers spent the least amount of time on caring can participate in the cultural events. Sonok farmers said for their cattle (P < 0.05). However, per animal, the time that sometimes they are looking for cattle whose appear- allocation was about the same in the three areas. This ance is well-suited to be paired with one of their own was because the madrasin farmers kept slightly fewer ani- cattle. Once they find a well-matched animal they are will- mals than the other farmers, and forage collection took up ing to offer a good price. Sonok cows are used not only for a large amount of time for madrasin cattle. the cultural event but also for breeding, whereas only excellent karapan bulls are used as breeding bull. Once karapan bulls cannot be used for racing anymore, for Physical characteristics of the cows instance they are already old or injured, they will be sold Table 4 shows that cows in the sonok area were higher and for slaughter. During the survey, only two karapan bulls heavier (P < 0.05) than karapan and conventional Madura were used both in the cultural event and for breeding. cows in the madrasin area. Madrasin cows were signifi- Table 5 presents the comparison of cattle prices and input cantly higher and heavier compared with their convention- costs among purposes and areas. The selling prices of ani- al Madura herdmates. Sonok cows, however, were mals below 1.5 years indicate that animals used for cultural

Table 4. Physical characteristics of Madura and madrasin cows in each study area.

No. Physical characteristic Area

Karapan Sonok Madrasin

Madura (N = 52) Madura (N = 61) Madura (N = 53) Madrasin cross1 (N = 18)

1. Age (year)ns 4.5 ± 1.9 4.3 ± 1.3 4.9 ± 2.7 4.5 ± 0.2 2. Height at the withers (cm) 116.4 ± 4.9a 128.4 ± 6,1b 119.2 ± 10.7c 125.7 ± 6.2b 3. Body weight (kg) 294.3 ± 43.0a 392.3 ± 60.4b 279.1 ± 89.0a 400.1 ± 92.6b 4. Body condition score (BCS) 2.6 ± 0.5a 3.9 ± 0.7b 2.6 ± 0.8a 3.8 ± 0.7b

a,b,cDifferent superscripts indicate signiﬁcant differences between areas (P < 0.05). 1Madrasin cross: cross of Madura (♀) × Limousin (♂). Unique cultural values of Madura cattle 149

purposes fetch much higher prices (P < 0.05) than their tourism as karapan and might be preferred in the future herdmates not used for cultural events. Also, the cash because there are concerns about cruelty to karapan cattle inputs are considerably higher (P < 0.05) for animals parti- and gambling during racing. Even though there is a rule cipating in cultural events. At (relatively) the same age and about not hurting the bulls during racing, in practice, to body weight, Madura cattle used for cultural events were get the bulls to run faster, people use nails or thorns, caus- valued at prices that were 2–3.5 times higher than cattle ing wounds and then even pour salt or lemon into the not participating in cultural events. Karapan and sonok wounds (Kuswadi, 2009, personal communication). On farmers spent considerably more money on cattle older Madura, spiritual leaders are already criticizing these prac- than 1.5 years, which participate in the cultural activities, tices and the gambling during karapan events. 86 and 87 percent, respectively, than on their herdmates The cultural practices have a marked effect on cattle hus- with no cultural function. In particular, the additional bandry. Cattle used for cultural events are well cared for feed costs were high for cattle participating in the cultural and trained. This gives the owners an overall sense of well- events. being. Most writers about Madurese society emphasize the The price of madrasin cattle usually depends on body size. attention that is given to the animals. The men are said to Even when an animal wins the madrasin contest, the be more devoted to their animals than to their wives (De slaughter price is based on its body weight. In 2009, a Jonge, 1990). Karapan and sonok farmers give intensive 800 kg cross-bred bull won the contest, and a butcher daily care and exercise to their animals (about 3 h/day bought it for 22 Million IDR (1 US$ = 9200 IDR), a nor- per animal) to improve their physical appearance and per- mal price per kg live weight. formances. They also spend considerably more money (on The profits per animal aged more than 1.5 years for cattle average 46.5 percent) on better quality feeds for the ani- participating in karapan and sonok were 3 and 2 times mals that participate in the cultural events. higher, respectively, than for their non-cultural herdmates Alderson (2003) mentioned that breed characteristics that (Table 5). Madrasin cattle did not show significantly higher are important for socio-cultural practices generally relate profit than their conventional Madura herd-mate cattle. to the appearance of an animal, and traits that define the Cattle not participating in the cultural event in the sonok market value of an animal. Also, in Madura, cultural area gave a more-or-less similar profitasmadrasin cattle. values have a strong economic component. Most karapan and sonok farmers argued that they like the race or contest because it keeps the prices of cattle high. Karapan and Discussion sonok farmers received more than 50 percent higher prices for breeding stock than the normal market prices. Not only The introduction of cross-breeding in Madura has high- the prices of karapan and sonok cattle that participate in lighted that the Madura cattle population is not homo- the contests were high, but also the overall input costs geneous. There are three cattle production systems, each (Table 5). In general, farmers in the sonok area are rela- in a more-or-less separate area, with now four types of tively well-off due mainly to the income generated from cattle: karapan, sonok, madrasin and conventional tobacco (Smith, 2011). Part of this income will be invested Madura cattle. The main source area of karapan cattle is in their sonok cattle and the participation in sonok events. Sapudi Island. Sonok cattle are found mainly in the northern part of Madura Island. Madrasin cattle and the conven- Effect of cross-breeding on culture and cattle tional Madura cattle can be found in the rest of Madura farming Island. This zonation reflects different cattle production systems in terms of cattle management practices and cul- Farmers outside the karapan and sonok areas kept conven- tural activities. These differences also emphasize the tional Madura cattle before the introduction of Limousin need for different conservation and breeding strategies semen. These were relatively small in size. This made it for the different Madura cattle sub-populations. easy for the government to introduce Limousin semen at that time, as the farmers wanted cattle as big as the sonok cattle. Almost all madrasin farmers said that they Synergy of culture and cattle farming in Madura took more pride in keeping madrasin cattle than conven- Karapan and sonok events as well as Madura cattle are tional Madura cattle. Table 4 shows that cross-breeding unique, as markers of the traditions and culture for increased the body size of cows to about one-third. The Maduranese people (De Jonge, 1990; Smith, 2011). The body condition score of madrasin cows was also much events are famous, not only in Madura itself but also higher than that of their conventional Madura herdmates. throughout Indonesia and even abroad. Thus, they are Cross-bred animals were fed better; this resulted in higher also a tourist attraction. The total number of spectators in feeding costs for madrasin cattle (Table 5). Nevertheless, all events in a year is approximately 20,000–25 000 the selling prices minus the total cash costs for the period people. Karapan is much older and better known than an animal is kept, show that the cash profits of madrasin sonok, but the enthusiasm for sonok seems to increase animals were on average, 55 percent higher than from con- year by year. Sonok may become as important for rural ventional Madura herdmates (Table 5). The economic 5 ...Widi T.S.M. 150

Table 5. Age, period of keeping, selling price, input costs and cash proﬁt per period of keeping of Madura and madrasin cattle based on the age group, area, purpose of keeping cattle and type of cattle (Million IDR)1.

Age group/ Area/category of cattle al. et

Karapan Sonok Madrasin

Animals in the event Animals not in the event Animals in the event Animals not in the event Madura breed Madrasin cross2

<1.5 year (N =13:♂ =13;♀ =0) (N = 22:♂ = 10;♀ = 12) (N =12:♂ =0;♀ = 12) (N = 18:♂ =6;♀ =12) (N =4:♂ =0;♀ =4) (N =16:♂ =13;♀ =3) Age (year) 0.60 ± 0.29 0.36 ± 0.21 0.75 ± 0.29 0.51 ± 0.29 0.85 ± 0.44 0.68 ± 0.42 Period of keeping3 (year) 0.47 ± 0.27 0.37 ± 0.20 0.68 ± 0.30 0.48 ± 0.25 0.77 ± 0.43 0.64 ± 0.30 Selling price (Million IDR) 14.81a ± 8.1 2.85b ± 0.87 20.38a ± 5.27 5.88b ± 2.26 5.93a ± 2.79 7.65a ± 5.53 Total input cost per period of keeping (Million IDR) 0.90a ± 0.80 0.04b ± 0.11 2.83a ± 2.16 0.12b ± 0.25 0.24a ± 0.36 0.23a ± 0.28 • Forage 0.04 ± 0.10 0.04 ± 0.09 0.09 ± 0.14 0.06 ± 0.41 0.14 ± 0.27 0.11 ± 0.18 • Additional feed4 0.80 ± 0.69 (88.9%) 0.00 1.12 ± 0.93 (39.6%) 0.06 ± 0.11 0.11 ± 0.88 0.11 ± 0.14 • Veterinarian 0.01 ± 0.02 0.00 0.01 ± 0.03 0.00 0.00 0.00 • AI/mating 0.00 0.00 0.00 0.00 0.00 0.00 • Training, racing/contest 0.07 ± 0.08 (7.8%) 0.00 1.50 ± 1.29 (53.0%) 0.00 0.00 0.00 • Labour 0.00 0.00 0.03 ± 0.09 0.00 0.00 0.00 Profit per period of keeping (Million IDR) 14.2a ± 8.1 2.6b ± 0.9 18.1a ± 4.3 5.8b ± 2.1 5.7a ± 3.0 7.4a ± 5.2 >1.5 year (N =10:♂ =10;♀ =0) (N =6:♂ =0;♀ =6) (N =10:♂ =0;♀ = 10) (N = 11:♂ =0;♀ =11) (N =9:♂ =0;♀ =9) (N =29:♂ =16;♀ =13) Age (year) 2.86 ± 1.21 3.00 ± 1.38 2.21 ± 0.94 2.85 ± 1.34 3.27 ± 0.87 2.65 ± 0.89 Period of keeping (year) 1.03 ± 0.08 1.83 ± 0.91 1.27 ± 0.34 2.01 ± 0.70 1.3 ± 0.62 1.58 ± 0.49 Selling price (Million IDR) 21.90a ± 8.70 6.17b ± 0.68 25.44a ± 4.40 11.59a ± 4.43 6.59a ± 1.66 12.15b ± 4.89 Total annual input cost (Million IDR) 4.30a ± 1.80 0.08b ± 0.10 4.66a ± 1.60 1.03b ± 0.50 0.45a ± 0.39 0.91a ± 0.40 • Forage 0.34 ± 0.16 0.07 ± 0.10 0.24 ± 0.22 0.36 ± 0.23 0.19 ± 0.25 0.47 ± 0.27 • Additional feed 3.70 ± 1.5 (86.0%) 0.00 2.35 ± 0.94(50.4%) 0.66 ± 0.37 0.24 ± 0.19 0.42 ± 0.22 • Veterinarian 0.006 ± 0.02 0.004 ± 0.01 0.009 ± 0.02 0.00 0.00 0.003 ± 0.009 • AI/mating 0.01 ± 0.01 0.01 ± 0.01 0.006 ± 0.01 0.007 ± 0.01 0.06 ± 0.02 0.006 ± 0.02 • Training, racing/contest 0.25 ± 0.20(5.8%) 0.00 2.07 ± 0.83 (44.4%) 0.00 0.00 0.00 • Labour 0.00 0.00 0.00 0.00 0.00 0.05 ± 0.18 Total input cost per period of keeping5 (Million IDR) 4.4 ± 1.8 0.17 ± 0.3 6.01 ± 3.1 2.1 ± 1.6 0.6 ± 0.7 1.5 ± 0.9 Estimated profit per period of keeping6 (Million IDR) 17.6 ± 7.8 5.9 ± 0.7 19.4 ± 5.5 9.4 ± 4.3 5.9 ± 1.6 10.6 ± 4.7 Annual profit7 15.9a ± 9.6 3.9b ± 1.6 17.8a ± 11.5 5.1b ± 2.5 4.2a ± 2.2 7.5a ± 4.3 a,b Different superscripts indicate significant differences between different categories of cattle between areas (P < 0.05). 1One US $ = 9,200 IDR. 2Madrasin cross: cross of Madura (♀) × Limousin (♂). 3Time needed to keep cattle from purchasing or being born until sold. 4Feeds that are added into the basal diet (forage), such as concentrate, rice bran, maize bran, palm sugar, herbs, etc. 5Total input cost per period of keeping (Million IDR) for cattle with age group >1.5 years was calculated by multiplying total annual input cost (Million IDR) and period of keeping (year). 6Profit per period of keeping (Million IDR) was calculated by deducting total input cost per period of keeping (Million IDR) from selling price (Million IDR). 7Annual profit (Million IDR) = Profit per period of keeping (Million IDR)/period of keeping (year). Unique cultural values of Madura cattle 151

profit estimates in Table 5 ignore the other benefits cattle Nevertheless, the future of the karapan event is at stake. provide, such as manure, draught power and security. A There is no strict adherence to the government rules for comprehensive economic analysis is needed to calculate the karapan race, and the numbers of bulls that can join the overall benefits of the different types of cattle keeping a karapan race is going down. (Smith, 2011). Inbreeding is strictly avoided in the sonok area, by circulat- Also, cross-breeding contributes to the overall pride resulting the bulls between villages whereas, in the karapan ing from keeping cattle that perform well. The madrasin area, which is an isolated island, selection to avoid contest was initiated by the government and seems a top- inbreeding is not done. This may have contributed to the down initiative. It attracts in particular farmers; no tourists small size of karapan cows. This requires urgent attention, attend. The main motivation of farmers for enrolling their as it can be a threat to karapan as well as the capacity for animals in the contest is the expectation that it can result ploughing on Sapudi Island. in a fair price for their animals, since all the participants’ AI was introduced in the late 1990s in Sapudi Island, but it cattle are weighed and the price per animal is based on was never a success. The main reasons were that infrastruc- the actual weight. In general, when the cattle are sold, the ture was poor, the availability of AI personnel was limited, prices that traders offer are based on visual appraisal only. farmers considered AI a sin and an abnormal way of mating Cross-breeding is not only influencing the feeding prac- their cattle, and farmers never believed that the Madura tices and selling prices of cattle, but also influences the semen used for AI was from a karapan bull. cattle trade flows. Key informants explained that most The phenotypic performances of sonok cattle are compar- small-scale slaughter-men in Madura do not like to slaugh- able with the performances of madrasin cattle. This shows ter cross-bred animals as they cannot handle and sell the that through good breeding practices, farmers have main- products within one day. Therefore, cross-bred slaughter tained good performances of sonok cattle. This guarantees animals are transported as live animals to the abattoir in successful conservation of Madura cattle in the sonok area Surabaya, the nearest major city in Java. in the future. It can be speculated that, in the future, the great majority of Conservation and breeding developments the conventional Madura cattle will be crossbreds. Farmers The different farming systems and the different cultural outside the sonok and karapan areas prefer the madrasin events, and consequently the different types of Madura cattle over conventional Madura cattle due mainly to their cattle, show the importance of considering local variation heavier body weights and consequently higher market when developing conservation strategies. Government prices; although farmers said that they are aware that there bodies that are concerned with livestock conservation is a risk that the second and third generations of cross-bred focus primarily on numerical factors (Alderson, 2009). cattle will show poorer performances than the present F1 Also, in Indonesia, the inventory of animal genetic crossbreds. Research is needed on the performances of future resources for the FAO world’s animal genetic resources generations of crossbreds to inform farmers about potential gives only the total number of Madura cattle losses in production with continuing cross-breeding. (Anonymous, 2003). It does not distinguish between The governmental breeding centre is promoting cross- different Madura types of cattle. breeding in the south of Madura and Madura cattle in Although Sapudi is designated by the government as an the north. Unfortunately, it does not distinguish between in-situ conservation area for Madura cattle (Anonymous, the sonok and karapan sub-populations. The present top- 2003), until now, no specific conservation strategies have down approach towards implementing the conservation been implemented, except that cross-breeding is not of Madura breed has to be turned into support approaches allowed. In the whole of Madura, there are no reliable that involve the livestock keepers themselves. data on numbers of different cattle. There is an urgency to correct that deficiency, especially in the face of the rapid change in the genetic structure of the island’s cattle Assessment of cultural values herd because of the high adoption rate of cross-breeding Gandini and Villa (2003) developed a framework for in the major part of the island. assessing different cultural contributions of livestock Field observations and the discussions with key stake- breeds in Europe with the aim of strengthening their con- holders indicated that the populations of cattle in the kar- servation. The present study illustrates the complexity apan and sonok areas are relatively small: about 45 000 behind assessing cultural values of local breeds in and 50 000 animals, respectively. However, there is no countries such as Indonesia. Cultural assessments should immediate threat that these populations will be reduced not be restricted to identifying cultural functions and esti- by cross-breeding because of the continuing importance mating market value of cultural products and services; they of the cultural values of the karapan and sonok events. have to include the effect of cultural functions on manage- It is unlikely that the ban of importing other cattle, in ment and specific requirements for animals that have cul- particular crossbreds, in Sapudi Island will be lifted. tural value, and the social and economic contribution of 152 T.S.M. Widi et al.

these animals to the livelihoods of the households farmers community as well as all farmers. Last but not concerned. least, the first author would like to thank the undergraduate The Madura example shows that the cultural value of a students for their support and assistance in collecting the specific breed can be related to a complex mix of area- data. specific cultural functions that are associated with different animal performances, management practises and economic benefits. Consequently, participatory research approaches are needed to unravel such complexities, References with information collection at animal, farm and regional Alderson, L. 2009. Breed at risk: definition and measurement of the levels. factors which determine endangerment. J. Livestock Sci. 123: 23–27. Anderson, S. 2003. Animal genetic resources and sustainable livelihoods. J. Ecological Economics. 45: 331–339. Conclusions Anonymous. 2003. National Report on Animal Genetic Resources Indonesia; a Strategy of Development. Department of Agriculture, The Madura cattle example shows that the cultural values Jakarta. of livestock can be a main driver of maintaining a local Anonymous. 2007. Riwayat singkat tentang kegiatan sapi cangkean/ breed. The enthusiasm of people and farmers for maintain- pajangan dan sapi sono di kecamatan Waru. Kabupaten ing the cultural values attached to Madura cattle, and con- Pamekasan, Pamekasan. sequently the high prices for cultural cattle are keys to Barwegen, M. 2004. Browsing in livestock history; large ruminants conserve the Madura cattle within the areas where the cul- and the environment in Jawa, 1850–2000. In P. Boomgaard, & tural events are maintained. The future of sonok cattle is H. David, eds, Smallholder and Stockbreeders; Histories of better than karapan cattle in terms of performances, and Foodcrop and Livestock Farming in Southeast Asia. KITLV Press, – general preference of people. Cross-breeding will not Leiden: 283 306. directly influence the cultural events and management De Jonge, H. 1990. Of bull and men: The Madurese aduan sapi. – practices of Madura cattle in the areas where the cultural Bijdragen tot de Taal-, Land-en Volkenkunde 146(4): 423 447. events are important. In the areas where the cultural events FAO. 2007. Global Plan of Action for Animal Genetic Resources and the are not important, farmers, however, take more pride in Interlaken Declaration. FAO, Rome (available at http://www.fao.org/ keeping madrasin than conventional Madura cattle, mainly docrep/010/a1404e/a1404e00.htm). because of their growth performance and the resulting FAO. 2012. Phenotypic Characterization of Animal Genetic Resources. financial benefits. Therefore, cross-breeding is an immedi- FAO, Rome (available at http://www.fao.org/docrep/015/i2686e/ i2686e00.htm). ate threat to the future of conventional Madura cattle in the main part of Madura. Monitoring and characterization of Gandini, G.C. & Villa, E. 2003. Analysis of the cultural values of local – animal genetic resources studies have to consider cultural livestock breeds: a methodology. J. Anim. Breed. Genet. 120(1): 1 11. values of livestock breeds and how these values interact Mohammad Noer, D. & Maduratna, Ch. 1975. Kerapan sapi: permai- with management practices, animal performances and nan dan kegemaran rakyat di kepulauan Madura. Kinta, Jakarta. economic benefits. Ott, R.L. & Longnecker, M. 2001. An introduction to Statisctical Methods and Data Analysis. Duxbury, Thomson Learning, Inc, Pacific Grove, CA. Payne, W. & Rollinson, D. 1976. Madura cattle. J. Z. Tierzüch Acknowledgements Zuchtsbiol 93: 89–100. The first author was supported by a PhD grant from the Smith, W. 2011. Ecological anthropology of households in East Madura, Directorate General of Higher Education, Department of Indonesia. Wageningen University, Wageningen (Ph.D. thesis). Education, Republic of Indonesia. The first author Sommerfeld, K. 1927. Das Madurarind, ein hervorragender Typ der acknowledges the support of the DGHE, Gadjah Mada Java-Madura-Sumatra-Rasse, eine kombinationszuchtung zweier Rinderarten, des Balirindes (domestizierten Bantengs) und des vorder- University and Animal Production Systems, Wageningen indischen Zebus und seine wirtschaftliche und sonstige Verwendung University, the Netherlands, staff members of Livestock und Bedeutung. J. Zeitschrift fur Tierzüchtung und Zuchtungsbiologie Bureaus in Madura, the leaders of sonok and karapan 8: 65–112. Animal Genetic Resources, 2014, 54, 153–162. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000083

Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia

Sandip Banerjee, Mohammed Beyan and Hiwot Bekele School of Animal and Range Sciences, Hawassa University, Hawassa Ethiopia

Summary Livestock husbandry plays an important role in the rural economy of the country in a tangible or intangible way. The livestock are selected for their functional efficiency and also for their adaptability to the particular agro climate. The study was conducted at two selected woredas (districts) of Southern Ethiopia and was based on individual interviews; focus group discussion and questioner survey. The results indicated that livestock are selected based on some traditional methods by the elders of the society. The respondents indicated that they maintained informal pedigree for cattle and horses based on their dam line, while phenotype of an animal viz. presence or absence of hair whorl, udder and teat length, length of the legs, body length played an important role in selecting the bovines. Tail type, presence of wattles, leg length and coat colour were considered while selecting the small ruminants. Chickens with single comb are preferred over the other types. Coat colour played an important role in selection for functional efficiency of cattle. The study indicated that most of these selection attributes had correlation with the production and reproduction efficiency of the livestock as was explained by several authors. It is suggested that further studies should be carried out to validate the selection methods and incorporate them in region-wise livestock selection strategies.

Keywords: livestock selection, Southern Ethiopia, traditional methods

Résumé L’élevage joue un rôle important dans l’économie rurale du pays de façon tangible ou intangible. Le bétail sont sélectionnés pour leur efficacité fonctionnelle et aussi pour leur capacité d’adaptation au climat agro particulier. L’étude a été menée dans deux woredas sélectionnés du sud de l’Ethiopie et a été basée sur des entretiens individuels, groupes de discussion sondage de discussion et poser des questions. Les résultats indiquent que le bétail sont sélectionnés sur la base des méthodes traditionnelles par les anciens de la société. Les répondants ont indiqué qu’ils maintenaient pedigree informel pour les bovins et les chevaux en fonction de leur lignée maternelle, alors que le phénotype d’un savoir sur les animaux. La présence ou l’absence de spirale de cheveux, la mamelle et les trayons longueur, la longueur des jambes, la longueur du corps ont joué un rôle important dans la sélection des bovins. Type de queue, la présence de barbillons, la longueur des jambes et la couleur du pelage ont été considérés lors de la sélection des petits ruminants. Poulets avec un peigne simples sont préférés aux autres types. Robe a joué un rôle important dans le choix de l’efficacité fonctionnelle du bétail. L’étude a indiqué que la plupart de ces attributs de sélection avait corrélation avec la production et l’efficacité de reproduction du bétail comme cela a été expliqué par plusieurs auteurs. il est suggéré que d’autres études doivent être menées afin de valider les méthodes de sélection et de les intégrer dans les stratégies de sélection du bétail sages de la région.

Mots-clés: sélection du bétail, méthodes traditionnelles, Sud d’Éthiopie

Resumen La ganadería desempeña un papel importante en la economía rural del país ya sea de forma tangible o intangible. El ganado es seleccionado en base a su eficiencia funcional y también de acuerdo con su capacidad de adaptación a unas condiciones agroclimáticas particulares. Este estudio fue llevado a cabo en dos woredas del Sur de Etiopía y para su realización se emplearon entrevistas individuales, grupos focales de discusión y una encuesta. Los resultados indicaron que el ganado es seleccionado por los ancianos del lugar en base a métodos tradicionales. Las personas encuestadas señalaron que controlan, de manera informal, la genealogía del ganado bovino y de los caballos, siguiendo la línea materna. Rasgos fenotípicos del animal como, por ejemplo, la presencia o ausencia de remolinos de pelo, el tamaño de la ubre y la longitud de los pezones, la longitud de las patas y la longitud del cuerpo son criterios importantes para la selección del ganado bovino. El tipo de cola, la presencia de mamellas, la longitud de las patas y el color de la capa son aspectos que se tienen en cuenta para la selección de los pequeños rumiantes. En cuanto a las gallinas, se prefieren aquéllas con cresta sencilla. El color de la capa juega un papel destacado cuando la selección va encaminada a mejorar la eficiencia funcional del ganado bovino. El estudio mostró que la mayoría de estos atributos de selección estaban correlacionados con la eficiencia productiva y reproductiva del ganado, tal y como ya había sido observado por varios autores. Se sugiere que los próximos estudios

Correspondence to: S. Banerjee, School of Animal and Range Sciences, Hawassa University, P.O. Box 05, Hawassa, SNNPRS, Ethiopia. email: sansoma2003@ yahoo.co.in; [email protected]

153 154 S. Banerjee et al.

deberían ir dirigidos a validar los métodos de selección para, a continuación, incorporarlos en las estrategias de selección de ganado de la región.

Palabras clave: selección del ganado, métodos tradicionales, Sur de Etiopía

Submitted 5 September 2013; accepted 27 February 2014

Introduction Materials and methods Livestock husbandry plays an important role in shaping the livelihood among all traditional societies of the The study was conducted at Arbegona and Lok Abaya wor- world. Since the advent of agriculture, livestock rearing edas situated in SNNPRS region of Southern Ethiopia. The ′ ′ has been an integral part of rural development. The live- Arbegona woreda is situated between 6.47 N and 6°82 N ′ ′ stock species being reared and the type within a species latitude 38°59 E and 38°84 E longitudes. The agro ecology have been largely selected by rearers keeping in mind of the woreda comprises 92percent (highland) and 8 per- the needs and prevailing agro climate in the region. cent mid land although Lok Abaya woreda is situated ′ ″ ′ ″ Such a delicate balance between need, adaptation and West of Lake Abaya (6°17 25 N 37°49 44 E) and the availability of the livestock species govern the develop- major part of the woreda is in the lowlands. These two wor- fi ment of a particular type which is further selected both edas were purposively selected as they were identi ed by naturally and by the inhabitants to suit their specific the Bureau of agriculture to have a large section of agrarian requirements. In the last century, many temperate live- society practicing traditional livestock husbandry practices. stock breeds were introduced in Asia, Africa and Latin The two woredas are in contrast to each other as the former America which altered the traditional livestock husbandry is largely highlands and midlands whereas in the latter a practices and also the introduction of modern selection major part is the lowlands; based on the agro climatic con- criteria has led to the dilution of the traditional attributes ditions the livestock selection methods too were expected to of livestock selection in these regions. Result of a study be different. The study was conducted using stepwise pur- by Mohanty (2004) indicated that selection of livestock posive selection methods. The respondents were selected has been an age old practice in India and the methods from the community based on their history of livestock are often made simple by integrating it with the folk tra- rearing followed by their association with livestock hus- ditions, similar observations have also been reported by bandry methods, furthermore the selection was narrowed Hiwot (2013) from Ethiopia, where she recorded folk down to people who practiced traditional methods of selec- fi songs sung by the livestock rearers of Southern tion for the livestock they owned. Thus, ve kebeles (vil- Ethiopia indicating the virtues of good cattle, in both lages) were selected based on their proximity to all cases the authors have observed that the selection meth- weather roads, three villages (Toga, Rooko and Hafurssa ods are mostly based on the phenotypic description of Nameto) were selected from Arbegona woreda, whereas the livestock. As indicated by Jena (2007) indigenous two villages (Falka and Abaya Zurya) were selected from technical knowledge (ITK) is an integral part of the cul- Lok Abaya woreda. ture and history of local communities. It has evolved The study was based on a semi-structured questionaire fol- through many years of regular experimentation on the lowed by focus group discussions and individual inter- day-to-day life and available resources surrounding the views. The respondents were also requested to practically community, and is essential for maintenance of the gen- show the selection criteria they followed for their live- etic resources for continued survival of the breed/type. stock, all the interviews were video recorded. The data Indigenous knowledge has significance in the designing were analysed using simple descriptive statistics (percent- of sustainable farming systems including animal husband- age and ranking). ry practices thereby improving the livelihood of rural populations who accept, develop and maintain innova- tions and interventions (Singh and Kumar, 2012). Ghosh and Sahoo (2011) were in the opinion that ITK Results is essential for the maintenance of genetic resources (both plant and animal) for the well being, sustenance The results of the study indicated that almost all the and development of the community. respondents were elders (aged above 50 years) and most of them did not have any formal education, these observa- The present study highlights some of the traditional live- tions are in accordance with the findings of Mesfin and stock selection methods as practiced by the inhabitants Obsa (1994), Fekadu (2010) and Mishra and Patro of two woredas of the Sidama zone of Southern Ethiopia. (2010) who also reported that the harbourers of traditional Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia 155

Table 1. Various reasons for rearing livestock as indicated by the fermented curd) are considered as delicacies in Ethiopian respondents. cusine, thus requiring a lot of dairy inputs. The results Purpose of keeping livestock Arbegona Lok Abaya also indicate that the respondents in Arbegona rear livestock to supplement their income while those of Lok Milk and milk products 1st 1st Abaya prefer to rear livestock as a source of tribal pride nd rd Income source 2 3 and insurance against unforeseen calamities. However, Pride and insurance 3rd 2nd Others (hide, skin, traction, meat etc.) 4th 4th the study also indicates that the sale of livestock products such as hide, meat etc., is ranked the lowest as there is a 1 denotes the most and 4 the least important criteria. lack of market opportunities for the sale of such value added livestock products. It was also observed during the study that the respondents of Arbegona woreda still barter knowledge are the elders among the community. The dairy products for household commodities like salt, the results from Table 1 indicate the various reasons for findings are in accordance with the results of Fekadu which the respondents prefer to rear livestock. The results (1994), from southern parts of Ethiopia and also from indicate that the primary reason is to meet their require- Amhara region by Azage et al. (2010). The uses of livestock ments of dairy products. Butter is an important product asasourceofincomeandforfulfilling emergency needs are fi both for culinary uses like niter kibbeh (clari ed spiced in accordance with the observations of Iqubal (2013). butter used in almost all the dishes) and also as a hair moisturiser for women. Moreover, Ayib (Ethiopian cottage The results from Table 2 indicate the criteria used by the cheese prepared from fermented milk) and Ergo (naturally respondents for selection of livestock. The results also

Table 2. Criteria used for selection of livestock by the respondents in the study areas.

Criteria A (percent) L (percent)

Selection for milking cow and heifer ✓ Small head, thin slender neck in cow 55 49 ✓ Long teat, well-structured udder 89 91 ✓ Good milk veins 45 86 ✓ Large, hollow body cavity 32 27 ✓ Thin skin denotes high milk yields 6 66 ✓ Fine tail of a cow, not too long 54 56 ✓ Docility, hence good lactation and fertility in cows 44 52 ✓ Hair whorl at the back 75 4 ✓ Straightened rear ﬂank 63 55 ✓ Hind loin no ﬂesh 56 54 ✓ Prominent pelvic girdle and tail bone 62 44 Selection of bull ✓ Long penis and short and thick neck denotes good fertility in bull 86 78 ✓ Straight and strong hump 68 78 ✓ Large scrotum 89 90 ✓ Strong voice denotes fertility 67 77 ✓ Broad ear 23 18 ✓ Long tail 56 43 ✓ Lengthy and good temperament 77 86 Selection of cattle for fattening ✓ Straight backbone 78 86 ✓ Straight legs denote good meat production 54 63 Selection of ox ✓ Body length and muscularity 76 68 ✓ Castrated 45 61 ✓ Not very fat 11 31 Sheep and goat ✓ Big udder in goat milk production – 78 ✓ Long penis and large scrotum denote good fertility 45 37 ✓ Well-developed wattle 90 97 ✓ Flat and broad ear 56 68 ✓ Fat and broad tail in sheep while short for goats 54 44 ✓ Long and thin legs – 100 ✓ Colour 44 Selection of horse ✓ Thin and long leg 89 ✓ Aggressive and muscular 78 ✓ Long tailed 56 ✓ Sire line 49 ✓ Broad hoof 87 Selection of poultry ✓ Single comb 78 67

A, Arbegona L, Lok Abaya. 156 S. Banerjee et al.

based on its body length and muscularity, it should be castrated with not much fat, the rankings are consistent; however, the degree varied between the respondents of the two woredas. The result of the study further indicated that the goats are selected based on the presence of wattles, where goats with wattles are preferred over those not having one, the observation was consistent in both the woredas. The respondents of Lok Abaya woreda preferred to select does with good udder capacity, whereas those of the Arbegona woreda preferred to select bucks with longer penis. The respondents of Lok Abaya preferred to select goats with long and thin legs followed by flat and broader ears while it was the primary criteria for selecting goats by the respondents of Arbegona worda. The study further indicates that respon- Figure 1. Dairy traits were used by the respondents of Lokabaya to select a cow. dents of both the woredas preferred to rear sheep with fat and broader tails, while for goats short tail was preferred. Selecting small ruminants based on their coat colour was indicate that while the primordial criteria for the selection practiced only among the respondents of Arbegona woreda. of milch cattle as was reported by the respondents from The study further indicated that horses were reared only in both the studied woredas was the length of the teat and Arbegona woreda, where they were selected based on their well-structured udder (Figure 1), followed by prominent long and thin tail and broad hooves followed by their mus- milk veins in Lok Abaya and hair whorls on the back in cularity, tail length and sire line. The study further revealed Arbegona. The third important criteria as mentioned by that the respondents (irrespective of the woredas) preferred the respondents in Lok Abaya was the presence of thin single combed chickens. skin among the cattle while those from Arbegona preferred to give importance to straight rear flank and prominent pel- The results as presented in Table 3 indicate the respondent’s vic girdle and tail bones. perception towards selection of livestock based on their coat colour. The study indicates that while cattle with black coat The study further indicated that the bulls were selected colour is preferred in Arbegona woreda (Figure 2), the based on the size of their scrotum with larger scrotum respondents of Lok Abaya woreda preferred rearing cattle denoting better fertility this criterion was similar in both with light-coloured coat (Figure 3). The results further indi- the studied areas, this was followed by the length of the cated that the respondents in both the studied locations pre- penis where it was considered by the respondents that ferred to rear cows with reddish brown or red colour coats bulls with longer penis had better fertilizing ability, this for they believe that such cattle are good milk providers, was followed by selection of docile bulls (for better hand- whereas the respondents in Arbegona woreda prefer to ling) and also the bulls with longer body length were con- rear cattle with grey colour coats which they consider to sidered to have a better libido. The results also showed that have beef with high fat percentage. the respondents selected bulls with hoarser voice, which they thought denoted muscularity. The selection of fattening livestock depended on the two Discussions prime criteria which was consistent in both the studied woredas, the first being straight backbone and legs, while The respondents of the present study preferred to rear the respondents indicated that an ox is usually selected native livestock breeds over the crossbreds, they indicated

Table 3. Selection of cattle based on their coat colour as indicated by the respondents in the study areas.

Coat colour (Percent) Remark

Arbegona Lok Abaya

Black 49 – ✓ Strong survive the dry season, do not emaciate ✓ High market access White (creamy) – 80 ✓ Can resist the hot environment, shiny appearance ✓ High market demand Dalecha (grey) 7 –– Bure (reddish brown) 14 10 ✓ Can give more milk Red 11 6 – Blue 10 4 ✓ Have fatty meat Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia 157

The study further indicates that the respondents still prefer the age old phenotypic selection methods and that the better animals are identified based on the coat colour. The coat is correlated with the production traits and other similar phenotypic characters, these observations are similar to those reported by Karthikayan and Gajendran (2005). The phenotypic selection methods (both qualitative and quantitative traits) as observed in the study find similarity with the observations of Ouma, Abdulai and Drucker (2006), Wurzinger et al. (2006), FABRE (2006), Mbuku, Isaac and Alexander (2006), Shiferaw (2006), Kakar (2009), Endashaw et al. (2012) and Shigdaf (2011). The study also indicated that the mating of the livestock is Figure 2. Dark coat color of cattle reared in the Arbegona Woreda. by and large panmictic, this is because most of the mating takes place in the communal grazing lands. The observations are in close agreement with the reports of Shiferaw that the native breeds play a multi-functional role in rural (2006) for Kereyu cattle reared in the Fental district of livelihoods, contributing not only to cash products but Oromia region and also of Agere (2008) for Horro cattle also to manure, and have a cultural importance. The from the Horro Gudru area. respondents in the study reported that the crossbred and exotic breeds of livestock are unable to thrive in the area The two prevalent methods of livestock selection are the due to lack of proper veterinary facilities and feed avail- phenotypic selection and also an oral tradition of keeping ability throughout the year. They are also prone to com- the pedigree of the calves based on their maternal lineage. monly occurring diseases prevailing in the region, while Formal pedigree recording in livestock has been a very old on the other hand the native breeds have much better tol- tradition in many parts of the globe with reports from erance of diseases and are also able to thrive under envir- Winchester (1977) indicating the presence of such lineage onmental stress; during the dry season (October–March) of horses on clay tablets of Mesopotamia. The study also they are mostly under fed and receive water once in several indicates that the respondents were aware of the heritable days. The quality of the feed they receive is also very poor, attributes of the livestock, which included physical, pro- just enough for their sustenance. duction and behavioural traits. The method of pedigree in the traditional breeding system has also been reported The results are in consonance with the observations of by Mathias, Ilse and Jacob (2005) from other parts of Kunene and Fossey (2006) who reported that the farmers Ethiopia, and also by Aujla, Jasra and Munir (1998), in the Natal region of South Africa prefer native (Nguni) Kunene and Fossey (2006), Mbuku, Isaac and Alexander livestock (cattle, sheep and goats) over the exotic breeds (2006), Raziq and Younas (2007), Raziq, Younas and because of better adaptability and the ability of the native Kakar (2008) and Kakar (2009) from different parts of breeds to thrive on locally available feed resources. Similar the world. observations as indicated in this study too have been The respondents were well aware of the problems asso- reported by Raziq and Younas (2007), Raziq, Younas ciated with inbreeding in any form and try to avoid the and Kakar (2008), Bavikatte and Jonas (2009), Kakar same in cattle and horses as they believe that the offspring (2009). born of such breeding have a weak physical conformation, are susceptible to diseases and have retarded growth thereby affecting their survivability. The results presented in Table 2 indicate that the major criteria for the selection of cows is the presence of long teat and well-structured udder, the long teats facilitate milking and studies by Morse et al. (1988) have indicated that cows with longer teat tend to suffer less from mastitis, the reason being that the teat canal serves as an effective barrier towards any external infection, as the teat canal closes within a few minutes of milking and thus shutting off the link between the external environment and the udder tissues, the results, as reported by the respondents, too find similarity with the findings of Kuczaj (2003), Alphonsus, et al. (2010) and Yakabu (2011). However, results of studies by Berry, Harris and Winkelman Figure 3. Light coat colour of cattle selected at Lok Abaya woreda. (2005) and Yamazaki et al. (2009) indicate that cows 158 S. Banerjee et al.

with long teats tend to hurt themselves while grazing or (1984)), thereby the milk productivity of such animals is even while lying down and are therefore more prone to expected to be less in the lowland areas besides it may mastitis, whereas Tilki et al. (2005) reported that there also affect the reproductive efficiency of the cattle. was a negative relationship between teat length and milk yield. The presence of well-developed milk veins was As observed in the study, coat colour too plays a primor- also considered as a selection criterion and the results dial role in the selection of small ruminants by respondents are in accordance with the observations of Bonsma of Arbegona wordea; the observations are in accordance fi (1980). The respondents in Lok Abaya woreda reported with the ndings of Shigdef (2012) who indicated that that the cattle with thin skin have better milk producing white and red coat colour are preferred in sheep and ability, the observations are in consonance with the results goats in the Washera and Ferta districts whereas black is obtained by Bonsma (1980) and Hamid et al. (2000). the least preferred one. The selection of goats on the However, the results of studies by Hammond, Ian and basis of their coat colour too has been reported by Terence (1971) indicate that cattle with thinner hide tend Mbuku, Isaac and Alexander (2006) by pastoralists of to be more susceptible to tick borne diseases as thinner Kenya. skins with high vascularity have blood capillaries near The preferences of selecting cattle based on their coat col- the skin surface, which in one way helps in dissipation our as reported by the respondents have been presented in of heat but on the other hand are exposed to infections Table 3. The results indicate that coat colour is considered transmitted by stinging pests thereby leading to suscepti- an important criteria for selection of cattle especially for bility towards many external parasite borne diseases. their potential to adapt and production capacity. Coat col- Kshatriya, Trivedi and Dhami (2009) in their study con- our as a selection criterion for cattle has also been well cluded that both biometry of udder and skin thickness documented in the folklores of India as reported by were related to milk yield and therefore could be consid- Mohanty (2004). The respondents in this study indicated ered as one of the criteria for the selection of dairy that coat colour is related to the adaptability and influences cows. All the above selection criteria are in accordance the disease tolerance of the livestock, they also reported with the findings of Endashaw et al. (2012) who reported that coat colour also influences the ability of the livestock that the Musi and Bodi pastoralists of Ethiopia select cows to ward away the evil eye, meat of animals of a particular based on their milking ability, coat colour, fertility and coat colour are considered to be tasty and certain coat col- udder size. ours have more market demand than the others. This might The docility and fertility are of course the primary concern be a fallout, as an animal with a particular coat colour has for the rearers as a cow which is docile can be easily better ability to tolerate the environmental variations; these handled by all the family members followed by good fer- findings are in close agreement with the results presented tility ensuring a higher frequency of lactation and also an by Ebozoje and Ikeobi (1998) and Sanusi (2008) who increase in herd size. The observations are in accordance reported better survivability. The observations are in con- with the reports of Raziq and Younas (2007), Raziq, sonance with the findings of Tekleyohannes et al. (2012) Younas and Kakar (2008) and Kakar (2009) from from Ethiopia and Adedeji et al. (2012) from Nigeria. Baluchistan where docility was considered an important Results of a study by Bonsma (1980) indicated that at criteria for the selection of livestock by traditional camel, higher altitudes there are short wave or ultraviolet rays cattle and sheep breeders of tribal areas of Pakistan. The fl which are tolerated well by those with dark coat colour, traits like hind loin which is devoid of esh with promin- like black and brown whereas the cattle with white coat ent hip and pin bone, as was reported by the respondents, colour tend to suffer from skin burns associated with ultra- may be attributed to safe calving and less chances of dys- violet rays. Although in the lowlands the infrared or long tocia, the observations are in accordance with the observa- wavelength rays predominate which is reflected by the ani- tions of Bonsma (1980) and Hammond et al. (1971). mals having white coat colour, similar observations too Result of a study by Peters, Horst and Kleinheisterkamp have been reported by Finch and Western (1977); Olson, (1982) indicated that coat colour pigmentation can assist Lucena and Chase Junior (2003). Studies by Makokha in productive and reproductive ability of various livestock et al. (2006) indicate that animals with dark coat colour species especially in the humid and sub-humid environ- attract the Tse Tse flies and are therefore not preferred in ment. Thin skin of cattle is correlated with good milking the lowland areas such as Lok Abaya and places infested ability especially in the lowland areas, where the climate by stinging insects, which in this case is at Lok Abaya is usually quite hot. This might be attributed to the low woreda. Similarly the results as obtained in the study body fat deposition and the fact that most of the energy finds consonance with the observations of Stewart is utilized for milk production. Studies by Foster, Fourie (1953), Hutchinson and Brown (1969), Finch (1986) and and Neser (2009) indicated that hide thickness is correlated Hansen (1990) who reported that lighter coat colour with rectal temperature, high rectal temperature depresses reflect the sunlight and hence beneficial for cattle reared the let down of milk. Similarly, cattle with thick and in the lowlands . This result goes in agreement with the dense hair coat can exacerbate the effects of heat stress finding of Ndumu et al. (2008) who reported that the farm- (Berry and Shanklin (1961), Finch, Bennett and Holmes ers prefer red colour in Ankole longhorn cattle over other Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia 159

coat colours, Shiferaw (2006) too reported high ranking for with thick bones especially those of scapulae and white colour preference for Kereyu cattle owners, the metatarsus. findings are in consonance with the present study for cattle Results of studies by Hammond et al.,(1971) indicate that selected in the Lok Abaya woreda. bulls with thick shoulders have large capacity in the rib- The influence of coat colour on milk yield has also been cage (where heart and lungs are located) also indicating reported in folklores by Mohanty (2004) who indicated masculinity in bull, thicker and coarser muscles associated that cattle with black coat colour are supposed to be better with thick shoulders, the observations are in accordance in milk yield, the importance of feather colour in selection with the reports of Bonsma (1980). Large and well- of chickens have also been reported by Kassahun (2013) developed scrotum and testes is also a good indicator for from Ethiopia and also by Karma (2005) from Bhutan fertility amongst male livestock, hoarse voice is usually and Banerjee (2012) from India. associated with androgen activity and is an indicator of muscularity as indicated by Hammond, Ian and Terence The hair whorl is also one of the important aspects to be (1971), broad and long ear is a breed character and it is selected according to the respondents. It is also known usually associated with high vascualrity, which may be that a cow with high hair whorl tends to reach puberty earl- related to better dissipation of body heat and thereby better ier and that they also have high total milk and total solid feed efficiency and also grazing ability. yield (Young, McNaughton and Spelman, 2007). Results ’ of a study by Broucek et al. (2007) also indicate that ani- Long tail is usually associated with an animal s ability to fl mals with more hair whorl may grow faster than animals get rid of external parasites especially ies and other sting- with a middle or low hair whorl; which can be correlated ing insects which if the animal is able to overcome can be to the early development of mammary gland. According correlated with good health of the livestock as a whole. to the results of a study made by Bharadwaj, Sastry and The results pertaining to selection of livestock for fattening fl Yadav (2011), coat colour, shape of head and face, tail indicates that those with at spinous process and long switch, shape of udder and teats and skin thickness were femur are better for fattening, a common indicator for well- some of the traits which were associated with milk produc- developed longissimus muscle, and a long and thick femur ’ tion. The short and straight femur as reported by the is an indicator for an animal s ability to walk long dis- respondents (straight thigh) can serve as an indication of tances for fodder and also a thick bone provides large sur- its good grazing ability and endurance while grazing or face area for the attachment of muscles, thus muscle reaching the watering points. The results are in accordance bundles are large and thick under such circumstances, with the observations of Kedija (2007) who reported that the observations are in accordance with the results farmers preferred cattle with long and straight thighs in obtained by Hammond, Ian and Terence (1971) and the Mieso district of Ethiopia. Kedija (2007). As indicated by the selection of bulls based on penile The selection of small ruminants and equine with long legs length, the farmers correlated long penis with good fertility is a criteria to select animals which have endurance in trav- as they indicated that if the bull had a short penis there will elling long distances. Results of studies by Hafez (1968) be problems associated with its ability to breed, the obser- indicate that longer femur and metatarsus are adaptive vations are in accordance with the results obtained by means for livestock reared in hot climates, long legs in Bruce (2012), Bonsma (1980) and Hammond, Ian and one hand help the livestock travel long distances besides Terence (1971). The indication of long penis and fertility they help protect the vital organs against the solar radiation may be associated with the proper function of the cremas- from the hard soil surface. ter muscle which is responsible for the movement of the Fat tail of sheep is a breed character but among the animals penile sheath and therefore frequent urination and libido of a particular breed those having fatter tails are able to tol- are correlated with the high level of androgen in bulls; erate drought better as they are able to have enough Kedija (2007) from his study in Meiso district too reported reserves in the form of fat, even the livestock with fat that the farmers selected bulls with long penis. In addition, deposits have usually better feed conversion efficiency as scrotal circumference is a measure of the daily and total excess of energy in feed is converted to fat deposits. sperm production of a bull. Studies agree that neck con- Broad hoof in horses is associated with its stability as a formation in bulls is very important by which a good pack animal especially in the highlands of Ethiopia bull should have a thick crest over his neck as well as a where the equines are generally employed as pack animals short neck. Studies by Hammond, Ian and Terence and hence strong and broad hoof is associated with stron- (1971) too indicate that short neck in bulls are a good indi- ger planter cushions and tendons and cartilages which are cator of libido and high testosterone levels, as well as lar- in accordance with the observation of Hafez (1968). ger scrotal circumference (which relate to higher levels of hormones) and wider, more muscular shoulders. The selection criteria for small ruminants (sheep and goats) Short-necked bulls tend to sire wide-rumped (more depth as observed in the present study agree with the finding of in the hindquarters) steers and daughters that mature Estifanos (2010) in the Umbulo watcho water shed of early, higher levels of testosterone are also associated Southern Ethiopia, the presence of wattle as a selection 160 S. Banerjee et al.

criteria for goats as observed in the study are in accordance Aujla, K.M., Jasra, A.W. & Munir, M. 1998. Socioeconomic profile of with the findings of Odubote (1994b) who reported that camel herders in South-western mountainous areas of Pakistan. In goats with wattles are docile and thermoregulatory mech- Proc. Third Annual Meeting Animal Products under Arid Conditions, vol 2, pp. 154–174. anism of the goats with wattles are better than those without wattles. Results of a study by Casu, Boyazoglu and Azage, T., Belete, A., Fekadu, B. & Berhanu, E. 2010. Cattle milk and Lauvergne (1970) indicated that ewes with wattles had bet- meat production and marketing systems and opportunities for market orientation in Fogera woreda, Amhara region, Ethiopia. Improving ter fecundity when compared to the non-wattled. Although Productivity and Market Success (IPMS) of Ethiopian Farmers Richard et al.,(1990) reported that there might be a genetic Project, International Livestock Research Institute (ILRI), Addis linkage between the presence of wattles and disease toler- Ababa, Ethiopia, Working Paper No.19. ance with Alphine and Saanen goats. The goats sporting Banerjee, S. 2012. Morphological characterization of indigenous chick- wattles had significantly lower faecal output of strongyle ens of Sikkim and West Bengal, India. Anim. Genet. Res., 51: 57–71. eggs than those without wattles indicating a form of genet- Bavikatte, K. & Jonas, H. 2009. Bio-cultural community protocols: a ic tolerance to helminthoses. The result, as assessed in the community approach to ensuring the integrity of environmental law present study pertaining to the selection of goats having and policy. Natural Justice, United Nations Environment Programme wattles and of a particular coat colour, is also in accord- (available at http://www.unep.org/communityprotocols/PDF/commu- ance with the observation of Adedeji et al. (2012). nityprotocols.pdf). Berry, I.L. & Shanklin, M.D. 1961. Physical factors affecting thermal insulation of livestock hair coats. Missouri Agricultural Experiment Conclusion Station Research Bulletin. 802. Berry, D.P., Harris, B.L. & Winkelman, A.M. 2005. Phenotypic asso- The results from the present study indicate that the respon- ciations between traits other than production and longevity in New – dents in the studied areas practice selection of livestock Zealand dairy cattle. J. Dairy Sci., 88: 2962 2974. based on some predefined physical attributes most of Bharadwaj, A., Sastry, N.S.R. & Yadav, M.S. 2011. Lactation and which have proper scientificjustifications and are best suited mammary system influences on milking behaviour of buffaloes–a – to their agro-climatic conditions. Hence, these criteria can note. Indian J. Anim. Prod. Manage., 3: 129 133. be included in developing selection mechanisms which Bonsma, J.C. 1980. Livestock production: a global approach. Cape can be further coupled with modern selection techniques. Town, South Africa, Tafelberg Publishers. The age old system of livestock selection using phenotypic Broucek, J., Peter Kisac, S., Tefan, M., Anton, H., Michal, U. & markers can be further refined using molecular markers; Vladimir, T. 2007. Hair whorls of holstein friesian heifers and affects hence, it is of primordial importance to properly document on growth and behaviour. Arch. Tierz. Dummerstorf, 50(4): 374–380. the oral traditions from the elders who practice them. Bruce, E.E. 2012. Bull, In Male Theriogenology (available at http:// www.vetmed.lsu.edu/eiltslotus/theriogenology-5361/bull.htm), assessed on 3rd September 2013. Acknowledgements Casu, S., Boyazoglu, J.G. & Lauvergne, J.J. 1970. The inheritance of wattles in the Sardinian breed of sheep. Annals. Genet. Sel. Anim., 2: – fi 249 261. Contribution to animal production and management. The authors acknowledge the nancial assistance received Review and. Scientific Tech. Off. int. from Coordinator NORAD project. The authors also acknowledge the assistance received from the elders and Ebozoje, M.O. & Ikeobi, C.O.N. 1998. Colour variation and reproduction in the West African Dwarf goats (WAD). Small Rumin. Res., 27: traditional livestock rearers of both the woredas, the 125–130. authorities of the agricultural bureau, head and staff members of the School of Animal and Range Sciences, Endashaw, T., Tadelle, D., Aynalem, H., Wudyalew, M. & Okeyo, M. 2012. Husbandry and breeding practices of cattle in Mursi and Bodi Hawassa University. pastoral communities in Southwest Ethiopia. Afr. J. Agric. Res., 7 (45): 5986–5994. Estifanos, T. 2010. Sheep production system and on farm evaluation of References dry season Ethiopia. Hawassa university. p.104 (M.Sc. Thesis). FABRE 2006. Sustainable farm animal breeding and reproduction A Adedeji, T.A., Ozoje, M.O., Ojedapo, L.O., Fasoyin, O.A., & Ogundipe, vision for 2025. FABRE technology platform. February 2006. 30 R.I. 2012. Indirect selection amongst small scale holders of West p. (available at http://www.fabretp.org). African Dwarf (WAD) goats based on coat pigmentation and wattle types in Ogbomoso, Nigeria. Iran. J. Appl. Anim. Sci., 2(1): 21–25. Fekadu, B. 1994. Present situation and future aspect of milk production, milk handling and processing of dairy products in southern Ethiopia. Agere, M. 2008. Assessment of feed resources, feeding system and milk Farm made milk products in southern Ethiopia: 1. Chemical and production in Horro district, western Oromia, Ethiopia. Presented microbial quality. Department of Food Science, Agricultural to the School of Graduate Studies of Hawassa University, Hawassa, University of Norway. (Ph.D. Thesis). Ethiopia (M.Sc. Thesis). Fekadu, F. 2010. Ethiopian medicinal plants in veterinary healthcare. A Alphonsus, C., Akpa, G.N., Oni, O.O., Rekwot, P.I., Barje, P.P. & mini-review. Ethiopian e-J. Res. Innov. Foresight, 2(1) 48–58. Yashim, S.M. 2010. Relationship of linear conformation traits with bodyweight, body condition score and milk yield in Friesian × Finch, V.A. 1986. Body temperature in beef cattle: its control and rele- Bunaji cows. J. Appl. Anim. Res., 38: 97–100. vance to production in the tropics. J. Anim. Sci., 62: 531–542. Some traditional livestock selection criteria as practiced by several indigenous communities of Southern Ethiopia 161

Finch, V.A. & Western, D. 1977. Cattle colours in pastoral herds: natural Kenya. Contributed paper prepared for presentation at the International selection or social preference. Ecology, 58: 1384. Association of Agricultural Economists Conference, August 12–18, Gold Coast, Australia (available at http://ageconsearch.umn.edu/bit- Finch, V.A., Bennett, I.L., & Holmes, C.R. 1984. Coat colour in cattle: stream/25752/1/cp060765.pdf); accessed on, September 5 2013. effect on thermal balance, behaviour and growth, and relationship with coat type. J. Agric. Sci. (Camb.), 102: 141–147. Mathias, E., Ilse, K.-R. & Jacob, W. 2005. Pastoralists, local breeds and fi ’ Foster, L.A., Fourie, P.J. & Neser, F.W.C. 2009. Effect of heat stress on the ght for livestock keepers rights. Prepared for the PENHA 15th “ six beef breeds in the Zastron district: the significance of breed, coat Anniversary Conference Pastoralism in the Horn of Africa: ” colour and coat type. S. Afr. J. Anim. Sci., 39: 224. Surviving against all odds on Thursday, 29 September 2005. London. fi Ghosh, P.K. & Sahoo, B. 2011. Indigenous traditional knowledge. Mbuku, S.M., Isaac, S.K. & Alexander, K.K. 2006. Identi cation sys- Orissa Review. 65–70 (available at http://www.orissa.gov.in/ tems and selection criteria of pastoral goat keepers in northern Kenya – e-magazine/Orissareview/2011/Jan/engpdf/66-71.pdf). implications for a breeding programme. Conference on International Agricultural Research for Development. University of Bonn, October Hafez, E.S.E. (Ed.) 1968. Adaptation of domestic animals, Philadelphia, 11–13: 2006. Lea & Febiger. Mesfin, T. & Obsa, T. 1994. Ethiopian traditional veterinary practices Hamid, M.A., Husain, S.M.I., Khan, M.K.I., Islam, M.N. & Biswas, and their possible contribution to animal production and management. M.A. 2000. Skin thickness in relation to milk production in crossbred Rev. Sci. Tech. Off. Int. Epiz., 13(2): 417–424. cows. Pak. J. Biol Sci., 3(9): 1525–1529. Mishra, D. & Patro, L. 2010. Ethno veterinary practices among the rural Hammond, J. Jr., Ian, L.M. & Terence, J.R. 1971. Hammond’s farm people of Ganjam district Orissa (India) : a case study on some com- animals. U.K., Edward Arnold. mon veterinary ailments. Bioscan., (3): 739–746. Hansen, P.J. 1990. Effects of coat colour on physiological responses to Mohanty, R.B. 2004. Traditional wisdom on livestock selection and man- solar radiation in Holsteins. Vet. Rec., 127: 333. agement in folk proverbs of Orissa. Indian J. Tradit. Knowl., 3(1): 92–95. Hiwot, B. 2013. Assessment of indigenous knowledge in livestock rear- Morse, D., DeLorenzo, M.A., Wilcox, C.J., Collier, R.J., Natzke, R.P. ing practices at Arbrgona and Lokabaya woreda of Sidama Zone of & Bray, D.R. 1988. Climatic effects on occurrence of clinical mas- Southern Ethiopia. Submitted to School of Graduate Studies, – Hawassa University, Ethiopia (M.Sc. thesis). titis. J. Dairy Sci., 71: 848 853. Hutchinson, J.C.D. & Brown, G.B. 1969. Penetrance of cattle coats by Ndumu, D.B., Baumung, R., Hanotte, O., Wurzinger, M., Okeyo, A. radiation’. J. Appl. Physiol., 26: 454–464. M., Jianlin, H., Kibogo, H. & Sölkner, J. 2008. Genetic and morphological characterisation of the ankole longhorn cattle in the Iqubal, M.A. 2013. Livestock husbandry – a sustainable livelihood in African great lakes region.Genet. Select. Evol., 40: 467–490. Ethiopia. Int. J. Econ. Manage. Soc. Sci., 2(8): 603–607. Odubote, I.K. 1994b. Influence of qualitative traits on the performance of – Jena, M. 2007. Community Health Knowledge Register. Tradit., 5: 6 10. West African Dwarf goats. Nigerian J. Anim, Prod., 21: 25–28. Kakar, A.R. 2009. Assessing the potential of the indigenous livestock Olson, T.A., Lucena, C. & Chase Junior, C.C. 2003. Evidence of a breeds of Baluchistan. A Drynet Science and Technology Expertise. major gene influencing hair length and heat tolerance in Bos taurus 59 p. (available at http://www.comap.ca/kmland/display.php? cattle. J.Anim. Sci., 81: 80–90. ID=144&DISPOP=VRCPR) assessed on 3rd Sep 2013. Ouma, E., Abdulai, A. & Drucker, A.G. 2006. Pastoralists preferences Karma, N., Penjor, D.P., Gurung, R., Arasta, P. & Moran, C. 2005. for cattle traits: letting them be heard. pastoralism and poverty reduc- Genetic structure of the indigenous chickens of Bhutan. SAAR tion in East Africa: Nairobi, Kenya, A Policy Research. J. Agric., 3: 69–89. Peters, K.J., Horst, P. & Kleinheisterkamp, H.H. 1982. The import- Karthikayan, S.M.K. & Gajendran, K. 2005. Traditional technologies ance of coat colour and coat type as indicator of productive adaptabil- in the improvement of breeds of livestock in Tamil Nadu. Indian J ity of beef cattle in a sub-tropical environment. Prod. under Arid Tradit. knowl., 4(3): 303–306. Conditions., 2: 154–174. Kassahun, H. 2013. Evaluating poultry production system and egg qual- Raziq, A. & Younas, M. 2007. Socioeconomic profile of camel in ity parameters of village chickens at Hawassa zuria woreda of southern Ethiopia. Hawassa University, Ethiopia. (M.Sc. Thesis). Suleiman mountainous region of Balochistan, Pakistan: in Recent trends in camelids research and future strategies for saving camels. Kedija, H. 2007. Characterization of milk production system and oppor- In: Proc. International Camel Conference Rajasthan, India, February tunity for market orientation: a case study of Mieso district, Oromia 16–17, pp. 123–128. region, Ethiopia. Haramaya University. (M.Sc. Thesis). Raziq, A., Younas, M. & Kakar, M.A. 2008. Camel – a potential dairy Kshatriya, P.S., Trivedi, M.M. & Dhami, A.J. 2009. Association of animal in difficult region of Balochistan, Pakistan: in recent trends in udder biometry and skin thickness with milk yield in Kankrej and camelids research and future review Ethiopian e-journal for research – Crossbred cows. Indian J. Field Vet., 5(1): 11 13. and innovation foresight. Health Issue, 2(1) 48. Kuczaj, M., 2003. Analysis of changes in udder size of high yielding Richard, S., Cabaret, J. & Cabourg, C. 1990. Genetic and environmen- cows in subsequent lactations with regard to mastitis. tal factors associated with nematode infection of dairy goats in north Electron. J. Pol. Agric., 6(1), #02. (available at http://www.ejpau. western France. Vet. Parasit., 36(3–4): 237–243. media.pl/volume6/issue1/animal/art-02.html). Sanusi, A.O. 2008. Effects of coat colour genes on heat stress and toler- Kunene, N.W. & Fossey, A. 2006. A survey on livestock production in ance to Haemonchus contortus among West African dwarf sheep. some traditional areas of Northern Kwazulu Natal in South Africa. University of Agriculture, Abeo (M.Sc. Thesis). Livestock Research for Rural Development. Volume 18, Article #113. Retrieved September 3, 2013, (available at http://www.lrrd. Shiferaw, G. 2006. In-situ phenotypic characterization of Kereyu cattle org/lrrd18/8/kune18113.htm). type in Fentalle district of Oromia region, Ethiopia. M.Sc. Thesis. Presented to the School of Graduate Studies of Haramaya Makokha, S.N., Karugiab, J., Staalc, S., & Kosura, O. 2006. Valuation of University, Haramaya, Ethiopia. cow attributes by conjoint analysis: a case study in Western 162 S. Banerjee et al.

Shigdaf, M. 2011. Performance evaluation of Washera, Farta and their Winchester, A.M. 1977. Heredity: an introduction to genetics. New crossbred sheep in the Western highland of Amhara Region, Delhi, India, Oxford & IBH. Ethiopia. Bahir Dar University, Ethiopia (MSc. Thesis). Wurzinger, M., Ndumu, D., Baumung, R., Drucker, A., Okeyo, A.M., Singh, G. & Kumar, J. 2012. Traditional knowledge on some less Semambo, D.K., Byamungu, N. & Sölkner, J. 2006. Comparison of known wild edible plants used among Munda tribe of Jharkhand. production systems and selection criteria of Ankole cattle by breeders Ecoscan., 6(3&4): 153–155. in Burundi, Rwanda, Tanzania and Uganda. Trop. Anim. Health Prod., 38: 571–581. Stewart, R.E. 1953. Absorption of solar radiation by the hair of cattle. Agric. Eng., 34: 235. Yakabu, A. 2011. Path analysis of conformation traits and milk yield of Bunaji cows in small holder herds in Nigeria. Agricultura Tropica et Tekleyohannes, B., Jamroen, T., Sayan, T., Girma, A., Asrat, T. & Subtropica., 44(3): 152–157. Somkiert, P. 2012. Purposes of keeping goats, breed preferences and selection criteria in pastoral and agro-pastoral districts of South Yamazaki, T.H., Takeda, A., Nishiura, & Togashi, K. 2009. Omo Zone. Livestock Research for Rural Development. Volume 24, Relationship between lactation curve and udder disease incidence in Article #213. Retrieved September 5, 2013, (available at http://www. different lactation stages in ﬁrst lactation holstein cows. Anim. Sci. lrrd.org/lrrd24/12/berh24213.htm). J., 80: 636–643. Tilki, M., Inal, S., Colak, M. & Garıp, M. 2005. Relationships between Young, S.K., McNaughton, L.R. & Spelman, R.J. 2007. Hair whorl ˙ milk yield and udder measurements in brown swiss cows. Türk patterns are related to age at puberty and milk-production traits in Veterinerlik ve Hayvancilik Dergisi. 29: 75–81. dairy cattle. Proc. N. Z. Soc. Anim. Prod., 67: 130–135. Animal Genetic Resources, 2014, 54, 163–170. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000034

Phenotypic characterization and description of production systems of autochthonous sheep breeds in Kosovo

H. Bytyqi1, R. Baumung2, H. Mehmeti1 and B. Fuerst-Waltl2 1University of Prishtina – Faculty of Agriculture and Veterinary, Prishtina, Kosovo; 2Department of Sustainable Agricultural Systems, Division Livestock Sciences, University of Natural Resources and Life Sciences (BOKU), Vienna, Austria

Summary Considering about 13.9 percent of total surface under pastures and meadows and big demand for milk and meat, sheep are considered as an important livestock species in Kosovo. Hence, the overall objective of this study was to provide the characterization of autochthonous sheep breeds and sheep production systems in Kosovo. Breed characterization and proper economic management schemes, selection for most economically important traits and best animals in the flock will assist farmers in increasing the feeding efficiency, reproduction and productive traits and thus profit. Including Bardhoka (BAR), Balusha (BAL), Sharri (SHA) and Kosva (KOS) sheep breeds, for a period of 1 year (September 2009–August 2010) the study was concentrated in 20 sheep farms in different regions of Kosovo. The least-squares means show that for milk yield and milk content, breed differences were significant (P < 0.05). The highest average daily milk yield for BAR (0.63 kg) exceeded the milk yield of BAL, KOS and SHA with 0.09, 0.14 and 0.18 kg milk per day, respectively. For milk butterfat and protein SHA ewes obtained the highest rate compared with other three breeds for 9.4–19.9 percent. Milk dry matter among different breeds ranged between 18.4 and 19.3 percent. Voluntary disposal of ewes mainly occurred at lactation 4 and 5. Considering a standard yield at 1st lactation specific to the breeds, the highest production is reached at 3rd lactation.

Keywords: breed characterization, economic traits, sheep breed, least-squares mean

Résumé Compte tenu du fait que près du 13.9 pour cent de la superficie totale correspond à des pâturages et des prairies et qu’il existe une grande demande en lait et viande, les moutons constituent une espèce animale importante au Kosovo. Ainsi l’objectif principal de cette étude a été de caractériser les races ovines autochtones et les systèmes de production ovins au Kosovo. La caractérisation des races, la description des schémas de gestion économique les plus pertinents et la sélection des paramètres les plus importants sur le plan économique et des meilleurs animaux du troupeau serviront aux éleveurs pour augmenter l’efficience alimentaire, les performances reproductives et productives, et par là même les bénéfices. L’étude, qui a duré un an (de septembre 2009 à août 2010) et a été menée sur 20 exploitations ovines de différentes régions du Kosovo, a compris les races ovines Bardhoka (BAR), Balusha (BAL), Sharri (SHA) et Kosva (KOS). Les moyennes par moindres carrés ont décelé des différences significatives (P < 0.05) entre les races pour la production laitière et la composition du lait. La production moyenne journalière de lait la plus élevée de la race BAR (0.63 kg) a dépassé de 0.09, 0.14 et 0.18 kg journaliers de lait la production laitière des races BAL, KOS et SHA, respectivement. Pour ce qui est des taux butyreux et protéique du lait, les brebis SHA ont atteint les pourcentages les plus élevés, qui ont dépassé de 9.4 à 19.9 pour cent les trois autres races. La matière sèche du lait a varié entre 18.4 et 19.3 pour cent selon la race. La plupart des brebis sont retirées du troupeau à la quatrième ou cinquième lactation. Si la première lactation est considérée comme étant la production standard propre aux races, la production la plus élevée est atteinte à la troisième lactation.

Mots-clés: race ovine, caractérisation raciale, moyenne par moindres carrés, paramètres économiques

Resumen Teniendo en cuenta alrededor de un 13.9 percent de la superficie total bajo pasturas y praderas y gran demanda de leche y carne, ovino son considerados como una importante especie de ganado en Kosovo. Por lo tanto, el objetivo general de este estudio es proporcionar a la caracterización de las razas ovinas autóctonas y sistemas de producción ovina en Kosovo. Sistemas de gestión económica adecuada caracterización de razas y la selección de rasgos económicamente más importantes y sus mejores animales en el rebaño serán ayudar a los agricultores a aumentar la eficiencia de la alimentación, la reproducción y las características productivas y por lo tanto los benefi- cios. Incluyendo Bardhoka (BAR), Balusha (BAL), Sharri (SHA) y Kosva (KOS) razas ovinas, por un período de un año (09 2009 hasta agosto 2010), el estudio se concentró en 20 granjas de ovejas en diferentes regiones de Kosovo. Los medios mínimos cuadrados muestran que para la producción de leche y el contenido de la leche, las diferencias entre razas fueron significativas (P < 0.05). La producción diaria de leche promedio más alto para BAR (0.63 kg) superó la producción de leche de BAL, KOS y SHA, con 0.09, 0.14 y 0.18 kg de leche por día, respectivamente. Para la grasa de mantequilla leche y proteína de ovejas SHA obtener la tasa más

Correspondence to: H. Bytyqi, Faculty of Agriculture and Veterinary, Animal Science Department, University of Prishtina, Rr “Lidhja e Pejes” 10000, Prishtina, Kosovo. email: [email protected]

163 164 H. Bytyqi et al.

alta en comparación con otras tres razas de 9.4 a 19.9 percent. Leche de materia seca entre diferentes razas varió entre 18.4 y 19.3 percent. Entrega voluntaria de las ovejas se produjo principalmente en la lactancia 4 y 5. Considerando un rendimiento estándar en primera lactación especíﬁca a la cría, la producción más alta se alcanza a tercera lactancia.

Palabras clave: Raza ovina, breed caracterización, por mínimos cuadrados medios, los rasgos económicos

Submitted 2 October 2013; accepted 7 January 2014

Introduction production difficult. Owing to small size of land owned Maintaining local sheep breeds is a complex problem. It by the sheep farmers, about 1.5 ha land per farm integrates economic, social, environmental and technical (MAFRD, 2012), cultivation of such feedstuffs is often – aspects, and involves many different actors. These actors fragmented in 6 10 plots and located at large walking dis- are the direct users, such as farmers and consumers and tances from the house. Following their normal seasonal other indirect users to benefit, such as tourism industry, state of lambing, sheep in Kosovo are seasonally bred extension services, etc. Domestic animals make a major (summer) and offspring are born at the time of contribution to human requirements for food in the form December and January, mostly. So far, limited attempts of meat, milk, milk products, eggs, fibre and fertilizer for were made to change this breeding practice in Kosovo. crops as well as draught power. Approximately 4 500 There is an obvious lack of education and awareness con- breeds drawn from 40 or more animal species are listed cerning sheep breeding and proper management despite in this context (FAO, 2007). Nowadays, it is well known the efforts made by local authorities and some of donor (Dillon et al., 2003; Hayes et al., 2003; Bytyqi et al., organizations in this regard. Farmers need to be trained 2005) that animals best adapted to their environment, to be familiar with breeding schemes specific to sheep show the best overall trait performance (i.e. survival, pro- farms in Kosovo in order to be qualified to manage their duction, reproduction, etc.). However, even though flocks and, in particular, to distinguish productive from Kosovo is a rather small area, most breeds of sheep that non-productive animals. originated in the different regions perform very differently There are not yet, proper breeding and selection schemes and have gone through diverse selection focuses. Breed that will assist farmers in increasing the feeding efficiency, characterization, including selection and breed perform- reproduction and productive traits. Thus, their breeding ance reactions to environment should thus be a part of strategy was focused clearly to the “methodical selection” all management and breeding decisions in sheep produc- criteria. All sheep breeds in Kosovo are classified in the tion enterprises. long tail group and share triple purpose breed character- Kosovo has a long tradition in sheep farming, but like istics, milk–meat–wool (Bytyqi and Mehmeti, 2006). other sectors of Kosovo economy it has experienced a per- Milk production is in the focus of all sheep breeds in iod of decline. Sheep production reached a peak during the Kosovo but performance testing does not exist yet. Thus, beginning of the 1960s (644 000 sheep). A dramatic a survey including detailed milk performance testing was decrease started after 1996 and accelerated during the carried out in eight sheep farms to describe the production war declining to a minimum of less than 90.000 head, in level of ewes. Additionally, to receive a first insight into the year 1999. Currently, the number of sheep has started growth and non-production traits, 20 sheep farms (includ- to grow again and it is estimated to be of about 127 000 ing the eight farms with milk recording) were integrated in animals (Ministry of Agriculture Forestry and Rural an investigation regarding various other traits in sheep Development-MAFRD, 2012). flocks. Hence, the overall objective of this study was to provide the necessary information regarding characteriza- Normally, the sheep are kept in the barn from the end of tion of autochthonous of sheep breeds and sheep produc- November until the end of March (winter period) inde- tion systems in Kosovo. pendent of breed. The grazing period for sheep normally starts in April, close to farms. Mountain grazing time takes place from end of May until September, where sheep production is solely Material and methods dependent on grass feed (summer period). After that, Sheep production and management sheep graze on low land close to the barns again. Concentrated feed is given to the lactating sheep only dur- characteristics in Kosovo ing the winter period at the amount of 0.20–0.50 kg per After cattle, the sheep production is the most important ewe. Small diversity of feed sources, poorly constructed segment of the economy in rural households in Kosovo. barns, and low level of hygiene make intensive sheep The sheep production includes various breeds and Production systems of autochthonous sheep breeds in Kosovo 165

categories of sheep. To date, sheep production in Kosovo can be clearly deﬁned as an industry consisting of small and large farms ranging from 40 to 1 000 ewes producing for home consumption and the market. The sheep production of Kosovo has many characteristics typical for sheep production in developing countries, as it depends almost exclusively on the resources locally available on the farm.

Production characteristics Most of the quantity of sheep’s milk produced in the Kosovo is turned into cheese, and is rarely consumed as fresh milk. Therefore, the quality of the sheep milk should refer to its ability to be transformed into milk products. Milk is frequently being processed following the traditional way of milking ewes only after the weaning period, which occurs about 90 days after lambing. Very little investments were made in regard to milking and processing infrastructure. Lambs and culling ewes are usually sold as a live animals and no specific commodity has so far been developed to increase added value of sheep meat. There is almost no farm that has any connection to milk or/and processing unit or market chain. The milk is being processed at farm gate and sold locally (green mar- Figure 1. Living area for different sheep breeds in Kosovo. ket), just a small proportion of farmers sell their products to retailer shops or gross shop houses. The usual price for cheese ranges from 5 to 6 €/kg. Almost, there is no 42.5 μm (Mehmeti et al., 2007). The males are horned sheep meat process habit among farmers in Kosovo. and the females are polled, usually. This strain is well sui- Mostly, lambs (a part of those kept for breeding) are ted to a grazing production system, calm, easily milked sold at an age of about 150 days, at spring time before and easy to work with. the mountain-grazing period. The price ranges from 3 to 4 €/kg live weight. Mature animals are being sold at no specific period at lower prices about 2.5 €/kg live weight. BAL sheep Minimum attempts were made to promote sheep products The area in which these sheep are kept and bred is also in as a tool to increase profitability and production. the southwestern part of Kosovo sharing the same region with BAR breed, named Dukagjini plane (Figure 1). The name originates from the word (Albanian language) History of breeds of sheep in Kosovo “Bale-Spot”. The total number of this breed is about 9 The sheep population in Kosovo mainly consists of triple- 900 animals bred in Kosovo, and covers about 7.8 percent purpose breeds of widely different body sizes and produc- of the total sheep population (Table 1). As in the case of tion capacities, i.e. Bardhoka (BAR), Balusha (BAL), BAR breed, BAL sheep have also been going through Sharri (SHA) and Kosva (KOS). breeding selection with special focus on milk production. This strain is well adapted to natural grazing abilities and easy milking. BAL is bred as triple purpose sheep BAR sheep For centuries this sheep breed has been bred in the Southwest of Kosovo, mostly (Figures 1 and 2). The name originates from the word (Albanian language) “Bardhe-White”, as the fleece colour of this sheep is com- pletely white. These sheep have a share of 19.6 percent of the total number of sheep bred in Kosovo (Table 1). The BAR sheep are kept for triple purpose, milk–meat–wool. However, through natural and selective breeding this sheep breed became one of the most yielding sheep for milk among the long tails sheep in the Balkan Peninsula (Cinkulov et al., 2008; Bytyqi, 2009). On average, the height at withers for mature animals is 64 cm. Although the grease fleece weight ranges between 2.5and 3.5 kg, it is of very poor quality with an average fibre diameter of Figure 2. Bradhoka sheep breed. 166 H. Bytyqi et al.

Table 1. Breed structure, number of sheep per breed and breed proportion.

Sheep breed Number per breed Proportion (%)

SHA 72.628 57.37 BAR 24.780 19.57 KOS 10.649 8.41 BAL 9.894 7.82 Others 8.641 6.83 Total 126.592 100.00

Source of data: Identiﬁcation and registration unit of Kosovo 2011.

(milk–meat–wool) of the medium size. The height at Figure 4. SHA sheep breed. whithers for mature animals is on average about 65 cm. BAL sheep have a black head (Figure 3), whereas the col- total population of sheep this breed represents about 8.4 fl fl our of the eece and the legs are white. The grease eece percent (Table 1). This strain has advantages of natural – – weight for females and males averages 2 2.5 and 3 3.5 hardiness and grazing abilities. The Kosovo sheep is char- fi kg, respectively. The quality of the bre is coarse with acterized with its triple milk–meat–wool, purpose. The tail fi an average bre diameter of 43 μm. Usually, the males is long; the colour of the head and the legs is black, while are horned and the females are polled (Bytyqi, 2009). the fleece is of white colour (Figure 5). Some individuals can also be found having black or grey spotted faces. SHA sheep The grease fleece weight for males and females averages This strain is well adapted to natural hardiness, grazing 1.5–2.5 and 2.5–3.5 kg, respectively. The mean fibre diam- abilities, and is well adapted to the cold mountain climates. eter is 37 μm (Bytyqi and Mehmeti, 2006). In accordance The breeding region for these sheep is mainly in the Sharri to the other sheep breeds of Kosovo males are normally Mountain, in the western part of Kosovo, from where the horned while females are polled. The mean height at with- name originates (Figure 1). This breed represents approxi- ers for mature animals is about 62 cm. mately 53.4 percent of the total sheep population (Table 1). Other sheep represent about 6.8 percent of the total popu- SHA sheep are primarily kept for milk production, although lation, mainly composed by unknown crosses. quality of carcass and wool gives this strain a triple purpose status. Regarding body size, SHA is one of the smallest breeds in Kosovo. The mean height at withers for mature Data recording in selected herds animals is about 62 cm. SHA sheep have a white colour – fl For a period of 1 year (September 2009 August 2010), of the eece,thefaceandthelegs(Figure 4). The grease eight sheep farms in different regions of Kosovo were fleece weight for females and males averages 1.5–2and – fi included in a study regarding dairy performance. Each 2.5 3 kg, respectively. The mean bre diameter is 37 μm sheep breed (SHA, KOS, BAR and BAL), was represented (Bytyqi and Mehmeti, 2006). The males are horned and by two farms, respectively. The average herd size was 114, the females are polled, usually. 94, 92 and 143 animals for SHA, KOS, BAR and BAL breeds, respectively. The milking sheep in the flock were KOS sheep selected randomly, resulting in 1 010 milk test day yield The name of this breed comes from the Kosovo plane, (TDY) samples, which were taken during evening and where it has its area of distribution (Figure 1). From the morning milking. An amount of 40–50 ml milk was put

Figure 3. BAL sheep breed. Figure 5. Kosova sheep breed. Production systems of autochthonous sheep breeds in Kosovo 167

in sterile preserved bottles (Azidol), stored in mobile cool- Results ers at a temperature of 4 °C and transferred to the laboratory for analyses of milk contents. Recording cards In Table 2, the results of the descriptive analysis for the containing information (i.e. animal ID, lambing date, four Kosovar sheep breeds in 20 different farms and differ- lamb weight at different stages, test day milk yield, still- ent disposal reasons are shown. The proportion of ewes in birth, lambing characteristics, etc.) were developed and different lactations depended on the percentage of culling distributed in order to enable farmers to record further for infertility, for involuntary or voluntary reasons. With traits. Two types of scales (from 1 to 30 kg and 30 to small variations among different breeds, the ewes’ volun- 100 kg) were used to record lamb live weight at different tary culling mainly occurred at 4th, 5th and 6th lactation, stages (birth, 3, 6 and 12 months of age). The milk yield ranging from 4.6 to 26.0 percent. Culling for involuntary was measured individually by manual milking. For accur- reasons occurred in 1st, 2nd and 3rd lactation, respectively ate performance recording, farmers were trained and mon- and ranged from 2.4 to 7.8 percent. Culling for infertility itored monthly. tends to be constant across all lactations, ranging from In order to receive more information for relevant traits (e.g. 0.1 to 2.0 percent. Most of ewes in lactation >6 were dis- posed for voluntary reasons. milk production, lamb growth rate, lambing characteristics, fertility traits, survival rate, mating type, feeding, price, Arithmetic means for milk yield and their standard devia- costs, culling characteristics, etc.), farmers of 20 sheep tions for lactations are listed in Table 3. The highest yield farms (including the eight previously mentioned farms) in the first lactation was recorded for BAR, followed by were interviewed. Results were processed and stored BAL, SHA and KOS breeds, yielding 0.62, 0.54, 0.42 electronically. and 0.41 kg/day milk. For all sheep breeds, highest milk yields were registered for the 3rd lactation with 0.72, 0.65, 0.51 and 0.49 kg/day for BAR, BAL, SHA and Statistical data analysis KOS sheep, respectively. Comparing milk yield between All results referring to daily milk yield and component ewes in the 3rd and those in the 5th lactation, a decrease traits refer to actually recorded milk samples at the eight of about 22.9, 33.8, 33.3 and 20.6 percent for BAR, farms. All other averages and proportions represent results BAL, SHA and KOS breed becomes obvious. The of interviews made in all 20 farms. least-squares means results (not shown) for lactation number differences, were significant (P = 0.0103) For milk and milk composition the JMP-starter business unit program of SAS, (Institute SAS Inc. Sall, Lehman Based on milk yield results in eight farms and the inter- and Creighton, 2004) was used to analyse the data for views in 20 sheep farms, results of a descriptive analysis the least-squares means and significant variances. are stated in Table 4. Lactation production length ranged Variations between different variables were tested using from 190 to 280 for different breeds. Owing to longer pro- the Duncan test (Steel and Torrie, 1980). The same ewes duction, BAR and BAL ewes produce more milk in their were sampled for milk production during the whole year. 1st lactation (174.0 and 135.0 kg) compared with KOS Owing to small number of farms included, the only and SHA (82 and 80 kg, respectively). The age at first breed and lactation effects were included in the milk lambing and minimum days open were about the same model. in all breeds.

Table 2. Proportions for different disposal reasons in the four sheep breeds of Kosovo according to interviews in 20 Kosovar farms.

Breed Reason for disposal Proportion in lactations (%)

123456>6Total

BAR Voluntary culling 2.1 0.5 4.3 10.7 18.9 16.2 8.1 60.8 Involuntary culling 7.0 7.8 6.4 5.4 8.1 2.7 0.0 37.4 Infertility 1.1 1.5 1.1 0.1 0.1 0.0 0.0 3.9 Total culling 10.8 9.8 11.8 16.2 27.1 18.9 8.1 100.0 BAL Voluntary culling 1.5 1.1 2.2 11.3 17.1 9.1 2.2 54.5 Involuntary culling 2.7 2.4 3.3 6.3 16.2 6.1 0.0 37.0 Infertility 0.4 0.1 0.8 2.2 5.0 0.0 0.0 8.5 Total culling 4.6 3.6 6.3 19.8 49.0 15.2 2.2 100.0 SHA Voluntary culling 1.8 1.3 1.6 4.6 14.0 26.0 7.3 56.6 Involuntary culling 3.8 4.5 4.1 5.7 7.7 8.9 0.6 35.3 Infertility 1.0 0.6 0.7 1.2 0.3 1.1 0.0 4.9 Total culling 6.6 6.4 6.4 11.5 25.1 36.0 7.9 100.0 KOS Voluntary culling 1.7 2.3 2.8 10.6 22.3 16.8 3.0 60.1 Involuntary culling 5.7 5.7 6.0 5.7 7.7 7.0 0.0 37.8 Infertility 0.6 0.4 0.1 0.4 0.6 0.0 0.0 2.1 Total culling 8.1 8.5 8.7 16.9 30.6 23.8 3.0 100.0 168 H. Bytyqi et al.

Table 3. The average daily milk yield and its standard deviation (SD; in brackets) per lactation in the breeds.

Breed Average milk yield (SD) in kg/day in lactations

123456>6

BAR 0.62 (0.01) 0.68 (0.02) 0.72 (0.02) 0.61 (0.01) 0.51 (0.01) 0.37 (0.02) 0.37 (0.02) BAL 0.54 (0.02) 0.61 (0.02) 0.65 (0.03) 0.52 (0.03) 0.43 (0.02) 0.36 (0.01) 0.33 (0.01) SHA 0.42 (0.04) 0.47 (0.05) 0.51 (0.05) 0.47 (0.04) 0.34 (0.04) 0.29 (0.04) 0.29 (0.04) KOS 0.41 (0.02) 0.43 (0.04) 0.49 (0.02) 0.43 (0.04) 0.34 (0.02) 0.31 (0.02) 0.31 (0.02)

Depending on breed, stillbirth rate was reported to range BAL and BAR with 6.23, 6.05 and 5.51 percent, between 1.8 and 2.5 percent, whereas survival rate respectively. between born live and mating for breeding lambs was The SHA sheep also had significantly higher protein con- about 94 percent. tent in milk compared with other three breeds (P < 0.05; The proportion of male lambs kept for breeding for differ- Table 5). SHA ewes had the highest protein rate on milk ent breeds ranged between 6.8 and 20.0 percent. (5.49 percent), followed by BAL (5.29 percent), BAR Proportion of single/twin/multiples born was for BAR (5.07) and KOS sheep (4.54 percent). (78.5/21.0/0.5 percent), BAL (77.0/20.0/3.0 percent), For milk dry matter content breed differences were not SHA (84.5/15.0/0.5 percent) and KOS (82.6/16.4/1.0 per- significant (Table 5). However, BAL sheep produced cent). Survival rate of ewes from the 1st to 2nd lactation slightly more (5.02, 4.92 and 3.88 percent) dry matter was higher than 93.7 percent. Birth weight of lambs varies milk content compared with the other three breeds. depending on breeds, type of lambing (single/tweens/triple) and gender, and ranges from 2.5 to 4.2 kg. For all breeds, the body weight is considered to be medium. The mature female individual on average weighs 65, 73, Discussion 67 and 65 kg for BAR, BAL, SHA and KOS, respectively. Even though there have been some movements in improv- The ram mature live weights are 110.0, 113.0, 107.0 and ing overall management in recent years, sheep farms are 95.0 kg. managed mostly traditionally and new technologies take In Table 5, the least-squares means with corresponding place at very minimum in the sheep industry in Kosovo. standard errors and level of significance on test-day milk It should be pointed out that among all four breeds, defini- yield records are given. For milk production, breed differ- tion of clear breeding objectives will be a need for the near ences were significant (P = 0.0033; Table 5). The highest future. average daily milk for BAR, exceeded the milk yield of Ewes can survive in the flock till the seventh lactation at BAL, KOS and SHA with 0.09, 0.14 and 0.18 kg milk maximum. This is a rather empiric selection method, per day, respectively (Table 5). which might be associated with a decrease in milk produc- Breed showed a significant effect on butterfat content in tion and incapability of sheep to graze well (long walking milk (P < 0.05; Table 5), with the highest percentage distance) in their later lactations. Therefore, due to trad- obtained by SHA ewes (6.68 percent), followed by KOS, ition, sheep farmers in Kosovo prefer to cull their sheep

Table 4. Descriptive statistics per trait in the breeds based on interviews in 20 Kosovar farms.

Trait (unit) Breeds

BAR BAL SHA KOS

Min. lactation length (day) 280 250 190 200 Age at ﬁrst lambing (day) 365 365 365 365 Min. days open (day) 85 95 120 120 1st lactation yield (kg/lact.) 174.0 135.0 80.0 82.0 Stillbirth rate (%) 2.5 2.3 2.2 1.8 Single/twin/multiples (%) 78.5/21.0/0.5 77.0/20.0/3.0 84.5/15.0/0.5 82.6/16.4/1.0 Male lambs for breeding (%) 15.0 20.0 9.0 6.8 Fattening period (day) 150 150 150 150 Male birth weight single/twin/multiples (kg) 4.1/3.5/2.6 3.8/3.3/2.7 4.2/4.0/3.1 4.1/3.6/2.5 Female birth weight single/twin/multiples (kg) 3.9/3.3/2.5 3.4/3.1/2.5 4.1/3.6/2.8 3.9/3.5/2.5 Ewe mature weight (kg) 65.0 73.0 67.0 65.0 Ram mature weight 110.0 113.0 107.0 95.0 Survival rate between born live and mating (%) 93.0 94.0 94.0 94.0 Survival rate ewes from 1st to 2nd lactation 93.7 96.0 95.2 93.7 Production systems of autochthonous sheep breeds in Kosovo 169

Table 5. Number of observations and least-squares means ± SD for breeds and the traits daily milk yield, as well as fat, protein and dry matter percentage (within a line different letters indicate signiﬁcant differences, P < 0.05).

Traits Breed (N )

BAR BAL SHA KOS Pr > F

Milk yield N 258 233 278 241 Milk yield (kg/day) 0.63 ± 0.04a 0.54 ± 0.05ab 0.45 ± 0.04b 0.49 ± 0.05b 0.0033 Milk contents N 252 227 268 233 Fat (%) 5.51 ± 0.23b 6.05 ± 0.22ab 6.88 ± 0.20a 6.23 ± 0.24ab 0.0375 Protein (%) 5.07 ± 0.39ab 5.29 ± 0.36a 5.49 ± 0.33a 4.54 ± 0.41b 0.0481 Dry matter (%) 18.36 ± 0.73a 19.33 ± 0.68a 18.60 ± 0.66a 18.38 ± 0.71a 0.2760 voluntarily rather at higher ages than in other, more inten- Matutinovic et al., 2011). As Morand-Fehr et al. (2007) sively kept sheep populations (e.g. Fuerst-Waltl and pointed out, sheep production systems based on grazing Baumung, 2009). strongly affect milk composition and thus in turn the cheese produced. While an increased milk yield may be At present, no artificial insemination is used in Kosovo achievable by intensive farming, the “typical” cheese prod- sheep breeding. As a consequence, the breeding ram is uct may not be producible. Hence, farmers should aim at supposed to breed about 28 ewes per breeding period. sufficient milk yield while also including milk content However, due to small populations per breed and narrow traits in their breeding goals. Apart from the genetics, man- breeding replacement stock cycle (exchange of rams agement measures should be taken in order to guarantee between small number of flocks usually grazing together certain standards. for many years), inbreeding could be one area of concern. For some traits, i.e. lactation length, yield per lactation, there Compared with the milk yield at the 1st lactation, the high- were huge variations between breeds living in Dukagjini est production is assumed to be reached at the 3rd valley (BAR and BAL) compared with SHA and KOS lactation. breeds managed under other regions in Kosovo. Although Ewes of BAR, BAL, SHA and KOS breed tend to produce not far living distance from one region to another for all more milk in the 3rd lactation compared with the 1st for four breeds, according to Figure 1, it seems that a better about 16, 20, 21 and 5 percent, respectively. The decrease environmental conditions (more fertile soil, irrigation of in average daily milk yield from the 3rd to 4th lactation land) in Dukagjini valley has led BAR and BAL breeds to ranged from 7.9 percent for SHA to 20.0 percent for gradual milk yield improvement, for centuries. BAL sheep. Consistently, for all breeds lactation further Apart from dairy production traits, farmers were also ques- milk yield decreases occur after the 4th lactation. In tioned with regard to functional traits in their herds. In some dairy breeds, an increase of lactation yields was generally, stillbirth rate was reported to be low having a reported until the 4th or 5th lactation (e.g. Ploumi, positive economic impact on sheep farms in Kosovo. As Belibasaki and Triantaphyllidis, 1998; Ploumi and to other breeds of Pramenka (long tail) sheep in Balkan Emmanouilidis, 1999; Fuerst-Waltl und Baumung, 2009). Peninsula (Porcu et al., 2006), the twinning rate tends to No information was available on individuals regarding the be very low for all Kosovo breeds as well, therefore breed- milk samples, therefore was not able to include the per- ing improvements might take place in this trait, consider- manent environmental effect or a stage of lactation effect ing in details factors that affect this trait (inbreeding, for daily milk yield. feeding, etc.). The milk production data show that BAR and SHA ewes Some other traits (i.e. days open, seasonal lambing, lamb- fattening period, proportion of male lambs sold for produce more persistently all over the lactations compared with other two breeds (BAL and KOS sheep). Therefore, breeding, etc.) are set mainly due to sheep traditional man- breeding programme of selecting individuals for a better agement characteristics in Kosovo. The lambs for meat consumption are sold mostly as live animals and except persistent milk production in sheep could be one alternative of increasing profit of farmers of the latter breeds. a few visual indications and observations, no meat grading Besides, milk content traits are important under Kosovar rate is taking place in Kosovo. conditions as hardly any sheep milk but mainly cheese is consumed. The BAR breed was observed to produce the highest milk and SHA breed for protein and fat content. Conclusions However, it is still below other breeds in the Balkan region. For Dalmatian Pramenka, fat and protein contents This is first approach used to get a first insight into the of higher than 7 and 6 percent were reported (e.g. actual breed characterization and management situation 170 H. Bytyqi et al.

of sheep farms in Kosovo. Sheep production in Kosovo of the West Balkan Pramenka sheep types as revealed by micro sate- seems to be characterized by being predominately semi- lite and mitochondrial DNA analysis. J. Anim. Breed. Genet., 125(6): – extensive. In this environment, focus in breed selection 417 426. should not only be only on increased production, but Dillon, P., Buckley, F., O’Connor, P., Hegarty, D. & Rath, M. 2003. A also on how well the breeds are fitting to the local environ- comparison of different cow breeds on a seasonal grass-based system of milk production. 1. Milk production, live weight, body condition ment. The establishment of clear breeding objectives and a score and DM intake. Livest. Prod. Sci., 83: 21–33. recording system to enable breed characterization has been ’ a challenge for all sheep breeders in Kosovo. The present FAO. 2007. The state of the World s Animal Genetic Resources for Food and Agriculture – in brief, edited by Daffyd Pilling & Barbara study has considered describing a sheep-breeding system Rischkowski. Rome (accessible at http://www.fao.org/docrep/010/ that meets the current farmer demand and is a great contri- a1250e/a1250e00.htm). bution for development of the sheep sector and rural econ- Fuerst-Waltl, B. & Baumung, R. 2009. Economic values for per- omy as a whole. A further research for each trait is required formance and functional traits in dairy sheep. Ital. J. Anim. Sci.,8: in order to get a maximum profit from different sheep 341–357. breeds and breeding systems in Kosovo. Hayes, B.J., Carrick, M., Bowman, P. & Goddard, E.M. 2003. Genotype The sheep farmers showed detailed knowledge of the and environmental interaction for milk production of daughters actual situation, therefore the prerequisites for collecting of Australian dairy sires from test-day records. J. Dairy Sci., 86: 3736–3744. reliable data are considered to be very good. Matutinovic, S., Kalit, S., Salajpal, K. & Vrdoljak, J. 2011. Effects of flock, year and season on the quality of milk from an indigenous breed Acknowledgements in the sub-Mediterranean area. Small Rumin. Res., 100: 159–163. Mehmeti, H., Hysen, B., Skender, M. & Imeri, B. 2007. Breeding strat- The authors acknowledge the support of Austrian egy for small ruminants in Kosovo. Scientific session. “Indigenous Development Agency (ADA and Kosovo Austrian breeds traditional products, as potential for economic growth rural areas, Albania, Kukes 24–25 September. Institutional Partnership (KAIP) project for fanatical support. Special thanks to farmers, KVFA laboratory staff Ministry of Agriculture Forestry and Rural Development. 2012. – and others who contributed to this research. Mid-Term Evaluation Report of the ARDP 2007 2013. Livestock numbers, pp. 36–62. Morand-Fehr, P., Fedele, V., Decandia, M. & Le Frileux, Y. 2007. Influence of farming and feeding systems on composition and quality References of goat and sheep milk. Small Rumin. Res., 68: 20–34. Ploumi, K. & Emmanouilidis, P. 1999. Lamb and milk production traits – Bytyqi, H. 2009. Within the Kugler Waltrud rare breeds and varieties of Serrai sheep in Greece. Small Rumin. Res., 33: 289–292. of the Balkan Atlas. Monitoring Institute. Stuttgart, Germany, Heidehof Foundation, pp. 97–105. Ploumi, K., Belibasaki, S. & Triantaphyllidis, G. 1998. Some factors affecting daily milk yield and composition in a flock of Chios ewes. Bytyqi, H. & Mehmeti, H. 2006. Identification and Conservation of Small Rumin. Res., 28: 89–92. Animal Genetic Resourcesin South Eastern Europe. Catalogue of West Balkan Pramenka Sheep Types Breed – Bardhoka Strain. Porcu, K., Dzabirski, V., Popovski, Z., Andonov, S. & Tanaskovska, Skopje. ISBN 9989-845-23-9, pp. 45–53. B. 2006. Identification and Conservation of Animal Genetic Resourcesin South Eastern Europe. Catalogue of West Balkan Bytyqi, H., Klemetsdal, G., Ødega rd, J., Mehmeti, H. & Vegara, M. Pramenka Sheep Types Breed – Bardhoka Strain. Skopje. ISBN 2005. A comparison of the productive, reproductive and body condi- 9989-845-23-9, pp. 55–61. tion score traits of the Simmental, Brown Swiss and Tyrol Grey breeds in smallholder herds in Kosovo. Anim. Genet. Res. Inf., 37: 9–20. Sall, J., Lehman, A. & Creighton, L. 2004. JMP start statistics. A guide to statistics and data analysis using JMP and JMP IN software, 3rd Cinkulov, M., Popovski, Z., Porcu, K., Tanaskovska, B., Hodžic´, A., edition. USA, SAS Institute, p. 491. Bytyqi, H., Mehmeti, H., Margeta, V., Djedovic´, R., Hoda, A., Trailovic´, R., Brka, M., Markovic´, B., Važic´, B., Vegara, M., Steel, R.G.D. & Torrie, J.H. 1980. Principles and procedures of statis- Olsaker, I. & Kantanen, J. 2008. Genetic diversity and structure tics. New York, McGraw-Hill. Animal Genetic Resources, 2014, 54, 171–178. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000022

Evaluation des pratiques paysannes de conservation in situ du taurin Baoulé au Sud-Ouest du Burkina Faso

L.Y. Mopaté1, M.J-B. Kamuanga2, S. Hamadou3 and C-Y. Kaboré-Zoungrana4 1Laboratoire de Recherches Vétérinaires et Zootechniques de Farcha (LRVZ), B.P: 433 N’Djaména (Tchad), chercheur associé au LERNSE de l’UPB, Burkina Faso. Email: [email protected]; 2Consulting Agricultural Economist. 8082 Whispering Wind Lane. 20111 Manassas, VA (USA). Ancien Responsable de l’Unité Elevage et Environnement au Centre International de Recherche-Développement sur l’Elevage en zone Subhumide (CIRDES), Bobo-Dioulasso (Burkina Faso); 3Ancien chercheur à l’Unité Elevage et Environnement au CIRDES, Bobo-Dioulasso (Burkina Faso); 4Laboratoire d’études et de recherches des ressources naturelles et des sciences de l’environnement (LERNSE), Université Polytechnique de Bobo-Dioulasso (UPB), Burkina Faso

Résumé L’étude a évalué les pratiques paysannes mises en œuvre par les éleveurs pour assurer la conservation du taurin Baoulé. Elle a porté sur 52% des 636 producteurs recensés, répartis dans 54 villages ciblés qui constituent 7% du total des villages des provinces de Poni et de Noumbiel au Burkina Faso. L’enquête transversale et rétrospective doublée des observations des pratiques a été réalisée. Les modalités des fréquences des actes et des coefﬁcients qui y sont affectés ont été transformées en variables quantitatives. L’adhésion à un éventuel projet de conservation, à la vie associative et la scolarisation induisent des actes de conduite et de gestion, de soins de santé et des activités sociales positifs à la conservation mais négatifs à la reproduction (croisement). L’appartenance à l’ethnie Lobi contribue à la conservation du Baoulé. La possession d’un attelage favorise les actes de conduite et gestion, de soins de santé et des activités sociales conservatoires mais les actes de reproduction ne contribuent pas à la conservation. De plus, la possession d’un troupeau exclu- sif de Baoulé, la castration d’autres mâles non Baoulé, la prise individuelle des décisions de gestion et l’appartenance à l’ethnie Lobi sont favorables à la conservation du Baoulé. Le bilan global positif des pratiques mises en œuvre en pays Lobi burkinabè montre un faible développement du métissage du Baoulé. Les variables quantitatives obtenues par transformation constituent une nouvelle approche, permettant de quantiﬁer les pratiques de conservation des races animales en milieu naturel.

Mots-clés: Pratiques Conservation in situ, Taurin Baoulé, Pays Lobi, Burkina Faso

Abstract The study evaluated the small farmers’ practices implemented by cattle breeders to ensure the conservation of the Baoulé breed. The study involved 52% of the 636 known producers from the 54 targeted villages, which constitute 7% of the total villages of the Poni and Noumbiel provinces in Burkina Faso. The transverse and retrospective survey was carried out and was combined with direct observation. The modalities of the frequencies of various actions and the coefﬁcients that affected them were transformed into quantitative variables. Adhesion to a possible project of conservation, participation in farmers’ groups and other associations and schooling induce positive acts of conduct, health care and social activities for conservation but have a negative effect on reproduction (cross breeding). Membership of the Lobi ethnic group contributes to the conservation of Baoulé cattle. The possession of draught animals promotes the activities of conduct, management, health care and social activities that are protective but active management of reproduction does not contribute to conservation. Moreover, the possession of an exclusive herd of Baoulé cattle, the castration of males from other cattle breeds, the taking of individual decisions of management and the membership of Lobi ethnic group are all favourable to the conservation of Baoulé cattle. The positive total assessment of the practices implemented in the Lobi region of Burkina Faso shows the limited development of Baoulé cattle crossbreeding. The quantitative variables obtained by transformation constitute a new approach, making it possible to quantify the practices of conservation of the animal breeds in their natural environment (in situ).

Keywords: In situ conservation practices, Baoulé taurine, Lobi Region, Burkina Faso

Resumen El estudio ha evaluado las prácticas implementadas por los ganaderos para asegurar la conservación del ganado bovino Baoulé. El estudio fue llevado a cabo con el 52 por ciento de los 636 productores censados, repartidos entre 54 pueblos que constituyen el 7 por ciento del total de pueblos de las provincias de Poni y Noumbiel en Burkina Faso. Se realizó una encuesta transversal y retrospec- tiva pareada sobre las prácticas observadas. Las modalidades de acción y los coeﬁcientes correspondientes fueron transformados en variables cuantitativas. La adhesión a un determinado proyecto de conservación, a la vía asociativa y a la escolarización conlleva acciones de dirección y gestión, de atención sanitaria y de actividades sociales, positivas para la conservación pero negativas para la reproducción (cruzamiento). La pertenencia a la etnia Lobi contribuye a la conservación del ganado Baoulé. La posesión de una yunta favorece las acciones de dirección y gestión y de atención sanitaria y las actividades sociales conservacionistas pero las acciones

Correspondence à envoyer à: L.Y. Mopate, Laboratoire de Recherches Vétérinaires et Zootechniques de Farcha (LRVZ), B.P: 433 N’Djaména (Tchad), chercheur associé au LERNSE de l’UPB, Burkina Faso. Adresse électrorique: [email protected]

171 172 L.Y. Mopate et al.

de reproducción no contribuyen a la conservación. Asimismo, la posesión de un rebaño exclusivamente de raza Baoulé, la castración de los machos de otras razas, la toma individual de las decisiones de gestión y la pertenencia a la etnia Lobi ayudan a la conservación del ganado Baoulé. El balance global positivo de las prácticas implementadas en el territorio de los Lobi en Burkina Faso muestra un escaso mestizaje del ganado Baoulé. Las variables cuantitativas obtenidas por transformación representan una nueva aproximación que permite cuantiﬁcar las prácticas de conservación de las razas animales en su medio natural.

Palabras clave: prácticas de conservación in situ, ganado bovino Baoulé, territorio de los Lobi, Burkina Faso

Soumis: 18 L’octobre 2013; admis: 10 Le janvier 2014

Introduction Il est admis que l’étude et l’évaluation des pratiques concrètement mises en œuvre par les agriculteurs s’impo- En Afrique de l’Ouest, notamment au Burkina Faso, les sent et constituent un passage obligé, pour qui s’intéresse menaces d’absorption (par le sang zébu) des taurins à aux conséquences matérielles de l’activité agricole dans courtes cornes des savanes ont été évoquées par plusieurs un milieu donné (Landais et Balent, 1993). Il est donc auteurs (CIPEA, 1979; Thiombiano, 1993; Rege, Aboagye indispensable pour mieux appréhender la reproductibilité et Tawah, 1994; Camus, Landais et Poivey, 1981; Lhoste, de systèmes d’élevage du taurin Baoulé dans le 1995; Sigué et Kamuanga, 1997; Moazami Goudari et al., Sud-Ouest du Burkina, de mener des investigations dans 2001). Ces menaces s’inscrivent dans un processus global ce sens afin de mieux cibler les actions de développement. des perturbations climatiques intervenues depuis les L’objet de l’étude est d’identifier et d’évaluer les pratiques années 1970 et 1980. Des changements dans les zones pas- paysannes mises en œuvre par les éleveurs du taurin torales ont été observés, avec un déplacement progressif de Baoulé, dans la gestion individuelle ou collective des ani- ’ l élevage zébu des zones du Nord vers les zones du Centre et maux pour assurer la conservation de cette race. du Sud (Maillard et al., 1992). En plus des sécheresses, s’y ajoute la pression foncière croissante dans le Nord pour le Burkina Faso (Ouédraogo, 2002). La concentration impor- Matériel et Méthodes tante de la biomasse animale dans ces zones constitue une Site d’étude menace pour les races taurines adaptées à leur milieu. En effet, la cohabitation étroite entre zébus et taurins favorise La région d’étude au Sud-Ouest du pays est la partie des croisements passifs. De plus, pour accroître les formats burkinabé du pays « Lobi », nom de l’ethnie dominante des taurins sources de prestige aux yeux des paysans, à cheval entre le Burkina Faso, la Côte d’Ivoire et le améliorer l’aptitude au travail et le rendement en viande, Ghana. Elle est considérée comme le berceau du taurin les agriculteurs des savanes se sont lancés dans des croise- Baoulé appelé encore bovin « Lobi ». Elle est limitée à ments actifs depuis les décennies 1980 et 1990 (Landais, l’Est par le fleuve Mouhoun, frontalier avec le Ghana, à 1983; Chupin, 1994). Par ailleurs, la multiplication des ani- l’Ouest par la province de la Comoé, au Nord par celle maux trypanotolérants constitue l’une des stratégies d’utili- du Bougouriba et au Sud par la Côte d’Ivoire (Figure 2). sation et d’exploitation des parcours infestées par les Elle est située entre 9°25 et 10°35 de latitude nord et entre glossines (Achukwi et Musongong, 2009). 2°30 et 4° de longitude ouest et couvre 10 361 km², soit Les travaux d’identification des pratiques et stratégies pay- 3,2% du territoire national (Kienou, Sanou et van sannes de conservation in situ du taurin Baoulé (Figure 1) Bronswijk, 1996). Elle comprend les provinces du Poni et au Burkina Faso sont rares. Ce taurin constitue le seul bovin trypanotolérant bien implanté en pays Lobi de ce pays (Maillard et al., 1992; Thiombiano, 1993).

Figure 2.

Situation de la zone d’étude et principale aire de répartition du taurin Baoulé au Burkina Faso, en Côte d’Ivoire et au Ghana en Afrique de l’Ouest. (Fond Figure 1. Troupeau taurin Baoulé dans la zone de Gaoua (cliché: MOPATE). gratuit de Carte Google Map). Evaluation des pratiques paysannes de conservation in situ du taurin Baoulé 173

du Noumbiel avec respectivement Gaoua et Batié pour chefs- par un taureau Baoulé, d’une vache Baoulé par un taureau lieux. La région est à une altitude moyenne de 400 m. Le cli- Zébu ou Métis, un croisement favorisé de façon active mat est de type soudanien, caractérisé par l’alternance d’une entre Zébu ou Métis et Baoulé, tout comme l’abattage, la période pluvieuse de mai à octobre et d’une période sèche réforme, l ’achat de Zébu. Les croisements passifs de novembre à avril (Guinko, 1984). Les températures diurnes (mélange des troupeaux, divagation et confiage) ont été moyennes varient entre 24,9°C et 30,2°C, avec une amplitude concernés, tout comme les actes encourageant la conserva- moyenne de 5,3°C. C’est une zone agroclimatique tion du Baoulé par l’achat d’un taureau pur, la castration subhumide avec une pluviométrie variant de 900 mm à des mâles de race Zébu ou Métis du troupeau. 1400 mm entre 1992 et 2001 (Lankoandé, 2002). • Pratiques sanitaires: elles ont été orientées sur la participation à la lutte contre les glossines, les traitements aux Echantillonnage et méthodes de collecte de trypanocides, aux anthelmintiques, aux acaricides et aux soins traditionnels. données • Activités sociales liées à l’utilisation des animaux tels La base de sondage a été constituée de 636 propriétaires que l’abattage des taureaux pour les sacrifices, ceux détenteurs surtout de bovin Baoulé, recensés dans les remis en dot, en dons (divers), présentés aux concours deux provinces. Sur cet effectif, 52% ont été choisis au d’éleveurs, les vols enregistrés et les contacts ou hasard dans 54 villages ciblés, soit 7% du total de 769 vil- l’accessibilité à des agents d’élevage. lages dans les deux provinces. Les axes principaux à partir de Gaoua et Batié ont été concernés dans un rayon moyen de 25 km autour de chaque centre, afind’accroître la Analyse de données variabilité des situations tout en limitant le coût de la logis- Les saisies et les traitements ont été effectués par le logi- ’ fi ’ tique. L objectif nal a été d établir un échantillon assez ciel Statistical Package of Social Science (SPSS). Un large (>300 répondants) pour améliorer la représentativité codage en modalités de quelques variables issues des ques- de la population détentrice de bétail Baoulé et prendre en tions ouvertes a été effectué. Les pratiques (événements et compte les refus et absences au démarrage des enquêtes. actes) ont été quantifiées par des coefficients allant de – 3 L’étude a été réalisée par enquête informelle et formelle à + 3, selon le poids « objectif » que la pratique accomplie complémentaires et appropriées pour améliorer la connais- apporte en termes d’impact favorable ou pas pour la con- sance des systèmes d’exploitation agricole (Hubert, 1992; servation in situ de la race. Vianney et Palm, 1999). Ces enquêtes ont été menées en L’importance d’un acte est la variable quantitative résultante deux phases distinctes: de la multiplication du coefficient affecté à l’acte (variant de • Une pré-enquête pour recenser les propriétaires du taurin – 3 à + 3 suivant sa contribution à la conservation), par la Baoulé afin de disposer d’une base de sondage; consulter modalité de la fréquence de l’acte posé (aucune fois = 0; et discuter avec des groupes de personnes afinde moins fréquent = 1; fréquent = 2 ou très fréquent = 3). connaître le milieu et les pratiques de production et Les avortements, naissances vivantes et mortalités qui sont réaliser une étude bibliographique sur le milieu. considérés comme plus subis par les chefs d’exploitations • Une enquête (formelle) transversale et rétrospective, et les actes posés (conduite au pâturage et abreuvement) doublée des visites multiples pour améliorer les estimateurs. régulièrement réalisés dans la quasi totalité des exploitations ’ Des interviews et observations auprès d un échantillon et touchant également la quasi totalité des effectifs, n’ont pas ’ d exploitants propriétaires du taurin Baoulé ont été menées. été pris en compte. Seule l’importance des autres actes de Les questions fermées ont porté sur les pratiques d’élevage conduite et gestion (traite, complémentations, traction et liées à la conservation in situ.Ils’est agit ici d’identifier au parcage), de reproduction (croisements) et de conservation, ’ cours de l’année qui s’est écoulée (12 derniers mois), les de soins de santé et d activités sociales a été considérée. Une ’ actes posés par l’éleveur en terme de pratiques réalisées agrégation générale de l importance des ces actes posés ou des événements observés en relation avec la conserva- seuls et des pratiques (événements et actes compris) a été fi tion de son taurin Baoulé. Le nombre d’animaux concernés effectuée, a n de tirer des bilans généraux (positifs ou et les fréquences de ces événements et actes posés. Ainsi, négatifs) de la conservation au niveau de la région. quatre groupes de pratiques ont été visés: Un premier tri à plat a été effectué pour sélectionner des vari- • Pratiques de conduite et de gestion: elles renfermaient le ables à croiser, à soumettre en analyse de variance (ANOVA). pâturage, l’abreuvement, la traite, la traction animale, Le test de Bonferroni a été appliqué pour les comparaisons la complémentation minérale, en fourrages/résidus et le multiples (plus de deux groupes) des moyennes. parcage des animaux. La régression linéaire simple des déterminants de la par- • Pratiques de reproduction: elles ont été subdivisées, en ticipation ou contribution financière à un éventuel projet événements observés relatifs aux avortements survenus, de conservation et de l’importance des actes posés par aux morts-nés, aux naissances vivantes, aux mortalités les Chefs d’Exploitation a été effectuée. Pour cela, la naturelles et en un témoignage de l’éleveur. Ces méthode d’estimation par les moindres carrés ordinaires témoignages ont concerné la monte d’une vache Baoulé (MCO) a été utilisée. L’adéquation d’ensemble du 174 L.Y. Mopate et al.

Tableau 1. Importance des actes entre les Chefs d’Exploitation (CE) avec troupeaux purs et mixtes et entre ceux décidant seul de la gestion unique et à plusieurs.

Importance des actes Type de troupeau Centre de décision

Pur Mixte P Unique Multiple P N = 287 N = 38 N = 278 N = 47

Conduite 2,83 ± 3,18 7,81 ± 3,91 *** 3,65 ± 3,62 2,03 ± 3,49 ** Reproduction 3,27 ± 3,54 −4,73 ± 6,34 *** 2,16 ± 4,83 3,36 ± 4,83 ns Santé 1,04 ± 1,72 4,57 ± 2,07 *** 1,56 ± 2,17 0,82 ± 1,44 * Activités sociales 0,39 ± 2,06 1,89 ± 2,58 *** 0,70 ± 2,16 −0,23 ± 2,07 **

***p < 0,001; **p < 0,01; *p < 0,05; ns = non signiﬁcatif.

modèle est mesurée par le coefficient de détermination projet éventuel de conservation du Baoulé. Il en est de (R²), le R² ajusté et le test de Fisher. Le modèle fonctionnel même des actes de conduite et gestion et les soins de santé obtenu permet de légitimer l’impact des variables explica- des CE membres d’un groupement (Tableau 2). tives considérées prises individuellement sur la participa- ’ ’ tion ou l’importance des actes posés qui sont des Niveau d alphabétisation/d instruction des CE et variables à expliquer. On considère l’hypothèse alternative leur catégorie d’âge ’ fi qu au moins un des coef cients du modèle est différent de Les CE alphabétisés et scolarisés ont significativement les zéro. Les variables explicatives des déterminants de la par- moyennes les plus élevées pour les actes de conduite et fi ticipation nancière ont été choisies parmi les données gestion, de soins de santé et des activités sociales. En socio-économiques des CE et de leur exploitation. Quant revanche, la moyenne des actes liés à la reproduction ’ à celles des déterminants de l importance des actes des pour ce groupe est plus faible et négative que celle des CE, elles découlent des critères retenus pour leur analyse. non-alphabétisés (Tableau 3). Les CE âgés de moins de 40 ans posent plus d’actes des soins de santé que ceux Résultats de 40 ans et au-delà, dont les actes de reproduction sont significativement plus élevés. Importance des actes posés par les éleveurs Durée dans la pratique de l’élevage et appartenance ou pas à l’ethnie Lobi Troupeaux purs ou mixtes et prise des décisions de gestion unique ou multiple Les exploitations qui ont moins de 15 ans de pratique de l’élevage bovins ont des moyennes de l’importance des Les moyennes de l’importance des actes de conduite et actes significativement plus élevées en conduite et gestion gestion, des soins de santé et des activités sociales sont et en activités sociales (Tableau 4). Le groupe Non-Lobi significativement plus élevées dans les exploitations avec (migrants) pose significativement plus des actes de conduite, des troupeaux mixtes que dans celles avec des troupeaux de santé et des activités sociales. En revanche, les actes liés à purs. En revanche, les actes en matière de reproduction la reproduction des bovins y sont négatifs. Le groupe Lobi a ont une moyenne significativement plus élevée dans les la moyenne la plus élevée pour les actes de reproduction. troupeaux purs (Tableau 1); où il y a absence des croisements avec d’autres races ou des actes favorisant active- Possession ou pas d’un attelage dans ment ou passivement ces croisements. Pour la nature des l’exploitation et provinces concernées centres de décision, les moyennes sont plus élevées et Les chefs d’exploitation (CE) possédant un attelage dans significatives pour les décisions de gestion relevant leur exploitation ont significativement les moyennes les uniquement du propriétaire en matière de conduite et ges- plus élevées pour les actes de conduite et gestion, de tion, de soins de santé et des activités sociales. soins de santé et des activités sociales (Tableau 5). En ’ 1 revanche, ceux sans attelage ont la moyenne de l impor- Degré de participation à la conservation et tance des actes de reproduction plus élevée. appartenance ou pas à un groupement Selon les provinces, seuls les actes de reproduction mon- Pour les actes de conduite et gestion, des soins de santé ani- trent une différence significative. Noumbiel a effective- male et des activités sociales liées à l’élevage du Baoulé, ment la moyenne la plus élevée par rapport à Poni. l’importance de ces actes est significativement plus élevée chez les CE ayant annoncé une participation élevée à un Déterminants de l’importance des actes posés par les Chefs d’Exploitation

1 Le degré de participation à un éventuel projet de conservation du taurin Baoulé est Le test de Fisher (F) pour le modèle de régression linéaire la volonté exprimé par les chefs d’exploitation d’y contribuer de façon élevée (au ﬁ moins 1 000 F CFA) ou faible (moins de 1 000 F CFA) à ce projet dans les deux simple utilisé est signi catif au seuil de 1%. Ce qui indique provinces concernées. 1U$ = 500 F CFA que l’ensemble de coefﬁcients du modèle est Evaluation des pratiques paysannes de conservation in situ du taurin Baoulé 175

Tableau 2. Importance des actes entre CE annonçant une participation élevée et moins élevée et entre CE appartenant à un groupement ou pas.

Importance des actes Participation à la conservation Membre de groupement

Élevée Faible P Oui Non P N = 137 N = 188 N = 208 N = 117

Conduite 4,73 ± 3,82 2,45 ± 3,18 *** 4,11 ± 3,67 2,17 ± 3,24 *** Reproduction 2,03 ± 4,66 2,56 ± 4,77 ns 2,29 ± 5,13 2,41 ± 3,90 ns Santé 2,16 ± 2,44 0,95 ± 1,67 *** 1,83 ± 2,25 0,80 ± 1,54 *** Activités sociales 1,02 ± 1,92 0,22 ± 2,29 ** 0,57 ± 2,51 0,54 ± 1,41 ns

***p < 0,001; **p < 0,01; ns = non signiﬁcatif.

Tableau 3. Importance des actes posés entre CE alphabétisés ou pas et entre CE de moins de 40 ans et de 40 ans et plus.

Importance des actes Alphabétisation Catégorie d’âge

Oui Non P Moins de 40 ans 40 ans et plus P N = 47 N = 278 N = 208 N = 117

Conduite 6,19 ± 4,11 2,94 ± 3,34 *** 4,07 ± 4,12 3,29 ± 3,52 ns Reproduction −0,48 ± 6,55 2,82 ± 4,17 *** 1,09 ± 4,06 2,57 ± 4,81 * Santé 3,40 ± 2,54 1,13 ± 1,82 *** 1,92 ± 2,50 1,37 ± 2,00 * Activités sociales 1,65 ± 2,38 0,38 ± 2,09 *** 0,98 ± 1,70 0,48 ± 1,70 ns

***p < 0,001; **p < 0,01; *p < 0,05.

Tableau 4. Importance des actes des CE de moins de 15 ans de pratique d’élevage et de 15 ans plus et entre ceux Lobi et Non-Lobi.

Importance des actes Durée dans la pratique d’élevage Groupe ethnique

−de 15 ans 15 ans et + P Lobi Non-Lobi P N = 173 N = 152 N = 313 N = 12

Conduite 3,93 ± 3,85 2,82 ± 3,30 ** 3,25 ± 3,56 7,66 ± 3,31 ** Reproduction 2,30 ± 4,21 2,38 ± 5,26 ns 2,66 ± 4,39 −6,08 ± 5,58 ** Santé 1,63 ± 2,20 1,28 ± 1,96 ns 1,31 ± 1,98 5,16 ± 1,58 ** Activités sociales 0,79 ± 1,89 0,30 ± 2,4 * 0,50 ± 2,16 2,08 ± 2,19 **

***p < 0,001; **p < 0,01; *p < 0,05.

Tableau 5. Importance des actes entre CE possédant un attelage et ceux n’en possédant pas et entre les provinces.

Importance des actes Possession d’attelage Province

Oui Non P Poni Noumbiel P N = 121 N = 200 N = 168 N = 157

Conduite 6,45 ± 3,41 1,61 ± 2,35 *** 3,67 ± 3,72 3,14 ± 3,54 ns Reproduction 0,69 ± 5,59 3,31 ± 3,72 *** 1,76 ± 5,45 2,96 ± 3,72 * Santé 2,85 ± 2,16 0,63 ± 1,55 *** 1,43 ± 2,17 1,49 ± 2,01 ns Activités sociales 1,45 ± 2,12 0,03 ± 2,03 *** 0,38 ± 2,25 0,76 ± 2,07 ns

***p < 0,001; *p < 0,05; ns = non significatif. significativement différent de zéro. L’équation du modèle La valeur de R² ajusté montre que les variables considérées de l’importance des actes (IMPACTES) est: expliquent 28% de l’importance des actes (Tableau 6). Les variables attelage, participation et membre ont une influ- Impactes = ß0 + ß1 AGE + ß2 ATT + ß3 MEM + ß4 PTP + ence significative et positive sur l’importance des actes. ß5 TTB + μ Elles ont une influence individuelle respective de 99% et ’ ’ où les ß sont des paramètres à estimer et μ le terme d’er- 95% dans l importance des actes. En revanche, l âge et ’ fl reur ou de perturbation aléatoire. le nombre total de bovins de l exploitation ont une in u- ence significative mais négative. Ces variables influencent AGE = âge; ATT = attelage; MEM = membre; PTP = par- respectivement l’importance des actes dans 90% et 95% ticipation; TTP = total bovin des cas. 176 L.Y. Mopate et al.

Tableau 6. Estimation de l’importance des actes des CE dans conservation in situ de la race. Toutefois, ils gagneraient l’ensemble des deux régions. à être davantage améliorés par une investigation Variable Coefficients t de student préliminaire auprès des éleveurs et autres producteurs ou agents de terrain. Ce qui permettrait d’élucider le poids Constance 3,625** 2,418 réel accordé à tel ou tel acte préjudiciable ou favorable à − − Age (AGE) 0,040* 1,711 la préservation de la race, afin de parvenir à une certaine Attelage (ATT) 4,988*** 7,979 ’ fi Membre (MEM) binaire 1,402** 2,226 objectivité. L aspect quantitatif des coef cients permet Participation (PTP) binaire 0,004*** 4,559 de faire un bilan global (négatif ou positif) des pratiques Total bovin (TTB) −0,048** −2,188 conservatoires d’une race dans une zone, de discriminer F 22,342*** les acteurs, d’identifier les liaisons ou les déterminants R² 0,297 de certains actes posés et leur impact sur la conservation R² ajusté 0,283 en milieu naturel. ***p < 0,001; **p < 0,01; *p < 0,05. Le bilan global positif des actes seuls et celui des pratiques en pays Lobi burkinabè montre un faible développement Bilan des actes et des pratiques (événements et du métissage du Baoulé. Les actes de croisements interra- actes) au niveau régional ciales (Zébus/Métis X Baoulés) ou ceux favorisant active- ment ou passivement ces croisements y sont très faibles. ’ La moyenne de l importance générale des actes est de 7,78 La moyenne des actes de reproduction plus élevée à ± 6,28, avec un minimum = 5 et un maximum = 30. Celle Noumbiel a été en rapport avec la faible présence des des pratiques (événements et actes posés) est de 8,83 ± métis dans cette province. La conservation de la race 6,93, avec un minimum = 8 et un maximum = 34. Cette Baoulé y serait mieux assurée. moyenne générale positive indique que les actes posés ’ participent globalement à la conservation du Baoulé. L importance des actes (conduite, santé et activités ’ La différence de moyenne entre la province du Poni sociales) plus marquée chez les Chef d Exploitation (CE) (7,52 ± 5,88) et celle du Noumbiel (8,35 ± 6,68) n’est pas avec des troupeaux mixtes serait liée à leur pouvoir ’ fi significative au seuil de 5%. Il en est de même de d achat. En effet, seuls les moyens nanciers permettent celle des pratiques à Poni (8,56 ± 6,69) et à Noumbiel de complémenter (achat de minéraux et fourrages) les ani- (9,66 ± 7,44). maux, de côtoyer les agents vétérinaires pour leurs soins et de s’équiper en traction animale. Les corrélations positives entre les actes de conduite et de santé (r = 0,607), entre Corrélation de l’importance des actes entre eux actes de santé et activités sociales (r = 0,500) et entre la Les actes de conduite et gestion sont corrélés positivement conduite et les activités sociales (r = 0,379) expliquent aux actes de santé et des activités sociales. Il en est de même les liens entre ces actes. Les contacts pris avec les agents ’ ’ entre actes de santé et activités sociales (Tableau 7). de l élevage expliquent l importance des activités sociales En revanche, les actes liés à la reproduction sont de ceux avec des troupeaux mixtes. Une certaine similarité ’ négativement corrélés avec tous les autres actes. se dégage dans l importance des actes posés entre ce groupe et les migrants. Ces derniers ont surtout des troupeaux mixtes et disposent également des moyens. Cette Discussion possibilité explique aussi l’importance élevée de ces actes chez les CE qui possèdent un attelage. Néanmoins, Les variables quantitatives obtenues par transformation des les actes de reproduction plus faible dans ce groupe modalités des fréquences des actes et des coefficients qui y témoignent de la pratique du métissage. Il en est de sont affectés sont une nouvelle approche permettant de même des CE avec des troupeaux mixtes et les migrants, quantifier les pratiques des producteurs. Ici, il s’agit des où la cohabitation de plusieurs races dans les exploitations pratiques de conservation des races animales, notamment entraîne des croisements actifs et passifs, préjudiciables à le bovin Baoulé, en milieu naturel. Les coefficients la préservation. La préférence affichée des autochtones affectés bien que subjectifs, restent basés sur une connais- du pays Lobi burkinabè pour le Baoulé (Tano et al., sance des actes et leurs poids respectifs dans la 2001) et la réticence des Lobi ivoiriens (Camus, Landais

Tableau 7. Corrélation de l’importance des actes entre eux.

Actes Conduite/Gestion Reproduction Santé Activités sociales

Conduite/Gestion 1,000 −0,366*** 0,607*** 0,379 Reproduction −0,366*** 1,000 −0,391*** −0,231 Santé 0,607*** −0,391*** 1,000 0,500 Activités sociales 0,379*** −0,231*** 0,500*** 1,000

***p < 0,001. Evaluation des pratiques paysannes de conservation in situ du taurin Baoulé 177

et Poivey, 1981) et Burkinabé (Maillard et al., 1992)au de 15 ans et plus. Néanmoins, les actes de reproduction métissage contribuent à accréditer le faible penchant de faibles et négatifs des alphabétisés indiquent la pratique la grande majorité des Lobi pour le métissage. Toutefois, du métissage, préjudiciable à la préservation de la race l’émergence des autochtones avec des troupeaux mixtes malgré l’importance notoire des autres actes posés. (à visées spéculatives) constitue des menaces réelles qui L’influence de l’âge sur les actes de reproduction montre planent sur le Baoulé (Mopaté, 2003). En effet, des que les CE plus âgés (de 40 ans et plus) préserveraient études récentes dans les régions Nord de la Côte d’Ivoire mieux la race Baoulé que ceux de moins de 40 ans. On proche de la province de Noumbiel, indiquent que 86 à peut faire des rapprochements sur les actes de reproduction 90 % des élevages présentent des signes de métissage avec le groupe des CE alphabétisés qui se retrouvent sur- (Sokouri et al., 2007; Sokouri et al., 2009). tout dans la classe d’âge de moins de 40 ans. Ce groupe L’amélioration du format des Baoulé qui diffère significa- a une moyenne des actes de reproduction très faible et tivement du Méré (croisé Zébu et Baoulé) en est la raison. négative indiquant une tendance au métissage. Ce Métis utilisé comme animal de trait, se vend deux fois Les déterminants de l’importance des actes par l’analyse plus que le Baoulé. de régression renforcent les critères analysés et confirment La prise des décisions de gestion par une seule personne a que les facteurs socio-économiques de l’exploitation con- une influence positive sur les actes de conduite, de santé et ditionnent la conservation in situ de la race. les activités sociales. Ce qui confirme que les pratiques restent liées à l’opérateur (Teissier, 1979) qui agit selon les références qu’il s’est construites par expériences per- sonnelles et celles communiquées par les membres de sa Conclusion communauté (Landais, 1992; Pleine, Schmithüsen et Les pratiques d’élevage du taurin Baoulé mises en œuvre Sorg, 1995; Lalba, Badini et Kleene, 1999). On ressent en Pays Lobi du Burkina Faso participent globalement à cette influence globalement au niveau de la région où dom- la préservation de cette race, au regard du bilan positif ine le groupe ethnique Lobi. Les membres de ce groupe des actes et pratiques conservatoires. Les troupeaux appartiennent à la même communauté et gèrent seuls en Baoulés purs majoritaires en gardiennage individuel et la majorité leur troupeau. Les actes de reproduction positifs castration systématique des mâles d’attelage achetés sont chez les Lobi indiquent qu’ils préservent mieux la race des stratégies de conservation mises en œuvre. Baoulé du métissage, contrairement aux Non-Lobi L’importance des actes et leurs déterminants indiquent caractérisés par des actes de reproduction négatifs. Ces l’influence des facteurs socio-économiques qui concourent actes affectent gravement la conservation de la race dans à la préservation ou pas du Baoulé. Les actes de conduite, ce groupe. Les peuls gardiens des bovins Baoulé des auto- santé et activités sociales participent à la préservation chez chtones appartiennent au groupe Non-Lobi. La multiplica- les CE ayant annoncé une participation élevée, ceux gérant tion des centres de décision de gestion par des acteurs seuls leur élevage, ceux âgés et ceux membres d’un (propriétaire, copropriétaire et gérant) appartenant à groupement. Leur importance plus nette chez ceux avec différentes ethnies induirait des actes contradictoires qui des troupeaux mixtes, les migrants, les alphabétisés et ne favorisent pas la préservation de la race. ceux avec attelage ne favorise pas la conservation car l’im- L’importance des actes (conduite, santé et activités portance des actes de reproduction y est faible ou négative. sociales) plus élevée chez les CE ayant annoncé une par- La stratégie d’accumulation de richesse par la constitution ticipation élevée à la conservation du Baoulé a été proba- des troupeaux mixtes s’amplifiera dans la région, augmen- blement liée aux moyens, aux objectifs et à la préférence tant ainsi la menace d’extinction du Baoulé. Les comporte- accordée à cette race. Ces considérations liées à l’éleveur ments et intentions des chefs d’exploitation tendent vers influencent la gestion du troupeau (Lalba, Badini et l’adoption du métissage dans l’avenir. Le conservatisme Kleene, 1999). Les actes de conduite et de santé élevés Lobi a contribué à préserver le Baoulé du métissage. des CE membres des groupements contrairement aux Mais des changements socioculturels et économiques autres, montrent que l’ouverture à la vie associative est prévisibles modifieront le comportement de la jeune un facteur positif de conservation de la race Baoulé. génération vis–à–vis de ce bétail. Ainsi, l’élevage du Ceux qui y adhèrent marquent leur volonté de progrès Baoulé ne s’inscrira pas seulement dans une optique socio- social. On note l’influence de l’ouverture d ’esprit chez culturelle ayant contribué à sa préservation. Il faut à l’instar les alphabétisés qui se sont démarqués par des actes (con- du taurin N’Dama qui fédère le Mali, la Guinée, le Gambie duite, santé et activités sociales) plus fréquents que chez et le Sénégal autour d’un projet de développement et de les analphabètes, favorables à la conservation. Le souci préservation, envisager des actions similaires pour le taurin d’amélioration de l’élevage du Baoulé à travers les actes Baoulé en pays Lobi. Les Lobi, principaux éleveurs de posés explique probablement cette différence. Il en été taurin Baoulé, se partagent entre le Burkina Faso, la Côte de même des CE qui pratiquent l’élevage du Baoulé depuis d’Ivoire et le Ghana. Les facteurs ethnique et linguistique moins de 15 ans avec des actes de conduite et activités constituent un atout capital, favorisant l’organisation des sociales plus élevés, favorisant la conservation que ceux éleveurs autour du taurin Baoulé pour sa préservation. 178 L.Y. Mopate et al.

Remerciement Lhoste, Ph. 1995. Sélection génétique. In: Daget et Godron (coordonna- teurs), Pastoralisme: Troupeaux, espaces et sociétés. Hatier - Aupelf - – Les auteurs tiennent à remercier les différents bailleurs qui Uref (éds.), 322 324. ont contribué ﬁnancièrement à la réalisation de cette étude, Maillard, J.C., Congo, I., Bassinga, A., Cuveillier, J.F. 1992. ainsi que les responsables scientiﬁques du CIRDES et de Immunogénétique du taurin Baoulé en Pays Lobi (Burkina Faso). 1. ’ l’UPB à différents niveaux pour leurs appuis multiformes, Environnement de cette population trypanotolérante. Revue d Elevage et de Médecine Vétérinaire des Pays Tropicaux,45(1):63–68. les enquêteurs et les agro-éleveurs qui n’ont ménagés leur temps pour ce travail. Moazami Goudari, K., Belemsaga, D.M., Geriotti, G., Laloë, D., Fagbohoun, F., Kouagou, N. T., Sdibé, I., Codja, V., Crimella, M.C., Grosclaude, F., Touré, S.M. 2001. Caractérisation de la race Somba à l’aide des marqueurs moléculaires. Revue d’Elevage et de Références bibliographiques Médecine Vétérinaire des Pays Tropicaux, 54(2): 129–138.

Achukwi, M.D., Musongong, G.A. 2009. Trypanosomosis in the Doayo/ Mopaté, L.Y. 2003. Conservation In Situ du bovin Baoulé: pratiques pay- Namchi (Bos taurus) and zebu White Fulani (Bos indicus) cattle in sannes et stratégies en «Pays Lobi» du Burkina Faso. Mémoire du D. Faro Division, North Cameroon. Journal of Applied Biosciences 15: E.A en Gestion Intégrée des Ressources Animales, option « Systèmes 807–814. de Productions Animales ». Institut de Développement Rural (IDR), l’Université Polytechnique de Bobo-Dioulasso (UPB), Burkina Camus, E., Landais, E., Poivey, J.P. 1981. Structure génétique du chep- Faso, 89. tel bovin sédentaire du Nord de la Côte d’Ivoire. Perspectives d’avenir en fonction de la diffusion croissante du sang zébu. Revue d’Elevage Ouédraogo, D. 2002. Analyse socio-économique des pratiques de gestion et de Médecine Vétérinaire des Pays Tropicaux 34(2): 187–198. de la trypanosomose animale et les facteurs associés au développement de la chimiorésistance dans la province du Chupin, D. 1994. Le rôle de la biotechnologie de la reproduction pour la Kénédougou (Burkina Faso). Thèse de Doctorat Unique ès Sciences conservation des ressources génétiques animales. Animal Genetic Économiques, UFR/SEG, Université de Ouagadougou, 210. Resources Information 14: 13–26. Pleine, T., Schmithüsen, F., Sorg, J-P. 1995. Logiques paysannes et ’ CIPEA (Centre international pour l élevage en Afrique). 1979. Le reboisement: le cas d’Andohavary sur les hauts plateaux de ’ bétail trypanotolérant d Afrique occidentale et centrale. Tome Madagascar. Cahiers de Recherche-Développement, 42: 62–76. I. Situation général, Étude de systèmes, Addis - Abeba, Ethiopie, 155. Rege, J.E.O., Aboagye, G.S., Tawah, C.L. 1994. Shorthorn cattle in Hubert, J.P. 1992. Enquête informelle versus enquête formelle? Cas des West and Central Africa. I. Origin, distribution, classification and paysannats de Impo au Nord du Burundi. Annales de Gembloux 98: population statistics. World Animal Review 78(1): 2–13. 195–205. Sokouri, D.P., Loukou, N.E., Yapi-Gnaoré, C.V., Mondeil, F., Guinko, S. 1984. Végétation de la Haute-Volta. Thèse de Doctorat d’État Gnangbe, F. 2007. Caractérisation phénotypique des bovins à viande ès sciences naturelles. Université de Bordeau III, France, 318 + (Bos taurus et Bos indicus) au centre (Bouaké) et au nord (Korhogo) annexes. de la Côte d’Ivoire. AGRI 40: 43–53. Kienou, J.D., Sanou, H., van Bronswijk, P. 1996. Monographie provin- Sokouri, D.P., Yapi-Gnaoré, C.V., N ’Guett, A.S.P., Loukou, N.E., ciale du Poni, étude du milieu. Ministère de l’Économie et des Finances, Ouagadougou (Burkina Faso), 176. Kouao, B.J., Touré, G., Sangaré, A., Kouassi, A. 2009. Utilisation et gestion des races taurines locales sous la pression des Lalba, A., Badini, O., Kleene, P. 1999. Etablissement des priorités croisements avec les zébus dans les régions Centre et Nord de la géographiques de recherche et développement en élevage par combi- Côte d’Ivoire. Journal of Animal & Plant Sciences 5(2): 456–465. naison d’études de diagnostic informelles et formelles: Burkina Faso. ’ In: Roeleveld et Broek (éds.), les systèmes d’élevage: orienter la Sigué, H., Kamuanga, M.J.B. 1997. Le secteur de l élevage et le poli- recherche. Institut royal des tropiques et Centre technique de tique de la santé animale au Burkina Faso: Revue et perspectives. coopération agricole et rurale (Cta), 83–97. Document de travail 2, 36. Landais, E. 1983. Analyse des systèmes d’élevage bovin sédentaire du Tano, K., Kamuanga, M.J.B., Faminow, M.D., Swallow, B.M. 2001. Nord de la Côte d’Ivoire. Thèse Doctorat d’Etat IEMVT, Adoption and demand for trypanotolerant cattle in the subhumide Maisons-Alfort, France. 2 vol., 759. zone of West Africa. Journal of Agriculture and Environment for International Development 95(2–3): 213–235. Landais, E. 1992. Principes de modélisation des systèmes d’élevage: approches graphiques. Cahiers de Recherche-Développement 32(3): Thiombiano, D. 1993. La banque de semence du Centre de Recherche 82–95. sur la Trypanosomose Animale (CRTA). In: Chupin, Wagner et Wilson (éds.), amélioration génétique des bovins en Afrique de ’ Landais, E., Balent, G. 1993. Introduction à l étude des systèmes l’Ouest. FAO, Collection Production et Santé Animal 110. Actes de ’ ’ d élevage extensif. In: Landais (éd.), pratiques d élevage extensif : l’atelier tenu à Banjul (Gambie) en octobre 1992, 193–198. identifier, modéliser, évaluer. INRA, Études et Recherches sur les Systèmes Agraires et le Développement 27: 13–35. Teissier, J.H. 1979. Relations entre techniques et pratiques. Bulletin INRAP 38. Lankoandé, Y.F. 2002. Développement des bovins trypanotolérants au Burkina Faso: défis - potentialités - opportunités. Mémoire d’Ingénieur Vianney Labé, Rudy Palm. 1999. Statistique, empirique, informelle: de développement rural, option Elevage, IDR - Université quelle enquête pour la collecte d’information sur les exploitations Polytechnique de Bobo-Dioulasso (Burkina Faso), 68 + annexes. agricoles. Cahiers Agricultures 8(5): 397–404. Animal Genetic Resources, 2014, 54, 179–184. © Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/S2078633614000010

Inﬂuence of coat colour on Chamarita sheep phenotypes, behaviour, welfare and performances

M. Pascual-Alonso1, G.C. Miranda-de la Lama2, L. Aguayo-Ulloa1, M. Villarroel3, S. Alierta1 and G.A. Maria1 1Department of Animal Production and Food Science, Faculty of Veterinary Medicine, University of Zaragoza, Miguel Servet 177, E-50013, (50013) Zaragoza, Spain; 2Group of Animal Welfare and Sustainable Livestock Production, Department of Food Science, Metropolitan Autonomous University, UAM-Lerma, State of México, México; 3Department of Animal Science, E.T.S.I.A. Polytechnic University of Madrid, Madrid, Spain

Summary Many rustic breeds under low levels of genetic selection maintain variations in coat colour, which local farmers link to production traits, but few quantitative analyses have been made of the effect of coat colour on several zootechnical parameters of importance. The aim of the study was to describe differences in morphology, production, behaviour and welfare of adult sheep ewes (n = 50) in a Chamarita breed flock in terms of coat colour. The wool and skin colour of Chamarita sheep are quite variable, with white, black and galana sheep. Morphological measurements (weight, body length, height at withers and thorax circumference) were not significantly different for different coat colours but white sheep tended to be larger. Average litter size (1.29 lambs per ewe) was also similar, but white coated sheep gave birth to larger lambs. White-mother lambs also had a higher average daily growth during the milking period, although final weaning weight was similar to black ones. Most (>70 percent) of all the aggressive interactions observed were performed by white sheep, while affiliative interactions (friendly interactions that promote group cohesion) were similar among coat colour. Regarding the welfare assessment, most indicators were similar among sheep with different coat colour except for glucose in blood plasma, which was higher in white sheep. Chamarita ewes are relatively small compared with other breeds, and well adapted to their environment, including the production system and harder climatic conditions. Their performance is within the average of local breeds and still has margins for improvement, which underline the importance of their conservation.

Keywords: Chamarita ewes, coat colour, morphology, social interactions, stress response

Résumé Beaucoup de races considérées comme rustiques et ayant été soumises à une faible pression de sélection génétique conservent des variations dans la couleur de leur robe que certains éleveurs associent à des caractères productifs. Cependant, l’effet de la couleur de la robe sur des paramètres zootechniques d’importance n’a guère été étudié. L’objectif de cette étude a été de décrire les différences en morphologie, production, comportement et bien-être selon la couleur de la robe de 50 brebis adultes de la race Chamarita. Les mesures morphologiques (poids, longueur, hauteur au garrot, circonférence thoracique) n’ont pas différé significativement entre les couleurs de robe même si les brebis blanches ont tendu à être plus grandes. La prolificité (1.29 agneaux/brebis) a aussi été similaire mais les brebis blanches ont eu des agneaux plus grands. Les agneaux de mères blanches ont aussi eu un gain moyen quotidien plus élevé, bien que le poids final ait été similaire à celui des agneaux de mères noires. La plupart (>70 pour cent) des interactions agressives observées ont été réalisées par des brebis blanches, alors que le pourcentage d’interactions affiliatives a été similaire pour les deux couleurs de robe. En ce qui concerne l’évaluation du bien-être, aucune différence n’a été décelée entre les couleurs de robe pour la plupart des indicateurs, hormis le glucose sanguin qui a été plus élevé chez les brebis blanches. Les brebis Chamaritas sont relativement petites par rapport àd’autres races et sont bien adaptées à leur environnement productif. Leurs performances productives sont satisfaisantes et peuvent encore être améliorées, comme quoi la conservation de cette race s’avère importante.

Mots-clés: brebis Chamaritas, couleur de robe, morphologie, réponse au stress, interactions sociales

Resumen Muchas razas consideradas rústicas y sometidas a poca presión de selección genética mantienen variaciones en el color de capa que algunos ganaderos relacionan con caracteres productivos, pero se han realizado pocos estudios respecto al efecto del color de capa en algunos parámetros zootécnicos de importancia. El objetivo del presente estudio es describir diferencias en morfología, producción, comportamiento y bienestar de ovejas adultas (n = 50) de raza Chamarita en cuanto a color de capa. Las medidas morfológicas (peso, largura, altura a la cruz, circunferencia torácica) no resultaron significativamente diferentes entre colores de capa aunque las ovejas blancas tendían a ser más grandes. La prolificidad (1.29 corderos/oveja) también fue similar pero las ovejas blancas tuvieron corderos más grandes. Los corderos de madres blancas también tuvieron una mayor ganancia diaria (ADG), aunque el peso final fue similar al de los corderos de madres negras. La mayoría (>70 percent) de las interacciones agresivas observadas fueron ejecutadas

Correspondence to: M. Pascual-Alonso email: [email protected]

179 180 M. Pascual-Alonso et al.

por ovejas blancas, mientras que las interacciones aﬁliativas resultaron similares en ambos colores de capa. En cuanto a la evaluación del bienestar, la mayoría de los indicadores no presentaron diferencias entre colores de capa excepto la glucosa en sangre que fue mayor en las ovejas blancas. Las ovejas Chamaritas son relativamente pequeñas comparadas con otras razas, y están bien adaptadas a su ambiente productivo. Sus caracteres productivos son razonables con un buen margen de mejoría, demostrando la importancia de su conservación.

Palabras clave: ovejas Chamaritas, color de capa, morfología, respuesta de estrés, interacciones sociales

Submitted 25 September 2013; accepted 16 January 2014

Introduction The wool and skin colour of Chamarita sheep are quite Many local breeds under low selection pressure maintain variable, with white, black (roya) and rarely galana (a large variations in coat colour, which the farmers link to combination of white, black and brown skin with white production traits, but a few quantitative analyses have and black/roya wool) coats (Figure 1). The roya sheep is been made to substantiate the effect of coat colour on dif- born black but their wool turns in reddish-black with ferences in morphology, production, behaviour and wel- age, looking like brown. This type of coat is locally called fare. Chamarita sheep are a breed from La Rioja (Spain), roya. The black skin is mainly located in the face. It is like- included in the Official Catalogue of Spanish Livestock ly that, as with other breeds in the region, breeders have Breeds under Endangered Native Breeds. The Chamarita selected for hardiness (through fertility and morphological population size is about 10 000 sheep and it is officially characteristics) and adaptation to the environment and less recognized as an endangered breed to be preserved. The for pure white wool. Nowadays there are many black sheep herd book was created in 2007 by the Chamarita Sheep in herds, since it is associated with positive zootechnical Association of La Rioja (or AROCHA as it is known by traits (Doménech et al., 1992). Black animals usually its Spanish acronym), which has an official recognition to have a white spot on the nape (crowned), and another at promote and control the breed. Most herds are found in the distal end of the tail ( puntiblancos). Sometimes, the the mountains of lower La Rioja and in the Cidacos and spot of the neck extends over the head, forehead and Linares valleys (Warren, 2011). Animals are small in size nasal area (caretos). (adult rams 55–65 kg, ewes 35–40 kg; Barrio, Falceto Currently, Chamarita sheep are mostly used for meat and Doménech, 1991), probably as an adaptive response production, both for suckling lamb (12–14 kg live weight) to their hard living conditions and in their extensive and and a weaned lamb with a short indoor fattening period sustainable production system, which uses local resources (20–22 kg live weight; Doménech et al., 1992). The (Álvarez and Arruga, 2007). Chamarito lamb was recognized as a quality brand in

Figure 1. Differences in coat colour in Chamarita sheep. Inﬂuence of coat colour on Chamarita sheep phenotypes, behaviour, welfare and performances 181

April 2010 (PROCORCHA). Thus, farmers now used a Mouton®). A platform with a seat 3 m above the ground production system that ensures compliance with strict ani- wasusedtoobservetheflock from a distance. The ewes mal welfare health programmes, respecting the origin of were observed 6 h daily, from 8:00 to 10:00 h, 12:00 to sheep fed on extensive pastures and lambs produced in 14:00 h and 16:00 to 18:00 h for 16 consecutive days (96 La Rioja on their mother’s milk. In this study, we describe h of observation) by the same trained observer. A behaviour the morphological, productive and welfare traits of adult sampling technique was used to record all social interactions female Chamarita sheep and the influence of coat colour (agonistic and non-agonistic behaviour). Agonistic interac- on these traits. tions with contact included butts (when the ewe used the front of her head to make contact with another ewe), pushes (when a ewe used other parts of her body to make contact Materials and methods with another ewe) and bites (when one ewe bit another ewe’s body using her teeth). Agonistic interactions without A flock of 50 Chamarita adult ewes were transported from La contact included threats (when a ewe turned towards or Rioja to the Animal Experimentation Support Service (SAEA) approached another individual with her head down and of the University of Zaragoza, Aragón (41◦41_N). All sheep then lunged without making contact), and chase (when a used were raised, transported and slaughtered according to cur- ewe actively moved towards another individual, causing rent regulations of the European Community Commission the latter to walk or run away). Non-agonistic interactions (1986) for Scientific Procedure Establishments. All the proto- with contact included licking (when a ewe passed her tongue cols were approved by the Animal Experimentation Ethics over the body of another individual) and grooming (when a ’ Committee of the University of Zaragoza. This study was the ewe groomed another ewe s body using her teeth). fi collaboration between the Autonomous Regions of Aragon Non-agonistic interactions without contact included snif ng ’ fl and La Rioja (Spain). (when a ewe sniffed another ewe s body) and the ehmen response (when a ewe retracted the upper lip, wrinkled the nose and bared the gums in the presence of another ewe). Animals The adult multiparous ewes were housed in pens during pregnancy and lactation (2 m2 per ewe), fed twice a day Physiological welfare indicators (at 08.00 and 15.00) with pellet concentrate (11.5 MJ Blood samples were taken by jugular venipuncture with ME per kg DM and 15.5 percent crude protein; 0.3 kg vacuum tubes during the dry period to evaluate physiologic- per ewe) and ad libitum lucerne chaff (Medicagosativa). al responses to stress (two 10 ml tubes per animal, with The pen was equipped with a metallic water trough and without anticoagulant, Ethylenediaminetetraacetic acid (1.5 m × 0.60 m) and two metallic feeders (4.5 m × 0.80 EDTA-K3). Blood was sampled using the necessary precau- m, 27 cm per ewe) and a lick stone for minerals. tions to avoid sampling error on stress indicators. Samples Most ewes were inseminated by controlled natural mating were kept on ice for a maximum of 2 h and taken to the labora- before leaving the source farm and lambed at the tory for routine haematological measurements. EDTA plasma University a few weeks after arrival. The lambs were kept and serum were centrifuged at 3 000 rpm for 10 min and − together in the same pen with their mothers (2 m2 per aliquots were frozen and kept at 30 °C until analysed. sheep and lamb). Flock prolificacy was calculated as the An automatic particle counter (Microcell counter F-800 total number of lambs born per total ewes lambing. Lambs and auto dilutor AD-260, SysmexTM both) was used to were weighed at birth (BW) and at weaning (WW). count red blood cells (RBC) and white blood cells Pre-weaning average daily gain (ADG) was estimated by (WBC) (number per mm3), haemoglobin (g/dl) and haem- the difference WW−BW divided by the total milking period atocrit (percent). The leukocyte formula was estimated (30 days). from blood swabs on clean slides. Staining was performed by the rapid panoptic method using dyes from QuímicaClinicaAplicada Inc. With an optic immersion Morphological measurements of ewes microscope we counted and identified 100 leucocytes per One day before blood sampling, body length (BL), height sample (neutrophils, lymphocytes, eosinophils, basophils at withers (HW) and thorax circumference (TC) were mea- and monocytes). The neutrophil/lymphocyte ratio (N/L) sured as in Miranda-de la Lama et al. (2011), and each ewe was used an indicator of chronic stress (Lawrence and was weighed using a portable digital weighing scale. Rushen, 1993). Serum samples were used to determine the concentration of glucose (mg/dl, Ref. Glucose AE2-17), and the activity of creatine kinase (CK) (UI/L) (Ref. CK. Social behaviour NAC AE1-13) with a multianalyser ACE® (Clinical Social behaviour was evaluated by direct observation. All Chemistry System) and reagents from Alfa Wasserman. ewes were individually identified and marked with 30 Serum concentration of non-esterified fatty acid (NEFA) cm-high numbers and letters painted on the sides and rump levels was analysed by a multianalyser ACE® (Clinical with washable paint for sheep marking (PeintureMarquage Chemistry System of the Alfa Wasserman), with commercial 182 M. Pascual-Alonso et al.

kits (NEFA C Ref. 994-75409 of the Wako). The concentra- Table 1. Least-square means (±S.E.) for the effect of coat colour tion of cortisol was determined from plasma (EDTA-K3) by on morphological, physiological and behavioural traits in enzyme immunoassay using an “in home-kit” (validated by Chamarita sheep. Chacón et al., 2004). Each sample was determined in dupli- Whole flock White ewes Black ewes cate from 50 μl of plasma and the results were expressed in nmol/l, with the corresponding controls. Variation coeffi- Morphological measurements cients of the analysis, inter- and intra-assay, were 7 and 8 Bodyweight (kg) 44.26 ± 6.74 46.23 ± 1.72 43.36 ± 1.16 TC (cm) 104.05 ± 4.58 105.18 ± 1.11 103.47 ± 0.79 percent, respectively. The concentration of lactate was deter- HW (cm) 65.72 ± 3.7 66.55 ± 0.89 65.28 ± 0.64 mined using a Sigma Diagnostic kit (lactate no. 735-10) and BL (cm) 72.93 ± 4.25 73.82 ± 1.03 72.47 ± 0.74 spectrophotometer (Lambda 5, Perkin Elmer). Physiological measurements Cortisol (nmol/l) 27.59 ± 29.2 27.32 ± 7.62 27.72 ± 5.14 Glucose (mg/dl) 72.97 ± 33.91 87.2 ± 8.48a 66.51 ± 5.72b Data analysis Lactate (mg/dl) 18.57 ± 13.91 18.53 ± 3.63 18.6 ± 2.45 NEFA (nmol/l) 0.098 ± 0.098 0.07 ± 0.03 0.11 ± 0.017 Data were analysed using the least squares methods of the CK (UI/l) 131.89 ± 77.6 144.4 ± 20.13 126.21 ± 13.57 GLM procedure using SAS/STAT (9.1 SAS Inst. Inc., Ratio N/L 0.81 ± 0.37 0.80 ± 0.09 0.83 ± 0.06 fi WBC (103/mm3) 8.14 ± 4.0 8.42 ± 1.04 8.01 ± 0.7 Cary, NC, USA) by SAS (1998), tting a one-way 3 3 fi RBC (10 /mm ) 8.85 ± 1.28 8.96 ± 0.33 8.79 ± 0.23 model with a xed effect of coat colour (two levels) within HTC (%) 28.8 ± 4.27 29.45 ± 1.11 28.5 ± 0.75 the ewe data. The general representation of the model used Social interactions was: y = Xb+e, where y was an N × 1 vector of records, b Agonistic 163.88 ± 107.03 226.12 ± 23.78a 131.82 ± 17.06b denoted the fixed effect in the model with the association Affiliative 18.74 ± 21.55 19.58 ± 5.28 18.3 ± 3.79 matrix X and e was the vector of residual effects. A prob- Different letters (a, b) represent significant differences ( p < 0.05) between ability of P < 0.05 values was considered statistically groups. NEFA, non-esterified fatty acid; CK, creatine kinase; Ratio N/L, significant. neutrophil/lymphocyte ratio; WBC, white blood cells; RBC, red blood cells; HCT, haematocrit.

Results and discussion Assaf: 60–70 kg) is consistent with the morphology and weights of most breeds considered rustic and well The proportion of coat colours frequencies, according to adapted to a harsh environment (Catalogue Livestock the association of breeders statistics, are 45 percent white Breeds in Spain). Such is the case of Ojalada (between coat and 55 percent non-white coat (including the minority 35 and 45 kg of body weight), Mallorquina (30–40 kg of galana coat). According with the genetic study performed body weight) or Gallega (25–35 kg in the mountain eco- by Álvarez and Arruga (2007), the Chamarita phenotypes type). This adaptation is due to food shortages, thereby are considered to be as a unique population. reducing their energy needs (Álvarez and Arruga, 2007), allowing them to make use the resources at their disposal. In that regard, black sheep that tended to be smaller than Morphological measurements white sheep may be more adapted to a harsh environment Coat colour had little effect on morphological measure- than white sheep (Álvarez and Arruga, 2007). ments (Table 1). Live weight of the adult ewes averaged 44.26 (±6.74) kg with no significant differences between white and black sheep. The average TC was 104.05 Productive traits (±4.58) cm, HW 65.72 (±3.7) cm and BL from chest to The average litter size of the whole flock was 1.29 (±0.36) tail 72.93 (±4.25) cm. The average live weight and HW lambs born per ewe lambing with no significant differ- were higher than described by Barrio, Falceto and ences between coat colours (Table 2). Litter size is slightly Doménech (1991) (live weight 36.6 kg, HW 59.8 cm). higher than in Doménech et al. (1992), who found a ten- Chamarita sheep may have a lower body weight after dency for single births (1.1 lambs per ewe). However, food shortages in extensive systems, which was not the case in our experimental herd where sheep were fed twice daily with concentrate and had water and straw ad Table 2. Least-square means (±S.E.) for the effect of coat colour libitum. However, BL was shorter than in Barrio, Falceto on productive traits in Chamarita sheep. and Doménech (1991) (90.5 cm), but the latter authors Whole flock White ewes Black ewes did not find differences between body measurements of Prolificacy 1.29 ± 0.36 1.25 ± 0.09 1.3 ± 0.07 white and black sheep. Although body weight was not BW of lambs (kg) 3.63 ± 0.59 3.86 ± 0.16a 3.53 ± 0.1b fi a b signi cantly higher in white ewes, there was a tendency ADG of lambs (g) 215.56 ± 48.56 236.49 ± 13.04 206.18 ± 8.73 for black sheep to be lighter. WW of lambs (kg) 11.95 ± 2.12 12.46 ± 0.59 11.72 ± 0.39

The small size of Chamarita ewes compared with more Different letters (a, b) represent signiﬁcant differences ( p < 0.05) between productive breeds (e.g. Rasa aragonesa: 45–50 kg; groups. BLW, birth weight; ADG, average daily gain during the milking Manchega: weighing over 45 kg; Merino: 50–70 kg; and period; WLW, weaning weight. Inﬂuence of coat colour on Chamarita sheep phenotypes, behaviour, welfare and performances 183

the latter authors suggested that an increased food supply Physiological welfare indicators could increase prolificacy to 1.5 lambs per ewe and deliv- Physiological and haematological variables can be used to ery. These data confirm the potential for improvement of assess the welfare of animals subject to acute and/or chron- the Chamarita sheep when environmental conditions are ic stress (Barnett and Hemsworth, 1990). The results of optimal. The lambs weighed 3.63 (±0.59) kg at birth, physiological measurements are presented in Table 1. and those born to white mothers were significantly (P ≤ Cortisol levels averaged 27.59 (±29.2) nmol/l, glucose 0.05) heavier than these born (+9 percent) black ones. 72.97 (±33.91) mg/dl, lactate 18.57 (±13.91) mg/dl, Balda, Chavarri and Doménech (1981) reported a birth NEFA 0.098 (±0.098) nmol/l and CK 131.89 (±77.6) weight of 2.6 kg in pure lines and 3.6 kg in industrial IU/l. The N/L ratio averaged 0.81 (±0.37), which suggests crossings. In our study, the herd was pure and the crosses there was no immune suppression (Lawrence and Rushen, were with Chamarito males, so probably the increased 1993). None of the sheep had clinical signs of illness. The birth weight in lambs may be due to higher feeding rate WBC count was 8.14 (±4) 103/mm3, RBC count 8.85 in our experimental group. The fact that lambs from (±1.28) 103/mm3, and haematocrit percentage was 28.8 white females weighed more at birth than those from (±4.27). All physiological measures were within normal black females is probably due to larger body size of their ranges for sheep during routine handling procedures and mothers is associated with higher consumption of concen- do not suggest that welfare was compromised trate. This hypothesis partially corroborates the data on (Hargreaves and Hutson, 1990). Glucose was the only blood glucose levels and social interactions of the herd, physiological variable that was significantly higher (P ≤ which suggest that white sheep were dominant (see 0.05) in white sheep (by 31.1 percent) than black sheep. below). Similarly, lambs from white ewes had a higher Plasma glucose levels increase after acute stress when ADG during the milking period (+14.7 percent, P ≤ adrenaline induces muscle glycogen catabolism for its 0.05), compared with the ADG of the whole herd 215.56 use in gluconeogenesis in the liver (Apple et al., 1995). (±48.56) g. The ADG data support the hypothesis that However, in our case, the higher glucose levels in white white females could eat more and therefore produce sheep could have been the result of their higher feed more milk for their lambs. However, the average weaning intake. White sheep tended to be heavier and be the dom- weight of the lambs (11.95 ± 2.12 kg) was not significantly inant animals in the herd, as reflected in levels of aggres- different among different coat coloured mothers. That sive social interactions, all of which imply higher energy implies that a homogeneous product can be obtained at costs. However, we did not find a difference in stress levels weaning regardless of coat colour. between black and white sheep, probably since the group was quite stable and had abundant resources. Social behaviour Of the 163.88 (±107.03) aggressive interactions observed Conclusions (Table 1), most were performed by white sheep (+71.5 per- ≤ cent, P 0, 05). That suggests a higher dominance status for White ewes of the Chamarita breed are slightly larger than white sheep (Orgeur, Mimouni and Signoret, 1990;Barroso, the black-coated ewes and give birth to larger lambs. They Alados and Boza, 2000), tied with the fact that they were also behaved more aggressively and had higher plasma heavier and larger (Miranda-de Lama et al., 2011). concentrations of glucose, underlying their more dominant Dominance is a mechanism that regulates social behaviour character. Overall, the productive traits of Chamarita ewes and priority access to available resources, since high domin- and lambs were quite good compared with other sheep and ance animals have priority access to resources in intensive taking into consideration their small size. Owing to their production conditions (Barroso, Alados and Boza, 2000). strong adaptation to the harsh environment of La Rioja That coincides with the higher glucose levels in white sierra, using grazing resources within a sustainable sheep. The levels of aggressive interactions observed within production system, Chamarita breed should be preserved. a stable herd are consistent with the idea of Doménech et al. Appropriate selection and conservation practises will (1992)thatChamarita behaviour is more like goat behav- help to improve their performance and help contribute to iour, with a lively temperament and instinct for survival, rural development and sustaining the local agro-ecosystem, fi linked to foraging behaviour and browsing. Af liative inter- preserving a valuable cultural and gastronomic heritage of actions totalled 18.74 (±21.55) for the whole herd, with no the Community of La Rioja. significant differences regarding coat colour. Affiliations are a very important part of the social cohesion of a group (Miranda-de la Lama and Mattiello, 2010), but under conditions of social competition for limited resources, animals tend Acknowledgements to do without them (Miranda-de la Lama et al., 2011). It is likely that an adaptation to very scarce resources has led to This study was funded by the Spanish Ministry of Science Chamarita to favour aggressive interactions in detriment of and Innovation (projects AGL-2009-10794/GAN and affiliative ones in the context of the social group. AGL2012.37219). Many thanks to the Autonomous 184 M. Pascual-Alonso et al.

Community of La Rioja for the Scholarship PhD Programme Barroso, F.G., Alados, C.L. & Boza, J. 2000. Social hierarchy in the (Programa de AyudasPredoctoralesdelGobierno de la domestic goat: effect on food habits and production. Appl. Anim. – Rioja) of M. Pascual-Alonso. Special thanks to AROCHA Behav. Sci. 69: 35 53. and PROCORCHA associations for their support and their Chacón, G., García-Belenguer, S., Illera, J.C. & Palacio, J. 2004. invaluable help. This appreciation is special for the veterinary Validation of an EIA [enzyme immunoassay] technique for the determination of salivary cortisol in cattle. Span. J. Agric. Res.2:45–51. responsible of these associations, Don José Antonio González. Thank you very much to the staff of the Animal Research Doménech, J.M., Barrio, A.R., Falceto, M.D. & González Jiménez, J.A. Centre of the University of Zaragoza (SAEA). 1992. La oveja Chamarita. Serie Estudios no 25, Gobierno de La Rioja. Logroño, La Rioja, España. European Commission. 1986. Scientific procedure and breeding of animals for use in scientific procedure establishments. Directive 86/609/ ECC 1986. References Hargreaves, A.L. & Hutson, G.D. 1990. The effect of gentling on heart rate, flight distance and aversion of sheep to a handling procedure. Álvarez, L. & Arruga, M.V. 2007. Caracterización de la oveja Appl. Anim. Behav. Sci. 26: 243–252. Chamarita mediante marcadores microsatélites. Estructura genética, Lawrence, A.B. & Rushen, J. 1993. Stereotypic animal behaviour: fun- genotipado del gen Scrapie y creación de un banco de ADN. Beca damental and application to welfare. UK, CAB International. de estudios Arnedanos. Miranda-de la Lama, G.C. & Mattiello, S. 2010. The importance of Apple, J.K., Dikeman, M.E., Minton, J.E., McMurphy, R.M., Fedde, social behaviour for goat welfare in livestock farming. Small Rumin. M.R., Leith, D.E. & Unruh, J.A. 1995. Effects of restraint and iso- Res. 90: 1–10. lation stress and epidural blockade on endocrine and blood metabolite status, muscle glycogen metabolism, and incidence of dark-cutting Miranda-de la Lama, G.C., Sepúlveda, W.S., Montaldo, H.H., María, longissimus muscle of sheep. J. Anim. Sci. 7: 2295–2307. G.A. & Galindo, F. 2011. Social strategies associated with identity profiles in dairy goats. Appl. Anim. Behav. Sci. 134: 48–55. Balda, I., Chavarri, J.B. & Doménech, J.M. 1981. La oveja Chamarita I Jornadas Técnicas de ganado ovino en La Rioja-Logroño. Orgeur, P., Mimouni, P. & Signoret, J.P. 1990. The influence of rearing conditions on the social relationships of young male goats, Capra hir- Barnett, J.L. & Hemsworth, P.H. 1990. The validity of physiological cus. Appl. Anim. Behav. Sci. 27: 105–113. and behavioural measures of animal welfare. Appl. Anim. Behav. Sci. 25: 177–187. Statistical Analysis System Institute. 1998. SAS/STATS user’s guide (Release 6.03). Cary, NC, SAS Institute. Barrio, A.R., Falceto, M.D. & Doménech, J.M. 1991. La raza Chamarita: Estudio etnológico y morfométrico. XVI Jornadas Warren, J. 2011. Chamarito Lamb Promoted among Restaurant Owners Científicas de la S.E.O.C. and Butchers. http://www.foodsfromspain.com. Animal Genetic Resources, 2014, 54, 185. © Food and Agriculture Organization of the United Nations, 2014

Recent Publication

The camels of Kumbhalgarh. A biodiversity treasure I. Köhler-Rollefson, H. Singh Rathore and A. Rollefson Edited by K. Hardy Lokhit Pashu-Palak Sansthan, Sadri Published in 2013, pp 48. ISBN 81-90164-1-4 Available at http://www.lpps.org/wp-content/uploads/2013/10/ Camels_Of_Kumbhalgarh_web.pdf doi:10.1017/S2078633614000216

This publication describes the roles of the camels in the Kumbhalgarh region in Rajastan, India, ecologically sustainable camel herding system of unique biocultural heritage. The publication highlights the importance of the camel herding in the culture and livelihoods of the Raika people, traditional keepers of the camels. The two camel breeds present in the Kumbhalgarh are one-humped or dromedary camels (Camelus dromedarius), sub-set of the “Marwari” breed. The book provides descriptions of the camels, their feeding behaviour (the book includes an appendix of preferred forage plants containing their medicinal and other uses) and their roles in the functioning of the agro-ecological system, daily routine of the herders and the migration routes, as well as the importance of the traditional knowledge about the ecosystem. The annual cycle of the camel breeders and the management activities are presented, highlighting the unique character of this breeding system. However, a number of problems are system can exist as an integral part of the planned threatening this system and these problems are named National Park, attract visitors interested in ecological tour- and discussed. Due to declining incomes and camel popu- ism and provide a diversity of camel products such as lations, urban migration of the members of the herder fam- camel milk, camel milk soap and camel dung paper. The ilies and the plans of Rajastan government to declare the publication concludes with presenting the legal documents part of traditional herding lands as a National Park, this which are important to ensure the camel breeders’ rights system is threatened by the loss of its agro-ecological and recommendations for enhancing the potential of the and socio-cultural heritage values. However, the authors camels to contribute to the diversity of the Kumbhalgarh describe the options which can ensure that the herding Protected Area.

Recent Publication

Caractérisation phénotypique des ressources génétiques animales Directives FAO sur la production et la santé animales No. 11 Publication de la version française en 201, pp. 151. E-ISBN 978-92-5-207843-2 Available at http://www.fao.org/docrep/019/i2686f/i2686f.pdf doi:10.1017/S2078633614000228

Les Directives sur la caractérisation phénotypique des ressources génétiques animales ont été publiées en français en 2013. Ces Directives font partie d’une série de publication produites par la FAO pour aider les pays à mettre en œuvre le Plan d’action mondial pour les ressources zoogénétiques, qui reconnaît qu’«une bonne compréhension des caractéristiques des races est nécessaire pour guider la prise de décision en matière de programmes de développement et de sélection des animaux d’élevage» (FAO, 2007). Ces directives sur la caractérisation phénotypique des ressources génétiques animales s’adressent à la Priorité stratégique 1 du Plan d’action mondial: «Caractérisation, inventaire et surveillance des évolutions et des risques associés». Elles complètent, en particulier, les directives sur la caractérisation génétique moléculaire et sur la réalisation d’enquêtes et de suivi pour les ressources génétiques animales. En effet, les trois types d’informations – phénotypique, génétique et historique – sont nécessaires pour caractériser les ressources génétiques animales. Les informations fournies par les études de caractérisation sont primordiales pour planifier la gestion des ressources génétiques animales au niveau local, national, régional et ciblée et rentable qui contribue à l’amélioration de la gestion mondial. des ressources génétiques animales dans le cadre de la mise Ces directives permettent d’identifier des différentes races et en œuvre du Plan d’action mondial au niveau des pays. De de décrire de leurs caractéristiques externes et de production plus, un aperçu des concepts et des approches qui soutien- dans un environnement et un cadre de gestion donnés, en nent les études de caractérisation phénotypique est suivi tenant compte des facteurs socio-économiques qui les affec- de conseils pratiques sur la planification et la mise en tent. Pour cela, elles se concentrent sur la collecte et l’utili- œuvre des activités sur le terrain, ainsi que sur la gestion sation de l’information phénotypique. Par ailleurs, les et l’analyse de données. Enfin, les annexes comprennent activités de caractérisation phénotypiques sont difficiles à des formats génériques de collecte de données pour la mettre en œuvre d’un point de vue logistique et technique. caractérisation phénotypique des principales espèces Ainsi, les directives donnent des conseils sur la façon de d’élevage, ainsi qu’un cadre pour l’enregistrement de mener une étude de caractérisation phénotypique bien données sur les milieux de production des races.

Recent Publication

Cattle husbandry in Eastern Europe and China. Structure, development paths and optimisation Edited by A. Kuipers, A. Rozstalnyy and G. Keane Wageningen Academic Publishers Published in 2014, pp. 280 ISBN 978-90-8686-785-1 doi: 10.3920/978-90-8686-785-1 doi:10.1017/S207863361400023X

This book presents an analysis of the dairy and associated sector developments in Eastern Europe and China. The transition in cattle husbandry in Eastern Europe, resulting in enormous structural changes, but not in an increased production volume, is different from emerging countries, like China, where traditional small scale farming goes together with the remodelling or start of mega farms. Capacity building by means of cooperation appears to be hindered by historical experience. Farm development paths were studied in more depth in Poland, Lithuania and Slovenia using detailed farmer sur- veys. Farmers’ strategies, availability of resources, and opportunities and threats were analyzed, and interactive group trainings in strategic management were part of the analysis. The results are presented here. Although farmers showed similar wishes concerning farm development, the local environment and policies determines the degree of achievement. Farm fragmentation, the high percentage of rented land and availability of suitable labour are major constraints. Future EU policies are a concern but new technologies are embraced. This book is a must for those interested in the transition in marketers, companies, farm leaders and government Eastern Europe. It is indispensible to consultants, ofﬁcials in agriculture.

189 20786336_54-0_20786336_52-0 04/06/14 3:47 PM Page 1 54 2014 CONTENTS Editorial...... i 54 Phenotypic and morphological characterization and reproduction attributes of native pigs in Bangladesh C.H. Ritchil, M.M. Hossain and A.K.F.H. Bhuiyan...... 1 Caractéristiques phénotypiques de la chèvre du sahel au Niger par analyse des indices de