Study area

The study area included 53 communities from 17 municipalities and 221 producers. The geographic coordinates are 18° 54’ and 19° 56’ N and 97°o17’ and 98° 37’ W. The municipalities sampled in the state of Puebla were: Acajete, Aquixtla, Atempan, Atzizintla, Chignahuapan, Chignautla, Cuyoaco, Ixtacamaxtitlán, Libres, Satillo la Fragua, Tetela de Ocampo, Tlachichuca, Tlahuapan and Tlatlauquitepec. In the state of Tlaxcala, the municipalities of Nanacamilpa and Calpulalpan were selected. The altitude had a range of 1621 to 3160 m (^{INEGI, 2014}). In the region there are three types of climate: a) C(w)(w), humid with abundant summer rains, b) C(w1)(w), sub-humid temperate with summer rains, and c) C(w2)(w), sub-humid semi-cold with summer rains (^{García, 1998}). The temperature varies from 5 to 26 °C. The annual rainfall has a range of 400 to 1200 mm. The type of vegetation that predominates is open grassland, crassicaule shrub in the dry climate, and pine-oak-sacred fir forest and shrub in the sierra and high mountain zones. The dominant soils are Andosol, Litosol, Regosol (^{INEGI, 2014}).

Data registry

For the registry of information a sample of sheep production units was selected through simple random sampling and maximum variance (^{Mendenhall et al., 1987}), due to not knowing the number of production units that breed sheep in the study area. With a margin of error of 6.5 % and a maximum variance of 0.25, the size of the sample calculated was 227 producers from the states of Puebla and Tlaxcala.

The sheep producers in the study zone do not have a record of their exploitation, as was described by ^{Acero et al. (2004}) for Spain. In the region, since 2006, the authors of this study have carried out studies of the sheep production systems per zones (^{Vázquez-Martínez et al., 2009}; ^{Galaviz-Rodríguez et al., 2011}); the sheep population was characterized (^{Vargas et al., 2012}) and the growth of lambs was monitored (^{Galaviz-Rodríguez et al., 2014}). The technical staff from the previous studies was the same that collaborated for the recording of field data from direct interview with the owner of the production unit and direct observation of the flocks. In the design of the field work and the elaboration of the questionnaire to record information, the methodologies that use direct surveys with the owner and inside the farm were reviewed (^{Toro-Mujica et al. 2011}; ^{Gaspar et al., 2011}; ^{Milán et al., 2011}), as well as the direct observation for data recording (^{Riveiro et al., 2013}). In the social part the questions were about the owner and the family. In the technical part, the size of the flock, land, crops cultivated, and productive and reproductive aspects of the sheep were recorded. In the economic part, the costs of inputs, workforce, veterinary services and total earnings were recorded. In the registry of sheep weight and the productive parameters of production units, databases were used from the study of sheep populations that was carried out in parallel to the survey, and when the production unit was not in the sample the data were obtained directly by the technical staff and the producer for each flock.

Statistical analysis

The database was captured in Excel and exported to Statistical Analysis System (^{SAS, 2003}), version 9.4 for Windows. The factors of the production system were estimated with a factorial analysis; with this, linear combinations were obtained to form reduced groups of standardized variables and which explained the higher proportion of variance of the original data. The typology of the sheep farms was determined with cluster analysis. The objective of this analysis was to group the production units based on individual homogeneity and heterogeneity among different groups. As a measure of linkage, the square Euclidian distance and Ward method was used. The cluster analysis grouped the sheep farms by the lowest variance within the group and the differences with other groups. The description of each typology was carried out with the SAS GLM procedure for social, productive and income variables and the means from each grouping were compared with the adjusted Tukey test (^{SAS, 2003}). The productive and economic indexes were determined with what was mentioned by ^{Acero et al. (2003)} and ^{Vázquez-Martínez et al. (2009)}.

Characterization of the sheep production system

The owners are 52.0±0.9 years of age, primary school studies (5.6±0.2 years) and experience in sheep production of 18.4±0.7 years, which is common in family subsistence systems (^{Pérez-Hernández et al., 2011}). The schooling and experience were related to the increase in the productivity and the adoption of new technologies, as was pointed out by ^{Morales et al. (2004)}.

The producers own 6.1±0.4 ha of land, where maize (Zea mays L) (99.5 %), oats (Avena sativa L) (30.7 %) and barley (Hordeum vulgare L) (21.7 %) are cultivated. There is a relationship between sheep and crops; the sheep take advantage of residues from the harvests and the stubbles from the agricultural system and provide manure (^{Rivas et al., 2014}).

The flocks have 64 sheep in average, of which 55.6 % are ewes; 2.2 %, studs; 25.9 %, young ewes; and 16.3 %, fattening lambs. The size of the flock is higher than the one found by ^{Mavule et al. (2013}) in South Africa. The high proportion of sheep in the flock is related to the orientation of lamb production. The sheep are a cross of Creole x Suffolk (51.1 %), Creole (20.8 %), crosses of Hampshire (10.4 %) and crosses of Suffolk x Hampshire (3.6 %).

The management of sheep is simple; extensive grazing is used in cultivation areas after the harvest (58.8 %), on the edges of paths (31.6 %), in forest areas (8.0 %) and induced grasslands (1.5 %). The use of cultivation areas is common in sheep breeding, as was pointed out by ^{Galanopoulos et al. (2011)} in Greece. The duration of grazing is 2 to 8 hours day^-1. Supplementing in the farmyard is carried out by 42.8 % of the producers; this practice is common in shepherding systems of small ruminants (^{Kosgey et al., 2008}).

The males stay in the flock all the time. The ewes have their first birth at the age of 15.5±0.2 months, similar to what was reported by ^{Milán et al. (2011}) in sheep systems of northeastern Spain. The mating takes place in June-August and with a birth rate of 80.4o% in the months of October to January; the interval between births was 11.6 months, which is considered as a long period and is attributed to the seasonality of ewes from the photoperiod (^{Arroyo et al., 2007}).

The lambs are weaned at 3.3±0.1 months at age and weight of 16.6±0.2 kg. The fattening period is 9.7 months, with a weight of 46.4±0.5 kg and average sales price of $39.1±0.2 kg^-1, with a range of $28.0 to $48.0 kg^-1 for standing lamb.

The factors of the sheep production system

With the factorial analysis, the following factors of sheep production in the study zone were identified (Table 1).

Factor I. Size of the production unit. This factor explained 31.4 % of the total variance. The size of the flock (0.96), the production costs (0.94), the total expenses in workforce (0.93), and the total from sheep earnings (0.86) were the variables with highest weight in the factor.

Factor II. Socioeconomic profile of the producer. This factor explains 11.6 % of the total variance of the data. The variables that had the highest contribution are age (0.95) and experience of the producers in sheep breeding (0.43).

Factor III. Production of agricultural inputs. Their contribution in the variance explained was 9.7o% and the variables with highest weight in the factor are the surface sown with maize (0.95) and the yield of maize grain (0.46).

Factor IV, VI and VII. Lamb production. This production component explained 18 % of the total variance. Factor IV contains 7.8 % of the total variance and the variables with greatest contribution in the factor were the final weight of the lambs (0.80) and the duration of the fattening (0.31). In turn, factor VI explained 5.6 % of the variance and the weight at birth had the highest contribution to the factor (0.51). Factor VII had an explanation of 4.6 % of the total variance and the main variable that contributed in the factor was the sale price of lambs (0.54).

Factor V. Means of production. This factor explained 6.4 % of the variance and the small-scale property land was the variable with highest contribution to explaining the factor (0.95).

Typology of sheep production units

The cluster analysis generated three groups of sheep production units for the study zone, which are described next:

Cluster 1. Family subsistence production units. This group includes 60.6 % of the sheep flocks. The purpose of these farms is sheep breeding as a complementary activity to others, and the sheep are a source of income in times of economic need. This group did not differ in the variables of the producer and family from Clusters II and III (Table 2), but it did with the lowest values in the use of family workforce (1.1 shepherds), total land (4.5 ha) and barley cultivation (0.5 ha). From maize production, an average of 2.3 tons of stubble is obtained (Table 2). The flock is 34.1 sheep (Table 3), which is very small but is common in shepherding systems of semiarid climates in countries with a situation of poverty (^{Mavule et al., 2013}); or else, in the mixed farmyards of developed countries, where sheep coexist with more animal species in the same farm (^{Gaspar et al., 2008}). The diet depends on grazing, which is described as a way of reducing the production costs (^{Gelasakis et al., 2012}) and purchasing fodders is scarce. The farms do not have infrastructure for animal management, as has been pointed out by Gaspar et al. (2008) and ^{Riveiro et al. (2013}), for farms with better socioeconomic conditions in Spain. The productive parameters indicate a low efficiency of the farms. The fattening period of lambs is 10.7 months and lambs of higher weight are sold (47.3 kg). Table 4 shows the costs of production and earnings. The main cost is the use of family workforce ($17 432.80) to carry out grazing and other management practices. As was indicated by ^{Acero et al. (2004}), in the family farms the workforce is a cost of opportunity and is not included in the accounting of losses and profits. In this sense, the producers in the region do not take into account the time devoted to caring for the flock and this is an occupation without economic compensation. In average, 5.6 lambs are sold per year, which is related to low economic results ($10 621.30 of annual gross income), but which covers the value of the rural food basket of one person (^{CONEVAL, 2016}). The farms in this group are the most common in the region, where sheep production is a way of contributing to the family income and using agricultural stubbles.

Cluster 2: Sheep-cereal association. This group concentrates 71 farms (32.1 %) located in the valleys and areas with moderate slopes. The farms use 2.1 labor days for sheep handling, 8.4 ha of land and 2 ha of barley sown, which indicates that they have the resources for sheep production, if compared with Cluster 1. The average flock is 84.8 sheep (Table 3), which gives it an intermediate size between family subsistence farms (Cluster I) and extensive mountainous farms (Cluster III). Sheep-cereal association is a way of producing sheep to take advantage of stubbles after the harvests (Ruíz et al., 2008; ^{Rivas et al., 2014}). The byproducts harvested from cereals are used in the diet of the lambs and lactating ewes, which agrees with the observations by ^{Caballero (2001}). The duration of lamb fattening is 8.4 months in this group and 2.2 months shorter than in Cluster I. The production costs are $44 588.00, where the workforce represents 83 % of them. The annual income is $26 159.00, which does not cover production costs, mainly of the workforce (Table 4). However, in face of the lack of employment opportunities in rural areas the income from sheep production in this group covers the value of the annual rural food basket for 2.4 people, and it covers the welfare line for 1.3 members of the family, according to estimations by ^{CONEVAL (2016)}. Sheep production associated to the production of cereals in the study area is complementary activity to agriculture and, with an approximate flock of 100 sheep earnings are obtained from the sale of lambs that contribute to the family’s food security.

Cluster 3: Extensive mountainous farms . This group includes 16 farms (7.2 %). For sheep production, three people and 8 ha of land are used (Table 2). Land ownership is communal in mountainous areas, which allows producers a greater access to forest zones for grazing. Grazing lasts 6.5 h día^-1. The average flock is 223.6 sheep (Table 3) and it is the group with largest size of the production units, even when the flock size is similar to farms classified as subsistence in countries with sheep production tradition. The average cost of production is $94o482.50 (Table 4), which is high compared to groups I and II. The main production cost is workforce (87.1 %) and the purchase of external inputs was scarce (9.2 %). The production form in the mountainous production units is dependent on the use of native grasslands and does not agree with what is described for farms with similar dimensions studied by ^{Milan et al. (2011}) in Spain, where it was found that 68 % of the farms purchase half of the fodders and 87 % purchase more than half of the concentrates.

Because of the size of the production unit, the fattening of weaned lambs was carried out for 7.5 months, until reaching a weight of 42.9 kg (Table 3), using the agricultural inputs produced. The total annual gross income is $68 503.40 for the sale of 37 standing sheep, representing a low efficiency of the production units. However, with the income from the sale of lambs the value of the annual rural food basket is covered for three members of the family (^{CONEVAL, 2016}) and, likewise, it covers the income from the sale of workforce of two people with a salary of $120.00 day^-1. The production units in this group can be supported in their management process to improve the efficiency of production because they have the resources and a flock size that is within the minimum dimensions found in countries with sheep production tradition, and are in possibilities of covering the welfare line of members of the families devoted to sheep production.