Introduction

The conservation agriculture (AC) is an alternative to improve both sustainability and profitability requirements, risk and externalities of agricultural crops. The sustainability of agriculture as ^{Kassam (2009)} is challenged by the rising cost of food and energy, water scarcity, degradation of ecosystems, climate change, biodiversity, and the financial crisis. In this context (^{Scopel et al., 2013}) are required to be more profitable crops resistant to adverse environmental conditions, low risk, based on low-cost technologies and minimum negative externalities.

For some authors (Cadena et al., 2009; ^{Stagnari et al., 2010}; ^{Kassam and Friedrich 2011}; ^{Ndah et al., 2013}) the AC is a combination of agronomic, biological and mechanical measures that improve soil quality; for others (^{Knowler and Bradshaw 2007}; ^{Hobbs et al., 2008}) it is the formal integration of practices by farmers prior to the emergence of the AC as a form of production. Thus, the AC can be understood as an innovation in the production process, or as the integration of common practices carried out in isolation.

The three practices that integrate the system under AC are the minimum soil disturbance, permanent soil cover and crop rotation (^{Govaerts et al., 2009}; ^{Moreno et al., 2011}). The minimum clearance refers to direct seeding to avoid mechanical soil disturbance. The permanent coverage is the use of crop residues and cover crops to prevent erosion and moisture loss. The crop rotation guides the selection of the main crop species for the purpose of fixing nitrogen and breaking the cycle of pests and diseases. The success of AC system is related to the simultaneous implementation of the three principles that integrate and time in which they have been applied. Perform only one of the practices mentioned it is not considered in this research as a process innovation but as the adoption of a technological practice.

Globally the adoption of conservation agriculture (AC) is 7% of the agricultural area as ^{Kassam (2009)}; in southern Latin America, the AC is practiced in 60% of the agricultural area, mainly between commercial and large farmers; in Mexico the adoption of AC has been limited, and to the 2008- 2009 cycle covered approximately 1% of the agricultural area, equivalent to about 22 800 ha.

Among the factors contributing to the low adoption of AC on producers is their business orientation, objectives expected in the short term, economic constraints of farmers, as well as the norms, culture and perceptions of farmers (Derpsch and Friedrich 2008; ^{Baudron et al., 2009}; Stagnari and Abstract, 2009). The decision of farmers to adopt new agricultural technology in preference to alternative depends on complex factors and the perception of the producer to adopt another technology that has been little studied. ^{Negatu (1999)} notes that knowledge of the factors that influence perceptions facilitate improved development and transfer of appropriate technologies.

According Feder and Umali (1993), the adoption process is based on a sequence of decisions that individuals make to adopt or reject an innovation. For some authors (^{Knowler and Bradshaw 2007}; ^{Hobbs et al., 2008}; ^{Baudron et al., 2009}; ^{Moreno et al., 2011}) there are many factors that influence the adoption of agricultural innovations, of which the following are mentioned: farmer's age, education, generational renewal, capital, management capacity, availability of equipment, type of land ownership, farm size, crop yields, crop yield and soil type. To these factors we can add access, both in quantity and quality of information. Increased access to quality information will increase the adoption of an innovation.

Most investigations have focused the study on the adoption of innovations, according ^{Ghadim (1999)} have been concerned about answering questions: (a) what determines whether a particular producer, adopt or reject an innovation; and (b) that determines the pattern of diffusion of innovation through the population of potential adopters. ^{Negatu (1999)} notes that the conceptual models used to explain the decision of farmers to adopt new technologies can be classified into three groups: (i) the model of innovation-diffusion in which technology is transferred from its source (the systems research) to end through an intermediary agent (extension systems) and its diffusion potential users-users communities depends mainly on the personal characteristics of individual potential user; (ii) the model of economic constraints where the distribution of the allocation of resources among potential users in a country or region determines the pattern of adoption of technological innovation; and (iii) context model technology user characteristics which integrates approaches assume that the characteristics of the agro- ecological, socio-economic and institutional contexts of technology users to play the central role in decision-making and process broadcast.

The theory of diffusion of innovations Rogers (2003) defines diffusion as the process by which an innovation is communicated in time and broadcast by certain channels, among members of a social system; and adoption as the process in which an individual or production unit goes from having a first knowledge of innovation, the formation of an attitude toward her, the decision to approve or reject your application, and confirmation this new idea. Engel and Salomon (1999) point out that the key to innovation, including the dissemination and use of innovations arising from others is the quality of the interaction between farmers, businesses, donors, researchers and governments which cannot be considered as an individual competition, nor as the sum of a series of individual skills; instead, it should be seen as a social competition, something shared by all those individuals, institutions and organizations interested in rural development.

In this approach Koschatzky (2002) points out that the network approach innovation explicitly recognizes that innovation, production and marketing of a product cannot be carried out by a single company, but only in collaboration with other agents and as a result the interaction thereof. An innovation network is a set of actors (individuals, companies and institutions) located in a territory to promote individual and collective development through the pooling of resources for the generation of value. A network consists of nodes representing actors and ties representing social and trade relations techniques. The network approach is used in recognition that innovation is a social process.

The aim of this study was to analyze the relationships between producers to identify both the diffusion of innovation related conservation agriculture as the source of knowledge of these innovations. This analysis aims to identify whether producers seeking more information and adoption the AC, they are also producers referred by their peers as relevant sources of information.

Materials and methods

The state of Tlaxcala is divided into three Rural Development Districts (DDR); the Distrito163 based in Calpulalpan, with 11 municipalities, representing about 25% of the area with maize; 164 district based in Tlaxcala with 36 municipalities 35% and 165 district located in Huamantla which includes 13 municipalities 40%. The state of Tlaxcala, is referred to as a state with the greatest genetic diversity of native maize and in turn as a state with the highest percentages of degradation soil.

The information used in this research corresponds to the database composed of 461 observations generated mapping innovation networks in the Hub highland underwritten the program MasAgro adopted the AC system. The information was collected in the months of september to december 2012 by certified technicians and technical MasAgro working in the Support Program for the Productive Chain Maize and Bean (PROMAF). The survey was divided into three sections; He considered the first producer identification and cultivation: name, age, level of education, culture, varieties, planting density, yield, soil type and irrigation, use of machinery; the second section corresponded to the type of innovations that are used and who were learned (technical network); who considered the third producer (social network) as well as who buys inputs and who sells his production (commercial network) relates.

Field information was integrated into a database in Microsoft Excel^® version 2010. The data were analyzed using descriptive statistics using the SAS statistical package.

For the analysis of the network typology of producers in which four categories were defined according to three basic principles of AC, integrating a total of 461 observacioens it was developed. The typology of producers as ^{Duch (1998)} is a conceptual and analytical means, which includes the agricultural production units in sets with similar characteristics, identifies and accurate technical, economic and social problems of each type of producer and contributes to understanding the structure and regional organization for agricultural production, its relationship to society and state agencies.

For this research, the type A corresponds to the application of the AC line with the three principles that govern it. This population therefore deserves special attention within the research. Type B performs two of the three practices AC and producers are considered in the process of the complete adoption of the AC. In type B producers who made two of the three practices, regardless of which of these practices they are adopted integrated.

For network analysis, centrality measures (Koschatzky, 2002) were applied, and key players (^{Borgatti, 2006}). The degree centrality measures used are the degree of input and output level. The degree of output indicates the number of relationships that actors say they have with the rest of the network. The degree of input is the number of ratios referring to an actor by other actors. Both indicators can be expressed in standardized form, indicating the relationships present in relation to the possible.

An actor with high output is identified as an information seeker; an actor with high entry refers to a source of information. In terms of diffusion of innovations, an actor with high output could refer to a time convenient to integrate it into a process of access to knowledge such as demonstration events or training in field situation producer. Meanwhile, an actor with high input is a source of information, which other producers say learn. Both indicators, input and output level, refer to the immediate ability of the actors to collect or provide information. However, it should be considered that both indicators do not consider network coverage supplier or search for information.

Table 1. Types of producers on the basis of the three basic principles of AC. 

The indicators of key players based on ^{Borgatti (2006)} allows the identification of source nodes (harvest) and collector nodes (diffuse) considering coverage; that is, these nodes consider their access or scope to the entire network. The source actors are those who receive information requests from other actors, demonstrating prestige as information providers to present the highest coverage from its entry level relative to its position in the network. The collectors actors are identified from the output degrees depending on their position in the network. A collector is an actor in search of information for decision-making or validation of achievements. For identification of source and collectors actors, being a network indicator and not only node, it is necessary to estimate the coverage of each actor or group of actors. To calculate coverage actors source and collectors proposal ^{Borgatti (2006)} under the following calculation method was used.

Where: R= abbreviation range; d_mj= sum of the inverse of the distances between each actor (d_{mj- 1}) and the rest of the network; and N= total number of nodes in the network.

The network data were captured in Microsoft Office Excel 2010 and codified in Microsoft Notepad version 6.1. The file generated in the notebook was exported directly to the plotter NetDraw 2,097. For the calculation of the indicators of centrality was used Ucinet 6,288. Actors for identifying the software Key Player 2. Statistical analysis was performed using the SAS program using routines comparing means for the six types of producers was used.

Results and discussion

The analysis of the database reflected an average age of 47 years and an average schooling of 7 years. The planted area is 4.3 ha with a minimum of 0.5 ha and a maximum of 28 ha. The 98.3% of the surface is temporary and 1.7% of irrigation. The average yield is 0.86 t with a minimum of 0 and maximum of 7.5 t. The 74.07% of producers have ejido land, 87.36% do not perform any cycle AC, 12.20% of the producers are in year zero (start conversion process) AC and 60.57% of the machinery uses is rented and only 23.75% own. These characteristics of the producers located as producers typical of the area, even though not a statistical sampling method was used. The selection of producers interviewed producers are assigned by the program as population MasAgro meet criteria according to geographical location and the integration of a list of producers willing to receive technical assistance.

The degree of output standard-type producersA(performs all three practices AC) it is 65%, reflecting that such producers are looking for more technical information and are classified as collectors actors or information seekers. In turn, these players receive 3% type A entries or information queries. That is, are the actors more information search (p< 0.05), but not similar in character as sources of information (p< 0.05).

His character information seekers contributes to understand why present three innovations related to AC. As noted by Engel and Salomon (1999) interaction with other actors favors an innovation to be adopted. However, in type A players may question their ability to spread the AC in both its degree of consultations by other producers is similar to other producers.

Producers of the type D (practice makes no AC) are those that reflect the lower level of coverage as collectors of information (p< 0.05), which ranks as which type of producers with lower information search. The tillage practices that perform are conventional and have been learned by parents, relatives and others close to them producers.

The Ssources of information producers for the adoption of AC no statistically significant differences (p<0.05) between types of actors. The main sources were the providers of professional services (42%), family (15%), and educational and research institutions (12%). This dependence on external sources of information (professional service providers and educational and research institutions) reflects on the one hand an immature to rely on external information for network development. On the other hand, it realizes a structure with weak bonds that targets a heterophilic network that allows access to new knowledge.

In a broadcast network innovation, the role of producer organizations and government institutions, as organizations representing both sectors could play a more active role in the diffusion process. However, when analyzing the information sources AC process, it is observed that these are the minor mention. Information sources AC with further reference by the producers are the National Institute of Forestry, Agriculture and Livestock (INIFAP) and the International Center for Maize and Wheat Improvement (CIMMYT).

^{Rahm and Huffman (1984)}; ^{Westra and Olson (1997)} point to sources of information that positively influence the adoption may include other farmers, media, meetings and extension officers (PSP). With respect to the latter source, ^{Agbamu (1995)} shows that if information dissemination is ineffective, inaccurate or inappropriate by the extension agent no adoption of new practices.

Table 2. Measures of central and key player by type of producers in the diffusion of innovations. 

Figure 2. Sources of information for learning practices AC. 

Figure 3. Institutions participating in the adoption of AC. 

The interaction between producers (^{Parra-López et al., 2007}; ^{Calatrava and Franco-Martínez, 2011}) is considered as a factor favoring the adoption of an innovation. This research highlights that the current network structure geared more to dependence on external entities, so the interaction between producers is an element to be developed.

Conclusions

The adoption of agriculture conservation occurs in producers with greater search capabilities information. A greater interaction with external actors such as teaching and research center providers trained and professional services, the adoption of the three principles of conservation agriculture is higher. However, these producers searchable information and adoption of these principles, are not mostly referred by their peers as sources of learning. This contradiction between "that more does not teach" is the main challenge to the dissemination and appropriation of innovations.

Future research could focus on understanding why a producer adopter is not referred to by their peers as a source of information. This analysis will help design extension process indicators as to how to operate programs with an extension component design and targeting strategies to increase the adoption of innovations.