The traditional agriculture of México

México devotes 32 million hectares (md ha) to labor land, of which 6.3 million have irrigation and 25.7 million are cultivated under rainfed conditions. 66% of these labor lands are managed in small production units with less than 5 ha each (^{Robles, 2007}). In them, mestizo producers and from 62 ethnic groups practice the paradigm of traditional agriculture (PAT) -also known as peasant or subsistence agriculture- that focuses on the cultivation of maize. The seed of maize planted is almost exclusively native and self-produced, although “creole-like” varieties are also planted -product of the genetic interaction among native maize and improved varieties of different type, after a selection attached to criteria favored by the same producers.

The PAT has two sedentary variants and one itinerant: 1) simple maize cultivation, often in monoculture; 2) the historical milpa (cultivation of the association of maize with beans and zucchini and other species including arable); and (3) the roza-tumba-quema system, which is itinerant and is managed as a simple maize crop or as historical milpa. Typically, the two sedentary variants are handled annually, fertilized and agrochemicals are applied, albeit in a restricted manner.

The maize grain produced is self-consumed as human food and as fodder and stubble is used as fodder or as fuel. This agricultural paradigm is repeated throughout the national territory, including a great diversity of edaphoclimatic conditions (agro niches) associated with human activity. For each of these agro niches there are one or more agronomically adapted native maize breeds, which have also been developed organoleptically and nutritionally as human food. Altogether, the grains of these native breeds are irreplaceable for the elaboration of more than 600 preparations of the pluricultural Mexican cuisine. Also, the native maize breeds are, until now, agronomically irreplaceable in marginal agroeniches of low quality due to its soil and its climate.

Small producers are the mainstays of Mexican maize biodiversity, expressed in 60 native breeds and its varieties, which count for many thousands. Since the advent of maize, more than 6 750 years ago (Matsuoka et al., 2012) ago, 200 or more generations of Mesoamerican farmers have taken over its genetic improvement. They have been able to manipulate and conserve the vast biodiversity of the species genetic reservoir, both that inherited through the teocintle and its ancestors, as that which appeared from its interaction with the Mesoamerican inhabitant. The elements of this autochthonous genetic improvement (MGA) of maize have been described by Hernández X. in several publications (^{Hernández X. 1985}, ¹⁹⁸⁷, and ¹⁹⁹³). The MGA includes four key elements: 1) knowledge of the phenology of the native maize varieties and the crossing mechanism; 2) the introduction of allopatric parent materials (with which local maize would not cross due to geographical distance); 3) the selection of the seed made by the woman who is guided by morphological criteria of the grain, the cob and the totomoxtle -bracts that cover the cob- as proxi of agronomic adaptation and of quality for its consumption as food; and 4) the exchange of seeds within the local environment (sympatric parent materials), a practice that helps to avoid the inbreeding of the native variety.

Overall, Mexican farmers sow, observe and select the seeds every year from the productive yield among 10¹¹ and 10¹² (one hundred billion to one trillion) different maize genotypes in the agro global national ecosystem, the number of genotypes planted annually, is similar to that of maize seeds preserved in all germplasm banks in the world. This activity, which has been repeated since ancient times, can also be seen as a mega putative experiment of genetic improvement “in parallel”, where all genetic reservoir of the species intervenes in México. The authors of this paper presume, that is sought to develop materials more adapted to agroclimatic niches and with higher organoleptic and nutritional qualities, without sacrificing biodiversity.

This is one of the undervalued facets of traditional Mexican agriculture, but taken advantage of by the scientific community dedicated to the Mendelian and biotechnological genetic improvement of maize. However, in the near future, the Mexican maize biodiversity reserve and the MGA would be key to addressing the challenges of climate change, for food security in México and the world.

Almost half of the arable land of the country under rainfed conditions lies on hillsides with slopes from moderate of 4 to 10% to steep of more than 40% (^{Turrent et al., 2014}). Except for the roza-tumba-quema historical system, traditional hillside agriculture has historically been managed in México with an extractivist model. Water erosion has been a central factor in the degradation of hillside agricultural soils. Readings of soil loss by erosion have been reported of the order of 40 kg of soil per kg of maize produced in a slope of the humid tropics of México (^{Uribe et al., 2002}).

This hillside has a slope of 14.5%, is cultivated with rainfed maize, yearly plowing and burning the stubble; no protection practices against water erosion are applied; the average annual rainfall is 1 633 mm. This extractive soil management has been repeated for some time in the large majority fraction of hillside soils of the country, placing it in situations of fragility against torrential rain, high temperatures and disaster droughts that comes along with climate change (^{Easterling et al., 2000}; ^{Ahmed et al., 2009}). 23% of the cultivated land with maize under rainfed conditions are slopes with deep soils and 39% are slopes with shallow soils (^{Turrent et al., 2014}).

It is of high priority for the country to stop the degradation of these agricultural soils, as a critical option to aspire to food security in the face of imminent climate change. Along with the crisis of soil degradation by erosion, the economic crisis coexists, which is equally compelling for traditional small-scale agriculture. However, any technological option to modernize this type of agriculture would have to pursue multiple objectives, responding advantageously and simultaneously to ecological, socio-economic and cultural criteria.

MasAgro conservation agriculture

SAGARPA signed along with CIMMYT a founding agreement of the MasAgro Program, with duration of 10 years counted from April 2011 (^{Del Toro, 2012}). The project agreed on the collaboration between the two sides to achieve the following objectives in the second line of action, “international strategy to increase maize yield”: 1) sustainable modernization of traditional agriculture, turning it into the paradigm of conservation agriculture (PAC); 2) to increase domestic production of maize under rainfed conditions, among 5 to 9 million tons annually, and to increase yield from 2.2 to 3.7 t ha^-1; and 3) to replace between 1.5 and 3 md ha to native maize currently grown by improved maize varieties that are more yielding, drought tolerant and resistant to lodging. There are three more lines of action that can be consulted in ^{del Toro (2012)}. SAGARPA would contribute with $ 1 656 million pesos during the project period and would also support with its technical field staff. CIMMYT would provide the leadership and training of SAGARPA peasants and professionals participating in the project.

On April 2017, six years of implementation of the MasAgro program will have passed. Each year, CIMMYT publishes the progress of the program in some of its objectives. MasAgro had already included 592 municipalities in 30 states of the Republic in 2015, and had also certified 294 technicians, while 36 000 people had visited MasAgro’s electronic logbook. They had also obtained 6 800 maize genomic profiles from the genebank and 128 maize lines tolerant to drought had been identified (^{Martínez et al., 2016}). However, CIMMYT does not make publicly available or follow-up information relevant to the objective of increasing maize yield by 5 to 9 million tonnes per year in the traditional rainfed agricultural sector: i.e. a) coverage of small farmers (less than 5 hectares); and b) increases in maize yields and production, so central to the MasAgro-maize program.

Alternatively, Figures 1 and 2 show the public statistical series of national rainfed maize production, observed yields (^{SIAP, 2017}) and the projection of the positive trend of the years 1980-2010 (before MasAgro). This statistical series does not reflect further increases to this positive national trend in the first five years of the MasAgro program (2011 to 2015). What is shown is that the effects of MasAgro would be small compared to the annual variability associated with climate and other factors. In the same figures there are the projections of the explicit objectives of MasAgro-maíz as it was signed with the Mexican government. The difference is striking between what was projected in the Masagro program (low goal and high goal) and what is observed. It is obvious the disparity between what was offered and what has been achieved to this date. Such are the visible results that depend on a) the magnitude of the surface treated with MasAgro; b) the yields obtained; and c) its significance on the national scale. It could have occurred that the reached yields of rainfed maize were as expected or even greater, but that the treated surface was not significant at the national scale, the treated area was significant but the yields did not exceed the trend values or some combination of both.

Figure 1 National production of rainfed maize in the period 1980-2015 (^{SIAP, 2017}) and projections to 2020 year. 

Figure 2 National yield of rainfed maize in the period from 1980 to 2015 and projections to 2020 year. 

It should also be noted that CIMMYT’s publications minimize the technological demand of rotating crops with maize, which are required by the PAC. In order to meet such requirement, it would be necessary to make that technology explicit, since rotating crops would have to be as financially successful as maize, so that the PAC would become adoptable. In its absence, more than PAC, MasAgro would be impelling in practice the paradigm of conservation tillage, which is outside of the commitment.

Milpa intercalated in fruit trees (MIAF)

The development of the MIAF technology is a product of the collaboration of more than 30 years between INIFAP and the Colegio de Postgraduados in Agricultural Sciences. There are over 60 scientific, technical and broadcast papers on MIAF, where technology and advances are detailed (^{Cortés et al., 2005}; ^{Cortés et al., 2007}; ^{Francisco et al., 2010}; ^{Cortés et al., 2012a}, ^b; ^{Camas et al., 2012}; Santiago, 2014; Salinas, 2015; ^{Albino et al., 2016}; ^{Torres, 2016}). MIAF technology has two precursors, one is the intensification model of the historical milpa (MH) developed by traditional producers of the San Martín Texmelucan-Huejotzingo region in the state of Puebla, which introduces the cultivation of fruit trees in interaction with the Milpa (^{Cortés and Turrent, 2012a}); the second is the technology for living wall terrace slopes (^{Turrent et al., 1995a}, ^b, ^c). MIAF is a multi-objective technology that seeks to intensify the PAT to: 1) significantly increase net income and family employment, while continuing to produce its staples; 2) to protect the soil against erosion, without eliminating its breakage, except under special conditions; 3) to foster the interaction between the component crops, for a greater economy of the use of natural resources and the imported inputs to the plot; and 4) to increase the capture of atmospheric carbon.

We will examine the set of hypothesis and assumptions on which MIAF is based due to its relevance to the PAT’s intensification topic, particularly in relation to the first three objectives. In MIAF some properties of the historical milpa are adopted: i) its relative soil efficiency (ERT) (^{Mead and Wiley, 1980}) superior to the unity; ii) its biodiversity as a factor of cultural and agro-climatic resilience; and iii) the cultivation of native seeds as a source of the family staple food.

Relative soil efficiency (ERT) greater than unity. Compared to a single crop, the structure of the vegetative canopy of the MH is complex, consisting of several foliar strata occupied by interacting species, adapted to each of them and with adequate population densities to achieve a high joint benefit of incident photosynthetically active radiation (RFA). A wider exploration of the soil is also achieved. This system gives the MH a relative efficiency of the soil greater than unity. In MIAF, the same effect is approached with a foliar canopy structure of three interacting strata: an epiculture (fruit), a meso culture (maize) and a single crop (shrub beans or other species of low size). Population densities have to be optimized for this system.With the aim of collating the hypothesis that higher ERT values than the unit can be achieved with an ad hoc topological arrangement different from the MH, two experiments on MIAF were conducted in the Campo Experimental Valle de México, one under rainfed and the other under irrigation, in the period 2002 to 2005.

The total space of a MIAF module was assigned in thirds of six rows each, peach, maize and shrub beans. The central third was assigned to a row of peach and the two flanking thirds to maize and beans, these, in six alternating strips of two maize rows and two of beans. Figure 3 shows the topological arrangement of MIAF and the penetration of RFA in maize, favored by the flanking bean strips.

Figure 3 MIAF-peach tree module with topological arrangement of two alternating maize and bean rows and a peach row under irrigation. CEVAMEX, 2004 

Table 1 shows that ERT values above the unit can be achieved in the MIAF. In the annual crop component of MIAF under irrigation, the ERT averaged 1.494 -addition of the partial relative efficiencies 0.875 in maize and 0.619 in beans- and under rainfed it was equal to 1.539 -addition of 0.906 plus 0.632 (Table 1). The implication for the rainfed case is that 0.906 ha of simple maize crop, plus 0.632 ha of simple shrub bean cultivation, would be required to match maize grain and bean crop yields of one hectare grown with the crop component of MIAF. Several authors have reported ERT values above unit in MIAF under different conditions: ^{Cortés and Turrent (2012a)} under rainfed in Puebla, ^{Albino et al. (2016)} under irrigation in CEVAMEX, ^{Camas et al. (2012)} in Chiapas under rainfed and ^{Torres (2016)} under rainfed in Oaxaca.

Table 1 Average yields of maize grain and bean and relative efficiency of the soil: MIAF in relation to simple maize and bean crops, under irrigation and rainfed in 2002-2005. Campo Experimental Valle de México. INIFAP. 

† = cálculos a partir de datos de los Cuadros 2 y 3 de ^{Turrent (2005)}; ERT= calculó ignorando al epicultivo, debido a que entró en producción hasta el último año; ‡ = los rendimientos de MIAF se expresan sobre la base de media hectárea: t (0.5 ha)^-1 que es el área efectiva ocupada por cada uno de los dos cultivos en el arreglo topológico de tiras de dos surcos de maíz, alternando con tiras de dos surcos de frijol arbustivo; § = cultivo simple, los rendimientos de cultivo simple se expresan en t ha^-1, que es el área ocupada por cada especie. §§= eficiencia relativa parcial del cultivo, dividiendo el rendimiento en MIAF en media hectárea entre el rendimiento del cultivo simple en una hectárea; ERT= eficiencia relativa de la tierra que es la suma de las ERP de ambos cultivos.

MIAF with native maize. ^{Torres (2016)} conducted a MIAF experiment under rainfed for six years in the Sierra Mixe of Oaxaca. The experimental site is a hill with slope of 29.8%. The soil order is Acrisol, with pH less than 5, with very low content of organic matter and traces of assimilable Phosphorus. The Olotón maize breed, adapted to this edaphic condition, was cultivated. Lime was not applied to correct soil pH. The topological arrangement of maize and beans was of a maize groove alternating with a bean groove. Maize occupied 30% of the space of the MIAF module, bean 30% and peach 40%. The experiment included five population densities of maize, from 13 333 to 26 700 total plants per hectare, occupying only 30% of the space.The author did not report lodging problems. Optimal yields ranged according to the quality of rainfall and the relative position of the groove on the terrace, 1.48 (0. 3) t ha^-1 to 3.02 (0.3) t ha^-1.

Protection against erosion. The elements of living wall terrace technology are also adopted in MIAF to protect the soil against erosion: i) trees planted in contour 1m apart, as a support for the terraces; ii) runoff filter to reduce runoff water velocity and promote sediment deposition; and iii) “downward spreading” to encourage the gradual development of terraces. ^{Uribe-Gómez et al. (2002)} -in an experiment on alive wall terraces- and ^{Camas et al. (2012)} -in an experiment on MIAF-, compared runoff and sediment dynamics under traditional tillage (LT), living wall terraces (TMV) and conservation tillage (LC), or under MIAF on slopes of Mexico’s sub-humid tropics. ^{Uribe-Gómez et al. (2002)} reported drainage coefficients (CE) of 31%, 15% and 17% respectively for LT, TMV and LC in an Entisol with slope of 14.5%. The sediment losses were 199 t ha^-1 year^-1, 3 and 1, respectively.

On a 3 years study, ^{Camas et al. (2012)} compared MIAF, with live barriers and with LC in micro-watersheds of 50 or more meters in length and slopes greater than 30%, under non-breaking conditions. Reporting EC of 12.4%, 13.15% and 18.6% respectively for MIAF, living barriers and LC, and erosion losses of 5.8 t ha^-1 year^-1, 6.3 and 16.8. The MIAF efficiency to stop water erosion is not affected by the slope length. This does not happen with conservation agriculture (^{Liu et al., 2000}; ^{Roose and Barthes, 2001}). Figure 4 shows views of a permanent experiment on rainfed MIAF, established in 2004 on a Vertisol with slope of 18%. The dimensions and topological arrangement of maize and native beans in the spring summer cycle are the same as those described in the experiment of CEVAMEX (Figure 3). In the autumn-winter cycle, native early maize is grown under single cultivation.

Figure 4 Research-demonstration site with MIAF-chicozapote in 2013, planted in 2003, on a hill vertisol with slope of 18% in the Los Tuxtlas region, Veracruz. 

Increase in family income. The inclusion of fruit trees cultivation in MIAF enriches the milpa in three areas: a significant increase in net family income, soil protection against erosion and localized increase in soil organic matter content -the latter associated with the anual installation of a runoff filter. By design, the fruit trees occupied a third of a hectare of MIAF on shallow slopes, in which is located a population of 695 fruit trees. In the remaining two-thirds of the MIAF hectare 30 000 maize plants and 80 000 bean plants are located. The total populations of the epiculture, mesoculture and sotoculture approximate the total populations of two hectares managed in a conventional way, one with fruit trees and the other with milpa. With this intensification there is an advantage of the interaction between botanically different species and complementary to each other, in the use of space and time.

The inclusion of fruit trees substantially increases the value of production as explained by ^{Cortés and Turrent (2012a)}. An economic and financial analysis ex ante the application of MIAF in the Northern Sierra of Oaxaca, covering a period of 15 years has been published (^{Jiménez et al., 2016}). These authors report the following parameters of the MIAF financial analysis per hectare, in a microbasin of the Sierra Mazateca, Oaxaca: VAN (net present value) equal to $53 714 (constant pesos of 2004), TIR (internal rate of return) equal to 20.68% and R (B/C) (benefit/cost ratio) equal to 1.49. The parameters of the financial analysis were calculated from costs and income of MIAF additional to those of the traditional milpa. The most expensive item of MIAF was the purchase of 1 000 grafted peach seedlings, equal to $ 30 000.00. However, this investment could be largely replaced by self-employment and time. The cost of the same 1 000 seedlings was $ 5 080.00, self-producing them in miniroots and grafting them a year later in the field.

MIAF’s technology transfer strategy. This technology is designed to intensify the paradigm of traditional agriculture with simultaneous emphasis on sustainability, family income and employment, biodiversity and interaction with peasant technological resources. Its technological domain is that of small production units. The technology is intense in knowledge and its transfer is complex and long term.Even so, the response of small producers has typically been enthusiastic, when they have had access to observe the MIAF unit in full operation. The fruit-culture component of MIAF and at the same time, the emphasis on producing its native crops on the same plot has been the intensifying element that has attracted them the most.

It also operates in favor of the adoption of MIAF, a) its compatibility with the paradigm of traditional agriculture; and (b) promising not only sustainability of resources but also substantial increases in income and family employment. Just like all technological knowledge in the process of development and scaling, it is necessary to accompany it with adaptive research of agronomic, social and yield types, as it is derived from the experience of Plan Puebla (^{CIMMYT, 1974}). There are several formidable challenges for the transfer of MIAF technology, but they are not insurmountable, i.e. a) the intensity of technological knowledge; b) the investment in fruit trees at the beginning, and consistent and adequate sources of financing and technical assistance; and c) access to the fresh fruit market. MIAF transfer experiences can be consulted in ^{Ruíz-Mendoza et al. (2012)}; Zambada-Martínez et al. (2013). These are cases in which MIAF research was associated with its transfer.

There are also cases of spontaneous adoption of MIAF by producer organizations that already had experience in financing and marketing, and that were looking to intensify their production or to change the type of their agricultural activity. Table 2 shows relevant information on three cases of spontaneous adoption of MIAF in the state of Chiapas, México. These adoptions have occurred as a result of visits by producers and managers of organizations to MIAF research-demonstration sites.

Table 2 Cases of spontaneous adoption of MIAF by pre-existing organizations of producers in the state of Chiapas. 

§ = Colectivo ISITAME, A. C. Transformando realidades. www.isitame.org. MC. Yolanda Romero Alvarado. Tel. (01) 961 1470630. Prolongación de la 2d ª Poniente Norte 1804. Col. Penipak Norte, Tuxtla Gutiérrez, Chiapas. CP. 29039. §§= Ramal Santa Cruz, S. P. R. de R. I. Gerente Rigoberto Velasco. Av. Julián Grajales Núm. 430, Chiapas de Corzo, Chiapas. CP. 29160. † = Proasus S. A. de R. L. Lic. Adolfo Ocampo Guzmán. Tel. (044) 9671141667. Cerrada Jovel núm. 3-1, Col Fstse 2000. San Cristóbal de las Casas, Chiapas.

Subsequently, in response to requests, the INIFAP-COLPOS cooperative program has provided technical advice on MIAF. This adoption path is based on the merit of technology and its affinity with the paradigm of traditional agriculture. The collaborative program of INIFAP and COLPOS has established 12 MIAF research-demonstration sites in several regions of the country, in collaboration with producers and in two cases, on INIFAP’s land. The establishmet period of these sites has been mostly from 4 to 8 years. Only two of them have remained for more than 10 years. The permanent site of Los Tuxtlas (Figure 4) inspired the Isitame project in Chiapas, while the Santa María Tlahuitoltepec, Oaxaca, experimental site, inspired the two Ramal Santa Cuz and Proasus, Chiapas projects, (Table 2). There are lessons from this adoption case, which could support the design of an ambitious State program to intensify the PAT at the national level.