Study area

Fieldwork was conducted in El Ejido la Encrucijada 3^rd Section (Las Calzadas), located in the municipality of Cardenas, Tabasco, located at 18° 15' 25'' north latitude and 93° 33' 16'' west longitude, it has 1 505 inhabitants. The topography is a floodplain, with an altitude of 5m and its current vegetation is grassland (^{INEGI, 2010}).

The climate is warm humid (Am) with abundant rain in summer, this climate covers the Tabasco plain, both the base and the east slope of the Sierra Madre Oriental and on the Pacific slope of the southeastern portion of the Sierra Madre Chiapas (García, 1973). The average annual temperature is 26 °C with an average annual rainfall of 2 500 mm. The dominant soils correspond to GLvr/2 (Gleysol Vertic) group of medium texture (^{INEGI 2012}). The agricultural potential use is for corn, beans, rice, watermelon and its livestock potential for cultivated pastures with agricultural machinery (^{INEGI, 2012}).

Sample size

In a preliminary tour in the locality, informal interviews with local authorities and use of sampling frame, a population of 92 family orchards was obtained. The sample size was obtained using the formula of maximum variance proposed by Abdoellah et al. (2006) to obtain 29 family orchards.

Where: n= number of samples; N= number of homes with orchards in the study area (92); Z= value of a normal distribution Z_α/2 (1.96) for a confidence level of 95%; p= probability of success (0.5); and d= sampling error (0.15).

Subsequently it convened a meeting with the 29 families in a place known for the locals to seek their support and consent, explain research objectives, stages and activities to do in their orchard. To locate the selected sampling units, a freehand map was developed with the help of the local authority.

Data collection

Data collection took place from 2013 to 2014. First a tour to each orchard was conducted in the company of family head to delimit the area of the orchard with respect to other production systems; then measurements were made in each orchard, including orchard surface and dasometric mearurements to trees, shrubs, perennial grasses and palms. Total height (m), diameter at breast height (1.3 m from the base of the trunk or stem) and crown cover (m²) were measured (^{Matteucci and Colma, 1982}; Lopez et al., 2012). Upon completion of measurements in each family orchard, a plant mapping was performed to identify potential management areas, which consisted roughly on locating each species in the garden, taking as reference the owner's house (Gliessman and Somarriba, 1984).

In the second stage a quadrant of 16 m² was established on each family orchard, consisting of areas destined to annual herbs combined with other life forms, measuring only the overall height and canopy cover. The criterion to determine the size of the quadrant was: plant density, which was supported with a basic question addressed to the owner: What is the main area of herbs? the area allowed to homogenize and compare (^{Mostacedo and Fredericks, 2000}).

Orchard surface was measured with a tape measure 50 m. The DAP was measured with a diameter tape, total height was measured with a graduated pole each meter for individuals under 4 m, and with Haga gun for individuals above 4 m. Crown cover was measured with a tape measure, considering two measurements (semi minor axis and semi major axis). Annual grasses were measured with a tape measure. Trees and shrubs that had several branches before 1.30 m from the ground, each branch is considered as an individual and was listed consecutively with an ink marker for subsequent measurement. The quadrants were drawn with a graduated pole each meter and the ends were marked with small stakes. Attribute such as common name and biological form were recorded. Plant measurements were carried out following the direction of clockwise around the owner’s house and temporary subunits or "tasks" to measure different size were established; to keep track on the record and avoid repeating mistakes. Observations on plants distribution and outstanding data were recorded in a field notebook.

Vegetation structure

The vegetation structure was analyzed through the importance value index (IVI). IVI was developed by ^{Curtis and McIntosh (1951)}, which is represented by separate relative dominance, density and frequency values for each species. The sum of the relative values gives a constant of 300 indicating the importance value, where few species achieve high levels of importance. This index measures the ecological importance of the species and their dominance hierarchy in a mixed stand (^{Zarco et al., 2010}).

Where: Dr= relative density; Domr= relative dominance; and Fr= relative frequency.

Dominance (biomass estimator: basal area, coverage) is obtained as follows:

Where:

Basal area (AB) of trees is obtained with the following expression:

Where: DAP= diameter at breast height (cm).

Relative density is calculated as follows:

Where:

Relative frequency is obtained with the following expression:

Where:

Coverage was estimated using the formula for an ellipse:

Where: A= semi higher axis; and b= semi lower axis.

Coverage of a species is the proportion of land occupied by the perpendicular projection of the aerial parts of the individuals of the species considered (^{Matteucci and Colma, 1982}).

At the end of data collection in field, botanical collections from plants not identified by their common name in field were taken for identification using taxonomic keys (Lot and Chiang, 1986) and with the invaluable support of Dr. Angel Sol Sánchez from the Colegio de Postgraduados, Campus Tabasco. The information was confirmed using scientific sources of information on family orchards.

Classification of family orchards

Family orchards have a size that does not exceed one hectare (^{Fernandes and Nair, 1986}), although there are exceptions (^{Henry et al., 2009}), indicating the nature of subsistence or traditional (^{Peyre et al., 2006}). Orchards areas recorded in this study are similar to that reported by ^{Romero (1981)} which was 300 m² to 2 268 m². However, it differs with the work done by Van der Wal et al. (2011) in five physiographic regions from the state of Tabasco, where the average was 1 710 m², thought noted that the orchards from the coast and highlands were larger than the flood plains and low hills.

Small orchards

These orchards cover an area of 293.93 m² to 494.41 m² and had an average species richness of 22.4. The most diverse botanical family was Musaceae with five species, such as rogue banana (Musa paradisiaca L.), square banana (Musa balbisiana Colla), Dominican banana (Musa paradisiaca L.) and apple banana (Musa sapientum L.) and being the most abundant Valery banana (Musa acuminata Colla). The diversity of bananas may be due to the following advantages: these are preferred species for the families, fast-growing and produce permanently, their production cycles overlap allowing to continuously count with fruit for consumption (Gliessman and Somarriba, 1984). The average coverage of the plants was higher than 70% (314 m²). Coverage seems high; however, when competing for space with houses, it shows the fragmentation process, which is related to family orchards for being located near the city. The total basal area of these orchards was 3 m², excelling timber species such as Cedrela odorata L. and Mangifera indica L. A correlation test between the number of species and the size of small orchards was positive, species richness increases with increasing size (Figure 1).

Figure 1 Correlation between richness / size of small family orchards. 

Medium orchards

The sample area was between 509.04 m² to 920.18 m², with an average richness of 19.42. The richest botanical family was Fabaceae with seven species, including chipilco (Diphysa robinioides Benth.) Cocoite (Gliricidiasepium (Jacq.) Kunth ex Walp.), Cuinicuil (Inga jinicuil Schltdl.), guayan (Cassia fistula L.), pata de vaca (Bauhinia variegata L.), tamarindo (Tamarindus indica L.) and zapatito (ternatea Clitoria L.).

The average vegetation cover was 335 m² which accounted for 45% of the total in these orchards. There were two family orchards with low coverage (12% and 4% respectively) affecting the average. The total basal area was 6.64 m², where 0.37 m² corresponded to timber species such as Cedrela odorata L., Tabebuia rosea (Bertol.) A. DC., and Mangifera indica L., representing 18% of registered plants. A correlation test between the number of species and size of medium family orchards was positive, the number of species increases with increasing size (Figure 2).

Figure 2 Correlation between richness / medium sized family orchards. 

Regular orchards

The sample area ranged from 1011.35 m² to 1498.7 m², an average of 20.8 species. The most diverse botanical families were Rutaceae and Malvaceae with seven species each. The Malvaceae family were cacao (Theobroma cacao L.) which was the most abundant, followed by tulip (Hibiscus rosa-sinensis L.), ceiba (Ceiba pentandra (L.) Gaertn.), cork (Ochroma lagopus Sw.), guacimo (Guazuma ulmifolia Lam.) majahua (Hampea macrocarpa Lundell) and sibil (Malvaviscus arboreus Cav). The Rutaceae family were lime (Citrus limetta Risso), lemon (Citrus limon (L.) Burm. F.), Mandarin (Citrus reticulate Blanco), orange (Citrus sinensis (L.) Osbeck), orange pork (Citrus aurantium L.) grey orange (Citrus grandis (L.) Osbeck) and rue (Ruta chalepensis L.).

From these orchards and with the presence of species like Ceiba pentandra (L.) Gaertn, Ochroma lagopus Sw., Hampea macrocarpa Lundell neighboring with the cocoa agro ecosystem, represented the association that make up both agro-ecosystems, as well as plots where corn, beans and grass is grown to feed cattle. The closeness between the different production systems allowed us to expand the richness and floristic composition of family orchards.

Plant coverage had an average of 709 m². Although this group consisted of more family orchards (10), its coverage was low (x-= 56%) compared with the other groups, indicating poor use of space. The basal area of all plants recorded in this group was 22 m², a very high value compared to other groups, this was due to timber and woody species were more abundant (5.01 m²) compared with the other groups, which included species like Mangifera indica L., Tabebuia rosea (Bertol.) A. DC.), Cedrela odorata L. and Ceiba pentandra (L.) Gaertn).

Large orchards

The surface of these orchards covered 1 545.78 m² to 1 756.8 m², with an average of 34.75 species, which made them the most diverse. Rubiaceae was best represented, comprised by species like gardenia (Gardenia jasminoides Ellis J.), garden (Ixora coccinea L.), coffee (Coffea arabica L.), castarrica (Alibertia edulis (Rich.) A. Rich. Ex DC.) jaule (Genipa americana L.) and noni (Morinda citrifolia L.). The presence of coffee (Coffea arabica L.), castarrica (Alibertia edulis (Rich.) A. Rich. Ex DC.) and jaule (Genipa americana L.) indicated that as orchards size increased, recorded less common species or rare from floristic components of the cocoa agro-ecosystem, as it happened in regular orchards. Rare or low abundance species indicate the ability of family orchards for biodiversity conservation (Kabir and Webb, 2008), but also it is due to farmers do not focus on the production of these species to be exploited.

The average vegetation cover was 82%, which showed high efficiency in the use of space compared to other groups of orchards. The total basal area of these orchards was 14.4 m² where Theobroma cacao L. contributed most due to its high abundance.

Very large orchards

The surface sampled ranged from 2 214.31 m² to 2 866 m² and on average 31.6 species, placing them as second group with the largest richness. The best-represented botanical families were Musaceae (six species), followed by Fabaceae and Anacardiaceae with five species each. Average crown coverage was 64%, indicating an efficient use of space and not in fragmentation process. The total basal area of all these orchards was 14 m², of which 2.59 m² corresponded to timber species, being the most abundant Tabebuia rosea (Bertol.) A. DC., as well as Cedrela odorata L., Ceiba pentandra (L.) Gaertn., Mangifera indica L., and Colubrina arborescens (Mill.) Sarg.

Vertical structure

The plants from family orchards exhibited a four layers vertical structure, being as follows: lower layer of 0-5 m, subdivided from 0 to 1 m, consisting of 311 individuals, 77% were annual grass, from 1.1 to 3 m, composed of 568 individuals, highlighting perennial grasses (46%) like valery banana and square banana as the most abundant, trees (35%) and bush (15%). A third sub-layer from 3.1 to 5 m, comprised by 558 individuals being tree the most abundant biological form (71%), represented by Theobroma cacao L. and Tabebuia rosea (Bertol.) A. DC. This layer recorded 164 species, therefore orchards with higher richness and abundance (1 437 individuals), representing 80% of 203 species and 73% of 1968 plants respectively. For the type of species that excelled it can be considered the most productive or at least their production cycle is short and constant management. ^{Das and Das (2005)} placed bananas and vegetables near the houses, and identified their management and harvesting is easier. A second layer from 5.1 to 10 m consisting of 387 individuals; from this layer, the trees were the most abundant (89%) and the physiognomy was complex and dynamic for light and shade requirements, highlighting species like Theobroma cacao L. as the most abundant, Annona muricata L., Mangifera indica L., Byrsonima crassifolia (L.) Kunth and timber such as Tabebuia rosea (Bertol.) A. DC., and Cedrela odorata L.

The third stratum, from 10.1 to 15 m, which was dominated by fruit trees such as Persea americana Mill., Inga jinicuil Schltdl, Mangifera indica L., Tamarindus indica L., Artocarpus altilis (Parkinson) Fosberg, Byrsonima crassifolia (L.) Kunth, Manilkara zapota (L.) P. Royen, timber as Cedrela odorata L. and palm (Cocos nucifera L.) and Roystonea regia (Kunth Of Cook). Finally a layer of 15.1 to 20 m also dominated by trees (75%) such as Artocarpus altilis (Parkinson) Fosberg, Manilkara zapota (L.) P. Royen, Inga jinicuil Schltdl., Annona muricata L., Spondias mombin L., Mangifera indica L. Tamarindus indica L., P. sapota (Jacq .) HE Moore & Stearn, Cedrela odorata L., Gliricidia sepium (Jacq.) Kunth ex Walp., Guazuma ulmifolia Lam., Tabebuia rosea (Bertol.) ADC, and Pterocarpus hayessi Hemsl., and palms (25%), among them Cocos nucifera L., Scheelea liebmannii Becc., and Roystonea regia (Kunth) Of Cook.

From the third to fourth layer, the plants do not require a continuous management in contrast with plants from the low and intermediate layer that require irrigation, weeding, pruning and clean. Van der Wal et al. (2011) reported that the species that dominate the top layer, may determine the establishment of other species in the lower layer and where the farmer seeks to harmonize the production cycle according to their ecological requirements to achieve constant succession. The four layers were not present in all family orchards and there were others where only dominated the upper layers and the dominance of a species, indicating a transformation process in its structure. Few species such as Mangifera indica L. were common in all strata and Bixa Orellana L., in all family orchards.

The number of strata coincides with that reported in other regions of the world (^{Fernandes and Nair, 1986}; ^{Perera et al., 1991}; ^{Das and Das, 2005}); however, may be in the process of transformation, even though did not show a pattern of uniform plantation or specialize in products of commercial interest that require spaces without shade, without a complex structure, high input and where diversity is not an imperative, because it favors the abundance of few species (Abdoellah et al., 2006).

Although distinct layers are formed by species without a uniform abundance, contrary to forest vegetation (^{Zimik et al., 2012}), these were not as complex as Mayan orchards in Yucatan which were characterized by their complex composition and diversity (^{Herrera et al., 1993}). 50% (988 individuals) of woody plants, perennial grasses and palms concentrated in the lower layers (0.1 to 15.51 cm) (Figure 3), it was for one of the three reasons; a) belonged to young plants, b) had low height and c) reached maturity.

Figure 3 Diameter categories of woody plants. 

^{Caballero and Cortés (2001)} reported that the species of low height allow easy handling and harvest, and generally complete their life cycle in short periods (months or few decades), while high bearing species may have life cycles up hundred years. Mix species of short and long cycle represent an art of good management which is to harmonize the space of the orchard to fulfill its function, either for selfconsumption or sale. This management principle determines species succession in the orchard and affects their floristic composition (Van der Wal et al., 2011).

Management areas

For plant distribution, two management areas were defined by their particular characteristics. The first group comprised by four family orchards excelling for their mixed area of perennial herbs, fruit trees and timber. Small areas of medicinal and ornamental grasses were observed but were dominated by one or two species. Species richness ranged from 4 to 30, being the most abundant apple banana in one of the orchards surrounded by pastures and neighboring plots. The size of these orchards was very varied, but a common denominator was the high percentage of common area (55.06%) which results in fragmentation, also being in flooded areas, reducing the space for plants, showing a poor structure, low density and low floristic richness. The crown cover of vegetation of the four family orchards was dominated by fruit trees (x-= 42.09%) and fewer other multipurpose trees (x-= 13.44%).

The second group consisted of 25 family orchards that had two management areas: one next to the housing area composed by medicinal herbs, ornamental plants and vegetables. The closeness had the advantage of constant use, easy access to management, it is appreciated by the family and visitors, care for them against theft (Mendez et al., 2001). This area was managed by the lady of the house, where sometimes children are involved. The area devoted to herbs ranged from 2 to 20 m², showing a clear demarcation, but without showing a pattern in their design and in it adapt or introduce new species obtained through purchase or exchange among friends and neighbors.

The second management area consisted of perennial herbs, fruit trees, timber, coconut palms and royal palms (Roystonea regia (Kunth) OF Cook). Like in the first group, here fruit trees also dominated, followed by timber, confirming the role of traditional and subsistence production for family consumption (^{Ortíz, 1979}; ^{Romero, 1981}; Abdoellah et al., 2006; Kabir and Webb, 2008; ^{Chi, 2009}). Management tasks were performed mainly by the lords of the house, although family is involved. The richness of the 25 family orchards ranged from 8 to 44 species being the most abundant bananas in eight orchards.

Fruit trees dominated coverage (x- = 44.29%), followed by multipurpose trees (16.94%) and herbs (0.07%). The use of space was efficient with a coverage x-= 61.31% vs x-= 38.69% of the common areas, this is similar to the study made by Mendez et al. (2001) which fruit also excelled, followed by residences. This types of orchards resembles in number of areas and components distribution reported by ^{Alvarez-Buylla et al. (1989)} and ^{Flores (1993)}, but not as that reported by ^{Das and Das (2005)} which were based on their function (example, timber, fodder or fuel), location and detailed species composition as it did ^{Herrera et al. (1993)}. The second area was denser, showing a mixed floral composition of several layers, usually adjacent to the cocoa agro-ecosystem, grasslands, road or natural vegetation. It was also observed that there was a small space where organic and inorganic waste was piled and where domestic birds feed freely (^{Coomes and Ban, 2004}).

A third area consisted of common areas devoid of vegetation, which included: yard (including areas for housework and personal hygiene), well for human consumption, storage cellar, bathroom, poultry pens or pigsty, the latter farther from the housing (^{Das and Das, 2005}; ^{Chi, 2009}). It was observed that the plants were distributed according to management intensity, ecological requirements and depending on the available space, but also for available family labor, preferences, tastes, and knowledge by genre type for its management, which had an impact on species composition and structure (^{Romero 1981}; Lok, 1999). The role of gender in plant management reaffirms what ^{Adhikari et al. (2004}); and ^{Lerner et al. (2009)} studied where men handle the trees and women herbs.

Importance values of woody species

Fruit trees were the biological form with higher importance value (Table 1). The values obtained indicated the ecological importance (best physiological attributes) and dominance of frequent species over other species that were less common.

Table 1 Woody species with higher importance value. 

Mangifera indica L. was the most frequent among sampled orchards and Theobroma cacao L., was very important species in the diet of households, from roasted seeds elaborates a traditional drink locally known as "pozol".