Description of the study area

The sampling site of the selected herbage was located in the community of Santa Ana Teloxtoc, in the municipality of Tehuacán, Puebla. It is within the common area, on the grounds of common use identified as Chantile where grazing goats regularly. It is located at coordinates 18° 19' 51" north latitude and 97° 33' 36" west longitude at 1 646 meters (^{CONAFOR, 2013}). The climate is semi-dry and semi-warm with rains in summer, climate key is BS1hw(w). The average temperature is 18.1 °C, with a maximum of 26.2 °C and minimum 10 °C, while the average annual rainfall is 485.6 mm (^{SMN, 2010}). The rainy season occurs from May to October, most likely between june and september (^{CONAFOR, 2013}).

Species selection

The selection of shrubs studied considered the opinion of several producers who graze goats within the site Chantile. The version of the goat farmers was verified by tours in the study area, taking as a criterion the frequency with which the goats browsed the bushes. The species selected were: Sierrecilla (Acacia subangulata), orégano (Lippia graveolens) and reed stick (Mascagnia parviflora). Once chosen the three most preferred forages for goats, the sampling area was delimited. To this end, a journey into the grazing area was performed to identify sites of natural vegetation with less deterioration, with regular presence of selected shrub species.

Once defined the sampling site, random selection of shrubs was performed within the total population of each species within the site. By this procedure they were selected and marked an A. subangulata of 50 individuals, M. parviflora of 50 individuals and L. graveolens of 50 individuals, placing a stone adjacent each bush a progressive number using indelible paint also performed each geoposition bush selected using GPS to better control the location of each individual sample. The experiment was conducted between May and October 2014, during the rainy season. The variables measured in the field were: plant density (number of individuals per ha), crown diameter bush (cm), plant height (cm), number of secondary branches and production of edible forage (kg MS per bush).

Forage production

To assess the forage resource of selected shrubs total dry matter production (kg) of each species per unit area was quantified. To do this, we proceeded to collect leaves and tender stems of each bush at a height of no more than 1.80 m, height simulating the grazing of goats (^{Meneses, 1993}). The plant material collected was dried for two days outdoors, and then placed in a forced air oven at a temperature of 70 °C for 72 h. At the end of the drying process, the foliage was weighed to obtain the dry matter production of each plant and estimate the dry forage yield per hectare.

In order to know the amount of dry matter production cutoff frequency (FC) and intensity of defoliation (ID) in the rainy season for the three shrubs under study; 5 treatments with 10 experimental units (bushes) per treatment were established. Treatment 1 (T1) considered 10 shrubs defoliated within 45 days of commencement of rains to an ID of 50%, which were defoliated by branch and alternately, leaving the bush the other 50% of the foliage produced. Treatment 2 (T2) consisted of defoliation of 10 shrubs of each shrub species 45 days after rains to a 100% of ID where was removed all the edible biomass (leaves and tender stems) present in every bush.

The third treatment (T3) was performed at 60 days with FC defoliation intensity of 50%. The treatment 4 (T4) was characterized by defoliation 10 to 60 days bushes FC and ID of 100%. The last treatment (T5) consisted of 10 shrubs defoliated a one-time 150 days after initiation of growth in the rainy season with ID of 100%. The foliage harvested at each cutoff frequency and intensity of defoliation was placed in paper bags clearly identified and was recorded its corresponding weight.Abushes treatments 1 and 2 underwent a second defoliation with FC accumulated 45 days and the intensity corresponding to each treatment, for the third and fourth treatment accumulated FC were intervals of 60 days with the established ID.

Statistical design

Astatistical design of randomized block factorial arrangement (3 x 5), where the first factor analysis was the shrubby species (A. subangulata, L. graveolens and M. parvif lora) was used, the second cutoff frequency factor (FC) and the third factor the intensity of defoliation (ID) (T1= FC 45, ID 50; T2= FC 45, ID 100; T3= FC 60, ID 50 T4= FC 60, ID 100; T5= FC 150, ID 100). The statistical analysis was performed using SAS (2009) GML procedure and comparison of treatment means by Tukey proposed with a confidence level of 0.05 procedure. An analysis of correlation (r) between forage production and morphological measurements of shrub species studied was also performed. Also, simple linear regression models were obtained to predict forage yield of each of the shrubby through morphological values to obtain prediction equations forage biomass variables.

Forage production

The yield of total dry bush foliage was statistically different for sierrecilla but similar to oregano and reed stick (p< 0.0001). The density value multiplied by forage production per plant and the amount of foliage available to be harvested per species per hectare (Table 1) was estimated.

Table 1 Plant density and dry matter production for errecilla (A. subangulata), orégano (L. graveolens) and reed stick (M. parviflora) in the Reserva de la Biósfera Tehuacán-Cuicatlán. 

Medias con distinta literal muestran diferencias entre las especies arbustivas en cada variable (p< 0.01).

In total, were found 796.3 kg MS for the three trees studied during the rainy season, corresponding to L. graveolens (oregano) 76.8%, A. subangulata (sierrecilla) 15.5% and M. parvifolia (reed stick) 7.7% of Total available supply of dry forage for goats (Table 1).

In a recent study, Race (2014) reported 175 kg ha-1, 255.36 kg ha-1 and 111.14 kg ha-1 for sierrecilla, oregano and reed stick respectively; i.e. 356 kg ha-1 less than oregano leaves than estimated in this study; while for sierrecilla and reed stick the MS production was 51 and 62% higher than that shown in Table 1. These differences in foliage production among the three bushes studied depend on weather conditions, soil and physiographic region and disturbance of vegetation in different grazing sites.

A. subangulata had the lowest percentage of individuals per ha, while the density of plants L. graveolens was five times greater than M. parviflora (Table 1). These results differ from those reported by ^{Carrera (2014)} with a density value for A. subangulata of 433 individuals per ha, 323 more plants per unit area to that observed in this research. La higher plant density (15 778 individuals) L. graveolens. was for. However, other research conducted in various semiarid regions of Mexico reported lower density values as found in this study; 6 000 plants ha-1 (^{Velázquez, 2005}); 3 891 plants ha-1 (^{Osorno et al., 2009}); 1 620 plants ha-1 (^{García, 2012}); 2 800 plants ha-1 (Carrera, 2014).

The only study with similar values to this research as reported ^{Sánchez et al. (2007)} in the range of 6 236 to 25 200 plants ha-1 for northern Jalisco. These differences in densities oregano plants show the adaptation of the shrubby species to various ecosystems, but also are a result of pressure by the systems of resource use that affect reproduction and abundance of plants per unit area.

The differences in yield of forage shrub species between treatments (p< 0.0001) were found. The MS yield of oregano in all treatments was higher compared to sierrecilla and reed stick. In the treatment of 5 (FC= 150, ID= 100) showed significant difference in the reed stick bush with respect to the other two species; with lower production. In the Table 2 shows that the maximum forage production accounted L. graveolens (560.12±132.5 kg ha-1) and M. parviflora (192.8±58.7 kg ha-1) for treatment 2; while A. subangulata had maximum production in T5.

Table 2 Production of dry matter (kg ha-1) in treatments for sierrecilla (A. subangulata), oregano (L. graveolens) and reed stick (M. parvif lora) in the Reserva de la Biósfera Tehuacán-Cuicatlán. 

T1= cada arbusto se defolió a 45 días de iniciadas las lluvias con ID de 50%; T2= cada arbusto se defolió a 45 días de iniciadas las lluvias con ID de 100%; T3= cada arbusto se defolió a 60 días de iniciadas las lluvias con ID de 50%; T4= cada arbusto se defolió a 60 días de iniciadas las lluvias con ID de 100%; T5= cada arbusto se defolió a 150 días de iniciadas las lluvias con ID de 100%. Medias con distinta literal en mayúsculas muestran diferencias entre las especies arbustivas (p< 0.0001). Medias con distinta literal en minúsculas muestran diferencias entre los tratamientos (p< 0.0001). ± e.e: error estándar de la media.

Lower production for A. subangulata was in treatment 1 (FC= 45 with ID= 50); L. graveolens and M. parviflora for lower production took into treatment 5, 150 days rains started and 100% defoliation. In the three species the behavior of cumulative production where treatments 1 and 3 received the lowest amount of foliage is observed. This is because in both treatments bushes were defoliated at 50% and the remaining forage was in the bushes prevented the issuance of new shoots, unlike those bushes that were defoliated 100% that were able to issue new shoots.

Differences between treatments 1 and 3 with respect to treatments 2 and 4 are explained by the fact that if the bushes are completely defoliated (ID of 100%) twice during the rainy season, the total removal of foliage present stimulates the emission of buds and new shoots, which allowed vigorous foliage growth after cutting 100%.

However, ^{Huss (1993)} provides that the complete defoliation during periods of inactivity of the plant can reduce carbohydrate reserves, adversely affecting plant growth next year and compromising the survival of the plant.

The treatment 5, which refers to the complete and only defoliation at 5 months of age of suckers, had maximum production due to the accumulation of foliage throughout the rainy season (150 days). Thus, forage yields were similar (p> 0.0001) than those obtained in treatments 2 and 4, suggesting favorable results to perform one or two removals of edible foliage goats for the three shrub species during the time of rain. However, the quality of forage harvested at different ages may be lower when the age of suckers is greater.

The MS production L. graveolens (orégano) was higher in treatments 2 (FC= 45; ID= 100) and 4 (FC= 60; ID= 100) compared to treatments 1, 3 and 5 (p< 0.0001; Figure 10), but the dry forage yield was similar (p> 0.0001) between treatments 1 (FC= 45; ID= 50), 3 (FC= 60; ID= 50) and 5 (FC= 150; ID= 100), indicating that the cutoff frequency affects the production of forage from oregano. Defoliation alternatives that offer better results during the rainy season are ID of 100% at intervals of 45 to 60 days with increased accumulation of forage without compromising the persistence of shrubby species.

M. parviflora (reed stick) had higher dry matter yield in treatment 2 (p< 0.0001) compared with other treatments, so the best cutoff frequency was 60 days after the onset of the rainy season with intensity 100% defoliation. This suggests that established the 60 days post reed stick rains had maximum production of foliage, then slow their growth until they lose their leaves during the autumn time.

This shows that every shrub species has morph- physiological adaptations accumulate nutrient reserves for the development of foliage during the rainy season and shedding leaves in the fall to save energy during the dry season, maintaining a state of low physiological activity (^{Passera et al., 2010}).

Correlation of morphological characteristics with forage production

Morphological characteristics obtained were used to estimate the correlation index (r) with forage production and obtain prediction equations forage biomass. The variables plant height, crown diameter and secondary branches number, showed differences (p< 0.0001) between the bushes A. subangulata, L. graveolens and M. parviflora. The sierrecilla (A. subangulata) was the highest bush, more glass area, but with fewer secondary branches, compared with oregano and reed stick (Table 3).

Table 3 Morphological characteristics of sierrecilla (A. subangulata), oregano (L. graveolens) and reed stick (M. parvif lora) in the Reserva de la Biósfera Tehuacán-Cuicatlán. 

Medias con diferente literal muestran diferencias entre especies arbustivas en cada variable (p< 0.0001).

In other investigations are reported height ranges similar to that observed in oregano plant in the present study: CONAFOR (2009) described as thin L. graveolens bush 70 cm to 2 m high, while ^{Sánchez et al. (2007)} report the height of the bush in a range of 45 cm to 1.8 m, indicating that the stems are branched at the top and tend to bare; in the lower parts, these morphological characteristics of oregano match the observed field. Regarding the other two forage species no prior information of their morphological characteristics was found, so that the values obtained are used are references for further research.

The dependent variable MS production was correlated with plant height, crown diameter and number of branches. The correlation coefficients revealed that plant height (r= 0.6842) and crown diameter (r= 0.7418) are positively associated with the performance of harvested foliage (p< 0.0001), but negatively with the number of secondary stems (p= 0.0007). The correlation coefficients for MS production with plant height and crown diameter were r= 0.4766 and r= 0.5519 for A. subangulata; r= 0.6888 and r= 0.7997 for L. graveolens; and of r= 0.779 and r= 0.7958 for M. parvif lora, respectively. Also, the correlation between MS production and number of secondary branches was negative and low (r= -0.2902) for oregano, positive and moderate (r= 0.5003) for sierrecilla and positive and low (r= 0.1417) for reed stick.

Prediction equations forage biomass

By regression analysis the effect of plant height (A), crown diameter (D) and number of stems (N) on the amount of dry fodder available (CF) was estimated. The coefficients of determination (R2) for each shrubby species accounted for 63 to 67% (p< 0.0001) of the variability in forage yield related to plant height, crown diameter and number of stems (Table 4).

Table 4 Prediction equations for sierrecilla dry matter (A. subangulata), oregano (L. graveolens) and reed stick (M. parviflora) considering the plant height, crown diameter and number of stems in Reserva 

R²: coeficiente de determinación; p: probabilidad ≤0,0001.

These results are similar to those obtained by Domínguez et al. (²⁰⁰³), when the prediction equations forage biomass for shrubs in arid R2= 0.7 were used as independent variables temperature, precipitation, livestock management and anthropogenic activities.