Mexico - part II

Main vegetation zones
The following description refers to regions as defined in section 3.

Northern Mexico
Shrubland (matorral) vegetation prevails in the drier areas of Northern Mexico. According to Cavazos and Medina (1995), characteristic species in the Sonoran Desert are Cercidum floridum, Larrea tridentata, Prosopis spp., Olneya tesota, Fouqueria spp. Yucca spp., Bursera spp., with chaparral type species of Adenostoma, Arctostaphylos, Simmonddsia chinensis, Ambrosia dumosa and Encelia farinosa, large cacti as Cereus giganteus and the teasel, Pachycereus pringlei. Characteristic species in the Chihuahuan Desert are Larrea tridentata, Prosopis spp., Euphorbia antisyphillitica, Flourensia cernua, Parthenium incanum, Agave lechugilla, Dasylirion, Nolina, and Myrtillocactus, including grass species of Bouteloua, Muhlenbergia, Sporobolus and Hilaria. The carrying capacity of matorral range is extremely low, most common figures are about 40 ha per Animal Unit (AU= an adult cow weighing 450 kg, in maintenance or gestation, with a daily liveweight gain lower than 0.18 kg and minimum exercise) varying between 15 ha AU^-1 in Acacia spp and Prosopis spp vegetation in flatlands in excellent condition and 70 ha AU^-1 in hills with shallow soils and poor condition (FIRA, 1986).

Native grasslands prevail in areas of higher rainfall, in plains and smooth hill country on relatively deep soils (Miranda and Hernández, 1985). Grasslands dominated by Bouteloua gracilis, typical of better soil and moisture conditions, are very important in Chihuahua and to a lesser extent in Coahuila and Durango (De Alba, 1976). In poorer soil and moisture conditions, grasslands are dominated by Bouteloua hirsuta or B. curtipendula and in conditions of high pH or salinity by Distichlis spicata, Sporobolus airoides and Hilaria mutica (Miranda and Hernández, 1985). The carrying capacity of these grasslands ranges between 7 ha per AU when in excellent condition in plains, lowlands and valleys and 30 ha per AU in poor condition on hillsides (FIRA, 1986).

Important proportions of N and to a lesser extent of NW and NE lie in the Mountain region described by De Alba (1976), in areas above 1,000 metres. Forest where Pinus spp., Quercus spp. and Abies spp. are dominant is the natural vegetation of this region. The understorey is mainly grasses such as Muhlenbergia spp., Festuca spp., Pïptochaetium spp., Bromus spp., Poa spp., Aristida spp. and others (Hernández, 1987).

Central Mexico

The natural vegetation of Central Mexico comprises natural grasslands dominated by Bouteloua gracilis and grasslands under varying densities of Quercus spp. including species of Bouteloua, Muhlenbergia, Lycurus, Stipa, Piptochaetium, Aristida, Panicum, Setaria, Andropogon and Elyonurus; the carrying capacity of these grasslands (between 10 and 18 ha per AU) is higher than that of grasslands in the north (De Alba, 1976; Cantú, 1990, Miranda and Hernández, 1985). However, in areas where the native vegetation remains are scarce there are some parts with a high density of population which were submitted to intensive agriculture before the arrival of the Spaniards (La Mesa Central) and some other parts have been submitted to intensive agriculture in the last century (El Bajío).

Southern Mexico
The Humid Tropical region was characterized by the presence of evergreen forest (originally in 80% of the area), including species such as Swietenia carophylla, Cedrela mexicana, Enterolobium cyclocarpum, Terminalia ama and Brosimum alicastrum (Jaramillo, 1994b). In some parts of this region the native forest vegetation has been replaced by sown pastures of introduced grasses such as Digitaria decumbens, Panicum maximum, Cynodon plectostachyum, Pennisetum purpureum (De Alba, 1976; Cantú, 1990) and more recently Brachiaria spp.; in some other parts this vegetation has been replaced by short grasslands of native species (locally known as “gramas”) dominated by Paspalum notatum, P. conjugatum and Axonopus affinis (Hernández, 1987).

Different types of deciduous forest are characteristic of the Dry Tropical (originally in 90% of the area) with Lysiloma spp., Leucaena spp., Bursera simaruba, Pithecellobium spp., Tabebuia rosea, Enterolobium cyclocarpum, Ipomoea intrapilosa and Prosopis juliflora as main species (Jaramillo, 1994c). According to De Alba (1976) grasses are rare in this native vegetation, however there are fair populations of legumes such as Leucaena spp., Desmodium spp., Macroptilium spp. and Centrosema spp. Cattle production systems are based on grazing of i) native vegetation or pastures sown to Hyparrhenia rufa in the drier areas (production of calves that are finished in the Humid tropical region), or ii) pastures sown to Panicum maximum or Cynodon plectostachyum and more recently with Andropogon gayanus (Dual Purpose System).

Areas of Veracruz, Puebla and Chiapas with annual rainfall higher than 1,000 mm are characterized by deciduous forests with Liquidambar styraciflua as the main tree; in some parts the forest has been replaced by pastures of Kikuyu grass (Pennisetum clandestinum) on which dairy and sheep production systems are based; these are more intensive and productive than ruminant systems in other parts of the Mountain Region (De Alba, 1976).

Photographs from the various agro-ecological zones [ click to view full images ]

Improved pastures and sown forages

Sown forages by agricultural regions
The area sown to different forages in the last few years are depicted in Figure 12, however care should be taken to consider areas sown to permanent rainfed pastures in absolute terms. Statistical data of SIAP (2004) on areas sown are rather incomplete, leading to underestimation. The area sown to permanent pastures has increased in the past few years due to governmental programmes that paid special attention to that issue; between 1996 and 2002 Alianza para el Campo achieved an investment of US $356 million for establishing pastures (SAGARPA, 2003), which led to a steady annual increase of 106,480 ha of area sown. The relationship between those figures leads to an average cost of US $ 478 per ha for the establishment of permanent pastures which is only 17% higher than the estimated cost of establishment of permanent pastures in an intensive dairy system in temperate Mexico (Amendola, 2002), which corresponds with the fact that the programme devoted part of the investment to the acquisition of equipment and infrastructure. SIAP (2004) kept good records of newly sown permanent pastures but records of those sown before 1996 appear to be incomplete. For instance Saldívar (1991) reports that by the beginning of the nineteen-nineties there were 500,000 ha in Tamaulipas sown to buffel grass (Cenchrus ciliaris) and according to Torres (1991) there was at the same time a relatively high proportion of sown pastures in Tabasco, in both cases those areas are not reported by SIAP (2004).Additional information on sown forages in NE Mexico is provided by Díaz Solis et al. (1998) available at http://cnrit.tamu.edu/cgrm/whatzhot/laredo/diaz.html.

Figure 12. Areas sown to forages in different regions of Mexico and dairy production in Northern Mexico, after SIAP (2004).

The area sown to forages increased strongly during the nineteen-nineties (Figure 12), a comparison between averages of the periods 1990 – 1996 and 1997 – 2002 shows that total area sown increased by 35%. A high proportion of that increase (56%) was in Northern and Central Mexico, particularly in the area sown to annual forages and irrigated pastures (41%); at the beginning of the twenty-first century this area represents 49% of the area sown to forages in Mexico. These changes appear to be related with others depicted in the description of regions: maize, the most important crop for human food in Mexico, is being displaced from north to south and animal production is displacing other types of agriculture in most regions.

Even though in the first half of the nineteen-nineties some institutions like FIRA (Torres, 1991) and researchers (e.g. Muñoz et al., 1995) had great expectations in the potential for animal production in tropical Mexico; animal production has been growing at much higher rates in Northern and Central Mexico under temperate and semi arid conditions. Dairy production has been much more dynamic than beef production which is the cause of higher rates of growth of animal production in Northern and Central Mexico where cattle production has a stronger emphasis on dairying than in the predominately tropical Southern Mexico. The growth of dairying in Northern and Central Mexico has been closely related to an increase in the area sown to annual forages and irrigated pastures (Figure 13). A regression equation of dairy production on area sown to annual forages and irrigated pastures for the period 1990 – 2002 (R² = 0.88, p < 0.01, CV=5.6%) shows that dairy production in Northern and Central Mexico increased by 2,488 litres for each hectare of increase of area sown.

Sown forages in Northern and Central Mexico under temperate and arid or semiarid conditions with irrigation
Between 1990 and 2002 on average 1,280,000 ha have been sown to forages (excluding permanent rainfed pastures) in Northern and Central Mexico. Rainfed and irrigated areas share equal proportions (Figure 13), but the rate of increase has been higher for rainfed forages (48,000 ha per year) than for irrigated forages (14,000 ha per year).

Figure 13. Distribution of sown forages in Northern and Central Mexico between 1990 and 2002.The category Other forages includes species mostly rainfed grown and merely locally important.

Within irrigated forages, lucerne  (Medicago sativa) has been the most important crop and its area increased at a rate of 4,800 ha per year (1.8% year^-1). In relative terms, the areas sown to forage sorghum (Sorghum sudanense and S. sudanense × S. bicolor), forage maize and permanent pastures grew at higher rates (8.5, 7.7 and 6.0 % per year, respectively) than the area sown to lucerne, while the area sown to annual ryegrass (Lolium multiflorum) decreased by 6.5% per year. Irrigated forages in temperate and dry climates are closely linked to dairying. Some areas such as Hidalgo state export forage to dairy areas and their crop rotation is not clearly defined (including maize for human consumption in some cases). In dairy regions such as La Laguna in the states of Coahuila and Durango, the crop rotation is clearly defined with exclusively forage crops, lucerne  as long-term crop, oats (Avena sativa) and annual ryegrass in winter and forage sorghum and maize in summer. The results of a survey (Chew and Santamaría, 2000) suggest that lucerne stands in La Laguna persist three years. Therefore, the proportions depicted in Figure 14 correspond with a dominant crop rotation of three years of lucerne  followed by two years with annual crops.

Figure 14. Areas sown to lucerne  and to winter or summer annual forages in Coahuila and Durango

The following descriptions of the range of yields and composition achieved with these forages in Mexican conditions are based on a survey of summaries reported in proceedings of the annual meetings of animal production research (Reunión Nacional de Investigación Pecuaria) between 1996 and 2003; the statistical data on area sown area are based on SIAP (2004).

Forage maize
Six dairy producing states (Jalisco, Chihuahua, Aguascalientes, México, Durango and Coahuila) concentrate 80% of the area under forage maize, both irrigated and rainfed. The rainfed area  has grown more rapidly than that of irrigated forage maize. Availability of irrigation water is low in Jalisco where about 40% of forage maize is rainfed. The use of irrigation for forage maize depends on amount and distribution of rainfall. In sub humid Central Mexico with rainfall of above 450 mm in the growing season, maize is irrigated twice after sowing to allow early sowing of higher yielding late varieties, while rainfed maize uses late-sown early cultivars. In most locations of the drier North with rainfall below 200 mm, maize is irrigated four to six times during the growing season with a total application of about 600 mm. In most dairy regions forage maize is ensiled but in the dairy region Altos de Jalisco a high proportion of maize is dried in the field with high labour input and thereafter ground.

Maize cultivars are classified as early (107 ± 8 days from sowing to harvest at Aguascalientes, 1,920 m of altitude), intermediate (134 ± 10 days between sowing and harvest) and late (about 150 days). Due to the decrease in average temperature, the length of the growing season increases by 3.2 ± 0.8 days per each increase in 100 m in altitude. Higher altitudes are also related to higher yield and better nutritional composition. Forage maize is regularly sown between mid April and mid May using about 80,000 plants ha (on average 80,294 ± 57 plants ha, n=18) and moderate levels of fertilization (187 ± 5 kg N, 85 ± 4 kg P₂O₅ ha and no K).

The range of experimental yields of forage maize is wide (Table 8) with a general mean of 19 tons DM ha^-1, and green yields of 61 tons ha^-1 ranging between 43 and 76 tons ha^-1, which is 30% higher than yield estimates of SIAP (2004) for commercial crops. That means that the use of intermediate and late hybrids, planting densities of 80,000 plants ha^-1 and harvesting with DM contents above 30% (common features of experiments reporting high yields) should enable substantial yield increases. The expected 4 tons DM higher yields with intermediate cultivars than with early cultivars arises from the longer growth cycle since average growth rates of both types of cultivars appear to be the same, ranging between 166 and 194 kg DM ha^-1 day^-1. Cultivars under evaluation are becoming shorter, while in 1996 V-107, a high yielding tall variety (3.78 m) was still used, average height in later years was 2.38 ± 0.31 metres.

Under experimental conditions, timing of harvest has been defined by the stage of maturity (between early dent and 2/3 milk line), with dry matter (DM) contents ranging between 28.6 and 38.4% (on average 33.2 ± 1.8, n=14; which corresponds with ½ milk line). No consistent differences in nutritional composition among types of cultivar have been detected but fibre contents decreased and accordingly digestibility and energy content increased with maturity at harvest. Mean crude protein, fibre and energy contents of experimental forage maize summarised in Table 8 are similar to contents reported by NRC (1989) for well-eared maize silage. In Central Mexico farmers harvest silage maize at the late-milk stage, with DM contents slightly above 22% with negative effects on the nutritional value of silage: less than 25% DM and 60 and 46% of NDF and ADF respectively (Amendola, 2002), which corresponds with reports of NRC (1989) for silage made from maize with few cobs. Nutritional composition of maize silage on commercial farms can therefore be improved by harvesting at later stages of maturity.

Estimates of potential milk production per ha, calculated by means of software packages that take into account aspects and yield and composition simultaneously, are providing a sounder basis for the selection of cultivars and agronomical practices.

Lucerne (alfalfa)  
Lucerne is sown in Northern and Central Mexico, with the exception of a small area grown in Valles Centrales of Oaxaca in Southern Mexico. Lucerne has been the traditional forage used by Mexican dairy farmers; therefore it is not surprising that the area sown evolved following fluctuations in milk price. Areas sown stagnated until the mid nineteen-nineties and since then have been increasing steadily, particularly in the dairy regions of Chihuahua and La Laguna.

Most lucerne  is used as cut-and-carry forage or hay, the proportion of lucerne managed as hay has remained at 26%, being important in the two northern states of Chihuahua and Sonora. The average DM conte nt of lucerne  forage in experiments is 21.9 ± 0.6%, while Guerrero and Winans (1997) report 83% DM in lucerne hay. SIAP (2004) estimates for the period 1990 ‑ 2002 average lucerne yields of 72 and 14 tons ha^-1 as fresh forage and hay respectively. Considering these DM contents, commercial lucerne yields would be 15.8 tons DM ha^-1 year^-1 with about 28% field losses in hay making.

Lucerne  seed production is inadequate to fulfil the demand. Therefore, even though there are local ecotypes derived from Spanish lucerne s (e.g. Oaxaca, Atlixco, San Miguelito), most seed is imported from the USA and Spain. Due to mild winters and the importance of winter forage production, cultivars used are non-dormant or semi-dormant (autumn dormancy scores 8 to 10). Moapa 69 and CUF 101, two rather old USA cultivars and Valenciana and Aragon, two Spanish ecotypes have been widely used in the past, nowadays many new cultivars are being introduced, mostly from the USA. In reports of cultivar evaluations, new introductions often outyield varieties chosen as standards (Cuf 101, Moapa 69 and Valenciana in most cases). In a review on this topic, Martínez and Meza (2001) conclude that local ecotypes (Atlixco, Oaxaca, San Miguelito) regularly outyield new cultivars in Central Mexico, while the opposite occurs in Northern Mexico.

(NRC, 1989). Crude protein and energy contents are higher during autumn and winter and lower during summer and spring. Lucerne stands regularly last three years; due to climatic conditions persistence is clearly linked with levels of disease resistance. Chew and Santamaría (2000) surveyed the incidence of crown rot (Fusarium sp., Rhizoctonia solani, Colletotrichum sp., and Phoma medicaginis). The authors report that incidences were 62.0, 87.1, and 98.4% in the 1st, 2nd, and 3rd year of stands and that the corresponding yield reductions were 11.6, 28.0, and 33.3%. This estimate of yield reduction corresponds with the summary of results reported in proceedings of the annual meetings of animal production research, depicted in Figure 15.

Figure 15. Average yield of lucerne in experiments taking into account one or more years of the crop, vertical bars depict standard errors.

Seasonal distribution of yield is a capital factor in the management of dairy systems where lucerne  is used; from reports it appears that 60% of growth takes place in spring and summer and 40% during autumn and winter. Taking into account a persistence of three years, one third of the stands should be in the establishment phase without useable production during their first winter; therefore, in terms of the system, the availability of fresh lucerne forage in the winter months is about 2.6 times lower than during summer. Concurring with the results of seasonal differences in growth rates, the average result of experiments on the effect of harvest frequency on yield is that best frequencies are 30, 40, 50 and 35 days for summer, autumn, winter and spring, respectively.

Levels of underground water used for irrigation are decreasing at rates higher than one metre per year in most regions, therefore the efficiency of irrigation water use is a matter of deep concern and lucerne is regularly considered an inefficient crop. In experiments, water applied as irrigation fluctuated between about 1,400 mm year^-1 (with drip irrigation) and about 2,300 mm year^-1(flood irrigation). The average amount applied was 1,840 ± 189 mm year^-1, which amounts to about 5 mm per day in regions where no irrigation is used during the rainy season. The average use efficiency of water applied with irrigation was 1.65 ± 0.25 kg DM m^-3 (n=7), ranging between very low values with flood irrigation (0.6 kg DM m^-3) and very high values with drip irrigation (2.5 kg DM m^-3); the increase in efficiency with drip irrigation is due to 32 to 51% lower water use and 16 to 23% higher yields (Godoy et al., 2003). Quality of irrigation water (in some regions sewage water is used) affecting levels of salinity of soils and quality of herbage has been a matter of research in the past few years (Robledo et al., 1998; Serrato et al., 2002).

Yields in experiments are summarised in Table 9. It appears that using proper varieties, irrigation techniques and harvest frequencies, there is scope to narrow the 58% gap between average commercial yields and average experimental yields.

Small grains and annual ryegrass
This group of forages is mainly used in two different situations, i) as winter annuals in irrigated crop rotations with forage maize or forage sorghum during the summer and lucerne  pastures as long-term crop and ii) as summer crops in rainfed conditions, particularly in regions where due to the low amount of rainfall and its erratic distribution, the success of maize become uncertain. In these conditions, when sowing has to be delayed due to the late onset of rains, the risk of drought and crop damage by low temperatures increases (Ochoa et al., 2002) and small grains have the advantage of a shorter growth cycle (e.g. 64 to 100 days for different oats cultivars).

Oats are the most important forage in this group. In the second half of the nineteen-nineties the area sown to forage oats increased sharply, 44,000 ha per year in rainfed conditions and 3,700 ha under irrigation. A high proportion is managed as hay (83% of rainfed oats and 60% of irrigated oats). Irrigated oats are important in dairy regions of northern and central Mexico, while 76% of rainfed oats are sown in three northern states (Chihuahua, Zacatecas and Durango).

Most research on oats has involved comparisons with other forages of this group; a summary of results is reported in Table 10. Irrigated winter annuals are sown between the end of September and mid December, most usually about 15 October. The length of the growth cycle depends on the way of utilisation in terms of number of harvests, which vary between only one harvest in a reproductive stage and multiple harvests in vegetative stage (at least in the first harvests). When multiple harvest are taken, the growth season usually lasts 200 days, until the end of April when, according to the usual crop rotation, maize has to be sown. Under this kind of use, yields of a mixture of oats and annual ryegrass are higher and more stable, profiting from the earliness of oats and the persistence of annual ryegrass. The first harvest is taken about 70 days after sowing and subsequently three to four harvest are taken at intervals of 35 to 40 days. When only one harvest is taken, delaying harvest from an early reproductive stage to a late milk stage has improved yield in 130 ± 17%; lowest yields of different species in Table 10  in most cases correspond with only one harvest taken at a relatively early stage.

Within this group, annual ryegrass has some particularities because it is not grown in rainfed conditions; besides being utilised as a cut-and-carry forage it is also grazed by growing heifers or steers and in some regions is over-sown in autumn into existing Bermuda grass (Cynodon dactylon) pastures. Yields of barley (Hordeum vulgare) are regularly lower than those of oats, but it has the advantage of being about 10 days earlier; new awn-less cultivars are being tested. The area sown to wheat (Triticum aestivum), rye (Secale cereale) and triticale (×Triticosecale rimpaui) has been very small (about 3,000 ha). Data in Table 10 do no clearly reflect the fact that when compared within single experiments, triticale regularly outyields oats. There are few data on yields of rye and wheat but when compared in the same experiment both species regularly outyield oats.

Data on rainfed conditions are still too scarce, particularly taking into account that local conditions may play an important role and that problems differ from those of irrigated winter crops; for instance in rainfed summer conditions, disease resistance becomes an important attribute (Salmerón, 2002).

Figure 16.Yield and crude protein content (CP) of rainfed oats at different stages. Adapted from Tovar et al. (2000).

Nutritional composition of forage of these crops is good with 16.5 ± 1.5 % CP (n=13) and 1.52 ± 0.05 Mcal NEL kg^-1 DM (n=7), provided they are harvested before the late milk stage when CP contents can be lower than 8% (Figure 16). Including ryegrass in mixtures with small grains improves the nutritional composition of forage (Amendola, 2002).

Forage sorghum
Sudan grass is included in this category together with hybrids (S. sudanense × S. bicolor). As with other forages, the area sown to forage sorghum increased sharply after 1995, with 15,750 ha per year in rainfed conditions (almost 230,000 ha sown in 2002) and with 5, 070 ha per year under irrigation (almost 145,000 ha sown in 2002). Five state of Northern Mexico share 70 % of that area, in Chihuahua, Sinaloa and Nuevo Leon predominantly rainfed and in Coahuila and Sonora predominantly irrigated.

In spite of the increase in area, reports on evaluation of forage sorghum are still scarce. Forage sorghum is important in dairy regions of the arid North because i) with an efficiency of slightly less than 3 kg DM m^-3 water, it is about 30% more efficient than maize and ii) is able to distribute forage production in summer more evenly than maize. However the nutritive value of forage sorghum is lower than that of other forages grown in those dairy regions (Núñez et al., 1997). In Northern Mexico under irrigation, forage sorghum is sown between the end of March and mid April, the first harvest is taken after 55 to 60 days and subsequent harvests are taken every 30 to 40 days when crops reach the boot stage; consequently about four harvest are taken until the beginning of October (Nuñez and Cantú, 2000). From experiments reported it appears that under such conditions, mean yield achieved reaches 16.6 ± 0.6 t DM ha^-1 (n = 20), CP and ADF contents average 9.9 ± 1.2 % (n = 6) and 43.8 ± 1.3 % (n = 4), respectively.

Annual legumes
Berseem clover (Trifolium alexandrinum)is a promising legume for use in mixtures or pure stand as a winter forage in dairy regions of Northern and Central Mexico (Nuñez et al., 1997). About 6,000 ha of common vetch (Vicia sativa) are grown in Michoacán but reports of research on the crop are scarce.

Irrigated long-term pastures
Dairy and meat production systems under grazing are becoming attractive options in order to reduce production costs, and irrigated long-term pastures are essential components of such systems. The qualifying adjective “long-term” is used instead of “permanent” or “perennial” because lack of persistence is a regular problem of these pastures (Amendola, 2002). The area sown to irrigated long-term pastures has grown at an annual rate of 5.6% from 82,000 ha in 1990 to 144,000 ha in 2002. Four states of Northern Mexico (Tamaulipas, Nuevo León, Sinaloa and Coahuila) with 56% and five states of Central Mexico (San Luis Potosí, Jalisco, Michoacán, Hidalgo and México) with an additional 31% concentrate most of that area.

Research (almost exclusively under irrigation) has been focused on pure stands or mixtures of lucerne, cocksfoot (Dactylis glomerata), perennial ryegrass (Lolium perenne), tall fescue (Festuca arundinacea) and white clover (Trifolium repens). Most evaluations have been carried out on first and second year pastures and exceptionally on third year ones; Améndola (2002) reports that due to low carrying capacity (Figure 17), the cost of grazing days is higher in pastures kept for a fourth year than on pastures only kept to their third year. Lack of persistence is particularly severe in lucerne and perennial ryegrass pastures. Mixtures of lucerne and grasses regularly outyield pure stands of grasses and mixtures of white clover with grasses (Table 11). Within grasses there is a trend of higher yields with tall fescue than with perennial ryegrass and cocksfoot. Grasses regularly dominate mixtures with white clover, but the legume dominates mixtures with lucerne and therefore bloat is a problem when grazing such pastures.

Figure 17. Carrying capacity of lucerne and orchard grass pastures of different age. Adapted from Amendola (2002).

Table 11. Yields of long-term pastures estimated from reports published in Mexico.
Pastures	Yield t DM ha-1 year-1	Standard error	Number of reports
N fertilized grasses	15.8	2.4	7
Mixtures of grasses with lucerne	23.2	1.3	4
Mixtures of grasses with white clover	15.2	1.3	6

Uneven seasonal distribution of herbage production is a common problem (Sosa et al., 1998; Velasco et al., 2001; Villegas et al., 2004). A high proportion of herbage accumulation takes place during spring and summer (66.2 ± 1.9%, n=9). Winter herbage production is lower with orchard grass mixtures than with perennial ryegrass mixtures and with lucerne  mixtures than with white clover mixtures, and is particularly low with Kikuyu (Pennisetum clandestinum).

Digestibility of herbage averages 72.4 ± 1.9% (n = 7) and CP content averages 18.8 ± 1.4 % (n=9), with highest values during winter and lowest values in summer. The digestibility of herbage from mixtures containing perennial ryegrass is regularly higher than that of mixtures containing cocksfoot, while herbage from mixtures with white clover shows higher digestibility than those with lucerne. Reports on performance of grazing cattle without supplementary feeding are scarce; daily weight gains range between 0.85 kg head^-1 on grass pastures and 1.20 kg on grass-legume mixtures; daily milk production per cow ranges between 17 kg cow-1 on grass pastures and 19.6 kg on grass-legume mixtures.

Rainfed long term pastures in temperate climate
There are very few reports available on evaluation of rainfed permanent pastures. In a humid temperate climate, the same forages have been evaluated as under irrigation, with the exception of lucerne, which due to low soil pH and high incidence of diseases is not suited to those conditions. In sub-humid conditions evaluations have included Kikuyu, buffel grass (Cenchrus ciliaris) and Rhodes grass (Chloris gayana).

Sown pastures in the tropics
Cattle production systems in the tropical regions are extensive and cattle feeding is based on grazing of native grass or introduced/naturalised grass pastures. Eighty percent of dual purpose system (DPS) farms have African Stargrass (Cynodon plectostachyus; C. nlemfuensis) in their pastures. Second in importance, mainly in the humid tropics, are native grass pastures (44%), and in drier areas Jaraguagrass (Hyparrhenia rufa, 20%-30%) and Guineagrass (Panicum maximum, 23%-55%). In recent years Llanerograss (Andropogon gayanus) a species released by CIAT, which is tolerant to drought, spittlebug attacks and low fertility soils, is making inroads into the DPS of the dry tropics and Brachiaria brizantha, also released by CIAT, is being adopted in the humid tropics. Other grasses used by DPS farmers are Pangolagrass (Digitaria decumbens, 15%-28%) on fertile soils, or Paragrass (Brachiaria mutica, 9%-24%) and Alemangrass (Echinochloa polystachya, 6%-17%) on temporarily or permanently flooded soils.

Pasture dry matter production is very seasonal. Forage grows very rapidly during the rainy season (June to November) and very little during the dry, winter season (December to May). Seventy percent of DPS farmers consider seasonal forage production as the main constraint to increasing animal yield. However, farmers do not take measures to overcome dry season feed shortages, with conserved forages and concentrate supplements. Half of the farmers never supplement their cattle; 40% supplement cattle only in the dry season. Only 4% to 5% use supplements all year round. Internal divisions of the pastures are scarce: four paddocks on the average. However, pasture division increases as cattle numbers increase. Only 10% of the DPS farmers fertilise their pastures. Of those that fertilise, 90% apply fertiliser only on some pastures or divisions. Three percent fertilise all their pastures. Weed control is the most commonly used pasture management practice by DPS farmers: 70% do it. Pasture technology can provide legumes that, besides improving the diet of cattle, also improve soil fertility through biological nitrogen fixation (BNF) ('t Mannetje, 2000) .

Calf suckling and the low nutritional level of the herd both have a pronounced negative effect on reproductive efficiency, leading to low apparent birth rates (50% - 60%) and reduced percentages of cows in milk (50% - 60%). Up to 65% of parturitions are between March to June, which results in a highly seasonal milk and calf production pattern.

These facts explain why tropical cattle husbandry has such a low economical efficiency.

Under the type of feeding described, daily milk production per lactating cow varies from 2.9 to 4.3 litres and lactation length varies from 180 to 237 days. Weaning age of calves is slightly more than eight months and weaning weights are in the 120 to 150 kg range. Considering the low stocking rates, between 0.5 to 1 cow/ha and the reduced number of lactating cows, annual milk production/ha seldom exceeds 500 litres.

Menocal et al. (1992 a, b, c) collected data from different experimental stations, which show the potential that an "improved" dual purpose system, to give an example, has with respect to the traditional one used by the average cattleman (Table 12).

The improved model of dual purpose system is many times superior to the current system. It is obvious that conditions for production are different among commercial DPS and those DPS developed at Experimental Stations. However, the latter have the great virtue of showing the milk production potential of tropical DPS, and can become the model to follow by innovative DPS farmers. However, it would be pretentious for the traditional DPS farmer to try to reach the production levels of the improved DPS in the shortest time. This should be a medium to long term objective, where the rate of adoption of technological practices must be in accordance with the farmer’s economical and educational possibilities.

Increases in stocking rate as well as in production per cow can lead to considerable increases in milk production per hectare in the improved system compared to traditional practice (Table 12). That increase may be ascribed mostly to better pasture technology, such as new pasture species and fertilization , even though lactating cows are supplemented at milking time throughout their lactation, but even discounting the effect of supplementation, the pasture is the major potential contributor to increases in milk production per animal and per hectare.

In the early nineteen-seventies, Mexican research institutions conducted pasture studies to identify high yielding grasses which, used in combination with N fertilization, could produce large amounts of meat and milk. The pasture model developed was simple and consisted of a slow rotation (3 to 5 paddocks), N fertilization around 150 kg/ha/year and, in some cases, supplementation with molasses or other by-products in winter. It was shown that this model was capable of doubling the stocking rate and could keep a milk production of around 6 kg/cow/day and average daily gains in the order of 0.4 to 0.6 kg/heifer/day (Fernández et al., 1993b) . Despite its high production levels, DPS farmers did not adopt this model.

Some trials indicated the benefit that pasture legumes can have on meat and milk production. Pangolagrass (Digitaria decumbens) alone or associated with the legumes Neonotonia wightii, Centrosema pubescens and Leucaena leucocephala were compared with respect to weight gain (3.3 heifers/ha) and cow milk production in a hot sub-humid climate, using a three-paddock, 42-day rotational grazing. Pastures were irrigated in the dry season. The associations produced from 25% to 38% more weight gain, and from 6% to 14% more milk, than the grass. Neonotonia and Centro produced more than Leucaena, but both disappeared from the pasture a year from the beginning of the trial (Garza et al., 1978; Portugal et al., 1979). At a stocking rate of 4 cows/ha, Leucaena was the most persistent and was still in use several years after the trial (Castillo, 1999). These trials show the potential level of productivity that can be attained under irrigation in the dry tropics, but the same is not within reach under current low input conditions of the Mexican DPS farms. Other experiments conducted in the dry tropics without irrigation indicated the superiority of the associations of introduced grasses with the shrub Leucaena leucocephala over monocultures of IG. Therefore this legume must be the choice to associate with introduced grasses in the dry tropics of México (Castillo, 2000).

Three important points were ignored when research on grass pastures was designed: i) farmers in the tropics of México do not fertilise their pastures; ii) once pastures were fertilised, the farmer necessarily needed more animals to consume the extra forage produced and this required capital, not always available; and iii) the choice of legume and grass species. Mexican researchers were using species like Bermuda Coast Cross 1 (Cynodon dactylon x C. aethiopicus) and Pangolagrass which show a superb performance when irrigated and fertilized, but disappear rapidly from the pasture if not. At the same time these grasses inhibit the establishment and persistence of herbaceous legumes.

As Pangolagrass, Guineagrass and Coastcross 1 are very susceptible to spittlebug (Aeneolamia spp.) attack farmers selected Stargrass (Cynodon plectostachyus), not the most nutritious of grasses, but resistant to spittlebug, able to grow on marginal soils, and capable to stand some degree of overgrazing.

Since the mid nineteen seventies, DPS farmers found themselves facing new problems. Highly productive Guineagrass and Pangolagrass started to die back. Although no Mexican study exists on the probable cause, it is considered that the decline in nutrient content of the soils, as a result of several decades of use, is the main reason. This hypothesis is supported by the high degree of weed infestation as well as the shift of pasture species from those that require high soil fertility, like Guineagrass and Pangolagrass to Stargrass, which does not.

The permanence of native pastures, known locally as “gramas nativas” or native grasses also supports the above hypothesis. In general, native grass pastures show low levels of productivity, which mean nutrients are taken up at slower rates than under more productive introduced grass pastures. In addition, native grass pastures contain variable amounts of native legumes, between 2.5% to 15.4%, which contribute to soil N (Bosman et al., 1990) , a trait not present in introduced grass pastures, where management aims for monocultures. Thus, native grass pastures tend to persist over long periods of time.

Mexican research institutions joined around the mid nineteen eighties the RIEPT (Red Internacional de Evaluación de Pastos Tropicales “International Network for the Evaluation of Tropical Pastures”) supported by the Programa de Pasturas Tropicales (“Tropical Pastures Program”) of CIAT, and began to follow the RIEPT’s philosophy of minimum pasture inputs along with its evaluation methods.

Agronomic regional trials in CIAT’s forage evaluation scheme, indicate that for the Mexican humid tropics the most promising legumes were, in order of dry matter yield: Centrosema acutifolium CIAT 5568, Pueraria phaseoloides CIAT 9900, Desmodium ovalifolium CIAT 350, and Arachis pintoi CIAT 17434. Surprisingly, by 1994, only A. pintoi CIAT 17434, had shown itself as a promising legume to associate with introduced grasses, which agreed with reports from Costa Rica that showed its ability to persist under heavy grazing and also improve animal productivity.

In the last decade, some grasses from CIAT’s regional evaluation scheme have become commercially available. Llanerograss (Andropogon gayanus) is the choice for the dry tropics under low to medium fertility soils. In the humid tropics, Chetumal grass (Brachiaria humidicola) is recommended for very low fertility soils; while Insurgente grass (Brachiaria brizantha) and several Panicum maximum cultivars such as Tanzania, Mombasa, and Tobaitao, are recommended for medium to high fertility soils.

In the last three to four years other introduced grasses have became available. One is the first a hybrid between Brachiaria brizantha and B. ruziziensis, called Mulato grass (CIAT 36601). It shows considerable promise because, besides being more productive than other Brachiaria spp, it has a higher degree of palatability, good level of in vitro digestibility, and resistance to spittlebug attack. Nevertheless there is no research report on its performance in terms of milk and beef yield under grazing. The other is Andropogon gayanus ‘Tun-Tun’ which is shorter than the Llanero cultivar but 25% more productive in terms of milk and meat production.

It must be recognized that most of those grasses were made available without thorough testing in the Mexican tropics; the decision to make them available commerciallly was based on their performance in South America - Colombia and Brazil mainly.

Click to view photographs of improved pastures and fodder crops

Forage resources in the arid and semiarid region of Mexico
The arid and semi-arid region of Mexico is mainly north of 20^o N and with a mean annual rainfall between 350 and 550 mm; to the east (95 to 97^o W) annual rainfall is higher, up to 750 mm, but the Lang index (Annual precipitation/mean annual temperature) is still indicative of low water availability due to high temperatures. The total area is estimated at 54,000,000 ha or nearly 30 % of Mexico. Thick shrubby vegetation occupies 68 % of the arid and semi-arid region, 28 % is herbaceous vegetation (mainly short and medium grasses) with scattered shrubs and/or trees (oaks and pines) and 4 % (nearly 2,000,000 ha) of the region is now abandoned croplands with severe erosion and infested by annuals. At national level around 39 % of the region shows severe loss of top soil due to water and wind erosion (INEGI, 2000); however, for some specific sites up to 43 % of the land has been shown to have a serious loss of top soil (Royo et al., 2004).

North of 20^o N there are another 24,000,000 ha with annual rainfall below 350 mm, where the native vegetation is so poor in forage species and yield that estimated annual stocking rates (COTECOCA, 1978) range from 40 to 60 ha per animal unit (non-lactating mature cow of 450 kg live-weight). Besides, many studies have shown the sowing of grasses of higher forage potential to be unsuccessful. It can be concluded that, under Mexican conditions, below 350 mm annual rainfall there are no reliable forage resources to sustain high animal outputs from livestock systems based on cattle, sheep and goats. On the other hand, there is an urgent need for research to increase animal species that have proved to be successful desert dwellers, one example is the Pronghorn (Antilocapra americana) subspecies Sonoran, Mexican and Peninsular, this last subspecies survives in the Viscaíno Desert with a mean rainfall of 50 mm per year, which includes years with no rain at all.

Along the two major mountain chains, the East and West Sierras Madres on the continental or inner slopes, there are an estimated 12,000,000 ha of forest (oaks and pines) and reforested (eucalyptus) lands with potential to grow forages (Aguirre et al., 1996). Annual rainfall ranges from 550 to 900 mm, altitude goes from 2,300 to 3,000 m above sea level, mean annual temperatures are from 11 to 18 ^oC, in some cases with one or two light frosts in summer. As conservationist groups and forest professionals have promoted the passing of legislation to ban grazing on these lands, on the grounds that livestock grazing is damaging the forest, limited research on forages has been done. However, smallholders keep livestock on these lands once they are cleared of trees. Felling of trees is done to grow crops, maize at lower and potatoes at higher elevations; once top soil and natural soil fertility are lost, the land is abandoned and livestock are introduced. Although pine needles and some oak leaves can be poisonous to livestock (e.g. Denogean et al., 2004; Hatch and Pluhar, 1993; Schmutz and Hamilton, 1986; Schmutz et al., 1974; Stubbendieck et al., 1992), it is mainly the trampling and browsing of seedling tops and bark chewing that will be the main impact of cattle and sheep on the oaks and pines. Therefore planning tree planting and grazing along with sound animal feeding should be enough to prevent these injuries.

The following description of forage resources comes from studies carried out in the Mexican arid and semi-arid region as defined above. First native plants are discussed followed by introduced grasses.

Native plants
An issue that has not been extensively settled regarding native plants is: does livestock grazing improve plant biomass accumulation compared to non-grazing since grazing might promote a higher rate of plant tissue turn-over. This issue is ecologically relevant as some native plants are very difficult to restore. 

Grazing can reduce forage biomass yield compared to non-grazing because animals graze some species more intensively than others allowing the latter to become dominant. Aguado (1994) found that after four years of moderate grazing (mean removal of 40 to 60 % of the annual accumulated above ground biomass) some sites showed from 16 to 80 % less forage yield than no-grazing, while in other sites, the grazed area showed up to 19 % more forage yield than no-grazing. The negative or positive effect on forage yield of grazing compared to no-grazing was explained by the different botanical composition of the sites. In the sites with a low proportion of plant species highly preferred by cattle and a large percentage of plant species not eaten by cattle, grazing decreased the forage yield because cattle removed plant tissue from the few species leaving others slightly disturbed or undisturbed, which in turn yielded less forage. While 80 to 90 % of the above ground biomass of Bouteloua spp (preferred grasses) was removed by cattle, less preferred grasses like Microchloa kunthii had 12 % or less of their above ground biomass removed. The higher defoliation of Bouteloua spp. led to their contribution to the botanical composition decreasing from 31 to 7 % and a sharp increase in the average contribution of shrubs to the botanical composition from less than 4 to as high as 34 %. On sites where most of the plant species occurring were grazed with similar intensity, grazing promoted higher forage yield than no-grazing.

Recurrent droughts mean that, on commercial lands, if livestock removes 60 % or more of the year’s forage accumulation, forage biomass will show a constant decline with sharp increases in shrubs in the botanical composition as well as of bare ground; this, in turn, will lead to overgrazing and decline in animal production. Grazing strategies should incorporate annual estimation of proper stocking rate along with controlled grazing to keep forage removal between 40 to 60 % of the year’s forage accumulation.

Native grasses
The genus Bouteloua has over 22 species of high to medium forage production and quality, distributed all over the arid and semi-arid region of Mexico, from the drier to the more humid, from shallow to deep soils, and from valleys to slopes. Most Bouteloua species are liked by cattle and can represent up to 80-95 % of the diet of grazing cattle, provided that there is no restriction on forage availability. However, this preference can lead to heavy overuse of these species and removal of 80 to 92 % of the above ground biomass has been reported. If grazing is not controlled as to prevent this high level of use subsequent forage yield and persistence decline sharply.  

Among Bouteloua species described as a good forage resource are:  
B. curtipendula (Michx.) Torr, B. chondrosioides (Kunth) Benth. ex S. Watson, B. dactyloides (Nutt.) Columbus, B. eriopoda (Torr.) Torr, B. gracilis (Kunth) Lag. ex Griffiths and B. hirsuta Lag. Research has concentrated on B. curtipendula and B. gracilis as they are widespread in the region in addition to having shown higher forage yield and quality than other Bouteloua species.

Bouteloua curtipendula has a large genetic pool with over 77 accessions collected from Canada to Mexico. Mexican specialists have come up with at least three outstanding cultivars, unfortunately seed is not readily available as commercial seed production is scarce. It is widely recommended for annual rainfall between 300 to 550 mm, to be sown in abandoned cropland; seeding rates go from 2 to 5 kg ha^-1, seedling emergence can go from 0 to 50 % and plant survival a year after sowing is from 10 to 20 %. Since establishment by sowing can be very poor and seed expensive, planting has been studied as an alternative to sowing; plants are grown from seed in a greenhouse or nursery for 70 to 90 days then transplanted at the onset of the rainy season. Rate of success usually is over 60 % or higher if some soil modifications to catch rain is done along with the planting. Evaluated planting densities have been variable and in most of the studies rate of plant survival after planting has been the variable of interest rather than searching for an optimum population density.

Bouteloua gracilis tends to show lower forage yield but better quality than B. curtipendula, however there is one recent study (Bravo et al., 2004) in which at low soil moisture availability B. gracilis showed a better yield performance than B. curtipendula. As was the case with B. curtipendula Mexican specialists have come with better cultivars but commercial seed availability is scarce and establishment from sowing could be very poor.

Native shrubs
The arid and semi-arid region of Mexico is very rich in shrub species with high to medium forage availability and quality. An excellent review on the biology and uses of Mexican shrubs was prepared by McKell and Garcia (1989). Local producers have for many years been aware of the use of shrubs by livestock. However, systematic published studies on their agronomy started only in the second half of the nineteen-nineties.  Before that published studies dealt with ways of eradicating all shrub species, after some years of research focused on the identification of shrubs eaten by cattle, sheep and goats and season of the year when livestock consume them. Then, the paradigm changed from indiscriminate shrub eradication to strategies for selective shrub control that incorporate maintenance and/or propagation of some specific shrubs.

Among the Fabaceae the genus Acacia with over 15 species distributed all across Mexico,  Eysenhardtia with two species E. polystachya (Ortega) Sarg (synonym: E. amorphoides Kunth) and E. orthocarpa (A. Gray) S. Watson and Dalea bicolor Humb. & Bonpl. ex Willd are among shrubs with proven forage potential for rainfall between 400 to 500 mm. Many of the Acacia shrubs have thorns making their handling in the nursery difficult and this also reduces the quality of the forage available to cattle. In the Fabaceae, at similar rainfalls, there are other widespread shrubs which are toxic to livestock, among them: Astragalus mollissimus Torr, Lupinus argenteus Pursh, Lupinus montanus Kunth, Oxytropis campestris (L.) DC and Oxytropis deflexa (Pall.) DC.

Research on shrubs of the Chenopodiaceae has concentrated on one species: Atriplex canescens (Pursh) Nutt. Within the annual rainfall range of 250 (with some terracing for rain harvesting) to 500 mm, A. canescens has shown the highest survival after planting (90 to 100 %), a leaf crude protein concentration between 15 to 23 % with a mean of 18 % and annual forage yield 1.9 to 4 times higher than other shrubs. Reported annual forage yield under clipping has been from 1.9 to 5.4 t DM ha^-1 with plant densities ranging from 1,250 to 10,000 plants ha^-1; some research estimates that an optimum population should be around 2,500 to 3,500 plants ha^-1 in the majority of cases. Also, some tolerance to saline soils has been indicated. Forage yield was determined by branching and plant height; as these two attributes increased so did forage yield. There are female and male plants, so in plots for seed production around 40 to 50 % of the plants would yield seed. Seed yield per plant ranges from 300 to 600 g. A. canescens is well accepted by cattle, sheep and goats; however, long term performance of this shrub under different grazing strategies and herbaceous understorey has not been extensively researched. 

Two other Chenopodiaceous shrubs recently brought to the attention of researchers are Atriplex acanthocarpa (Torr.) S. Watson (tolerant to saline soils and readily eaten by cattle, Sierra and Melgoza (2004) and Atriplex polycarpa (Torr.) S. Watson that can grow on very dry soils shallow and rocky) in the Baja California peninsula.

The genus: Opuntia, of the family: Cactaceae, subfamily: Opuntioideae and tribe Opuntieae includes many shrub-type species that have been used by local producers as sources of forage and water for livestock. The photosynthesis pathway in these species is the Crassulacian Acid Metabolism (CAM) which enables them to survive longer without added water than other shrubs; however, many specialists question the relevance of Opuntia as a forage resource, mature cattle fed only Opuntia shrubs can not maintain  live weight.  These shrubs are low in dry matter (8 to 10 %) and crude protein (3 to 9 %) and many of them should be scorched to get rid of thorns so that cattle can consume enough of them to survive. Another major limitation of Opuntia is its slow regrowth after harvesting. Some wildlife use Opuntia only as a source of water.

Local cattlemen refer to Opuntia shrubs as ‘nopales’ or ‘choyas’and common practice was to give these wild shrubs to cattle to help them survive in a dry year. Nowadays over 60 % of cattlemen surveyed (Luna and Chávez, 1994) offer Opuntia to cattle year after year indicating that native grasses are constantly overgrazed and then not able to supply the required amount of forage regardless of the amount of rainfall received. The constant offering of Opuntia to cattle has decreased the density of wild populations in many areas in the region. In the face of this trend, state and federal programmes have been established to promote the planting of Opuntia. The reduction of some wild populations of Opuntia shrubs has an ecological impact, as many of these shrubs are endemic to Mexico and might be considered as endangered species.

Some species that have been used as forage for livestock are: Opuntia azurea Rose, Opuntia engelmannii Salm-Dyck ex Engelm, Opuntia engelmannii Salm-Dyck ex Engelm. var. cuija Griffiths & Hare, Opuntia engelmannii Salm-Dyck ex Engelm. var. engelmannii , Opuntia engelmannii Salm-Dyck ex Engelm. var. flavispina (L. D. Benson) B. D. Parfitt & Pinkava , Opuntia engelmannii Salm-Dyck ex Engelm. var. flexospina (Griffiths) B. D. Parfitt & Pinkava, Opuntia engelmannii Salm-Dyck ex Engelm. var. lindheimeri (Engelm.) B. D. Parfitt & Pinkava, Opuntia engelmannii Salm-Dyck ex Engelm. var. linguiformis (Griffiths) B. D. Parfitt & Pinkava, Opuntia leucotricha DC, Opuntia rastrera F. A. C. Weber  Opuntia ficus-indica (L.) Mill and Opuntia robusta H. L. Wendl. ex Pfeiff.

Introduced grasses

After years of evaluation the following introduced grasses are the most widespread in the region: Cenchrus ciliaris L., Panicum coloratum L.,  Eragrostis curvula (Schrad.) Nees , E. curvula var. conferta (Schrad.) Nees, E. lehmanniana Nees and E. superba Peyr.

Cenchrus ciliaris is the introduced grass with by far the largest sown area in the region, estimated between 1,250,000 to 2,000,000 ha in the last seven years. Sowing of Buffel grass has been grouped in four major areas: Northeast (states of Tamaulipas and Nuevo Leon), Northwest (states of Sonora and Sinaloa), South-centre (states of Zacatecas and San Luis Potosi) and the Dry Tropic of the South and South-east of Mexico (states of Michoacan, Morelos, Oaxaca and Yucatan, mainly). Buffel grass was brought from Texas in 1954 to the state of Nuevo Leon, thereafter the expansion of area sown to Cenchrus grew sharply. The grass showed excellent establishment and high forage yield. Annual stocking rate of native shrubland was from 12 to 27 ha per Animal Unit; when the land was cleared and sown to Buffel grass annual stocking rates were from 3 to 8 ha per Animal Unit. In the range of 400 to 550 mm annual precipitation Buffel grass has shown better establishment and higher forage yields than most native grasses.

Low winter  temperatures have been the major constraint to the expansion of Buffel grass in the arid and semi-arid regions as extensive winter kill occurs in areas with monthly average of absolute minimum temperatures of 5 ^oC or lower. Some commercial cultivars claim better tolerance to low temperatures: Frio, Llano and Nueces. More tropical cultivars like Molopo, Biloela and Gayndah may show low persistence in most of the arid and semi-arid region because they are highly susceptible to low temperatures.

In the last 10 years another constraint to the expansion of the area seeded to Buffel grass and a threat to its persistence in areas already sown has been building-up; it is the spittlebug (Aeneolamia albofasciata Lall.) which had been considered a problem only in the tropics. In 2000, the Ministry of Agriculture reported that around 50,000 ha in the state of Sonora were affected by this pest; in this area annual rainfall is around 400 mm and frosts are common in winter.

The excellent performance of Buffel grass under rain-fed conditions has led to its being evaluated under irrigation; forage yield can be increased up to five times compared to rainfed conditions; however, annual nutrient yield (protein and digestible dry matter)  ha^-1 of Buffel grass has not been as high as reported in some studies with lucerne (Medicago sativa) (Ibarra et al., 2004; Palomo et al., 2004).  

 Eragrostis curvula has been sown in two areas within the region, one is in the North-centre (state of Chihuahua) and the other is in the South-west (states of Jalisco, Zacatecas and Aguascalientes). This grass is an alternative to Buffel grass in areas with lower winter temperatures. Seeding rates are from 1 to 2 kg ha^-1. However, the rate of forage quality reduction as the plant reaches maturity and heavy accumulation of dead matter that limits regrowth in the following season have been pointed out as two major concerns in the management of Eragrostis curvula pastures (Chavez et al., 1996).

Panicum coloratum has been proposed as an introduced grass to be sown on abandoned croplands in the annual rainfall range of 450 to 550 mm. It is high forage yielding, but of slow establishment so local producers do not use this species extensively.

From some unpublished data it appears that the annual Eragrostis tef (Zuccagni) Trotter is worth evaluating as a source of high quality hay.

A major threat to animal production in the arid and semi-arid region of Mexico is the loss of herbaceous cover, increasing bare ground, shrub encroachment and top soil loss. Native grasses with higher forage potential are preferred by livestock which lead to overuse if grazing is not controlled; sowing is limited because of poor seed availability and establishment success. The native shrub with the greatest body of research is Atriplex canescens. Cenchrus ciliaris is the more widespread introduced grass in the region; its use has increased carrying capacity of the land, but winter kill and insect pests restrain its expansion and persistence.

Crop residues
Annual amounts of crop residues and by-products produced in Mexico are presented in Table 13. Other residues and by-products such as citrus pulp, pineapple process residues and waste bananas are locally used as feedstuffs in some regions (Jiménez, 1989). While by-products are commonly used in animal feeding, in spite of their high availability the use of crop residues as feedstuffs is not widespread. The relatively high costs of transport and the low feeding value are the main constraints. During the nineteen-eighties and nineteen-nineties the use of chemical treatments to improve feeding value of crop residues was a matter of research, however the adoption of these practices has been disappointing, probably due to their low cost-effectiveness. Crop residues are mainly used as feedstuffs in regions where crop and animal production coexist such as in Central Mexico; it is a usual practice of small farmers that grow crops and own a few animals. The fact that intensive-commercial agriculture and animal production are rarely integrated in mixed systems is a constraint for a more widespread use of crop residues as feedstuffs. Jimenez (1989) estimates that about 60% of maize residues are used - with varying efficiencies - in animal nutrition, the highest proportion is grazed or cut and fed in rudimentary feeding yards (common practices of small farmers), while about 10% is marketed as bales. The author reports that use of crop residues in animal feeding reaches 10% in the case of sugarcane tops (exclusively in sugarcane producing areas) and 30% in the case of wheat straw.

Photographs of crop residue use (click to view full image)

Mexican farmers have serious competitiveness difficulties facing the increasing import of subsidized products from developed countries or from countries of the Southern Hemisphere which produce at very low costs. Besides the unavoidable role of forages as a means to increase the sustainability of Mexican ruminant farms, increases in competitiveness should also come from improvement of farmers organisation (particularly in the field of meat marketing channels) and paying attention to aspects of modern animal production that until now have been set aside such as food safety, effects on the environment and animal welfare (Pérez, 2004).

Although intensive sheep production systems based on grazing sown pastures are being developed in all regions (since the high price of lamb makes this an attractive option), such as in the past few years, improvement of pastures resources in the near future will be closely linked to the development of competitiveness of the cattle dairy and beef sectors. The evolution of the dairy and the cattle meat production sectors have been different and therefore the role of forages has to be discussed for each sector.

After a severe crisis in the mid nineteen-nineties the dairy sector has grown rapidly. Most growth has been within the Specialized and Semi-Specialized dairy systems of the Plateau and North of Mexico, and nowadays these systems account for more than 70% of national production. For these systems, reduction of feeding costs is urgent, since this represents a high proportion of the theoretical world milk price. Efficient forage production and utilization might become the key issue to solve problems of high feeding costs and uneven seasonal distribution of dairy production.

Data in section 5 suggest that dairy farmers are increasingly relying on rainfed annual forages. This shift is seen as a means to solve two problems simultaneously, the reduction of production costs and the decreasing availability of irrigation water. However, priorities in research on temperate forages have not been closely based on the areas sown, particularly referring to irrigated and rainfed conditions, since more than 80% of research was on irrigated forages. Therefore the opportunities for improvement of forages for the Specialised and Semi-specialised dairy system needs the input of researchers placing priorities on varieties and management of annual rainfed forages. Within irrigated forages for these systems it appears that attention should shift from attainment of highest yield to such topics as water use efficiency and adjustment of forage alternatives to the requirements of systems e.g. adequacy of sowing and harvest times to dominant crop rotations, seasonal distribution of growth and persistence.

In the present circumstances, the evaluation of the economic feasibility of new forage technologies become a pressing need. Dairy systems based on grazing and the use of moderate amounts of conserved forage and few concentrates are an essential component of the competitiveness of the new leaders of the dairy world market. Grazing-based dairying might become an alternative for low-cost production in the Plateau and North of Mexico. The availability of sown forages for permanent pastures in sub-humid temperate conditions and in subtropical humid conditions is very low; finding appropriate options should lead to substantial increases in productivity and profitability in these conditions.

In contrast with the very well organised farmers of the Specialised Dairy System, the degree of organisation and integration of farmers of the Tropical Dual Purpose System is very low. Therefore, in the case of these farmers, improvement in organization and consequently integration becomes inevitable. Quality of product and seasonal variation of production appear to be the main constraints for this improvement. Forage researchers and governmental programmes must play an essential role in the reduction of the seasonal variation in production, where feeding strategies for periods of scarcity are crucial. Even though this is identified as a capital issue, Alianza para el Campo (a very extended governmental programme) has focused on sowing new pastures whereas forage conservation received very little attention. The evaluation of the economic feasibility and the adequacy to system requirements of different forage conservation alternatives should become research priorities in the near future.

In the case of beef farmers, experiences in Veracruz and Tabasco show that organisation and integration is a sound way to increase competitiveness. The reduction of feeding costs is also urgent in this case. Taking into account shifts in consumers preferences towards meat from feedlot animals, the development of efficient pre-finishing systems by grazing sown pastures in all regions come into sight as an appealing alternative. Due to high liveweight gain goals, strategies of supplementary feeding of grazing animals should become a key issue in the efficiency of these systems.

Dealing with problems of very small dairy farmers of the Family-based systems requires special policies. Their most important characteristic is that production and selling the product is usually not their main activity. Therefore, participatory research appears as the proper way to develop and introduce technological innovation, which should be focused on increasing their nutritional safety, enhancing opportunities for them to remain in their own communities, and improving ecological sustainability of their agricultural practices.

Research on pastures and forages is carried out by INIFAP (National Institute for Forestry Agricultural and Animal Production Research) and public universities. Research of INIFAP, is organised in eight Regional Research Centres, corresponding in general terms with regions depicted in Figure 1.

Northern Mexico
In Northern Mexico there are three Regional Research Centres of INIFAP: i) CIRNO (Northwest) covering Baja California, Baja California Sur, Sinaloa and Sonora (143 researchers, of whom 21 have a PhD degree), ii) CIRNOC (Central North) covering Aguascalientes, Chihuahua, Durango, La Laguna and Zacatecas (193 researchers, of whom 38 have a PhD degree) and CIRNE (Northeast) covering Tamaulipas, Nuevo León, San Luis Potosí and Coahuila (122 researchers, of whom 12 have a PhD degree). Beef cattle, dairy cattle and goats are the main animal production systems considered for research. A great deal of research is related to extensive production systems based on grazing of native grasslands: rangeland ecology, management, conservation and rehabilitation of degraded rangelands, Atriplex spp. as important shrubs for salty soils, native grasses such as Bouteloua gracilis and Sporobolus airoides, pasture establishment in rainfed rangelands particularly buffel grass, green panic (Panicum maximum var trichoglume) and kleingrass (Panicum coloratum) and nutritional aspects of cattle grazing on rangelands. Research related to intensive production systems include forage production and grazing management of irrigated and rainfed small grains, irrigated annual ryegrass and berseem clover, lucerne  varieties and drip irrigation, maize and sorghum varieties and silage technology, rainfed summer legumes (Canavalia ensiformis, Mucuna pruriens and Lablab purpureus), and conversion of degraded cropland into sown pastures.

Six universities in Northern Mexico have research and graduate studies programmes dealing with animal production, pastures and forages: Universidad Autónoma de Chihuahua (UACH), Universidad Autónoma de Nuevo León (UANL), Universidad Autónoma de Tamaulipas (UAT), Universidad Autónoma Antonio Narro (UAAAN) in Coahuila, Universidad Juárez del Estado de Durango (UJED) and Universidad Autónoma de Baja California Sur (UABCS). Research topics in these universities correspond in general terms with those described for INIFAP. In Table 14 a brief summary is presented of personnel involved in research on pastures and forage related topics in Northern Mexico. 

Central Mexico
INIFAP has two Regional Research Centres in Central Mexico: CIRPAC (Central Pacific) covering Colima, Jalisco, Michoacán and Nayarit (174 researchers, of whom 22 have a PhD) and CIRCE (Cent