Appropriate dry matter degradability (DMD) results in the efficient absorption of nutrients into the blood stream of an animal. Therefore, this study was aimed to determine the effect of fibre content on the in vitro DMD of beef cattle finisher diets. The study procedure was approved by the ARC – AP (Irene) Research Ethics Committee, reference number: APIEC15/047. Stomach tube was used to collect rumen fluid from three Nguni steers (380 ± 25 kg live weight) fed a feedlot finisher diet. Rumen fluid was collected within two minutes per animal, which was then poured into a pre-warmed flask and sent to the laboratory within 20 minutes, where it was filtered through 2 layers of cheesecloth and kept at 39°C under constant flushing with CO2. Treatment diets contained with three crude fibre contents (41,54 %, 18,18 % and 10,77 %, referred as diet 1, 2 and 3, respectively) were used in the study. Dried feed samples were weighed (0.5 g) into nylon bags (5 cm x 10 cm; 40 µm pore size) and heat-sealed. Samples were incubated in ruminal fluid combined with a buffer solution (1:4, vol/vol) for 0, 2, 4, 8, 12, 16, 24 and 48 hours according to the in vitro (IV DaisyII) procedure. After incubation, bags with feed residues were rinsed thoroughly with tap water until water run clear, dried in a forced-air oven at 60°C for 48 hours and weighed. Dry matter disappearance data were fitted to the equation: Y(t) = a + b (1 – e-ct), where Y(t) = dry matter disappearance (%) at time (t) hours, a = soluble or rapidly degradable fraction, b = insoluble or slowly degradable fraction, c = fractional rate constant of degradation of b (1/h), t = incubation time and e = base for natural logarithm. There was a significant difference (P < 0.05) on the DMD of treatment diets for all incubation periods. A notable difference was observed in DMD (Mean ± SD) at 48 hours of incubation with 54,9 ± 0,71 % for diet 1; 61,6 ± 0,38 % for diet 2 and 71,5 ± 0,79 % for diet 3. A high fibre content such as in diet 1 (41.5 % CF) resulted in a marked decline in the DMD of the feedlot diets evaluated.

Winter cereals play a principal role in maintaining livestock condition during winter months in the Highveld. Due to the potential of these species to achieve high winter yields and forage quality, they remain the preferred species for green winter feed for highveld mutton producing animals. Despite harsh environmental conditions, the Highveld in Mpumalanga is a stronghold for South Africa’s agricultural production, boasting large herds of mutton sheep. Recent evidence shows that climate change, particularly trends related to increased seasonal variability of rainfall, has led to increased constraints on water availability in both rain-fed and irrigation-based agriculture. However, the effect of climate change, particularly those leading to below mean annual precipitation (MAP) in the Highveld, on the productivity of winter cereals is unknown. The aim of this study was to assess whether below average rainfall would adversely affect the production potential of winter cereals the Highveld area. Assessing whether resistance of winter cereals to drought is associated with seed size of cereal cultivars was a secondary aim of this study. Three winter cereals species (Avena sativa cv. Overberg (Oats), Triticale hexaploides cv. Cloc 1 (Triticale) and Secale cereale cv. LS62 (Stooling rye)), were evaluated. Each treatment was replicated three times in a randomised experimental design (plot = 1 ha) and greenhouse experiment that were planted mid-March 2014 at the Athole Research Station. The study was carried out seasonally from 2014 to 2017 in order to take into account the effect of seasonal variation on productivity. The ten year average for total autumn rainfall at the Athole Research Station is 320 mm. The actual rainfall during the four year study period was 70, 269, 325 and 439mm for 2014 to 2017, respectively. Seed size of the varieties evaluated were categorized according to seed weight as large, medium and small for A. sativa, T. hexaploides and S. cereale, respectively. Production potential (kg DM ha-1) was evaluated a week before grazing during the second week of June each year using an exclosure cage, a 0.16 m2 quadrat and falling plate. Productivity was associated to rainfall seasonality. There was a significant positive correlation between drought level (rainfall/soil moisture availability) and yield two of two species (A. sativa, R2=0.90 and S. cereale, R2=0.86). Therefore, yields for consecutive years of study for the respective species were significantly different (p< 0.05). Emergence rate was evaluated as percentage of seed germinated for each species at regular day intervals (7d - 42d) and simultaneously vigour was inferred. Avena sativa and Secale cereale had significantly higher (p< 0.05) emergence and vigor than T. hexaploides during seasons with above normal rainfall, with the opposite true (P> 0.05) when rainfall decreased below normal (320 mm). The species with either large or small seeds (oats and stooling rye respectively), showed a greater ability to maintain yield potential during periods of moisture stress resulting from below average rainfall. In contrast, triticale was able to emerge, but not maintain production, during periods of short term water stress. We thus conclude that the results support the research hypothesis of winter cereals productivity varying with drought level. We recommend that producers should consider interchanging winter cereals based on the predicted or apparent rainfall, rather than being reliant on one species alone. Alternatively, T. hexaploides should be diversified with the other two cultivars during any one year to maintain forage production during unexpected drought conditions.

This study investigated the influence of season, distance from homesteads and topography on species composition and biomass production in two communal rangelands surrounding homestead areas, in Dyamdyam and Machibi in the Eastern Cape. At Dyamdyam, rangelands are set on a flat slope. Therefore rangelands were divided according to distances from homesteads into near (up to 1 km), middle (> 1–2 km) and far (> 2–3 km). At Machibi rangelands are set on steep slope. Therefore, rangelands were divided into position on hill-slope into bottom, middle and top sites. In each site a Homogeneous Vegetation Unit (HVU) of 50 m x 20 m was used and six 0.25 m2 quadrats were laid randomly on each HVU. Herbaceous species found within a quadrat were identified, counted and their height and tuft diameter were also measured. Herbaceous species were also harvested, bulked and placed into well labelled brown paper bags and oven-dried for 48 hours at 60 oC to determine biomass production. Herbaceous species were classified according to their palatability, ecological status and life form. Data were collected for two seasons over 2014/15 (winter and summer). A total of 20 herbaceous species were identified in the study areas. Of these, 17 were grass species. The most common or dominant grass species were Themeda triandra, Cynodon dactylon, Erograstis capensis, Erograstis plana and Sporobolus africanus. Themeda triandra had a higher frequency of occurrence at the far site than at the middle and near homesteads sites at Dyamdyam. The frequency occurrence of S. africanus at Machibi was similar in all the study sites. Density of herbaceous species was significantly lower on near and bottom sites at Dyamdyam and Machibi, respectively. In both winter and summer, far and top sits at Dyamdyam and Machibi respectively, had higher (p < 0.05) biomass production. The biomass production was significantly higher in the summer than the winter across the study sites in both study areas. It can be concluded that season, topography and distance from homesteads are important factors that affect biomass and grass species composition. Therefore, rangeland management programs in communal grazing lands should consider these factors. Adoption of rotation grazing in communal rangelands is recommended to improve grass species diversity and biomass production.

During spring, kikuyu pasture has high content of non-structural carbohydrates (NSC); however, dry matter and fibre content are low. The low fibre content affects fibrolytic enzymes-that affect, nutrient digestibility and subsequently intake, health and performance. Kikuyu is also high in oxalates, which reduce absorption of calcium and exacerbates metabolic disorders in lactating cows. High energy and high fibre feed and forage hay are critical for sustainable milk production in spring and early summer when nutrient supply from pasture is limited. Supplementing lucerne during spring would increase diet NSC, structural carbohydrates and calcium.The objective of the study was to determine effects of fresh lucerne (CP: 21.1 %; ME: 9.8 %; NDF: 42.7 5) and concentrate supplementation on milk production of lactating Jersey cows grazing kikuyu pasture (CP: 22.9 %; NDF: 27.8 %; ME 11.6 MJ/kg) during spring. Twelve mid-lactating primiparous Jersey cows (from the Dairy herd of the Bethel Agricultural College in Butterworth (Eastern Cape) were used in a randomised complete block design experiment. Three weeks before the start of the study, Jersey cows (DIM: 105 ± 11; BW: 398 ± 34 kg; 13.6 ± 8 kg) were assigned to one of six blocks of two cows, based on milk production during the two weeks prior. The two treatment groups were 1) control (not supplemented with lucerne), and 2) supplemented with 2.5 kg d-1 of chopped lucerne hay. Cows in both groups received 6 kg d-1 (as is) of concentrate. Milk yield was recorded daily and milk samples collected on one day weekly to determine milk composition. Body weight (BW) was measured at start and end of the experimental study (30 days). Milk yield averaged 14 kg d-1 and did not differ between the two groups. Milk protein and lactose did not differ between the two groups and averaged 3.71% and 4.79 % respectively, but milk fat concentration was increased (p< 0.05) by supplementing lucerne (4.5 vs 4.25 %). On average, both groups produced 0.64 and 0.53 kg d-1 of milk fat and protein, respectively. Milk urea nitrogen values were within acceptable ranges, but were slightly lowered (P= 0.04) by supplementing lucerne (13.5 vs 15.9 mg dL-1). The initial and final BW did not differ between the two groups. The absence of change in milk performance suggests that at the current feed intake, no rumen condition could justify supplementation of additional buffering fibre from lucerne hay.

There is a growing interest in Guinea grass (Megathyrsus maximus, formerly Panicum maximum) for its use as forage to feed livestock under a range of sub-saharan African (SSA) conditions. Traits related to a plant’s water use influence the fitness of genotypes for specific rainfall and evaporative demand environments. Atmospheric vapor pressure deficit (VPD) has been widely recognized as the force driving transpiration in plants, thereby influencing their efficiency in water use. The objective of this work was to test genotypic differences in water use efficiency (WUE: biomass produced per amount of water transpired), biomass production and leaf senescence of 126 accessions of M. maximus under three levels of VPD conditions (1.5, 3.0 and 4.0 kPa) and two irrigation conditions (soil kept at 90% and 60% of field capacity) under greenhouse conditions for two weeks. Overall, WUE, biomass production and leaf senescence were not affected in most genotypes when growing under moderate atmospheric and soil drought (3.0 kpA of VPD and 60% field capacity) when compared to plants growing under low atmospheric drought (1.5, kPa) and soil kept under 90% of field capacity. However, growth under high atmospheric drought (VPD of 4.0 kPa) resulted in reduced WUE (15% to 22% of reduction under soil kept at 90% and 60% of field capacity respectively) and biomass (26% to 35% under soil oil kept at 90% and 60% of field capacity respectively). Under moderate soil water content (60% of field capacity) and high atmospheric drought (VPD of 4.0 kPa), at least a two-fold increase in leaf senescence was recorded for all genotypes tested compared to the others treatments. Carbon starvation and hydraulic failure were likely the basis of reduced WUE and biomass and increased leaf senescence under high atmospheric drought and moderate soil water content in all genotypes tested. Our results suggest that the more productive Guinea grasses (15 genotypes showing biomass production of ~3 g/pot, which could be translated into an estimated productivity of above 25 ton dry matter/hectare under high fertility conditions) could be suitable options for forage production, without big yield penalties, in tropical agro-ecosystems across SSA where short, frequent and mild droughts and low atmospheric evaporative demand conditions are common (i.e., a range of humid to subhumid agroecosystems with annual precipitation above 900mm). Further research is needed to test promising genotypes under field conditions. This work was partially funded by CGIAR Research Program on Livestock (flagship on Feeds and Forages) and BMZ funded project "Improved forage grasses: Making the case for their integration into humid- to sub-humid livestock production systems in Kenya and Ethiopia".

The genus Indigofera is being evaluated for possible pasture potential of certain species occurring within the genus. There are at least 196 Indigofera species indigenous to southern Africa. Distribution records shows that species in the genus Indigofera occur primarily in the Savanna and Grassland Biomes. The majority of Indigofera species are herbs, followed by shrubs and dwarf shrubs. Even though nearly 12 Indigofera species are known to contain indospicine, a free amino acid that causes hepatotoxicity (Fletcher et al. 2015), a few species have been through a forage selection and evaluation process. These include I. alternans var. alternans and I. vicioides var. vicioides (contains indospicine) with Müller et al. (2017) prioritising I. meyeriana and I. nigromontana for further evaluations for water-limited pasture systems. The aim of this work was to describe the Indigofera species indigenous to South Africa, Swaziland and Lesotho, in terms of their adaptation or range of tolerance to mean annual rainfall, mean annual minimum and maximum temperatures, soil pH and phosphorus levels. A discriminant analysis was performed to determine which bioclimatic and/or edaphic predictor contributed mostly to species distributions, as recorded by PRECIS (2008). A high correlation between distribution and mean annual rainfall and soil pH were noted. Species such as Indigofera auricoma and I. exigua are listed only in one group, having several occurrences in the Kalahari Duneveld and Namaqualand Hardeveld Bioregions respectively. These species are therefore narrowly adapted to these bioclimatic and edaphic conditions. Overall, Indigofera species are found mostly in the Central Bushveld and Mesic Highveld Grassland Bioregions and fall in the 400–800 mm mean annual rainfall regions with a soil pH level of < 6.4. Species identified as having a high pasture value and adapted to low soil P, are I. alternans var. alternans, I. cryptantha var. cryptantha (contains indospicine), I. rhytidocarpa subsp. rhytidocarpa, I. torulosa var. torulosa and I. vicioides var. vicioides (contains indospicine). Of the few species adapted to low soil phosphorus levels and relatively low rainfall, Indigofera alternans var. alternans is the only herb recorded, occurring mainly in the Central Arid Region Group. These results shows that the genus Indigofera has numerous promising species available for further screening and pasture evaluations.

Planted grasses and legumes are less expensive feed sources compared to grain based concentrates and provide dietary fibre that is crucial for stimulating rumen activity. Fibre fractions (cellulose, hemicellulose, lignin, cutin, silica) and crude protein availability determine relative forage value (RFV). Legumes tend to have a higher nitrogen and lignin content per kg fibre than grasses, although the latter have a higher tensile strength. Reducing lignification of fibre through genetic manipulation affects plant morphology and viability; hence soil fertility practices that minimize lignification of protein and carbohydrate fractions are preferred. Inorganic fertilizers improve pasture growth and lower lignin content, but are not environmentally friendly. The objective of the study was to determine the effect of bio-digestate application on Vigna unguiculata (VU) and Pennisetum purpureum (PP) pastures and evaluate quality of lignified nutrient fractions. The hypothesis was that organic fertilization would improve the uptake of soil nutrients and minimise lignification of plant carbohydrates during the early growth period. A factorial experiment (factor 1: forage type- legume vs grass; Factor 2: three levels of digestate application) was done in a temperature-controlled environment using potted plants. Vigna unguiculata (VU) and vegetative material of PP were planted in pots that were fertilized using semi-solid digestate from a cattle/sheep abattoir. Samples of stems and leaves were harvested weekly from 21 days to flowering point in VU and 6 weeks for PP to assess carbohydrate lignification. Samples of stems and leaves were analysed for total fibre and crude protein, soluble and insoluble fractions and indigestible components following 300 hrs of in vitro fermentation. Interrelationships between fibre fractions and organic N and C were determined and comparative synthesis of variance in lignin complexes between VU and PP were assessed. Preliminary results show that application of the organic resin improved leaf N of both PP and VU by 12 and 14.5%, relative to control plots. Leaf and stem lignin content of both VU and PP were less than 2% at 21 days; however, rapid accumulation was noted in PP stems from week 5 reaching 18% NDF in dry matter (DM) and 5.3% unavailable carbohydrate. There was a similar trend in VU however; leaf lignin was lower for VU at flowering (4.1% DM) compared to the PP at all fertilization levels. At most, 9.1 % dry matter disappearance (DMD) was noted for PP lignocelluloses (high organic matter plot) disappeared by 240 hrs and no significant change was noted thereafter; materials from had significantly lower rates of disappearance. There was no variance in VU disappearance across treatments over the 300 hr fermentation period and maximum disappearance was 7.2% DMD, which was higher (P< 0.01) than control (4.4% DMD). Pasture fertilization with animal sourced bio-digestate did not result in significant sloughing of lignified cellulosic materials and further research on application methods and determining optimum bio-digestate application rates is essential to achieve higher nutrient supply from pasture.

Bidens pilosa is a pioneer weed listed as a global invasive species found in disturbed grazing lands and planted pastures, which are used for grazing. It forms dense stands that can out-compete, out-grow, and eliminate native vegetation of the lower strata. Its burrs are a nuisance to people, sheep and also a troublesome seed contaminant as they are difficult to separate once attached to wool during grazing. It creates challenges during both shearing and acts as undesirable foreign object during packaging of wool, thus negatively influences the yield and price of wool production. The coincidence of wool shearing season with the seeding period of B. pilosa in the Highveld negatively affects wool production. The Dohne Merino breed, the main wool producing sheep of the Highveld, is highly affected by contaminants of B. pilosa. The aim of this study was to determine the impact of B. Pilosa contamination on wool production yields of Dohne Merino sheep breeds in the Highveld of Mpumalanga Province. This study serves to quantify the damage on wool profits and highlight the significance of controlling B pilosa on grazing lands. The study was conducted at the Athole Research Farm in the Highveld region of Mpumalanga Province. The data was collected during three consecutive shearing seasons. Three groups of 150 Dohne Merino sheep were separated andgrazed on camps with various levels of B. pilosa infestation demarcated as none, low and high infested pasture camps. The groups of sheep were shorn, packaged and weighed separately to determine wool yields and analysed using excel. There was no difference in wool production yield between sheep in none infected and camps with low B. pilosa. Mean wool production yields per small stock unit of 3.0 kg, 2.6 kg and 2.0 kg were observed for none, low and high B. pilosa infested camps, respectively. The results demonstrate that wool production yield per sheep is inversely proportional to pasture infestation by Bidens pilosa, thus wool production yield increases with a concomitant decrease in infestation of the grazing camp. It is therefore recommended that wool producing sheep or any fleece producing livestock be prevented from grazing on camps infested with B. pilosa. The use of none fleece producing livestock at correct stocking rates will enhance re-establishment of indigenous vegetation and eliminate B. pilosa. These biocontrol mechanisms are economically viable and more sustainable than the use of highly expensive labour intensive physical and chemical control mechanisms.
 

The sustainability of forage production in tropical and subtropical areas is affected by environmental and climatic conditions. Certain varieties of forage species may show better adaptability to certain conditions and produce greater yields throughout the growing season. A study was conducted at Owen Sithole Research Station near Empangeni, KwaZulu-Natal to evaluate three Panicum maximum cultivars (Gatton, Tanzania, Mombaca), three Brachiaria brizantha cultivars (Marunada, Xaraes, BRS Piata), two Brachiaria humicola (Humicola, Lianero) and one Brachiaria decumbens cultivar (Basalisk) for in terms of dry matter yield for and use in livestock production systems to increase fodder availability. The area has a sub-tropical climate and is frost free. The trial was laid out in a fully randomized block design with three replications under dryland conditions and with a nett experimental plot sizeof 10 m x 2 m. The results for the 2016/17 growing season showed that some cultivars performed differently during different seasons i.e. summer and winter. The best summer yield was for P. maximum, cv. Gatton (3.3 t DM ha-1) and the lowest from B. brizantha cv. Xaraes (1.5. t DM ha-1). During the winter, the highest DM yield was for the B. decumbens cultivar Basalisk (4.3 t DM ha-1).  The P. maximum, cv. Gatton produced the second highest DM yield (3.3 t DM ha-1). The Brachiaria brizantha cultivars Marunada and PRS Piata, and Brachiaria humicola cultivar Lianero were dormant during winter. The winter and summer yields of all three Panicum cultivars compared well, with Gatton as best performer. The Brachiaria decumbens cultivar (Basalisk) performed well in both seasons and seemed to be, as the Panicum cultivars, tolerant of cold. The Brachiaria brizantha cultivars Marunada and BRS Piata seemed to be sensitive to cold, and Xaraes, which produced some material in winter, performed poorly in both seasons. The data produced can assist in cultivar choices for the farmer in terms of total yields and seasonal production trends. The trial will continue to collect data from three years of results and will be replicated under irrigation. Nutritional values will be also compared between cultivars and seasons.

Annual forage cereals including white oats (Avena sativa), black oats (A. strigosa), red oats (A. byzantina), rye / stooling rye (Secale cereale) and Triticale (x Triticosecale) are frequently sown during the autumn months under dryland conditions for subsequent grazing by livestock during late autumn, winter and early spring. An experiment was planted in April 2016 to compare dry matter yield of early-, medium and late-flowering forage cereal cultivars. This experiment was conducted on a Hutton soil type at the Cedara Research Station in the KwaZulu-Natal Mistbelt, South Africa (29º 32´S 30º 16´ E), at an altitude of 1075 m. The total rainfall recorded during the trial period from the planting date through to the final cut was 166.13 mm. Seed of 30 forage cereal cultivars of the above-mentioned species was sown at the recommended sowing rates with three replications in a 5 x 6 balanced lattice trial with a gross plot size of 2 m x 6 m and a nett plot size of 1.4 m x 4.6 m. All plots received 50 kg N/ha and 50 kg K/ha as 1:0:1 when the seedlings were approximately 10 cm high and the same amount again after each herbage cut. The plots were cut four times with a mower blade set at 5 cm above ground level when the plants reached grazing height. The fresh herbage was weighed within two to five minutes of cutting and the samples dried in an oven set at 75ºC. Early-flowering cultivars, for example, the rye cultivars LS 35 (1,05 t/ha) and SSR 727 (1,48 t/ha), generally yielded relatively well in late autumn (Cut 1: 07 June 2016) when they were in the early reproductive/flowering phase; while medium-flowering cultivars such as the stooling rye cultivars PAN 263 (0,96 and 2,79 t/ha) and NCD Grazer (0,65 and 2,97 t/ha) yielded relatively poorly during late autumn (Cut 1: 07 June 2016) but relatively well in late winter (Cut 3: 01 August 2016). On the other hand, late-flowering cultivars such as the Triticale PAN 299 (2, 64 t/ha) and the stooling rye cultivar LS 62 (2,89 t/ha) were amongst the highest yielding cultivars as they moved into the reproductive phase in late winter and early spring (Cut 4: 20 September 2016). ANOVA indicated significant differences (P < 0.05) among early, medium and late-flowing cultivars, although the LSD showed that mean dry matter yields for early vs. medium and for medium vs. late-flowering cultivars were not significantly different. The LSD and CV for the total dry matter yields of the 30 different forage cereal cultivars were 1.15 and 11.11 % respectively. While forage cereal cultivars that are highly ranked in total dry matter yield should be noted in deciding which forage cereal cultivars to utilize, careful consideration must be given to their flowering behaviour and their related fodder flow distribution.

Oversowing Italian ryegrass into kikuyu in autumn can be an aid to improving quality and quantity of pastures during the winter and early spring months. A trial conducted on Cedara research station in KwaZulu-Natal in 2016 and 2017, aimed to test the use of glyphosate herbicide at two levels and four different times of application vs the mechanical method of mowing at two levels to set back the kikuyu enough to oversow with Italian ryegrass (Lolium multiflorum). Glyphosate was applied at 400ml/ha or 800ml/ha either three weeks, two weeks, one week or one day before planting. The mowing treatments were to cut to a kikuyu residual canopy height of 5cm or 2cm one day before planting. A pure ryegrass control was attempted by applying glyphosate at 4l/ha (Lm control) for a complete kill of the kikuyu. In the second year no herbicide was applied, but all plots were mowed at 2 or 5cm before planting to determine any carry-over effects of the herbicide. The first year (2016) autumn results were erratic, with no clear indication of which herbicide concentration or time of application was superior, seemingly environmental effects vershadowed time of application or concentration of herbicide. In spring, there was very little difference in yield between treatments apart from the kikuyu control, which was the lowest yielding. Total yields across treatments in 2016 were not significantly different averaging 17.1 t DM/ha, with the exception of the unexpectedly low yielding ryegrass control. The second year (2017), no herbicide was applied and the trial was planted later in the season leading to a low yielding autumn. In winter all treatments were similar except for a significantly higher yield for the ryegrass control and lower yield for the kikuyu control. Total yields across treatments in 2017 were not significantly different averaging 14.2 t DM/ha, except for the ryegrass and kikuyu controls, which were lower. Year two had lower overall yields than year one, as it had a shorter season. The time of year that pasture was available and the species of grass dominating differed between treatments. What is important is that between 4 to 6 t DM/ha of pasture is available when it is needed more, in autumn, winter and early spring, and that it was of a better quality than pure kikuyu. Over two years it appeared that oversowing ryegrass into kikuyu gave a yield advantage over either species alone. The time of planting remains critical – an earlier planting, without herbicide, runs the risk of being swamped by kikuyu regrowth, planting too late may limit ryegrass production until late winter or early spring. The success and cost to this exercise is irrigation as it will not be successful if planted dryland.

Inadequate nutrition for goats remains a major constraint among smallholder communal farmers particularly in dry regions, where, ironically goats are of high importance. A number of forages have been tested and gave encouraging results as supplements for goats on-station. However, relatively less on-farm testing of the same forages as supplements for goats has been done. The present study was conducted to determine the effect of supplementing goats with forages: Bana grass (Pennisetum purpureum) and Velvet bean (Mucuna pruriens) hays in one of the driest parts of Zimbabwe, Beitbridge, during the dry season on growth, kid mortality, milk quality and milk offtake. The study lasted ten weeks. Farmers in eight of the 15 wards took part in the study. The trials followed a nested design that took into account sex, forage supplement, goat class and farmer. Each of the eight farmers acted as a replicate. Data on weekly weight changes, pre-weaning kid mortalities, milk composition and milk offtake were recorded. Repeated measures ANOVA of SAS and non-parametric tests of SAS were used to analyze for the effect of supplements on weekly weight changes, milk composition, milk offtake and pre-weaning kid mortality. Dietary supplementation resulted in greater weight gains than non-supplementing (P<0.05) but the difference was not significant between the supplemented groups (P>0.05). There was no significant effect of supplementation on milk composition, offtake was significantly higher in supplemented groups. Supplements had no significant effect (P>0.05) on pre-weaning kid mortality. Supplementing goats is therefore important during the dry months of the year as it ensures maintenance of goat weights and also availability of more milk for human consumption. Goat production is the most important economic activity in Beitbridge district throughout the year and maintenance of live weights will mean farmers can sell their goats throughout the year and get good returns per animal. We therefore concluded that supplementing goats with Bana grass or Velvet bean hay during the dry season in this arid region will improve weight gain and increase milk that can be available for household consumption; confirming results observed on-station.