Agriculture and Food Systems - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 74
  • Item
    Thumbnail Image
    Responses of young sheep to supplements when fed low quality roughages
    Rafiq, Mohammed ( 1999)
    The morphological components of which cereal straw is composed vary in composition and nutritive value. In a survey of relevant literature, stem materials (ST) are usually found to be consumed by ruminant animals at a slower rate than leaf material (LF) from the same crop residues when these are fed alone as separated fractions. This is attributed to the higher content of cell wall constituents (CWCs) and often lower content of N of ST. Because ST and LF can vary in proportions in the roughage fed as a basal feed, the response of animals to supplements might also be expected to be variable particularly when the basal feed is offered in excess that permits selection. In particular the response to supplemental N sources varying in ruminai degradability may vary. The objective of this thesis program was to investigate the interactions between the basal roughage and supplement measured as effects on digestion and LW responses of young sheep. The overall hypothesis was that across diets made up of different proportions of LF and ST fractions of cereal straw, the response to N supplements is dependent on CWCs concentration. The program was completed through a series of experiments conducted at the Mt Derrimut Field Satation of the University of Melbourne. In all experiments the animals were fed on a basal feed of LF or ST fractions of barley straw, with supplements including (a) Bar+USS, barley grain (Bar) fortified with urea solution prepared at a ratio of 5:1 (urea plus Na,SO4). (b) USS, urea solution alone added to the basal roughage and (c) FM, fishmeal. With each basal feed one group of lambs did not receive any supplement and served as a control group (CONT). Experiment 1 (Chapter 3) was conducted to evaluate chemical and nutritional characteristics of straw fractions of Parwon cultivar barley. Straw was separated into 4 fractions - stem (ST), leaf blade (LB), leaf sheath (LS) and broken fractions plus weeds (OT). The separated fractions were analysed chemically ( van Soest, 1974) and in vitro digestibility (Tilley and Terry 1964 ) determined. ST was the largest fraction and contained a significantly higher concentration of neutral detergent fibre (NDF; p<0.01) than LB, LS and OT (83.1, 78.6, 76.8,and 71.5 g/100g respectively). ST contained less hemicellulose (HC) than LB but more than LS and OT (37.5, 39.6, 36.3, and 35.7 respectively). N content was lower in ST than in LB, LS and OT fractions (0.4, 0.9, 0.6, and 0.7 respectively). Digestibility in vitro was significantly lower (P<0.001) for ST than for other fractions (38.5, 72.7, 60.1, and 63.0 respectively) while energy required for grinding (Chenost 1966) was much higher (P<0.001) for ST than for other fractions (121, 54.6, 64.2, 56.6 respectivly). In Experiment 2 (Chapter 4) ST and LF fractions of the same Parwon barley straw were fed as the basal feed to lambs and DM intake of ST was 15% lower than for LF (403 vs 473 g/d). When LF feed was supplemented with USS and FM, DM intake was greater by 28% and 25% respectively, while supplementation with Bar+USS resulted in 10% lower LF intake. In contrast, with animals fed ST as the basal roughage, only FM led to an increase in DM intake of only 10%. Supplementation with Bar+US and USS and FM improved overall digestibility, estimated metabolisable energy intake and N intake. Low N intakes on the basal roughages supported low ruminai ammonia-N concentrations (mg/1) immediately before feeding (ST, 20.4 ; LF 35.8), but these were improved where supplements had been fed with each of ST Bar+USS, 263.7; USS, 186.7; and FM, 151) and LF (Bar+USS, 219.5; USS, 62.5; and FM, 150). Six hours after feeding, ammonia-N concentrations (mg/l)were higher for ST (99) still low for LF(35) when fed alone, reduced below the prefeeding levels by supplements of Bar+USS (ST,167; LF 173) but raised by USS (ST, 201; LF 148) and FM (ST, 114; LF, 192). The concentrations of total volatile fatty acids (VFA) in rumen fluid (mMoUl) were not significantly different for ST and LF before feeding except where FM was the supplement, or six hours after feeding except where Bar+USS or USS were fed with LF (before feeding: ST, 51; Bar+USS, 55.7; USS,46.5; FM, 56 ; LF, 42.6; Bar+USS, 53; USS, 55.5; FM, 63.6; 6h after feeding ST, 55.5; Bar+USS, 70.2; USS, 62.1; FM, 50.4 ; LF, 55.7; Bar+USS, 69.5; USS, 60.9; FM, 60.9). Lambs on ST and LF alone lost weight (ST, -105; LF -98 g/d ). Rate of liveweight loss was less when Bar+USS (ST, -32.3; LF -2.4g/d) and USS (ST, -79.8; LF -31.2 g/d) supplements were fed, while FM promoted LW gain (ST, 37.5; LF, 72.4g/d). N retention data was consistent with these LW gains, except where Bar+USS was the supplement, in which case the animals were in positive N balance though losing weight. Wool growth was significantly improved (P<0.001) only by FM on both ST (+47%) and LF (+57%) basal roughages. In Experiments 3 and 4 (Chapter . 5 and 6) the objective was to investigate the factors responsible for low feed intake and poor performance of lambs fed ST compred to those on LF. The DMI (g/d) of lambs fed ST and LF were similar to those achieved in experiment 2; and were significantly greater for LF (P<0.001). Likewise supplements of Bar+USS, USS and FM had similar effects to those reported for experiment 2_ Changes in DMI, MEI ruminai environment before feeding or 6 and:12 hours after feeding were consistent with those recorded during experiment 2. However concentration of total VFA was significantly elevated at 12 hours after feeding. Differences in ruminal environment were evident in terms of VFA concentrations and the distribution of rumen digesta particulate material in different size fractions; both variables were affected both by the basal diet and the supplement. For LF, the proportions of particles >2mm and of very fine particles (0.125 mm) were greater and for particles between 0.5 and 1 mm less those for ST in all cases. Further, the proportion of particles >2mm was less where FM was fed than for any other feeding regime. The mean retention times of rumen fluid, measured from CoEDTA dilution rate, and calculated for rumen particulate material was longer (24%, P<0.01) for ST than for LF but there was no significant effect of supplement on this (Experiment 4, Chapter 6). The mean percentage of very fine particles in the faeces of lambs fed on LF was higher than for lambs fed ST alone or with supplements. Rate of ruminai degradation of OM of ST and LF as measured by nylon bag technique ( Experiment 3, Chapter 5) was similar at 12 and 24 hours but greater for LF than for ST at 48 and 96 hours of incubation. Bar (cracked whole grain) was degraded more rapidly and extensiveley than FM; in LF fed sheep this difference was more marked. Rate of degradation of acid detergent fibre (ADF) was influenced by the kind of supplement and was greatest in lambs given FM , and least in lambs given ST with no supplement. Only the FM supplement resulted in LW gains, though rates of LW loss were least and LW gains with FM were greatest with LF as the basal roughage. The responses are interpreted as flowing from the greater proportion of ADF and lignin in the CWCs content and the greater digestibility of ADF in the LF fraction. The ST feed fraction with higher concentrations of cell wall constituents (CWCs) as NDF was eaten at a slower rate (Experiment 5, Chapter 7) and digesta particulate material and, in these experiments, the fluid phase are retained longer in the rumen. LF showed not only an advantage over ST in these respects but also in terms of a number of important digestion parameters supported a greater response to supplements, particularly N supplements of low degradability. Thus FM is these experiments interacted with the roughage component of the diet. It provided more consistent ruminai ammonia concentrations supporting a better environment for microbial activity and growth. Microbial protein together with undegraded dietary protein together provide a balance of nutrients that allows LW gains on otherwise submaintenance basal feeds. The greater enhancement of performance with LF compared to ST and the particle size measurements suggest that greater fragmentability of LF may be a major contributor . In terms of technical improvement of livestock feeding systems, providing the animal with opportunity for selection of more leaf and less stem may improve the likelihood of responses to supplements but this was not demonstrated in Experiment 5. FM was used as the experimental supplement to provide slowly degraded and undegraded dietary protein of high biological value to the animal. FM is expensive and other crop byproducts and local feed materials with properties of slow degradability of protein and good amino acid balance need to be identified. An alternative strategy would be to provide a maximum opportunity for the selection of most digestible parts. If refusals are then collected, quality could be further improved with alkali treatment and necessary supplementation. This would provide a strategy for the use of morphological fractions which could be an economical approach for the efficient utilization of roughages.
  • Item
  • Item
    Thumbnail Image
    Nutritional modification of muscle long-chain omega-3 fatty acids in lambs : effects on growth, and composition and quality of meat
    Ponnampalam, Eric Nanthan ( 1999)
    Three experiments were conducted to investigate the effects of dietary supplements rich in omega-3 fatty acid on muscle omega-3 fatty acid deposition. The consequential effects on growth performance of lambs, colour and lipid oxidative stability of muscle over refrigerated display, and the sensory properties of cooked meat were also examined. A mixture of lucerne chaff : oaten chaff was used as basal diet, offered in different proportions were fed to lambs ad libitum (Expt. 1) or at 90% ad libitum (Expts. 2 and 3). Such mixtures of roughage diet support slow growth and provide a feed quality pattern similar to late spring to late summer pasture. In Expt. 1, fish meal (7%), canola meal (8%) and soymeal (7%) as natural feed supplements were compared in lambs fed low quality roughage diet. In Expt. 2, fish meal (9%) and oilseed supplements either in unprotected form (rapeseed - 7%) or in protected form (ground canola seed - 6%) were examined in lambs on medium quality roughage diet. Lipids and the proteins in the ground canola seed were treated (RUMENTEK) with aldehyde to protect them from the rumen microbial activity. Fish meal (9%), fish oil (1.5%), fish oil (1.5%) with sunflower meal protein (9%),' and sunflower meal protein alone (10.5%) (a commercial product of a protein supplement from RUMENTEK) were compared in lambs fed medium quality roughage diet in Expt. 3. Long-chain omega-3 fatty acids (eicosapentaenoic acid + docosahexaenoic acid) in muscle longissimus thoracis was increased modestly and markedly with fish meal and fish oil alone or with sunflower meal protein diet, respectively. These long-chain fatty acids were deposited in the muscle structural phospholipid rather than in storage triglycerides. All the diets mentioned above also significantly reduced omega-6:omega-3 fatty acid ratio in meat which is another beneficial effect, as the dietary recommendation in many countries has been to reduce the ratio of omega-6:omega-3 in human diet. Soymeal diet increased modestly both the omega-3 and omega-6 fatty acid content of muscle longissimus thoracis resulting in no differences in the omega-6:omega-3 ratio of the meat. A supplement of protected canola seed significantly increased the precursors of omega-6 (linoleic) and omega-3 (linolenic) but not the long-chain analogues such as arachidonic acid (omega-6) and eicosapentaenoic, docosahexaenoic acid (omega-3), respectively. The marked increase in linoleic acid content was in both triglyceride and phospholipid fractions of muscle longissimus thoracis but the modest increase in linolenic acid content was only in triglyceride fraction of meat. Supplements of canola meal used in Expt. 1, unprotected rapeseed used in Expt. 2 and protected sunflower meal protein used in Expt. 3 did not alter the fatty acid composition of muscle longissimus thoracis compared with lambs fed the control diet in that particular experiment. The increased level of long-chain omega-3 fatty acid and/or omega-6 fatty acid with the lipid supplements discussed above did not significantly affect the meat colour stability and lipid oxidative stability of fresh and vacuum packaged meat over the storage at refrigerated display. This suggests that the conditions under which the animals are grown (grazing vs grain fed or feedlot) and the species of animal are important in determining the oxidative stabilities of meat by altering the levels of muscle vitamin E concentrations at slaughter. The level of inclusion of lucerne chaff in the basal diet is an important factor in improving the redness of meat indicated by the a*-value; a higher level of lucerne chaff intake is more likely to be associated with increased intake of vitamin E. Thus colour and lipid oxidative stabilities of meat can be improved in red meat animals that are on poor quality diets by the inclusion of lucerne chaff in their diet. The sensory properties of cooked meat evaluated in the present study were not affected by the significant increase in muscle long-chain omega-3 fatty acid or omega-6 fatty acid content with fish oil and protected canola seed supplements, respectively. Addition of protected sunflower meal as a protein supplement together with fish oil significantly lowered the ratings of flavour and overall acceptability of meat compared with the control lambs. The results demonstrate that the common `lamby' and `muttony' flavour and aroma attributes were not hidden by any of the dietary treatments. These two characters associated with the species flavour and aroma were recognised by the panellists as a distinct attribute. Dry matter intake was not adversely affected by any of the lipid supplements used in the present study. Feed conversion efficiency was highest with fish meal diet on both low and medium quality roughage diets. At medium quality roughage-based diet, Feed conversion efficiency was modestly improved by protected canola seed diet but other supplements providing either natural (unprotected rapeseed) or protected protein (protected sunflower meal) did not support significant differences compared with basal diet. The significant increase in liveweight gain with fish meal diet reflected a significant increase in hot carcass weight compared with all other supplemented lambs either on low or on medium quality roughage diet. Protected lipid and protein offered by protected canola seed diet significantly and moderately increased liveweight gain and hot carcass weight from control diet but not different from unprotected rapeseed diet. The greatest muscle deposition was with the fish meal diet and is attributed mainly to the increased amount of protein and energy absorbed from the small intestine of those lambs. In addition to energy and protein absorption, the alteration of omega-3 polyunsaturated fatty acids in muscle membranes may have a further influence in lean meat production. In terms of carcass gain and intramuscular fat deposition of fishmeal and fish oil fed lambs, the results also lead to a hypothesis that modifying omega-3 polyunsaturated fatty acid of muscle membrane phospholipids may have an influence in improved muscle deposition in lambs by improving the insulin action at skeletal muscle site.
  • Item
  • Item
    Thumbnail Image
    The development of molecular markers for ascochyta blight resistance genes in Lathyrus sativus L
    Croft, Allison Mary ( 1999)
    Ascochyta blight is the most important disease of field pea in Australia. The most destructive pathogen causing this disease is Mycosphaerella pinodes. Extensive examination of the genus Pisum has failed to identify many sources of germplasm with effective levels of resistance. However, ascochyta blight resistance has been reported in the related genus, Lathyrus. The aims of this study were to examine, the genetic variation in the genus Lathyrus, the level of resistance in accessions of L. sativus (grasspea), the inheritance of resistance in L. sativus and to develop molecular markers linked to a gene(s) for ascochyta blight resistance in L. sativus. Random amplified polymorphic DNA (RAPD) analysis was used to examine genetic diversity in the genus Lathyrus, with particular focus on L. sativus. Phylogenetic analysis of the RAPD data showed groupings of the eight Lathyrus species that concurred with established morphological groupings of the species into Sections. Pisum sativum cultivar Pennant was included as an outgroup. However, when this restriction was removed, P. sativum clustered closest to the species in Section Clymenum (L. ochrus and L. clymenum). Plants from eight L. sativus accessions were used to examine genetic diversity in this species in more detail. RAPD analysis showed that L. sativus had a very low level of intraspecific genetic variation relative to interspecific variation observed in the genus Lathyrus. Four of the eight accessions exhibited variation within the accession that was similar to variation between the accessions, indicating that the accessions were not pure lines but mixtures from the collection site, or variable due to failure to control cross-fertilisation in the past. Resistance to Mycosphaerella pinodes, the most destructive of the pathogens that cause ascochyta blight of field pea, was examined in accessions/cultivars of L. sativus, L. ochrus, L. clymenum, P. sativum and P. fulvum. Accessions of the Lathyrus species were consistently more resistant to stem infection, whereas, resistance in Pisum varied depending on the accession/cultivar. Leaf infection was low in some of the Lathyrus accessions. However, in others it was similar to that observed in Pisum. Variation in the percentage of stem infection was observed between the L. sativus accessions. Two accessions were chosen based on their different response to infection, ATC 80878a (resistant accession) and ATC 80053 (susceptible accession). Under different inoculation conditions, in particular wetness duration after inoculation and concentration of pycnidiospores, consistent differences in response to infection of these accessions were observed. Inheritance of ascochyta blight resistance in the stems of L. sativus was examined using populations derived from a cross between a plant from the resistant accession, ATC 80878a, and a plant from the susceptible accession, ATC 80053. The fit of the results to Mendelian segregation models and the partitioning of the variability into genetic and environmental variance to predict the heritability of resistance were examined. The most simple Mendelian segregation model that fitted the phenotype segregation data for the F2 population from the cross was one in which resistance was conferred by two recessive genes, with interaction between the products of the two loci (additive effect between the two loci), and also codominance at one or both loci. Examination of F3 families descended from the F2 plants provided further evidence supporting this model. The variances in the parents, F2 and backcross generations were examined to predict the environmental and genetic components of resistance. The results showed that the highest level of variability was in the susceptible parental line. Heritability of resistance was estimated using the parent-offspring regression to be 46%, however, due to the low number of paired parent-offspring data points the estimate was not precise, with a standard error of �45%. Bulked segregant analysis was used with the F2 population from the cross between ATC 80878a and 80053 to identify RAPD amplification products specific to either the resistant or the susceptible bulk. Five amplification products that were potential markers were identified. Linkage between these markers and the ascochyta blight resistance trait was difficult to determine in the F2 population because more than one gene controlled the resistance phenotype, so the resistance genotypes of the F2 individuals could not be unambiguously determined. Short-range linkage mapping produced a map with four markers spanning 76.5 cM. Quantitative trait loci analysis identified OPU031090 as the marker most closely linked to one of the recessive resistance alleles. Further investigation of the inheritance of resistance to M. pinodes in L. sativus will enable identification of molecular markers more tightly linked to M. pinodes resistance genes.
  • Item
    Thumbnail Image
    A whole-farm study of draught animal power systems in West Java, Indonesia
    Basuno, Edi ( 1999)
    Draught animals continue to be used widely for land preparation in Indonesian smallholder agriculture despite the adoption of tractors in some other Southeast Asian countries. This implies that rearing and utilisation of draught animals can provide significant economic benefits to Indonesian farmers. This study attempts to define, explore and quantify these benefits and examine the outcome of selected changes to current animal management. Cattle and buffalo are the main providers of draught animal power. Most are raised on smallholder farms where they must compete with other enterprises for use of the farm's resources. They also generate income through producing outputs which are not draught-related. These factors made it necessary to look at draught-capable animals within the operation of the whole farm system, rather than assessing only their contribution to land preparation. A monitoring study was conducted over 14 months in Subang. West Java, Indonesia, with eighty respondent farmers. Data was collected on all inputs to and outputs from the respondent's farm, plus utilisation of the family's labour outside the farm. Linear programming was used to develop models representative of the Subang area based on data from the monitoring study. Three scenarios were simulated. Firstly, a manual labour farm with no utilisation of draught animals; secondly, the same farm but allowing for the hiring-in of draught animals to prepare land for irrigated rice production; thirdly, both rearing and utilisation of draught animals on the model farm. Two additional scenarios looked at the income effects of reducing labour requirements for cut-and-carry feeding and increasing calf output. The lowest Total Gross Margin of Rp2.4 million was recorded on the manual labour farm. A very small increase in TGM followed the hiring-in of draught animals; a larger increase in TGM resulted from rearing draught-capable animals on the farm. Reducing labour for animal feeding had little effect on TGM; increasing calf output had a greater income effect. The models indicate that most of the income benefits associated with cattle or buffalo enterprises derive from non-draught functions, with limited scope for obtaining increased income through changes to draught practices. This implies that the mode] simulated smallholders farm conditions with reasonable accuracy, as almost 90 per cent of farmers in the study area chose not to rear cattle or buffalo but most do rent-in draught animal power from the 10 per cent that rear large ruminants. The umbrella hypothesis of the study-that farmers are generally efficient in the allocation of their resources-remains to be disproved. In terms of future research, priority should be placed on reproductive strategies including a reduction in the calving interval and accelerated early calf growth. Both offer opportunities for obtaining substantial increases in net revenue, as the current levels of productivity achieved in the study area fall well short of management standards commonly achieved in other farming systems.
  • Item
  • Item
    Thumbnail Image
    Development of molecular markers for the genetic improvement of lentil
    Ford, Rebecca ( 1999)
    Molecular markers were applied in this study to explore two essential aspects of cultivated lentil breeding programs: 1) the genetic resources available within the species and within the Lens genus, including the wild relative species and 2) the genetics of resistance to foliar infection by the fungal pathogen Ascochyta lentis, examining both the host resistance mechanism and the diversity of the pathogen. 1) The genetic diversity within the Australian lentil breeding program was shown to be limited, representing a potential major constraint to the genetic improvement of lentil in Australia (Ford et al., 1997). Wild relative germplasm may be a source of variability for the transfer of desirable traits into superior breeding lines. A phylogeny of the Lens genus, based on nucleotide data from the 5S rRNA mutligene family, showed that L. culinaris ssp. orientalis was the closest and L. odemensis was the next closest related taxa to the cultivated species. Due to their close genetic relationships and crossibility with the cultivated species, these (sub-) species may provide new sources of useful germplasm. 2) The current major limiting factor to lentil production in Australia is ascochyta blight disease caused by Ascochyta lentis. A major strategy for disease control is breeding for resistance. However, pathogen variability is a threat to resistance breeding in that the pathogen may quickly adapt to overcome host resistance. A. lentis isolates in Australia were found to be as genetically variable as A. lentis isolates originating outside Australia. The genetics of resistance to foliar infection of A. lentis within cultivar ILL5588 was determined to be controlled by a single major dominant gene (AbRI). Molecular markers linked to the AbRI gene locus were identified with bulked segregant analysis and random amplified polymorphic DNA analysis. The two closest flanking markers being 6 and 14 cM away from AbRI. These markers were investigated for the potential to produce stable sequence characterised amplified regions (SCARS) for the identification of the AbRI locus in other lentil germplasm and the consequent use of the AbRI locus in marker assisted selection.
  • Item
    Thumbnail Image
    The development of molecular markers for shatter resistance in Brassica rapa L
    Mongkolporn, Orarat ( 1999)
    Shattering is a major problem in canola production worldwide and especially in Australia because the crop matures in summer under hot and windy condition. Little variation for shatter resistance is available in Brassica napus, which is the major canola species. Considerable variation for shatter resistance is available in B. rapa, which is a relatively minor oilseed species. Shatter resistance is being transferred from B. rapa into B. napus by backcrossing. Screening for shatter resistance in canola breeding programs is currently based on mechanical tests of individual siliqua strength. The test, which is time consuming, can only take place with completely dry and mature siliquae. This study aimed to identify molecular markers linked to shatter resistance in a population segregating for this trait, and to develop specific and robust primers to be used in a routine test, for screening for shatter resistance in canola breeding programs. Crosses were made between the shatter-susceptible Canadian cv. Torch of B. rapa L. ssp. oleifera (Metzg.) Sinsk and a shatter-resistant Indian line, DS-17-D of B. rapa L. ssp. oleifera var. Brown Sarson (Singh) Prakash. Shatter resistance was found to be a recessive trait controlled by two major genes, designated sh1 and sh2. The phenotypic segregation in the F3 population of this cross fitted a Mendelian ratio of 12: 3: 1 for shatter-susceptible (S), intermediate-shattering (M) and shatter-resistant (R) respectively. Further evidence from the phenotypic segregation in 19 F3 families was in agreement with the FZ ratio of 12: 3: 1. Bulked segregant analysis coupled with random amplified polymorphic DNA (RAPD) analysis was applied to both F, and F3 populations of the cross. Three RAPD markers, designated RAC-3, RX-7 and SAC-20, were identified from the F3 population. RAC-3 and RX-7 appeared to be linked in coupling to sh1 and sh2 at approximate distances of 13 cM and 20 cM respectively, whereas SAC-20 appeared to be linked in repulsion to both of these alleles at approximately 20 cM. The linkage of SAC-20 to both loci supported a theory of chromosome duplication during evolution of the B. rapa genome. Three pairs of sequence characterized amplified region (SCAR) primers designated SCAC-3, SCX-7 and SCAC-20, 21 to 24 bp in length were designed based on the sequences of RAPD markers RAC-3, RX-7 and SAC-20 respectively. Only the SCX-7 primer produced polymorphic markers corresponding to RX-7 amplified from 'shatter-resistant' DNA individuals, while SCAC-3 and SCAC-20 produced amplified products corresponding to RAC-3 and SAC-20 to both shatter-resistant and shatter-susceptible DNA individuals. SCAR SCX-7 was subsequently tested with other Brassica populations including B. napus and BC,F, of B. napus introgressed with B. rapa to investigate the presence of the sh2 gene. All lines from the backcross populations possessed the expected SCX-7 marker with the exception of one line where the marker was absent. This indicated that shatter-resistant gene sh2 was present in all these populations except for the one line.
  • Item
    Thumbnail Image
    The ensiling potential of tropical grasses
    Odunlami, Margaret Olanrewaju ( 1998)
    Tropical grasses are efficient users of sun energy which results in high biomass production. Their rapid growth, however, predisposes them to high lignification, dead material accumulation and therefore low digestibility when left in the field to mature. When harvested at appropriate maturity tropical grasses may be a reliable feed source for off-season feeding. High humidity in the tropics makes production of good quality hay, without mould growth very difficult. The alternative conserved form is silage, which is less weather dependent. However, tropical grasses are often regarded as poor silage material due to their low water soluble carbohydrate contents, high fibre contents and buffering capacity. This work examined the reliability of these factors as basis for classifying these highly productive tropical grass species as poor silage material. In three studies, the ensiling potential of a number of tropical grass species were examined. In a 5 x 3 x 3 factorial design, Panicum maximum, Chloris gayana and Digitaria decumbens were grown under the glasshouse condition at 17/24C day/night temperature and harvested at 2, 4, 6, 8 and 10 weeks. The grass material were wilted for 2-3 hours in bright sunshine and overnight in overcast days. The wilted herbage were divided into 3 treatments of molasses application 'at 0, 5, and 10%. The wsc ranged between 3-7.5, 4-6, and. 3-7%DM in Panicum, Chloris and Digitaria, respectively. WSC tended to increase with maturity: within the investigated intervals. Buffering capacity varied with species, 'Panicum showed lower resistance to pH change than Digitaria and Chloris (Panicum < Digitaria = Chloris). Addition of molasses reduced fibre contents and improved digestibility of silage in Panicum. Application of molasses at 10% improved digestibility over the control at 4, 6, but not at 8 weeks cut in Chloris. Digitaria had improved digestibility with molasses at all cutting intervals. Digitaria had the lowest NI-13-N among the 3 species. Silage made with Panicum cut at 8 weeks, Chloris at 6 weeks and Digitaria at 6 weeks gave the best silage quality. To investigate whether the inclusion of more stem in the harvested grass will produce higher water soluble carbohydrate (wsc) content in rhizomatous grass species, kikuyu (Pennisetum clandestine) was used in experiment 2. Using a dairy farm paddock at Berry in NSW, a 2 x 2 x 3 design was used. The grass was harvested at 3 and 6 weeks, with inclusion of long and short stem. Smaller amount of molasses, 0, 2.5, and 5% was used in study 2. Extraneous carbohydrate did not improve (P<0.005) the ensiling potential of kikuyu grass, but reduced the fibre content (P<0.005). Inclusion of more stem did not improve the wsc as hypothesized. Kikuyu grass showed low buffering capacity. The third experiment was carried out in the humid tropics (south-west Nigeria), where five of locally grown grass species were evaluated for their chemical composition, wsc and buffering capacity at pre-bloom, bloom and late bloom stages. Grasses and corresponding silages were analysed for their chemical composition; Neutral detergent fibre (NDF), nitrogen, organic matter, ammonia nitrogen, in-vitro digestibility and buffering capacity. The data were analysed using the Analyses of variance, General Linear Model and correlation. The locally evaluated grasses had wsc content of 3-5%DM and lower buffering capacity than the results in either of the previous studies. It is concluded that tropical grasses contain sufficient wsc to achieve good fermentation and therefore good quality silage with appropriate management practices, such as wilting and early harvest. The buffering capacity or the resistance to pH change is not a reliable yardstick to disqualify tropical grasses as silage material. Silage with reasonable digestibility and moderate fermentation can be produced from these species when harvested at 6 weeks and wilted to about 40% dry matter or higher.