A.G. KAISER and J.W. PILTZ

NSW Agriculture, Wagga Wagga Agricultural Institute, Private Mail Bag, Wagga Wagga, NSW 2650, Australia

[Paper presented at the XIII^th International Silage Conference, 11-13^th September, 2002]

Silage production is growing rapidly in Australia, and over the 10 year period from 1990 to 2000 increased from approximately 700,000 t/annum to approximately 3,000,000 t/annum as fed. Using individual State production statistics, we estimate that less than 30% of the silage is currently made in the summer rainfall dominant areas of northern Australia, mainly in southern Queensland and northern NSW. However, the use of tropical forages (C4 pathway grasses) for silage production is not confined to sub-tropical and tropical areas. For example on dairy farms, maize (Zea mays) is used in all States, forage sorghums and pennisetum are used in most States, and kikuyu grass (Pennisetum clandestinum) is used in coastal areas of NSW and southern Queensland, and in the south west of Western Australia (Kaiser and Evans, 1997). Conversely, temperate pastures and crops are used for silage production on dairy farms in Queensland and northern NSW, where silage is becoming an increasingly important component of the dairy feedbase (Figure 1). As herd sizes increase and production systems intensify, the proportion of silage in the diet is increasing while that of pasture is decreasing. In the beef industry, maize and forage sorghums are important silage crops in NSW and Queensland, particularly in the feedlot sector.

Figure 1. Seasonal change in feed intake for a dairy cow producing 5,200 litres of milk annually in a typical feeding system in northern Australia (Cowan and Lowe, 1998).

The rapid growth of tropical pastures during the wet summer and autumn period usually exceeds animal requirements, and results in poor utilisation of the available pasture. For example, Cowan et al. (1993) reported typical utilisation levels of only 30-50% of the available pasture in a dairy enterprise. The rapid accumulation of forage on offer is accompanied by a rapid decline in digestibility, but while ever the pasture is actively growing, the low grazing pressure allows grazing animals to select leaf, so the digestibility of the pasture consumed is usually considerably higher than that on offer. However, where the surplus growth has been cut and conserved as silage, the nutritive value has been low, with estimated metabolisable energy (ME) contents in the range 7.5-8.0 MJ/kg DM (Catchpoole and Henzell, 1971; Hamilton et al., 1978; Cowan and Kerr, 1984; Moss et al., 1984).

Where tropical pasture silages have been fed to dairy cows the milk production responses have been poor (Hamilton et al., 1978; Davison et al., 1984). Low silage digestibility is the main factor influencing the poor animal production response, although poor fermentation quality can also be a problem. These poor earlier experiences have led to considerable debate in Australia on the profitability of tropical pasture silages. If the nutritive value of the silage is low, the animal production response is unlikely to cover the cost of silage production. Clearly then the focus needs to be on cutting higher quality parent material. While tropical pastures are generally lower in digestibility than temperate pastures (Minson et al., 1993), an OM digestibility of 0.65 to 0.70 is achievable with rapidly growing tropical grasses if they are cut early at a leafy stage of growth. With kikuyu grass for example, OM digestibilities in the range 0.671 to 0.711 have been observed in 28-30 day regrowth (Kaiser et al., 1993 and 2000a). By comparison, the mean DM digestibility for 3626 temperate grass dominant hays over the five year period ending 2000/01 was 0.61 (P. Flinn, FEEDTEST Service, pers. comm.).

The following management regime is currently being recommended for silage production from tropical grass pastures:-

• Integrate silage cutting with grazing management to maintain the pasture at a vegetative stage of growth for as long as possible. The resulting pasture management benefits will contribute to the economic returns from silage.
• For the areas to be cut for silage, keep the pasture short and prevent the build-up of a low quality layer at the base of the pasture.
• Cut early at a vegetative stage of growth. The regrowth interval should be no more than 3-5 weeks in a rapidly growing pasture. Basing time of cut on phenological development of the grass rather than regrowth interval would be preferable and requires research.
• Wilt rapidly to a DM content of at least 300 g/kg. <>

This regime was developed for more intensively managed tropical grass pastures, and is based on research with kikuyu grass. It needs to be evaluated with grasses with a more erect growth habit such as setaria (Setaria sphacelata) and Rhodes grass (Chloris gayana). There has been some debate as to whether this more intensive cutting regime may depress total pasture yield. However, given the considerable wastage of forage with current grazing practices some loss of pasture production will probably be acceptable. The situation is likely to be quite different with tropical grass/legume pastures. To maintain the perennial legume content in the sward these pastures require a lenient grazing pressure. Hence cutting for silage may threaten legume survival - a marked decline in legume content over two years was observed by Moss et al. (1984) following a silage cut.

CROPS

Maize
Maize is an important silage crop in Australia and combines high yield with high nutritive value. With good management yields of 20-25 t DM/ha can be expected with irrigated crops. Maize can also be successfully grown under dryland conditions in higher rainfall environments, but is not suitable for use in lower rainfall environments where yield, grain content and hence nutritive value will suffer. The genetic base for maize in Australia is considerably more diverse than that used in Europe and north America. The hybrids available range from short, early-maturing types that can be harvested 85 days after sowing, through to tall, late maturing tropical types that can take as long as 150 days to reach silage harvest. Given favourable growing conditions the later-maturing hybrids are higher yielding, but they also have a lower grain content and lower digestibility (Table 1). In addition, the longer growing season may interfere with the subsequent cropping program for the paddock/field where the crop is grown. For these reasons it is generally recommended that farmers use early or mid-season varieties.

Management of the ensiling process for maize in Australia is generally similar to that in North America and Europe. One area of difference is aerobic stability, a significant problem in Australia - this will be discussed later in this review.

Table 1. Effect of hybrid maturity group on the OM digestibility of forage maize harvested at a milk line score¹ of 2 to 3 in two experiments (Piltz, 1993; Kaiser, Havilah and Piltz, unpublished).

There has been some debate as to whether the ME content of Australian maize silages is similar to that of maize silages grown in cooler regions in the northern hemisphere. It was initially believed by some researchers (Moran and Stockdale, 1992), that the ME content of Australian maize silage fell in the range 9-10 MJ/kg DM. However, there is now ample evidence to indicate that with good management silage ME should be in the range 10-11 MJ/kg DM. This is supported by one study where ME was directly determined (Margan and Moran, 1991), by digestibility studies with cattle given maize silages ad libitum with a urea supplement of 20 g/kg DM (eg. Kaiser and Piltz, 1994a - mean OMD 0.744 for 12 silages), by live-weight gain studies with young cattle (Kaiser, 1994 - mean 1.03 kg/day for 25 different maize silages with a urea supplement of 20 g/kg DM), and by on-farm studies of the digestibility of forage maize (Kaiser et al., 1994). In the latter study the mean estimated ME content of 60 farmers’ crops grown throughout NSW was 10.2 MJ/kg DM, and covered well and poorly managed crops. The sources of variation in the nutritive value of maize silage are now well understood (see Kaiser et al.,1993), and it is now widely used on dairy farms and in beef feedlots in Australia. However, differences between laboratories in procedures used to determine digestibility and calculate estimated ME content of maize silage are an on-going problem (Kaiser and Piltz, 1994b).

Sorghum, Forage Sorghum and Forage Pennisetum
The range of summer growing forage crops used for silage production are listed in Table 2. In Australia the term ‘forage sorghum’ includes Sudan grass, sorghum × Sudan grass, sweet sorghum and the multiple purpose grain sorghum types. Sorghum × Sudan grass, Sudan grass and forage pennisetum are dual purpose crops grown primarily for grazing, with silage being used as a management tool to utilise surplus growth. The most widely grown crop in this group is sorghum × Sudan grass, which is grown extensively in Queensland and NSW, and to a lesser extent in the southern States under irrigation.

Of the two single cut crops grown for silage production, sweet sorghum is more widely grown than grain sorghum owing to its higher yield potential. Sweet sorghum is most widely grown in Queensland and in northern and coastal regions of NSW. Grain sorghum is grown in drier inland areas considered too dry for maize in Queensland and northern NSW. Where irrigation is available maize is the crop of first choice owing to its high yield and high nutritive value.

Dual Purpose (Grazing and Silage) Forage Crops
These crops are very productive with a yield potential of 4.5-6.0 t DM/ha in dry areas through to 16- 22 t DM/ha with irrigation (Havilah and Kaiser, 1992). Under favourable growing conditions sorghum × Sudan grass hybrids can sustain growth rates above 200 kg DM/ha/day for an extended period. This rapid growth is accompanied by a rapid decline in digestibility (Kaiser et al. 1993). As a result, there is a small harvest window when farmers can graze or ensile the crop at an acceptable digestibility (ca. OM digestibility of 0.65 or better). The best stage of growth appears to be when the crop reaches a height of approximately 1.0-1.2m, which is usually 50-60 days after sowing for the first cut. This is supported by our data with irrigated sorghum × Sudan grass and forage pennisetum, and the results of a study with dryland Sudan grasses and sorghum × Sudan grass crops (Cole et al. 1996). In the latter study the mean estimated ME content of first and second cut forage harvested at a mean crop height of 1.1 m was 9.5 MJ/kg DM. At similar stages of growth forage pennisetum generally has a higher digestibility than forage sorghum (Stobbs, 1975).

Table 2. Sorghum, forage sorghums and forage pennisetums used for silage production in Australia.

When cut early at a vegetative stage of growth, the dual purpose forage sorghums and forage pennisetum are characterised by low DM content, low water soluble carbohydrate (WSC) content on a fresh crop basis and a moderate to high buffering capacity (BC, Table 6). Hence there is a risk of poor preservation without the use of wilting or silage additives (see later discussion).

More widespread use of dual purpose forage sorghums for silage production is limited by their moderate digestibility. This is now being addressed by plant breeders through the use of the brown-midrib (bmr) mutants. The bmr gene can cause a reduction in yield, but this is usually outweighed by the improvement in nutritive value. In studies in the USA forage digestibility has been shown to be significantly higher in bmr sorghum than in normal sorghum, and bmr sorghum silage sustained the same level of milk production as in cows given maize silage (Lusk et al., 1984; Wedig et al., 1987). In Australia, Cole et al. (1996) found that the estimated ME content of bmr Sudan grass was superior to that of three conventional Sudan grasses in first growth forage harvested 70 days after sowing (10.2 vs 9.7 MJ/kg DM), and in the second growth forage harvested 25 days later ( 10.0 vs 9.5 MJ/kg DM). In both cases the crop height at harvest was 1.0-1.2m. Where the first cut was delayed to 92 days after sowing, when crop height was 1.8-2.5 m, the estimated ME contents of the bmr and conventional Sudan grasses were 9.8 and 8.6 MJ/kg DM respectively. It is possible that the advantage in favour of the bmr forage is even greater when the harvest is delayed. The first commercially available bmr sorghum × Sudan grass hybrid was released onto the Australian market in 2001.

Another positive development is the emphasis that forage sorghum breeders are placing on the production of sorghum × Sudan grass hybrids with a higher WSC content. This is not only likely to improve the nutritive value of the forage but will also improve the silage fermentation.

Sweet Sorghum
Sweet sorghum is a high yielding crop with a similar yield potential to maize, and can also give good yields in drier environments unsuitable for maize (Cole et al. 1996). Its capacity to withstand dry seasonal conditions, combined with a long growing season means that it is able recover from a dry spell in response to rain and produce a reasonable yield. Another advantage of sweet sorghum is its large harvest window, with crop composition and nutritive value changes being quite small over a six week harvest period as the crop progresses from the milk to the hard grain stage (Table 3). Unlike maize, DM content falls only slowly with advancing maturity, and unlike other forage sorghums there is only a small change in digestibility with advancing maturity.

Table 3. Mean yield and composition of eight sweet sorghum crops harvested at four stages of growth (Havilah and Kaiser, 1992).

The WSC content of sweet sorghum is high (Table 3) making it an easy crop to ensile. There is cultivar variation in the content of WSC, and high WSC cultivars have been developed in the USA for ethanol production. Selecting cultivars with a higher WSC could improve the nutritive value of sweet sorghum forage, as in two studies we found that forage digestibility increased with WSC content (Figure 2).

Figure 2. Effect of WSC content on the digestibility of sweet sorghum forage (Kaiser, Piltz and Havilah, unpublished data).

The ME content of sweet sorghum silage is likely to be in the range 9-10 MJ/ kg DM and closer to the top end of the range if higher digestibility/WSC cultivars are selected. Clearly then maize, with its higher nutritive value, is the preferred crop under favourable moisture conditions. However, in drier Australian environments or where rainfall is unreliable, sweet sorghum will probably be the crop of first choice for silage production.

Grain Sorghum
Use of conventional grain sorghum crops for silage production is growing in Australia. To some extent this is being influenced by an awareness amongst producers that the use of these crops for silage may provide higher beef production (Havilah and Kaiser, 1992) and/or higher financial returns (Cole et al., 1996) per ha than from grain. Under irrigation grain sorghum can produce up to 18 t DM/ha which is usually less than yields attainable with maize. However, under dryland conditions grain production (and probably forage production) from sorghum can often be higher than that from maize (Muchow and Carberry, 1991). In the summer cropping regions of Queensland and northern NSW forage yields of 4-8 t DM/ha are likely to be achieved. The production and nutritive value of grain sorghum crops grown for silage production in two contrasting seasons are presented in Table 4. The results are presented for all cultivars grown in each year and for the common cultivars grown in both years.

Table 4. Forage yield and nutritive value for grain sorghum crops grown for silage production in wet and dry years (Cole et al., 1996).

Although yield was 33% lower in the dry year nutritive value was only slightly lower. This capacity to produce a silage of good nutritive value over a range of seasonal conditions is one of the advantages of grain sorghum. However its yield potential is considerably lower than sweet sorghum, and even though its ME content may be approximately 0.5 MJ/kg DM higher, it is likely to be a less profitable crop for silage production than sweet sorghum. Nevertheless grain sorghum provides farmers with the flexibility to harvest the crop for either silage or grain.

Grain sorghum is a relatively easy crop to ensile and be can managed in a similar way to maize. However one potential area of difference compared to maize is the digestion of the grain component of the silage. The data in Table 5 summarise the results of a number of studies at Wagga Wagga where maize and sorghum silages have been fed to young steers with a urea supplement of 20 g/kg DM. In each case the forage harvester was set at a short chop length, and most particles fell within a chop length range of 5-20 mm. With maize silages, a high proportion (74%) of the grain was damaged during the harvesting process, and of the remaining whole grain in the silage approximately 0.97 was digested, a result similar to that observed by De Brabander et al. (1987) for four maize silages fed to dairy cows. In addition the mean digestibility of starch in the maize silages was high at 0.914. So with short chopped maize silage, reduced digestibility due to poor utilisation of the whole grain fraction does not appear to be a problem. With sorghum silage however, the data in Table 5 indicate that a lower proportion (43%) of the grain may be damaged during harvesting, and that the utilisation of the whole grain fraction (digestibility 0.828) may not be as high as that observed in maize. Indeed Smith and Bolsen (1985) found that rolling sorghum silage cut at either the late dough or hard grain stage improved live-weight gain, feed efficiency and the digestibility of DM and starch. Hence, using forage harvesters fitted with a grain processor would be advantage with sorghum silage.

Legume Crops
Most of the interest in the use of legume crops for silage production is within the northern dairy industry, where there is a demand for supplementary protein during the winter/spring period. Low dietary protein content is a potential problem on farms where low protein silages such as maize, grain sorghum or sweet sorghum, comprise a significant proportion of the diet. Protein concentrates are expensive and hence legume silages, principally lablab (Lablab purpureus) and soybean (Glycine max), have been considered as an alternative. The possibility of intercropping these legumes with maize, grain sorghum or sweet sorghum to produce a mixed silage of higher CP content has been considered. However, this strategy has a number of limitations (see Kaiser et al., 1993). A better approach is the production of pure legume silages that can be incorporated into the diet as required.

Table 5. Digestion of the whole grain in maize and sorghum silages by young cattle (Piltz and Kaiser, 1994; Kaiser and Piltz, unpublished data).

Lablab, while capable of producing a silage with a high CP content (190 g/kg DM), has a low digestibility - Morris and Levitt (1968) reported an OM digestibility <0.60 for a several lablab silages fed to sheep and cattle. Soybean silage appears to have a higher nutritive value than lablab silage, although it appears that low yields will have to be accepted if high digestibility is targeted. Cutting at an early pod development stage can produce forage with a DM digestibility of 0.65-0.67, but potential yields are generally low (3-4 t DM/ha - Desborough and Ayres, 1998; Desborough, pers. comm.). Delaying harvest to the 50-65% pod fill stage is likely to double potential yield but DM digestibility is likely to fall below 0.65. Further research is required to identify better soybean silage cultivars that combine high yield and digestibility. However, one of the key problems with both soybean and lablab is their high stem content. What is really required is a multiple cut annual legume crop with a high leaf to stem ratio.

COMPOSITION OF TROPICAL FORAGES AND IMPLICATIONS FOR SILAGE PRESERVATION

Tropical forages vary widely in composition as can be seen from the Australian data summarised in Table 6. The WSC content in particular covers a wide range, from a mean of 30-40 g/kg DM in a number of tropical grasses up to >200 g/kg in sweet sorghum. For BC (using the method of Playne and McDonald, 1966), mean values for tropical forages are mostly in the range 250-450 meq/kg DM, and generally similar to the BC for temperate grasses reported by McDonald et al. (1991). An exception is sorghum × Sudan grass which had a higher mean BC that was similar to that observed with temperate legumes.

In addition to WSC tropical grasses may also contain starch. For example, Kaiser et al. (2000b) found that the starch content (mean 38.9 g/kg DM) of kikuyu grass was only slightly lower than its WSC content (mean 44.5 g/kg DM). The significance of starch in the ensiling process is not well understood. While most naturally-occurring lactic acid bacteria (LAB) do not utilise starch (McDonald et al., 1991), hydrolysis of starch by plant enzymes could release sugars for fermentation, perhaps replacing the WSC lost due to respiration during wilting process. Starch hydrolysis has been reported in lucerne (Medicago sativa, Knapp et al., 1973) and this may also occur with kikuyu grass. In two studies, we found that WSC content did not decline during a wilting period of 27.5h (39.4 vs 41.2 g/kg DM), whereas starch content dropped from 25.9 at mowing to 19.9 g/kg DM after wilting (Kaiser, Piltz and Havilah, unpublished data).

Table 6. Water soluble carbohydrate content (WSC) and buffering capacity (BC) of unwilted tropical forages grown for silage production.

While the higher WSC and/or DM content of maize, sweet sorghum and grain sorghum will generally result in good preservation, there is significant risk of poor silage fermentation in other tropical forages, most of which have a low DM content (<250 g/kg). The main problem with these forages is that their WSC content is often inadequate for successful low risk preservation - below the critical 25-30 g/kg fresh crop (Wilkinson, 1983). This is demonstrated in Figure 3, where we have drawn together the results from several studies with unwilted, precision-chopped kikuyu grass, sorghum × Sudan grass, forage pennisetum and sweet and forage sorghum silages. Both silage pH and ammonia-N content increased when WSC of the fresh forage fell below approximately 25 g/kg fresh crop, a result similar to that reported by Wilkinson (1983). This confirms that the generally accepted guideline of >30 g WSC/kg fresh forage for good silage preservation applies to both temperate and tropical forages. In Figure 3, the forages with WSC >30 g/kg fresh forage were predominantly sweet sorghum, but included sorghum × Sudan grass and forage sorghum cut at a later stage of crop development when both crop DM and WSC were higher. Taking account of the BC of the parent forage, by calculating a WSC/BC ratio, did not appear to improve the relationship with silage pH and ammonia-N content (Figure 4), although fewer data are available.

MANIPULATING THE SILAGE FERMENTATION

Owing to their low DM and WSC content, there is a significant risk of poor preservation when tropical grasses and legumes, and the multiple cut forage sorghum and pennisetum crops, are cut early to achieve a higher digestibility. In these circumstances two options are available to improve the silage fermentation quality - wilting or silage additives. Achieving an effective wilt (DM > 300 g/kg) can

Figure 3. The effect of WSC content of the parent forage on the pH and ammonia-N content of kikuyu grass, sorghum × Sudan grass, forage pennisetum, sweet sorghum and forage sorghum silages.

Figure 4.The effect of the WSC/BC ratio in the parent forage on the pH and ammonia-N content of kikuyu grass, sorghum × Sudan grass, forage pennisetum, sweet sorghum and forage sorghum silages.

sometimes be difficult as silage making often coincides with hot humid weather, which is associated with a high incidence of rainfall. So prevailing weather conditions are likely to influence the efficacy of the wilting option, and this is highlighted in Table 7, where the results of Australian wilting studies with tropical grass silages are summarised.

Wilting
Although limited data are available, there is an indication that where DM content was increased to >300 g/kg in approximately one day (24-26 h), wilting tended to improve preservation as indicated by a fall in ammonia-N content (Table 7). In the studies with kikuyu grass, longer wilting periods or ineffective wilting due to wet weather resulted in even higher ammonia-N contents than the poorly preserved unwilted controls. Clearly then increasing the rate of wilting is a key goal when ensiling tropical pastures. This may be achieved by changing management of the cut forage at mowing and by the subsequent use of tedding. In our studies with kikuyu grass we have found that leaving the mown forage spread wide at close to the full mowing width, compared to leaving the forage in a windrow occupying approximately 50-60% of the mowing width (common industry practice), will significantly enhance wilting rates. Tedding will further improve drying rates, although more research is required to establish the optimum timing and frequency of tedding operations for a range of drying conditions. Clearly if weather conditions are unfavourable for wilting, there will be little difference between wilting treatments, and the use of silage additives needs to be considered. Where wilting conditions are marginal a combination of a short wilt with additive use may be the best strategy and needs to be investigated.

Table 7. Influence of wilting on the preservation of tropical grass silages.

Silage Additives
Limited research has been conducted on the effect of silage additives on the preservation of tropical forages. Organic acid (or their salts) additives are not readily available in Australia and are not used on tropical forages. Molasses and inoculants are more widely used, although inoculants are not often used on tropical grasses and legumes, or the multiple cut forage sorghum and pennisetum crops. There is now ample evidence from Australia and other tropical regions that molasses is an effective additive for tropical forages (Muhlbach, 2000). Fermentation quality, as indicated by a lower pH and ammonia-N content and a higher lactic acid content, has been improved in a number of studies with tropical grasses (Catchpoole, 1966; Tjandraatmadja et al., 1994; Piltz and Kaiser, 2000) and tropical legumes (Morris and Levitt, 1968; Catchpoole, 1970b). In most cases an application rate of 40 kg/t fresh crop appeared to be adequate although a higher rate may be required under difficult ensiling conditions.

The application of LAB inoculants to tropical grasses with low DM and WSC content would not be expected to lead to a significant improvement in silage fermentation quality. This was confirmed in three studies at our institute with four kikuyu grass pastures (Kaiser and Piltz, unpublished data). In each case an attempt was made to lightly wilt the pastures, but all of the resulting wilted control silages were poorly preserved (ammonia-N, 140-440 g/kg total N). The inoculants failed to secure good fermentation quality (ammonia-N, £ 100 g/kg total N) under a range of wilting conditions:-

• a rain damaged crop that been left in the field for 48 h with no change in DM content (DM = 90 g/kg)
• a crop that had been rapidly wilted for 6 h (DM = 220 g/kg)
• a crop that been slowly wilted for 28.5 h (DM = 250 g/kg)
• a crop that had been slowly wilted for 52.3 h (DM = 300g/kg).

Poor aerobic stability is a significant problem in maize silage in Australia. It is also likely to be a problem with grain sorghum and sweet sorghum silages, but there are few aerobic stability data for these crops. The problem with maize has been highlighted in three studies at our institute, where aerobic stability was monitored over 7 days for 71 silages (Table 8). Aerobic stability was determined by placing silage samples in polystyrene containers covered with muslin cloth, storing them in a temperature-controlled room at approximately 20° C, and recording temperature with probes placed in the centre of each container. Silages were considered to be unstable once the temperature rose 2° C above ambient. The results showed that 69% of the silages were very unstable (stable <3 days) and only 11% were very stable (>7 days).

Table 8. The proportion of maize silages aerobically stable after different storage periods.

While there are no survey data to quantify the extent of poor aerobic stability in maize silages on farms and feedlots in northern Australia, there is sufficient anecdotal evidence to indicate that the problem is widespread, although it is often not readily recognised by farmers. The higher ambient temperatures during silage feeding, compared to those in Europe and north America, are thought to exacerbate the problem in Australia. For this reason a higher rate of silage removal is probably appropriate under our conditions so that the period of exposure of the silage at the feeding face to air can be minimised.

A longer-term solution may rely on the use of silage additives to improve aerobic stability. There is now sufficient evidence in the literature to indicate that homofermentative LAB inoculants are unlikely to improve stability and may even exacerbate the problem. This was confirmed in one experiment at our institute where two LAB inoculants increased beef live-weight gain/t silage DM by a mean of 11.4% but had no effect on aerobic stability (Kaiser and Piltz, 1998). However, the inclusion of heterofermentative LAB (eg. Lactobacillus buchneri) or organic acid salts in inoculants may be a successful strategy for improving stability (eg. Weissbach, 1996), and needs to be investigated under Australian conditions.

CONCLUSIONS

A number of key areas concerning the production and utilisation of silages from tropical forages require further research:-

• There needs to be a greater focus on cutting tropical pastures and the dual purpose forage sorghums and pennisetums earlier, at a higher digestibility stage of growth. The impact of this management regime on subsequent pasture/crop productivity needs to be investigated.
• The available legume forage crops do not generally satisfy the requirements of high yield and high nutritive value. Research should focus on selecting an improved forage soybean, or a multiple cut annual legume crop with a high leaf to stem ratio.
• The extent of field losses during wilting need to be investigated, as do the strategies that may be used to accelerate wilting, particularly under humid conditions with intermittent rainfall.
• A more selective approach needs to be adopted on the use of silage additives. In particular combinations of a short light wilt + molasses, and molasses + LAB inoculant need to be investigated.
• Poor aerobic stability, particularly with high value silages such as maize, is a significant problem that needs to be resolved.

CATCHPOOLE V.R. (1965) Laboratory ensilage of Setaria sphacelata cv. Nandi and Chloris gayana (C.P.I. 16144). Australian Journal of Agricultural Research 16, 391-402.

CATCHPOOLE V.R. (1966) Laboratory ensilage of Setaria sphacelata cv. Nandi with molasses. Australian Journal of Experimental Agriculture and Animal Husbandry 6, 76-81.

CATCHPOOLE V.R. (1968) Effects of season, maturity and rate of nitrogen fertiliser on ensilage of Setaria sphacelata. Australian Journal of Experimental Agriculture and Animal Husbandry 8, 569-573.

CATCHPOOLE V.R. (1970a) The silage fermentation of some tropical pasture plants. Proceedings of the 11^th International Grassland Congress, Surfers Paradise, Australia, pp. 891-894.

CATCHPOOLE V.R. (1970b) Laboratory ensilage of three tropical pasture legumes - Phaseolus atropurpureus, Desmodium intortum and Lotononis bainesii. Australian Journal of Experimental Agriculture and Animal Husbandry 10, 568-576.

CATCHPOOLE V.R. (1972) Laboratory ensilage of Setaria sphacelata cv. Nandi and Chloris gayana cv. Pioneer at a range of dry matter contents. Australian Journal of Experimental Agriculture and Animal Husbandry 12, 269-273.

CATCHPOOLE V.R. and HENZELL E.F. (1971) Silage and silage-making from tropical herbage species. Herbage Abstracts 41, 213-221.

COLE C.A., KAISER A.G., PILTZ J.W. and HARDEN S. (1996) An evaluation of sorghums for silage production in northern New South Wales. In: Proceedings of the 3^rd Australian Sorghum Conference, Tamworth, NSW. Eds M.A. Foale, R.G.Henzell and J.F. Kneipp. Australian Institute of Agricultural Science Occasional Publication No. 93, pp. 127-139.

COWAN R.T. and LOWE K.F. (1998) Tropical and subtropical grass management and quality. In: Grass for Dairy Cattle. Eds J.H. Cherney and D.J.R Cherney, CAB International, pp. 101-135.

COWAN R.T., MOSS R.J. and KERR D.V. (1993) Northern dairy feedbase 2001. 2. Summer feeding systems. Tropical Grasslands 27, 150-161.

COWAN R.T. and KERR D.V. (1984) Dairy feeding systems based on pasture and forage crop silage in the tropics and sub-tropics. In: Silage in the 80’s. Proceedings of a National Workshop, Armidale, NSW. Eds T.J. Kempton, A.G. Kaiser and T.E. Trigg, pp. 384-391.

DAVISON T.M., ORR W.N. and CLARK R. (1984) Changes in silage use on the Atherton Tablelands, North Queensland. In: Silage in the 80’s. Proceedings of a National Workshop, Armidale, NSW. Eds T.J. Kempton, A.G. Kaiser and T.E. Trigg, pp. 392-396.

DESBOROUGH P.J. and AYRES J.F. (1988) Cultivar and growth stage effects on the nutritive value of soybean hay. Proceedings of the Australian Society of Animal Production 17, 388.

DE BRABANDER D.L., VANACKER J.M., ANDRIES J.I., DE BOEVER J.L. and BUYSSE F.X. (1987) [The influence of roller maize processors and chopping length on the nutritive properties of silage maize for dairy cows.] Revue de l’Agriculture 40, 1517-1535.

HAMILTON R.I., CATCHPOOLE V.N., LAMBOURNE L.J. and KERR J.D. (1978) The preservation of a Nandi setaria silage and its feeding value for dairy cows. Australian Journal of Experimental Agriculture and Animal Husbandry 18, 16-24.

HAVILAH E.J. and KAISER A.G. (1992) Sorghums for silage: A review. In: Proceedings of the 2^nd Australian Sorghum Conference, Gatton, Queensland. Eds M.A. Foale, R.G.Henzell and P.N. Vance. Australian Institute of Agricultural Science Occasional Publication No. 68, pp. 338-354.

HAVILAH E.J., KAISER A.G. and NICOL H. (1995) Use of a kernel milkline score to determine stage of maturity in maize crops harvested for silage. Australian Journal of Experimental Agriculture 35, 739-743.

KAISER A.G. (1994) Opportunities for supplying roughages and silages to feedlots. In: Opportunities in Supplying Feedlots. Proceedings of a Special Industry Seminar, Cooperative Research Centre for the Beef Industry (Meat Quality), Armidale, NSW, pp. 85-91.

KAISER A.G. and EVANS M.J. (1997) Forage conservation on Australian dairy farms. Animal Industries Report No. 3, NSW Agriculture, Orange, 18 pp.

KAISER A.G., HAVILAH E.J., COLLESS J.M., LAUNDERS T.E. and PILTZ J.W. (1994) Forage maize quality on dairy farms in NSW. Proceedings of the 2^nd Australian Maize Conference, Gatton, Queensland, pp. 111-112.

KAISER A.G., HAVILAH E.J., CHOPPING G.D. and WALKER R.G. (1993) Northern dairy feedbase 2001. 4. Feeding systems during winter and spring. Tropical Grasslands 27, 180-211.

KAISER A.G. and PILTZ J.W. (1994a) Difference between sheep and cattle in the digestion of maize silage. Proceedings of the 2^nd Australian Maize Conference, Gatton, Queensland, pp. 232-233.

KAISER A.G. and PILTZ J.W. (1994b) Estimating the digestibility and energy content of Australian maize silages. Proceedings of the 2^nd Australian Maize Conference, Gatton, Queensland, pp. 230a-230e.

KAISER A.G. and PILTZ J.W. (1998) Effect of silage inoculants on cattle production from maize silages. In: Proceedings of Beef - the Path Forward. Eds J.F. Wilkins and W.A. McKiernan, NSW Agriculture, p.58.

KAISER A.G., PILTZ J.W., HAMILTON J.F. and HAVILAH E.J. (2000a) Effect of time of day on the WSC content of kikuyu grass. In: "Silage Making in the Tropics with Particular Emphasis on Smallholders". Ed. L. ‘t Mannetje. FAO Plant Production and Protection Paper 161, pp. 143-145.

KAISER A.G., PILTZ J.W., HAVILAH E.J. and HAMILTON J.F. (2000b) Kikuyu grass composition and implications for silage production. In: "Silage Making in the Tropics with Particular Emphasis on Smallholders". Ed. L. ‘t Mannetje. FAO Plant Production and Protection Paper 161, pp. 63-64.

KNAPP W.R., HOLT D.A., LECHTENBERG V.L. and VOUGH L.R. (1973) Diurnal variation in alfalfa (Medicago sativa L.) dry matter yield and overnight losses in harvested alfalfa forage. Agronomy Journal 65, 413-417.

LUSK J.W., KARAU P.K., BALOGU D.O. and GOURLEY L.M. (1984) Brown midrib sorghum or corn silage for milk production. Journal of Dairy Science 67, 1739-1744.

MARGAN D.E. and MORAN J.B. (1991) The energy value of an Australian maize silage. Proceedings of the 1^st Australian Maize Conference, Moama, NSW, pp. 128-129.

McDONALD P., HENDERSON N. and HERON S. (1991) The Biochemistry of Silage, 2^nd edition. Chalcombe Publications, Marlow, UK.

MINSON D.J., COWAN R.T. and HAVILAH E.J. (1993) Northern dairy feedbase 2001. 2. Summer pasture and crops. Tropical Grasslands 27, 131-149.

MORAN J.B. and STOCKDALE C.R. (1992) Maize silage for the pasture-fed dairy cow. 1. Effect of level of silage feeding, and responses to cottonseed meal while grazing perennial pastures in spring. Australian Journal of Experimental Agriculture 32, 279-285.

MORRIS J.G. and LEVITT M.S. (1968) The intake and digestibility of silages made from Dolicos lablab L. alone and with Sorghum cultivar Sugardrip. Proceedings of the Australian Society of Animal Production 7, 78-85.

MOSS R.J., BYFORD I.J.R., WINKS L., CHAMBERS G. and O’GRADY P. (1984) In vivo digestibility of tropical grass-legume pastures conserved as silage or hay. In: Silage in the 80’s. Proceedings of a National Workshop, Armidale, NSW. Eds T.J. Kempton, A.G. Kaiser and T.E. Trigg, pp. 137-140.

MUCHOW R.C. and CARBERRY P.S. (1991) Climatic constraints to maize productivity in Australia. Proceedings of the 1^st Australian Maize Conference, Moama, NSW, pp. 8-12.

MUHLBACH P.R.F. (2000) Additives to improve the silage making process with tropical forages. In: "Silage Making in the Tropics with Particular Emphasis on Smallholders". Ed. L. ‘t Mannetje. FAO Plant Production and Protection Paper 161, pp. 151-164.

PILTZ J.W. (1993) Digestibility of Australian maize silage. M. Rur. Sci. Thesis, University of New England, Armidale, 105 pp.

PILTZ J.W. and KAISER A.G. (1994) Digestibility of whole grain and starch in Australian maize silages. Proceedings of the 2^nd Australian Maize Conference, Gatton, Queensland, pp. 234-239.

PILTZ J.W. and KAISER A.G. (2000) The use of molasses to improve the fermentation of low dry matter kikuyu grass silages. In: "Silage Making in the Tropics with Particular Emphasis on Smallholders". Ed. L. ‘t Mannetje. FAO Plant Production and Protection Paper 161, pp. 165-166.

PLAYNE M.J. (1963) Buffering capacity of sweet sorghum: The effectsof nitrogen content, growth stage and ensilage. Journal of the Science of Food and Agriculture 14, 495-497.

PLAYNE M.J. and McDONALD P. (1966) The buffering constituents of herbage and of silage. Journal of the Science of Food and Agriculture 17, 264-268.

SMITH R.L. and BOLSEN K.K. (1985) Evaluation of grain sorghum as a silage crop. Proceedings of the 15^th International Grassland Congress, Kyoto, Japan, pp. 1020-1022.

STOBBS T.H. (1975) A comparison of Zulu sorghum, bulrush millet and white panicum in terms of yield, forage quality and milk production. Australian Journal of Experimental Agriculture and Animal Husbandry 15, 211-218.

TJANDRAATMADJA M., NORTON B.W. and MACREA I.C. (1994) Ensilage characteristics of three tropical grasses as influenced by stage of growth and addition of molasses. World Journal of Microbiology and Biotechnology 10, 74-81.

WEDIG C.L., JASTER E.H., MOORE K.J. and MERCHEN N.R. (1987) Rumen turnover and digestion of normal and brown midrib sorghum × sudan grass hybrid silages in dairy cattle. Journal of Dairy Science 70, 1220-1227.

WEISSBACH F. (1996) New developments in crop conservation. Proceedings of the 11^th International Silage Conference, Aberystwyth, Wales, pp. 11-25.

WILKINSON J.M. (1983) Silages made from tropical and temperate crops. 1. The ensiling process and its influence on feed value. World Animal Review 45, 36-42.