The VPIP calculates that within existing plantation leases there is about 20000 ha of bush that is likely to be developed over the next 10 - 20 years. Approximately 4000 ha of this total would be classified as secondary forest. A later section will show that

Regrowth burao with residual rainforest clumps

Secondary rainforest areas not recommended for pasture development

once pasture establishment costs including fencing and watering exceed 50000 VT/ha then returns on investment in pasture improvement are not commercially attractive. Some commercial plantations are developing Burao or Merremia dominated bush within this limit but the development of secondary forest will cost upwards of 75000 VT/ha. This message is widely understood by the ni-Vanuatu and expatriate plantation sector.

When the forest canopy has been grossly disturbed, most frequently by cyclones in Vanuatu, burao, and other fast growing pioneer regrowth species and creepers such as Merremia frequently invade in response to the full sunlight conditions on the forest floor and smother slower growing, low light requiring rainforest species. It is a commercial proposition to develop such areas for improved pastures, although still not as attractive, usually, as rehabilitating previously cleared areas or areas of degraded pasture.

Clearing of such bush and establishment of pastures can be undertaken provided all of the residual rainforest clumps are left undisturbed (ideally about 15 - 20% of the area) for nature conservation, future timber production and shade for cattle. Within residual clumps or strips of forest it is important to temporarily exclude cattle so that regeneration and development of young trees can occur to a stage where they are compatible with cattle.

This is an area where a growing number of smallholders are seeking advice. Strips of uncleared bush or connected clumps have the advantage of providing corridors for wild life movement between large areas of native vegetation. Clumps or strips should be at least 50 m wide. Leaving the required level of shade as single scattered trees, whilst attractive perhaps from a European perspective, tends to increase the losses associated with cyclones and reduces the chances of seedling regeneration.

Mechanical clearing with a bulldozer is most efficiently carried out in a two stage process - firstly to push over the target bush so that it starts to dry out and secondly to windrow the dry vegetation at a suitable later date.

Bulldozers fitted with root rakes for windrowing ensure soil movement into windrows is minimal. If valuable topsoil is moved then the sustainable productivity of subsequent pastures will be reduced. Windrows should always be aligned on the contour to minimise erosion and should not cross drainage lines. See Guidelines in Laefstock Newsletter 9, December 1991. Even without a root rake skilled operators can minimise soil movement which is particularly important on shallow soils.

The distance between windrows is a decision for on-farm management depending on denseness of bush, soil moisture, size of machine.

Following windrowing various techniques are used to produce a shallow, firm seedbed for sowing pastures. These include:

• dragging a heavy bar with anchor chain behind, often in two directions 90 degrees to each other. Some land levelling may be required which can cheaply be accomplished in this way. This technique is popular for very stony soils, for example around Rentabao on Efate.

• one or two passes with disc harrows followed by heavy chain harrows are frequently used, especially when there has been considerable weed regrowth through the pushed bush.

• reversing with a bulldozer fitted with a stick rake during or after the windrowing operation.

It is not necessary to perform deep cultivation as pasture seed should not be planted at a greater depth than 10-20 mm because of the small seed size. Seed can in fact be broadcast on the soil surface and covered by dragging a chain or chain + roller behind the seeder. A number of plantations are now using rubber tyred rollers with short anchor chain curtains in front of the roller for best results in lightly incorporating seed and compacting planted seedbeds. It is essential that prepared seedbeds are seeded as soon as possible to prevent/reduce weed invasion. The best form of weed control is vigorous pasture competition.

An alternative to the above is to clear strips of bush (10 m wide) with a dozer, say 20m apart, having already heavily grazed large animals to drastically reduce the cover of palatable creepers such as Merremia. Such strips should be sown, harrowed and rolled quickly. This technique has been successful in establishing pastures at Savoie Plantation, Santo for approximately half the cost of a conventionally cleared area. However the time required for complete improved pasture establishment was 1.5 - 2.5 years rather than 6 months. This approach suits a percentage of plantations which have limited capital and modest development targets.

The use of heavy grazing to virtually eradicate Merremia just prior to the end of the dry season and planting of signal/legumes at the beginning of the wet season is a low cost but risky option. If there is any more than 4 weeks between removal of the creepers and planting rain, then t-grass is likely to dominate and the successful establishment of signal/legumes would be jeopardised.

Having identified all residual secondary forest areas for conservation remaining regrowth bush areas can sometimes be effectively cleared using hand clearing and tree poisoning techniques. The latter technique has been shown to cost about 8000 VT/ha controlling up to 500 regrowth stems per hectare. e.g rain tree Tagabe. This technique is particularly relevant to small ni-Vanuatu plantations where capital is a major constraint. There are numerous examples of improved pastures successfully established from seeding or vegetative planting provided that planting immediately follows manual or chemical treatment of regrowth bush.

The major problem in most coconut plantations is that of replacing total weed infestations of pistache, blue rat's tail, lantana, pico, burao and other bush regrowth, guava, Nasasa, Korosol, with sown pastures, or of replacing native pastures with improved pastures or of establishing legumes into legume deficient improved grasses.

For plantations with access to machinery the same principles which apply to rehabilitating open weed infestations also apply to under coconuts, except that costs, are approximately 10% higher due to reduced efficiency of working machinery and the need for weeding around coconuts. In many cases the VPIP has found efficiently

used machinery used to establish vigorous improved pastures is the most cost effective option for rehabilitating shaded or open weed infestations. Again the linchpin of sustainable weed control is pasture competition.

Some examples of on-farm rehabilitation under coconuts with owner operated machinery (excluding depreciation) include:

Indigofera, burao under coconuts, Rentabao - windrowed, levelled with heavy bar, sown to glycine to smother regrowth, one high slashing in the first year, ungrazed for 18 months - 25000 VT/ha.

Lantana under coconuts, Undine Bay - fire in the dry season, heavy duty slash, 3 disc harrowings, heavy chain harrowing, seed sabi/legumes, spot spraying or hand weeding YR1 -20000 VT/ha.

Pistache under coconuts, South Santo - windrow with small dozer, slash, zero-till signal/legume mix, - 15000 VT/ha.

Plantation, Malekula - slash. burn, 2 discings, harrowing, broadcast signal or sabi with legume mix, followed by chain/roller - 15000 VT/ha

Santo - oversowing carpet grass - zero-tilling signal/legumes into coconut inter-rows -approximately 8000 Vt/ha.

Successful rehabilitation under coconuts without machinery involves the careful timing of various operations. In the case of lantana control on North Efate a hot fire towards the end of the dry season can save up to 30 mandays per hectare in hand slashing. Alternatively deferring grazing for long enough to allow climbing legumes such as siratro, glycine, centro, dolichos lablab, or puero to smother pico, white or narrow leaf broom, wild tobacco or blue rat's tail is an essential first step in the rehabilitation process which avoids the manual removal of the problem weed.

These techniques have been described in Technical Bulletin No.2 - Weed Control and in detail in Technical Bulletin 3 - ^"Lukaotem Gud Pasja mo Buluk long Vanuatu".

Regardless of the techniques used in rehabilitation, the choice of pasture species is especially important because of the need to use grass and legume species that will be productive under various shaded conditions. As pointed out in Technical Bulletin 3, signal, sabi and koronivia will perform well provided at least 70% of sunlight reaches the ground - i.e an average situation for a good stand of 60 year old coconuts. Given very careful management involving undergrazing, these grasses will probably persist down to 50% sunlight conditions. For replanted coconuts and particularly the hybrids at the recommended spacing of 9 m triangular, light transmission conditions will be below 50% from 5 - 40 years of age. Under such conditions buffalo grass is the best available option at this point in time, which combines high tolerance to shade and a growth habit which is resistant to overgrazing.

In the Solomon Islands and Western Samoa batiki blue grass has persisted at 50% light transmission levels under grazing at correct stocking rates. However batiki is adapted for acid soils which are not widespread in Vanuatu.

However, increasingly under conditions of depressed copra prices some commercial smallholders are showing interest in planting coconuts at lower densities so that the more productive improved grasses can persist under grazing.

Legumes like grasses differ in their tolerance of shade. Generally in legumes and grasses high levels of shade will encourage plants to become more etiolated where they grow taller in an effort to gain better access to available light. Also leaves become larger and thinner. This has the effect of decreasing the density of the pasture sward as well as the readily digestible fraction in the leaf. This is one reason why cattle, when given a choice of open and heavily shaded pastures of the same grass or legume species, will usually graze the open sites first.

In addition heavily shaded pastures (i.e less than 30 - 40% light transmission) have lower dry matter contents or higher water contents. This means an animal has to make more bites to achieve a certain level of daily intake to achieve a certain level of performance. It has been proven that once dry matter percentages of intake drop below 18% then the potential for ruminants growth drops off markedly. The dry matter percentages of shaded t-grass is frequently below 18%, especially under heavy shade.

On the other hand pasture quality for commonly available tropical species at moderate levels of shade (70% light transmission), on the basis of limited research work, appears similar to that grown under full sunlight. In some cases light shade will actually improve quality, for example protein content and digestibility of grasses.

Generally pasture production and sustainable stocking rates declines linearly with increasing shade. Different species decline at different rates. Those species whose production declines least rapidly compared with full sunlight levels of production are regarded as the most shade tolerant. What is most important in coconut plantations is shade tolerant legumes which are creeping and which persist and maintain good ground cover under moderate to heavy grazing.

The tolerance of the commonly available climbing smothering legumes to increasing shade is as follows: siratro < glycine < greenleaf desmodium <centro <puero < calopo.

The tolerance of creeping legumes under moderate to heavy grazing is as follows:

Indigofera spicata < Shaw vigna < Vigna hoseii, Teramnus < hetero, Arachis pintoi.

Arachis repens , Desmodium ovalifolium.

The production of a range of smothering and creeping legumes under 40% light transmission under coconuts is in progress at IRHO.

Where annual rainfall falls below 2000 mm in the humid tropics there have been instances where competition for soil moisture by understorey pastures has reduced copra yield when compared with a bare fallow situation. But this is of little relevance to Vanuatu. Buffalo grass being fairly intolerant of dry conditions, compared with Sabi or Koronivia, is less likely to compete for scarce soil moisture.

Refer to Technical Bulletin 3 for the discussion of recommended stocking rates for various types of shaded pasture under various rainfall and soil fertility situations.