Conventional feeds for commercial rabbit production in Vietnam are concentrates and grass. Increasingly, concentrates are becoming more expensive and they compete with human needs as they are based on cereals and protein meals such as soya bean and fish meal. A problem also with local production of concentrates is the absence of facilities to make "pellets" and rabbits do not like to eat feeds in the form of meal or powder because the fine particles cause respiratory and digestive problems.

This study is part of a research programme that aims to develop feeding systems that use locally available feeds accessible to poor farmers.

This study aimed to evaluate multi-purpose trees such as Mulberry (Morus indica) and Trichantera gigantea (introduced from the coffee-growing region of Colombia) as sources of protein; and molasses, cassava root meal and rice bran as sources of energy produced in the form of a "block" to avoid the "dust" problem.

Two experiments were carried out. Experiment 1 was in the reproduction/lactation phase with 20 New Zealand White does in individual cages. Experiment 2 was during growth and fattening (weaning to 2 kg body weight) with 60 weaned rabbits housed 5 per cage. The two treatments were the same in both experiments and were: concentrates and grasses (control) ; a molasses block and mixture of leaves from mulberry and trichantera (70%) and grasses (30%) (experimental).

The diet of molasses block and leaves from mulberry and Trichanthera gigantea resulted in better performance than the diet of concentrates and grass. During gestation and lactation the molasses/leaves diet supported higher body weight of does, heavier litters at birth and weaning, higher milk yield, better feed conversion and lower feed costs; and during growth and fattening, faster growth (by 30%), better feed conversion and lower feed costs. It was concluded that the new system was well suited to family scale production.

Rabbits are important in the rural economy as poor farmers can obtain income with little capital investment. The opportunity to consume animal protein is also a means of improving the diet of women and children.

Conventional feeds for commercial rabbit production in Vietnam are concentrates and grass. Increasingly, concentrates are becoming more expensive and they compete with human needs as they are based on cereals and protein meals such as soya bean and fish meal. A problem also with local production of concentrates is the absence of facilities to make "pellets" and rabbits do not like to eat feeds in the form of meal or powder because the fine particles cause respiratory and digestive problems.

In Vietnam molasses, rice bran, cassava root meal are cheap agricultural by-products which can be used instead of cereal grains for rabbits. New sources of protein are the leaves of trees such as Trichantera gigantea (introduced from the coffee-growing region of Colombia) and Mulberry which have been shown to have a high protein content and also are eaten readily by rabbits (Olsson and Sandberg 1995).

This study aimed to evaluate the leaves from multi-purpose trees such as Mulberry (Morus indica) and Trichantera gigantea as sources of protein; and molasses, cassava root meal and rice bran as sources of energy produced in the form of a "block" to avoid the "dust" problem.

Two experiments were carried out. Experiment 1 was in the reproduction/lactation phase with 20 New Zealand White does in individual cages. Experiment 2 was during growth and fattening (from 30 days of age to 2 kg body weight) with 60 weaned rabbits housed 5 per cage. The two treatments were the same in both experiments and were: concentrates and grasses (control) ; a molasses block and mixture of leaves from mulberry and trichantera (70%) and grasses (30%) (experimental).

The composition of the molasses blocks and the concentrate is shown in Table 1. Clay was used in the blocks to improve the "hardness" and texture. Molasses and cassava root meal were the main sources of energy in the blocks. In the concentrate the energy was from broken rice, maize meal and cassava root meal. The protein content of the concentrate was 14.7% (DM basis) and 9.723% in the block , half of which came from soya bean meal . The protein content (DM basis) was 11% for elephant grass and 13% for guinea grass, in the leaves of mulberry (only leaves) it was 25% and in Trichantera gigantea (leaves and petioles) 16%.

Table 1: Composition of the molasses blocks and the concentrates
	MB	Concent.
	----% air dry ----
Molasses	37
Cassava root meal	23	26
Rice bran	20
Clay	10
Soya bean meal	10	8
Broken rice		25
Maize meal		26.5
Fish meal		3
Groundnut cake	5
Mineral mixture	3
Bone meal		3
Salt		0.5

For Experiment 1 the does were started on the test diets immediately after mating and continued until weaning for one complete cycle of pregnancy and lactation. The experiment 2 on growing-fattening began with weaned rabbits at 30 days of age and continued for 60 days until 90 days of age.

Records were kept daily of feed offered and refused. The does were weighed at the beginning and at weaning. Litters were weighed at birth, at 21 days and at 30 days (weaning). Milk production was calculated as [(weight of litter at 21 days - litter weight at birth) x 1.18] g (Lebas et al 1986). Growing rabbits were weighed at the beginning, at 10 day intervals, and after 60 days (90 days of age).

The results for liveweight change and (calculated) milk production in Table 2 show significant benefits for the diet of molasses blocks and tree leaves compared with the conventional system of concentrates and grass. On the former the does gained 34 g during the production cycle while on the "conventional" system they lost 32 g (P=0.12). The litters from does fed the blocks were heavier at birth and at weaning (P=0.01) and this advantage was reflected in the 10% higher milk yield (P=0.01). Litter size was not affected by treatment. Conversion of feed dry matter and protein (N x 6.25) were significantly better on the molasses block and leaves than on the concentrate-fed control. Feed costs per unit weight of weaned rabbit were only one half that recorded for the control group.

Table 2: Effect on reproduction performance of rabbits of replacing concentrates and grass with molasses block and tree foliages
	MB	Control	SE/P
Liveweight, g
Initial	2,780	2,755	±84/NS
Final	2,802	2,736	±30/0.12
Weight change	33.8	-31.8	±30/0.12
Litter size
Birth	5.2	5.1	±0.15/0.65
Weaning	5.2	5.1	±0.15/0.65
Litter weight, g
Birth	298	278	±8.7/0.01
21d	1303	1188	±13/0.01
30d	2718	2547	±73/0.11
Milk yield, g	1808	1638	±44/0.01

Table 3: Mean values for feed intake of rabbits during pregnancy and lactation fed a molasses block and protein-rich leaves as replacement for concentrate and grasses
	M B	Con/grass	SE/ Prob
Feed intake (g/d)
Molasses block	186	0
Concentrate	0	163
Mulberry	347	0
T. gigantea	1262	0
Guinea grass	619	2205
Elephant grass	214	750
Total DM	203	244	±1.23/ 0.85
N x 6.25	23.4	34.6	±0.16/ 0.63
Conversion (kg/kg)
DM	4.92	6.23	±0.18/0.001
N x 6.25	0.57	0.88	±0.002/0.001
VND feed/kg weaned	4,974	10,804

The rabbits grew 33% faster (P=0.001) on the molasses block and leaves than when fed concentrates and grass (Table 4). Feed DM conversion was 24% better on the experimental diet (P=0.08) and there was more economical use of the protein (P=0.005). Feed costs on the molasses block and leaves were only half of those for the concentrate and grass diet.

Table 4: Mean values for growth and feed intake of fattening rabbits fed a molasses block and tree leaves or concentrates and grass
	MB	Concen.	SE/Prob
Liveweight, g
Initial	501	502	±13/NS
Final	1738	1429	±40/0.001
Daily gain	20.6	15.5	±0.67/0.001
Feed intake, g/day
Molasses block	65.5	0
Concentrate	0	60.2
Mulberry	123	0
T. gigantea	638	0
Guinea grass	206	605
Eleph grass	113	291
Total DM	72.1	71.3	±3.11/0.85
N x 6.25	10.4	11.0	±0.86/0.63
Conversion (kg/kg)
DM	3.51	4.61	±0.22/0.08
Protein	0.53	0.66	±0.04/0.005
Feed cost/kg gain
(VND)	3,920	8,055

The calculated daily intakes of protein and the proportions for the main diet constituents are summarized in Table 5. Does fed concentrates ate more total protein and a much higher proportion (59%) in the form of "protein meals" compared with those fed the blocks and leaves. In the latter case, "meals" accounted for 43% of the protein while the leaves supplied 42%. For growing rabbits, protein intakes were similar for both treatments but the proportion as "protein meal" was 27% for the concentrate diet compared with 15% for the molasses blocks. On the latter diet the leaves supplied 65% of the protein.

Table 5: Amounts of protein from different diet constituents
	MB		Concent.
	g/d	%	g/d	%
Does:
Meals	11.5	43	20	59
Leaves	11.2	42	0	0
Grass	4.0	15	13.7	41
Total	26.7	100	33.7	100
Growing:
Meals	5.1	15	7.7	27
Leaves	22.6	65	0	0
Grass	7.2	20	20.9	73
Total	34.9	100	28.6	100

In the first experiment with molasses blocks at this centre (Dinh Van Binh et al 1991) the blocks were based on a formula developed for cattle and contained 50% molasses, 20-30% bagasse pith and only 13% rice bran. The supplementary forage was elephant grass. Growth rates were between 9 and 15 g/day and less than on the control (17.5 g/day. The results from the present trial were much better with growth rates of 20.6 g/day for the blocks compared with 15.5 for the control. The differences were in the composition of the blocks (less molasses, no bagasse and inclusion of a small amount [10%] of soya bean meal) and the use of tree leaves (mulberry and Trichantera gigantea) replacing 70% of the forage. Probably both the more nutritional block and the protein-rich leaves contributed to the better performance in the experiments reported here. This finding is in agreement with the report of Olsson and Sandberg (1995) that weight gains on a nutrient-rich block (40% molasses, 35% rice bran, 15% cassava root meal, 10% clay) with 70% leaves and 30% grass were significantly higher (12.3 g/d) than when the forage contained 70% grass and 30% leaves (9.1 g/day).

The lower feed cost for the diet of molasses blocks and leaves was partly because of better feed conversion and also due to cheaper ingredients.

The actual proportions of leaves consumed were 82% of the total forage for the does and 66% for the growing rabbits. Thus for does there appeared to be more selection for the leaves and slightly less in the case of the growing rabbits (offer level was 70% in both cases).

The fact that leaves supplied most of the protein on the molasses block diets, and that the rabbits performed better on this treatment, provides support for our hypothesis that leaves from multi-purpose trees such as mulberry and Trichantera gigantea, which can be grown by the farmer, can replace the greater part of the dietary protein normally supplied by purchased concentrates.

• A diet of molasses block supplemented with tree leaves rich in protein can replace the traditional diet of concentrate and grasses for rabbits in all phases of the production cycle.
• Replacing concentrates with tree leaves also brings benefits to the environment and to biodiversity.
• It is recommended that the planting of multi-purpose trees such as mulberry and Trichantera gigantea be promoted so that this feeding system be applied at household level throughout Vietnam.

This research was supported by the International Foundation for Science through a grant to the senior author (B/1251-1).

(Received 31 August 1996)