A survey in the Eastern region of south Vietnam showed that cassava growing has increased in recent years due to the opportunities for sale of the roots to factories producing starch, sodium glutamate and cassava root meal. The varieties most popular with farmers were: Gon, "Japan", "India", KM60, KM94 and KM95. The crude protein in the leaves ranged from 23.9 to 34.7% and the fibre from 9 to 14% (both as % of DM).

Two digestibility and N balance trials, of double Latin square design, were done with growing pigs to evaluate levels of substitution of soya bean meal by cassava leaves (0, 15, 30 and 45% of the diet protein) in diets based on cassava root meal. The trials were done in successive periods first with ensiled leaves (Mean pig liveweight 40-50 kg) and then with sun-dried leaves (50-70 kg liveweight). There were significant decreases in apparent diet digestibility of all nutrients (from 94.3 to 85.1 for dry matter and from 88.2 to 65.7% for protein) when cassava leaf protein replaced 45% of that from soya bean. The decline appeared to be less marked for the ensiled compared with the sun-dried leaves. The linear regression between dry matter digestibility and percent cassava leaf protein predicted a digestibility of the leaf dry matter of 76.5% for ensiling and 72.8 for sun-drying.

Nitrogen retention as percent of N intake and N digested showed only a slight reduction up to the 30% level of soya protein substitution by the leaves, but decreased markedly at the 45% level. It was also observed that the pigs were reluctant to consume the diets with high level of leaf. There were no indications of cyanide toxicity on any of the diets.

Cassava is an important crop in the small-farm sector in Vietnam. Many researchers showed that cassava leaf has a high protein content (from 16.7 to 39.9%) with almost 85% of the crude protein fraction as true protein (Ravindran 1991). It is a readily available product at the time of harvesting the roots. However, in the rainy season, it is difficult to sun-dry and extending the drying period diminishes the nutritional quality of the product. It was considered that ensiling would be an attractive alternative way to conserve the product.

The aims of the research reported in this paper were to determine the nutritive value for pigs of:

• Different varieties of sun-dried cassava leaves
• Ensiling compared with sun-drying.

In both studies the cassava leaves were included as a replacement for soya bean meal in diets based on cassava root meal. Prior to these studies a survey was carried out on cassava production in the Eastern region of south Vietnam

The survey

The provinces of Dong Nai, Song be and Tay ninh in the Eastern region of South Vietnam were visited. Data were obtained on the areas planted to cassava.

Leaves were collected from six varieties 3-4 months after planting. They were sun-dried and samples taken for chemical analysis.

Two experiments were carried out: one with sun-dried cassava leaves and the other with ensiled leaves. Sun-drying consisted of spreading the leaves on the ground and turning them over while exposed to the sun during a 1-2 day period. The dried leaves were then ground (2mm sieve). For ensiling the leaves were chopped into small pieces (about 5 cm diameter) and mixed with molasses (5%) and salt (5 %) (both on fresh weight basis) prior to being stored in 50 litre containers. The contents were pressed and air removed with a vacuum pump prior to sealing.

In each experiment, the protein (N x 6.25) in both the cassava leaf meal and the ensiled leaves replaced the protein of soya bean meal at levels of 0, 15, 30, and 45% in diets for growing pigs based on cassava root meal. The dietary ingredients and their chemical composition are shown in Tables 1 and 2.

Table 1: Formulation of ensiled cassava leaf k (ECL)diets
	Protein from ECL, %
	0	15	30	45
Dry matter basis, %
Cassava root meal	74.0	70.3	66.0	69.0
Soybean meal	23.4	19.9	16.4	12.9
Cassava leaf	0	7.2	14.4	21.6
Bone meal	2.0	2.0	2.0	2.0
Salt	0.5	0.5	0.5	0.5
Premix	0.1	0.1	0.1	0.1
Chemical composition, % DM
Protein	12.7	13.3	13.5	13.7
Fibre	3.1	4.0	5.0	5.9
Ether extract	2.2	2.9	3.8	4.6
Ash	5.1	3.3	3.7	4.2

Table 2: Formulation of diets containing cassava leaf meal (CLM)
	Protein from CLM, %
	0	15	30	45
Ingredients, %
Cassava root meal	77.0	73.5	70.0	66.5
Soybean meal	20.0	17.0	14.0	11.0
Cassava leaf meal	0	6.5	13.0	19.5
Bone meal	2.5	2.5	2.5	2.5
Salt	0.5	0.5	0.5	0.5
Premix	0.1	0.1	0.1	0.1
Chemical composition, % DM
Protein	10.3	10.3	10.6	10.1
Fibre	2.7	3.5	4.1	5.1
Ether extract	2.7	3.2	3.6	3.9
Ash	5.2	5.7	6.1	6.4

A double 4 x 4 Latin square design with eight castrated male pigs (Large White x Landrace x Duroc) was used for each experiment. For the CLM diets the initial weight was 35-40 kg and for the ECL diets it was 40-50 kg. The experimental periods were 12 days, the first 7 for adaptation to the diet and the last 5 for total collection of faeces and urine. During the adaptation period the pigs were in concrete floor pens while for collection of excreta they were in metabolism cages. Feeding was ad-libitum during the adaptation period and restricted during the experimental period to the maximum eaten by the pigs in the adaptation stage.

For the ECL diets, the ensiled leaves were fed separately in quantities corresponding to their content in the diet adjusted to the daily amount of feed consumed by the pigs fed the control (zero ECL) diet.

During the experimental periods, feed and refusals were weighed daily. Urine was collected in a container with H₂SO₄ solution to keep the pH below 4 in order to preserve the N. Each day all the faeces and an aliquot of 10% of the urine were collected and stored in a freezer at -18 ºC. At the end of the collection period, faeces and urine were mixed thoroughly and representative samples were taken for chemical analysis.

Nitrogen was determined by the Kjeldahl method and dry matter by drying at 105 ºC on fresh sample. A sub-sample of faeces was dried at 60 ºC for other parameters such as: crude fibre, ash and ether extract conventional methods were used (AOAC 1984). HCN was determined by titration with AgNO₃ after boiling the sample and concentrating the HCN in KOH.

Interviews with farmers and extension agents revealed that there are two planting times: from October to December and from April to May with the respective harvests in March-April and October-December. The area planted to cassava has increased by 25% (from a total of 24,000ha to 31,000ha) from 1991 to 1994 (Anon 1995). The increase is due to the fact that there was an increase in deforestation to supply roots to the factories processing it for the manufacture of monosodium glutamate. In addition, in recent years, the price of cassava increased. All of these factors stimulated the farmers to plant more cassava.

Variety

The popular varieties of cassava were: Gon, Japan, India, KM60, KM94 and KM95. The yield of fresh leaves varied with the harvesting time (whether this was in the dry season or rainy season) from 2.0 to 3.6 tonne/ha. The composition of the sun-dried leaves from six varieties is shown in Table 3. On a dry matter basis the protein content (N x 6.25) varied from 23.9 to 34.7% and fibre from 9.68 to 14.6%. The content of HCN varied from 285 to 509 mg/kg DM and was reduced on average by 89% after drying. The levels in the fresh leaves were considerably lower than those reported by Ravindran (1991) who quoted a range of 800 to 3,200 mg HCN/kg DM.

Table 3: Chemical composition of some varieties of cassava leaves
	Protein	EE	Fibre	Ash	HCN*	HCN#
	-----------------% of DM-----------				--mg/kg DM--
KM 94	34.7	13.3	12.0	6.2	509	86.3
India	31.0	14.1	11.7	7.8	411	57.6
Gon	28.5	13.5	14.6	7.9	285	17.0
Japan	27.1	14.6	10.1	6.1	347	57.1
KM 60	25.4	14.4	9.70	5.0	490	23.0
KM 95	23.9	15.6	10.7	5.9	360	20.2

*Before sun-drying #After sun-drying

Table 4: Chemical composition of the dry and ensiled cassava leaves used in the experiments
	Ensiled	Sun-dried
% of dry matter
Protein (N x 6.25)	27.6	26.0
Fibre	17.1	16.1
Ether Extract	13.9	9.9
Ash	10.3	10.9
mg/kg dry matter
HCN	147	22.5

The crude protein contents of the ensiled and sun-dried leaves were similar and midway in the range of values determined in the field. After 6 weeks of ensiling the level of HCN was reduced to 147 mg/kg DM. The residual values after sun-drying were lower (23 mg/kg DM). It appeared that sun-drying was more effective than ensiling in reducing the HCN content; this is contrary to the experience of Bui Van Chinh (see Preston 1995) who reported values as low as 33 mg/kg DM for ensiling compared with 80 in sun-dried meal. Fermentation was reported by Limon (1991) and Nguyen Thi Loc et al (1996) to be an effective way of reducing the HCN content of cassava roots.

The mean values for apparent digestibility and N balance are presented in Figures 1 to 4 and in Annexes 1 and 2. The apparent digestibility of all nutrients decreased linearly (P=0.001) as the proportion of cassava leaves in the diet increased. The data for the 30% substitution level on the sun-dried meal are similar to those reported by Bui Huy Nhu Phuc et al (1995) for a substitution rate of 35% of the soya protein.

Figure 1: Effect of level of cassava leaf protein, in ensiled or sun-dried form, on apparent dry matter digestibility (values corrected by covariance for N intake)

Figure 2: Effect of level of cassava leaf protein, in ensiled or sun-dried form, on apparent protein digestibility (values corrected by covariance for N intake)

Figure 3: Effect of level of cassava leaf protein, in ensiled or sun-dried form, on apparent fibre digestibility (values corrected by covariance for N intake)

Figure 4: Effect of level of cassava leaf protein, in ensiled or sun-dried form, on N retention (values corrected by covariance for N intake)

The rate of decrease appeared to be less when ensiled leaves were used rather than the sun-dried material. The rate coefficients were higher (rate of decrease was faster) when sun-dried rather than ensiled leaves were used (Table 5). However, precise comparisons cannot be made as different batches of leaves were used for the two processes and the experiments with each were done at different times although the pigs were the same. On the basis of the regression equations it can be calculated that the dry matter digestibility of the leaf, if it was the sole component of the diet (X=100 in the equation), would be 76.5% for ensiled leaves and 72.8% for sun-dried leaf. Using the "difference" system (assuming that the 55% non-cassava leaf component of the diet had the same digestibility as the "control"), the corresponding values would be 74.9 and 72.7%.

Table 5: Linear regressions of digestibility and N balance (Y) on percent protein from cassava leaves (X).
	Ensiled leaf	r2	Sun-dry leaf	r2
Dry matter,%	94.0 - 0.175X	.92	93.8 - 0.211X	.99
Org matter,%	95.6 - 0.169X	.94	95.7 - 0.196X	.99
Protein ,%	89.5 - 0.415X	.97	87.5 - 0.528X	.94
Fibre,%	79.7 - 0.381X	.94	75.5 - 0.473X	.94
N retention
(g/d)	12.9 - 0.0847X	.95	14.8 - 0.109X	.95
% of N digest.	57.7 - 0.127X	.69	55.0 - 0.143X	.58

Nitrogen retention as percent of N intake and N digested showed only a slight reduction up to the 30% level of protein substitution by the leaves, but decreased markedly at the 45% level. It was also observed that the pigs were reluctant to consume the diets with high level of leaf. Similar findings were reported in an earlier study (Bui Huy Nhu Phuc et al 1995) where the optimum inclusion level was found to be 35% substitution of soya bean meal protein, with marked reductions in performance at 70 and 100% substitution.

There were no indications of cyanide toxicity on any of the diets which is in agreement with the recommendations of Twe (1991) that drying and ensiling are effective ways of reducing the toxicity of cassava products.

Both ensiled cassava leaves and sun-dried cassava leaf meal can be used to substitute up to 30% of the dietary protein in diets based on cassava root meal with no significant effect on nitrogen retention, although diet dry matter was reduced slightly.

It appears that the ensiled leaf was digested better than the sun-dried leaf meal and supported higher retentions of N at all levels of substitution. However, the confounding of treatment with time, weight of pig and origin of the leaves makes it unwise to draw too firm conclusions about the relative effects of ensiling versus sun-drying.

Ensiling would appear to be the more attractive method of processing when the harvest coincides with the wet season, but is likely to be restricted to an "on-farm" operation due to the high volume and short shelf-life (after being taken from the silo) of the ensiled product. For commercial purposes the leaf meal has obvious advantages. Experiences in two remote villages in Central Vietnam (Nguyen Thi Loc et al 1996) indicated similar advantages and disadvantages for ensiled cassava roots compared with the sun-dried root meal.

This study was supported partially by the Swedish Agency for Research Cooperation with Developing Countries (Project BIL-VIE-LM-04).

Received 15 August 1996

Annex 1: Coefficients (%)of apparent digestibility of diets containing ensiled cassava leaf
	Protein from ECL, %
	0	15	30	45	SE/Prob
Dry matter	94.4	90.1	89.9	85.7	.722/.001
Protein	89.0	82.9	79.0	69.5	1.60/.001
Organic	96.0	91.9	91.3	87.7	.573/.001
Fibre	81.1	71.3	69.2	62.8	2.30/.001
Ether extract	73.8	62.3	61.6	53.0	3.19/.001
Ash	68.8	64.6	66.3	53.1	2.62/.001

Annex 2: Coefficients of apparent digestibility (%) of diets containing sun-dried cassava leaf meal
	Protein from CLM, %
	0	15	30	45	SE/Prob
Dry matter	94.1	90.2	87.5	84.5	.669/.001
Protein	87.5	77.7	75.5	61.8	2.20/.001
Organic	96.0	92.5	89.7	87.2	.531/.001
Fibre	77.4	66.0	60.5	55.6	2.30/.001
Ether extract	73.7	59.7	59.5	49.3	3.19/.001
Ash	65.0	59.3	59.9	50.4	2.62/.007

Annex 3: Nitrogen balance in pigs fed diets containing ensiled cassava leaves
	Protein from CLM, %
	0	15	30	45	SE/Prob
Balance of N, g/d
Intake	24.72	24.29	24.19	24.08	1.05/.98
Faeces	2.57	4.07	5.06	6.98	0.38/.001
Urine	9.05	9.16	8.50	8.10	0.47/.38
N retention
g/day	13.09	11.06	10.63	9.0	1.01/.063
%of digest	58.8	53.5	55.2	51.9	3.09/.45
% of intake	53.0	45.5	43.9	37.4

Annex 4: Nitrogen balance in pigs fed diets containing sun-dried cassava leaves
	Protein from CLM, %
	0	15	30	45	SE/Prob
Balance, g/day
Intake	31.1	30.6	31.9	30.4	0.62/.323
Faeces	4.1	6.8	7.8	11.6	0.64/.001
Urine	12.0	11.8	10.7	9.9	22.2/.025
Retention
g/d	14.9	12.1	13.5	9.0	0.72/.001
% of digest	55.1	51.7	54.1	46.9	2.5/.121
% of intake	47.9	39.5	42.1	30.0