CHEMICAL COMPOSITION AND ENERGY UTILIZATION VALUES OF COMMON SRI LANKAN FEEDSTUFFS FOR GROWING PIGS

Thirteen balance trials, each involving eight growing pigs (average body weight, 32 & 4 kg), were conducted to determine the digestible energy values of commonly available local pig feedstuffs. The digestible energy values (Kcallg) of the feedstuffs on an as-fed basis were : maize, 3.64; sorghum, 3.53; rice bran no:l, 2.40; rice bran no:2, 2.09; cassava root meal, 3.69; coconut poonac, 2.54; gingelly poonac, 2.79; soybean meai, 3.16; cassava leaf meal, 2.13; ipil-ipil leaf meal, 1.54; poultry litter, 1.92 and fish meal (imported), 3.01. The energy utilization value of the local fish meal sample could not be determined, owing to its adverse effects on the animals. The proximate composition and mineral contents of the feedstuffs are also reported.


Introduction
Digestible energy (DE) is the widely used term for expressing the available energy in pig feedstuffs.Since available energy content is a .majordeterminant of the nutritive value of a feed, its assessment is of importance for efficient diet formulations.
While composition and energy values of common feeds in the temperate regions are well documented, corresponding data on materials indigenous to the tropics are limited.In Sri Lanka, digestible energy values of local feedstuffs for pigs have never been investigated.With the current interest in expanding the pig industry, such an evaluation would be timely.The aim of the present study was to determine the DE contents of thirteen local feedstuffs used for growing pigs.The composition and mineral profiles of these ingredients were also determined.

Feed Materials
All samples, except cassava (Manihot esculenta) root meal, cassava leaf meal, ipil-ipil (Leucaena leucocephala) leaf meal and poultry litter were purchased from commercial outlets in Kandy.Cassava root meal, cassava leaf meal and ipil-ipil leaf meal were prepared in bulk using previously reported methods.la3Poultry litter, about 12-month old and paddy husk-based, was collected from the deep litter poultry houses of the University Farm.The litter was sun-dried, heaped in a well ventilated roqm and allowed to mature for four weeks before use.

Experimental Design
Thirteen separz're balance trials, each involving eight growing, castrated male pigs (average body weight, 32 kg; range, 26-40 kg) were conducted.Different experimental designs were employed for the evaluation of energy asld protein supplements.

Energy Supplements
In each trial, eight pigs were grouped irlto two blocks based on body weight and the ouiccme groups were assigned to either a basal diet or a test diet.The basal diets consisted of 75% maize, 16% soybean meal, 5% fish meal plus 4% minera!and vitamin supplements, and supplied 16% crude protein.The test diet was formulated by repl2cin.g(w/w) the basal diet with 30% test feedstuff.
A 2 X 2 Latin Square change-over design was exnployed.Each experimental period consisted of an adjustment period (5 days), a preliminary period (5 days) and a total collection period (10 days).Thus each trial was of 40-day duration.By the conclusion of the trial, each diet has been fed to dl eight ~i g s .

Protein Supplements
Since evaluation of high-protein feedstuffs at dietary !.evels far in excess of normal levels of use may underestimate their energy utilization values: a modified desig was used for the protein supplements.In each trial, eight gigs were assigned to four outcome groups based c~n litter and body weight.The outcome groups were then allocated either to a basal diet (similar to that used for evaluation of energy stspplements) or to one of the three test diets where the basal diet was replaced (wh) with 6.6, 13.3 and 20% of the test feedstuff.The only exception was fish meal, where the replacement levels employed were 3.3,6.6and 10%.
A 4 X 4 Latin Square change-over disign was utilized.Each experimental period consisted of an adjustment period (5 days), preliminary period (5 days) and a total collection period (5 days).Each trial therefore lasted 60 days.By the conclusion of the trial, each diet had been fed to all eight pigs.Experimental Procedare Pigs were housed in individual metabolic cages which had facilities for separation and collection of faeces and urine.During the adjustment period, after being put into the cages, only the basal diet was given.~u r i n g the preliminary period, the dietary treatments were introduced and the pigs were fed twice daily with an amount calculated at 5% of their body weight.All diets were fed in a wet mash form (1:l feed:wate-ratio).During the collection period, faeces were collected daily from each cage, dried at 6 0 ' ~ for 4. 8 hours and weighed.Faecal and feed samples were bulked, ground to pass through a 60-mesh screen and representative samples were taken for the determination of gross energy by calorimetry.
The analyses on feed samples were conducted according to the methods of the Association of Official Analytical ~h e m i s t s .~ Cations were determined by Atomic Absorption Spectrophotometry and phosphorus by colorimetry, as described previously. 6

Statistical Analysis
Digestible energy contents of the energy suppl2ments were calculated by the difference method, whereas regression analysis ;vas employed for the protein supplements.7These calculations assumed that there were no associative effects between the basal diet and the test ingredient.

Results and Discussion
The chemical composition of the test feedstuffs are presented in Table 1.The average values of cereals, cassava root meal and imported fish meal are comparable to those documented by ~ohl: but differences were observed for the composition of by-product feeds.Rice bran, coconut poonac and gingelly poonac samples were .I found to contain relatively lower contents of crude protein and higher contents of crude fibre than those reported from elsewhere in the tropics.8The mineral contents of the feedstuffs are presented in Table 2.The calcium and phosphorus contents of the feedstuffs are very similar to those reported e l ~e w h e r e .~-l ~ 'Data reported in the literature for the trace mineral contents of feedstuffs are often variable, since significant differences can arise depending on soil and climatic conditions.The high levels of iron determined in all ingredients are probably reflective of contamination during post-harvest processing.Feedstuffs originating from plant materials contained very low levels of calcium, with the exception of gingelly poonac and the two leaf meals.These three feedstuffs were also found to be rich sources of trace minerals.The data also highlight the potential usefulness of poultry litter as a mineral supplement.Poultry litter contained 5.26% calcium, 2.40% phosphorus and rather high levels of trace minerals.Values for dry matter digestibility, gross energy and DE of the feedstuffs are shown in Table 3.The DE contents of maize and sorghum were determined to be 3.64 and 3.53 KcaVg, respectively.These values were remarkably similar to those reported by ~l l e n ~ and NRC" for samples from North America.Over 85% of the dry matt& as well as energy contained in both cereals were digested by growing pigs.Somewhat similar DE contents of maize and sorghum are of some practical relevance in Sri Lanka as sorghum, owing to its drought tolerant ability, can be -grown under solely rain-fed conditions to produce moderately good grain yields in areas where maize cannot be successfully cultivated.
In Sri Lanka, large quantities of rice milling by-products are available for use in local animal feed mixtures.Unfortunately, due to inefficient milling techniques, more than 50% of the by-products processed in Sri Lanka are unsuitable for non-ruminapt feeding.Most rice mills currently in use are of the out-moded huller type which removes both the husk and the bran in one operation, thus producing a bran with considerable admixture of husk.This is reflected by the high fibre and ash contents in the bran samples available in local feed market as rice bran no.1 and rice bran fis.2 (Table 1).The DE contents of rice bran no.1 and no.2 samples were determined to be 2.40 and 2.09 KcaVg, respectively.Only 58-65% of the energy contained in rice bran was digested by pigs.Digestible energy contents of local rice bran samples were considerabl lower than those reported for samples from Australia1' and North America?*''There lower values are related to the high fibre and ash contents of local The DE content of cassava root meal was found to be similar to that of maize.This finding is in agreement with that of Au&aitre12 and highlights the role that cassava root meal could play as a cereal replacer in pig diets.
Among the protein supplements, soybean meal had the highest DE content (3.16 Kcavg).The DE contents obtained for locally-produced coconut poonac4,1$ngelly poonac an.d soybean meal were lower than those reported eisewhere.This discrepancy is due to the relatively high fibre,contents of local sampies.
The crude protein content of locally-produced fish meal was analysed to be only 41.5%, far below the level of 70.6% determined for the imported fish meal sample (Table 1).When pigs were fed on diets contahing 6.6 and 10% local fish meal, performance was adversely affected.All animals had watery faeces and lost body weight.The balance trial had to be terminated after the first experimental period.The local fish meal used in the study appeared normal in colour and smelI, and there was no evidence of putrefacation.Subsequent analysis showed excessive saIt content to be the probable cause for the adverse effects.The sample contained 11.2% salt, which was much above the legally accepted level of 7%?3 The DE content of the imported fish meal was determined to be 3.01 KcaVg.
The inadequate supplies and r i s i ~g cost of traditional feedstuffs have focussed attention in the recent past on the need to explore the use of hitherto untapped feed resources.Cassava leaf meal, ipil-ipil leaf meal and poultry litter were evaluated for these reasons.The DE content of these t k e e feedstuffs was estimated fo be 2.13,1.54and 1.92 KcaVg, respectively.No comparable data is available OL tbe DE contents of these feed ingredients.Though the nutrient contents of cassava leaf meal and ipil-ipil leaf meal were remarkably similar, the energy digested 5y pigs was considerably lower (52 vs 38%) for the ipil-i~il leaf meal.The tannin and airnosine contents of ipil-ipil leaf meal may be responsible for this low digestible energy.3 The DE contents reported in this study were obtained using growing pigs.It is noteworthy, however, that the digestion coefficients of feedstuffs, particularly of high-fibre materials, could be expected to be higher for older pigs as reported by Fernandez et al. 14 in comparative studies with growing pigs and sows.Older pigs have a higher capacity for fermentation and utilization of fibre components Eil the hindgut than the growing animals.
Published data on the available energy content of feedstuffs of tropical origin are scanty.Consequently pig feed formulations in tropicai countries, including Sri Lanka, have traditionally used DE contents that have been estimated for feedstuffs originating from tempezate regions.The results of the present study showed that such usage may not be appropriate, in view of the relatively high fibre contents of tropical feedstuffs.This is particularly true for the agro-industrial by-products which form the feed base for pig productibn in tropical countries.The data presented will have applications in the formulation of accurate and efficient pig rations in Sri Lanka as well as in other tropical areas where these feed ingredients are commonly used.
meal (imported) Fish meal (local) per cent of dry weight

Table 1 :
Percentage chemical composition of the feedstuffs

Table 2 :
Mineral contents of the feedstuffs (70 dry matter basis)

Table 3 :
Dry matter digestibility (DMD), digestible .energy(DE) and energy digestibility of the feedstuffs for growing pigs.