Feeding value of grass ensiled with absorbents assessed in growing lambs

Sailorehusiilojen tayton yhteydessa ruohon joukkoon lisattiin kerroksittain (50 kg/1000 kg rehua) absorbenteiksi puristeleiketta (melassoimaton leike), sailoleiketta (75 % melassileiketta ja 20 % vehnanlesetta), litistettya ohraa ja melassileiketta. Sailontaaineena kaytettiin AIV II-liuosta seka granuloitua muurahaishappovalmistetta tai sailontaainetta ei kaytetty lainkaan. Puristenesteen ymparistoa kuormittavaa vaikutusta selvitettiin mittaamalla, kuinka paljon absorbentit pidattivat puristenestetta seka mika oli puristenesteen kemiallinen hapenkulutus (COD). Sailo-rehujen maittavuutta, rehuarvoa ja sulavuutta tutkittiin kasvavilla karitsoilla. Vapaan sailorehuruokinnan lisaksi karitsat saivat soijaa 200 g elainta kohti paivassa. Koko rehuannoksen sulavuus maaritettiin kayttamalla happoon liukenematonta tuhkaa rehun sisaisena merkkiaineena.Ymparistoystavallisin absorbentti-sailontaainepari saatiin rehulla, joka valmistettiin granuloidulla muurahaishapolla ja absorbenttina kaytettiin puriste-leiketta. Ohra ei sitonut puristenestetta lainkaan. Kokonaishapen kulutusta kasvattava tekija oli selvasti absorbenttien ravinnepitoisuus ja ravinteiden liukeneminen puristenesteen mukana. Absorbenttien kayttaminen ei vaikuttanut rehuannoksen sulavuuteen. Parhain orgaanisen aineen sulavuus (80,8 %) oli rehulla, joka oli valmistettu granuloidulla muurahaishapolla ja jossa absorbenttina oli puristeleike ja huonoin (77,3 %) AIV II:lla valmistetulla kontrollisailorehulla. Karitsat soivat eniten leikkeilla valmistettuja sailorehuja seka kasvoivat niilla myos parhaiten. Kokeellisen granuloidun muurahaishappovalmisteen kayttaminen AIV II-liuoksen sijasta ei vahentanyt puristenesteen erittymista, mutta se vahensi huomattavasti puristenesteen kemiallista hapenkulutusta sitoen siis ravinteita. Melassileike paransi sailorehun laatua ja vahensi puristenesteen eritysta verrattuna painorehuun, mutta se nosti puristenesteen kemiallista hapenkulutusta. Melassoimattoman puristeleikkeen kaytto absorbenttina osoittautui tuloksiltaan parhaimmaksi.

Silage made from direct-cut grass containing less than 300 g of dry matter (DM) kg ' produces a highly polluting effluent (Gordon 1967).One of the main strategies in seeking to reduce effluent production is to incorporate absorptive materi- als into the grass at ensiling.Cereal grains are generally considered less absorptive than fibrous materials (Dexter 1961), but marked absorption has been reported under both laboratory condi- tions (Spörndly 1986) and farm conditions (Done 1988, Jones et al. 1990).An ideal absorbent is expected not only to reduce effluent flow but also to have positive effects on silage feeding value and animal performance.A previous small-scale experiment identified materials which, when ensiled with grass, may reduce effluent losses (Mannerkorpi and Toivonen 1993).The purpose of the present study was to confirm these earlier findings using an ensiling technique which more closely resembles that used on a farm.Further, a common dietary energy supplement, barley, was tested as an effluent absorbent, and silage quality was examined using molassed sugar beet shreds (MSBS) without the widely applied for- mic acid silage additive.The effects of various absorbent materials on silage quality, effluent production, diet digestibility and the ad libitum intake and performance of growing lambs were tested.

Material and methods
The absorbents were sprinkled in layers at a rate of 2.5 kg of absorbent per 50 kg of grass fresh weight (50 kg f 1 grass).The control silages (Cl and C 2) were prepared without absorbents.With the exception of MSBS and the control (C2), all silages were ensiled using AIV2 solution (80% formic acid, 2% orthophosphoric acid) at a rate of 4.3 litres t 1 grass sprayed on grass during harvesting.Unmolassed sugar beet shreds were also ensiled with the same amount offormic acid in a granulated form (grFA) (Kemira Chemicals, Finland).The grFA (55% porous substance, i.e. damolin, 18% formic acid, 10%phosphoric acid) was added by hand in layers to the grass at en- siling (22.7 kg f 1).Each silo was tightly sealed with plastic after filling and weighted with con- crete blocks, 500 kg/m 2 .Silage effluent production was measured and sampled (10%) daily for the first two weeks and thereafter weekly until the silos were opened.Effluent samples were immediately frozen.The silage retention rate was calculated as follows (in kg): (control

Ensiling study
Seven silages were simultaneously prepared in identical 1-tonne capacity fibreglass-walled round silos (volume 1.5 m 3, diameter 1.4 m) us- ing second-growth direct-cut timothy-fescue (50% Phleum pratense and 50% Festuca prat- ensis) grass taken on 27th and 28th August 1993 by flail harvester.The silos were fitted with individual drainage systems for collecting and monitoring effluent production and composition.Samples of harvested material were collected from each load and the grass for each silo was weighed.The materials used to stem silage ef- fluent flow were molassed and unmolassed sugar beet shreds (MSBS, SBS), a commercial product (CP) and rolled barley (RB).The CP (a spe- cial compound feed manufactured by Suomen Rehu, Finland) was based on molassed sugarbeet shreds (75%), wheat bran (20%), malted barley (3.4%), plant oil (1.0%) and Na-benzoate (0.6%).

Animal trial
After a preservation period of 20 weeks, the silos were opened and the silages removed, mixed well, sampled and vacuum-packed in smaller portions for the feeding trial and stored about for 10 days in a sheep barn (maximum tempera- ture + 15°C) before the trial started.Fifty-six Finnish Landrace lambs with an initial weight of 38 kg (SD 5.1) and an initial age of 147 days (SD 5.6) were allocated by weight and sex into seven groups in a randomised block design.The lambs were housed individually in galvanised metal cages (measuring 1.2 x 2.2 m) with three feed-bins and a water nipple.In a preliminary study, daily grass silage DM consumption (mean 58 g kg' 1 live weight 075 ) was determined to es- tablish the daily portions to be fed to the ani- mals.Each group of eight lambs was offered its respective silage for 21 days.To balance the feed ration, the daily diet consisted of 200 g of soya- Vol.6(1997): 173-182.bean meal (crude protein 512 g kg 'DM) and si- lage ad libitum to allow refusal of 10% of the silage allowance.A mineral mixture (Ca/P=l.B) and salt (NaCl) were added at 20 g and 10 g lamb 1 day 1, respectively.The amount of silage offered and refused was recorded for each animal.Silage and feed refusal samples were collected daily during the last seven study days.Faecal samples were taken from each animal twice a day during the last five study days.Silage and faecal sam- ples were immediately frozen.

Chemical analysis
The DM content of the feeds, feed refusals and faecal samples was determined by oven drying at 105°C for 24h.The DM content of the silages was corrected as in Huida et al. (1986).The feed analyses of grass, silages, refusals and faeces were conducted according to standard procedures.In addition, the silages were analysed for total and water-soluble nitrogen by the Kjeldahl method, for ammonium nitrogen (McCullough  1967), pH and lactic acid (Barker and Summer- son 1941), for volatile fatty acid by gas chroma- tography (Huida 1973, Huida et al. 1986) and for water-soluble carbohydrates (Somogyi 1945) and ethanol (Huida 1982).The amounts of neutral detergent fibre (NDF) and acid detergent fibre (ADF) in silages were determined accord- ing to the methods of Robertson and Van Soest (1981).The digestibility coefficients for soyabean meal were taken from the feed tables (Salo  et al. 1990).The apparent digestibility of the diets was determined using acid insoluble ash as an internal marker on the last five days of the trial (Van Keulen and Young 1977).The digestibility of the silages was calculated by the differ- ence method (the diet -soya bean meal).The metabolisable energy (ME) of the experimental feeds was calculated according to MAFF (1975).The effluent samples were analysed for DM, crude protein by the Kjeldahl method, and wa- ter-soluble carbohydrates (Somogyi 1945).The chemical oxygen demand (COD) of the effluent was colorimetrically determined by the Hach method (Hach Company, Colorado, U.S.A.).

Statistical analyses
Data on organic matter (OM) digestibility, intake of OM, intake of digestible OM and the growth rate of the lambs were analysed with the GLM procedure (SAS 1992).The model was as follows: Y... =m+A .+ B+ C. + d..., where Y...
the observed response (e.g.OM digestibility), m is the overall effect, A the effect of block i (i=1..8), B. (j) the effect of animal j withinblock i (j=1..56), C k the effect of treatment k (k=1..7) and d... the residual error.The effect oftreatments !jk was evaluated using the following contrasts: SBS + RB + CP vs. Cl, MSBS vs. C 2, SBS(FA) vs. SBS(grFA) and SBS vs. CP.
The chemical composition of the absorbents is given in Table I.

Silage effluent and chemical oxygen demand
The results given in Table 2 indicate that none of the absorbents could totally retain the efflu- ent.The MSBS absorbent treatment without the use of formic acid additive was the most effective in retaining effluent (1.58 kg kg ') followed by SBS, SBS ensiled with grFA and CP.The absorption ability of the sugar beet shreds was SBS = unmolassed sugar beet shreds, RB = rolled barley, CP = a commercial product based on molassed sugar beet shreds (75%) and wheat bran (20%), MSBS = molassed sugar beet shreds, DM = dry matter thus better than in the study of Kennedy and Moore (1988) (1.0I effluent kg' 1 absorbent).
Rolled barley was the most ineffective absorb- ent for controlling effluent loss; effluent flow from RB silage was twice as high as that from MSBS silage and even higher than that from the control silage.These findings differ from those of Done (1988) and Offer and Al-Rwidah (1989), who found a significant reduction in effluent production when RB was incorporated into grass at ensiling.According to Jones et al. (1990), incorporation of RB may reduce effluent flow by up to 50%.Johnson (1992) found, however, that inclusion of 40 kg of RB had no effect on total effluent production, whereas rates of 60 kg and 80 kg showed progressively more effect on ef- fluent flow.Here the amount of RB (50 kg f 1 grass) used was, therefore, probably too low for efficient effluent retention.The use of formic acid in granulated form instead of liquid form did not affect the effluent retention rate.There were clear differences in chemical composition between the effluents produced by the different silages.The mean DM of effluent produced from silages treated with sugar beet shreds was 45 g kg 1 , which is in close agree- ment with that reported by Purves and McDon- ald (1963) for farm silos (40 g I' 1 ).With the ex- ception of RB silage, the absorbent-treated si- lages showed lower DM losses than untreated silages.However, effluent from the treated si- lages contained more sugar and crude protein than effluent from the corresponding control because the nutrients of the absorbents were also solubilized.The use of MSBS without any acid preservative resulted in a smaller amount of sugar in effluent than did unmolassed silage treated with formic acid.This was attributed to the hydrolytic effect of formic acid (compare Cl with C 2), which was used in combination with the SBS treatment.
The polluting strength of an effluent is de- scribed by its chemical oxygen demand (COD).The most environmentally friendly combination of absorbent and silage additive in terms of high retention rate and low COD content of effluent was SBS with grFA.The silage preserved with grFA had lower nutrient losses and COD than the silages treated with FA.Only the effluent of Cl silage had a lower COD content; this silage had, however, higher effluent production.The use of MSBS instead of SBS increased the COD of the silage effluent (MSBS 53.2 vs. SBS 43.4), which is consistent with the findings of a previous small-scale experiment (Mannerkorpi and  Toivonen 1992).The COD values of 33.7-53.2(mg 1 1 x 10 3 ) are in agreement with those pre- sented by Jones et al. (1990).

Chemical composition of silages
The chemical composition and quality of silages at silo opening are given in Table 3.The average DM content of the control silages during the ad libitum feeding period was 209 g kg 1 but that of the silages containing absorbents 219 g kg ' (Table 4).Incorporation of sugar beet shreds has been found to increase the silage DM content (Offer and Al-Rwidah 1989).Silages treated with sugar beet shreds had a lower crude protein con- tent than untreated or RB treated silages owing to the dilutive effect of the absorbents with markedly lower crude protein contents than in grass.Comparison of the formic acid-treated silages showed that SBS and CP inclusion raised the sugar content but that RB silage, in which etha- C 1 = control silage with formic acid, SBS = unmolassed sugar beet shreds, RB = rolled barley, CP = a commercial product based on molassed sugar beet shreds (75%) and wheat bran (20%), C 2 = control silage without any additives, MSBS = molassed sugar beet shreds, FA = 80% formic acid solution, grFA = granulated formic acid solution, DM = dry matter, FM = fresh matter, COD = chemical oxygen demand nol and lactic acid production were highest, had the lowest sugar content.Differences between silages in contents of crude fibre and N-free ex- tract were small.The ash content of SBS ensiled with grFA was conspicuously high.All silages were well preserved (Table 3 - 4).Only the control silage ensiled without any additive (C2) contained BO g kg ' total nitrogen.Butyric acid was encountered in small amounts in silage SBS ensiled with grFA.None of the absorbents incorporated in formic acid- treated silage (C 1) had a marked effect on silage quality with the exception of SBS in combina- tion with formic acid, which improved silage fermentation owing to the lower content.
The results for RB are consistent with those of Johnson (1992) and Done (1988), who found that RB used as an absorbent did not improve fer- mentation quality in silage prepared without any additives.In the untreated control silage (C 2), fermentation quality as reflected by lower am- monia N/total N and acetic acid content and pH was clearly increased by MSBS inclusion.Be- cause of the more extensive lactic acid fermentation in silages not treated with formic acid, pH values were generally higher in silages treated with formic acid than in untreated silages, as was found also by Offer and Al-Rwidah (1989).

Digestibility, intake and growth rate
In general, diet organic matter digestibility was not affected (p>0.05) by silage absorbent treat- ment (Table 5).Digestibility of SBS silage was, however, increased when grFA was used as an additive (p=0.0003)owing to the nutrient absorption capacity of the grFA product.This silage also showed the highest digestibility (80.6%).The formic acid-treated control silage (Cl) had the lowest digestibility (77.1%).In the study of Mannerkorpi and Toivonen (1993), inclusion of SBS and MSBS increased the in vitro organic matter digestibility of silages.Offer and Al-Rwidah (1989) likewise found that silage digestibility in vitro was improved by the addition of Table 3. Quality of silages at silo opening (C 2 sample is taken 28 days earlier) with standard deviation.In each figure pair the upper figure is the mean and the lower figure the standard deviation.Cl = control silage with formic acid, SBS = unmolassed sugar beet shreds, RB = rolled barley, CP = a commercial product based on molassed sugar beet shreds (75%) and wheat bran (20%), C 2 = control silage without any additives, MSBS = molassed sugar beet shreds.FA = 80% formic acid solution, grFA = granulated formic acid solution.DM = dry matter, WSCs = water-soluble carbohydrates MSBS and RB.Lambs offered absorbent-treat- ed silage had significantly higher OM intakes of ensiled material than those offered untreated silage (SBS+RB+CP vs. Cl, p=0.004,MSBS vs.

Silages
C 2, p=0.0001) (Table 5).Intake of diet organic matter was about 29% higher for lambs receiv- ing silages treated with sugar beet shreds than untreated silages.This finding suggests that the increased intake for treated silages was not a result of better silage digestibility but of either palatability or the effect of absorbent treatment on rumen fermentation kinetics.In general, in-take was higher for the absorbents with the most efficient effluent retention rates.
The observed positive effects of absorbents on intake are consistent with the findings of Jones et al. (1990), Johnson (1992) and Davies and Perrott (1991), but not with the observation of Kennedy (1988), who reported that intake by cattle receiving 2 kg ofsupplementary feed head' 1 day 1 was unaffected by absorbent treatment.
In the study of Done and Appleton (1988), MSBS, but not RB, treatment improved the dai- ly DM intake.In our study, intake of silage treat- Cl = control silage with formic acid, SBS = unmolassed sugar beet shreds, RB = rolled barley, CP = a commercial product based on molassed sugar beet shreds (75%) and wheat bran (20%), C 2 = control silage without any additives, MSBS = molassed sugar beet shreds, FA = 80% formic acid solution, grFA = granulated formic acid solution, DM= dry matter, NDF = neutral detergent fibre, ADF = acid detergent fibre, WSCs = water-soluble carbohydrates ed with RB was also noticeably lower than that of silages treated with sugar beet shreds, most likely due to silage fermentation quality.Further- more, the RB silage had an unpleasant mucous consistency and a particularly sweet smell.Vol- untary feed intake can be reduced by a high lac- tic acid content of silage (McLeod et al. 1970).
The high lactic acid content in MSBS silage without formic acid treatment may partly explain why intake of this silage was lower than that of other sugar beet shred-treated silages.The most palatable SBS-grFA silage included butyric acid.Experience has shown that butyric acid does not necessarily affect silage intake (Sormunen-Cris- tian 1984).The use of sugar beet shreds as absorbents improved the diet ME value (p=0.003) and live weight gains of lambs (p=0.02).Daily growth of lambs was highest with SBS silage (124 g), next highest with SBS ensiled with grFA (108 g) and lowest with RB (36 g).Incorporation of sugar beet shreds has yielded a tendency towards higher lamb live weight gains (Done 1988), but the improvement in silage ME value has been slight (Offer and Rwidah 1988).One of the lambs in the group fed RB silage lost weight and two of them did not grow at all.All the above findings refer to an absorbent- treated grass silage compared with pure grass si- lage.If the objective is to establish whether there is a differencebetween absorbent/concentrate added at ensiling and that fed separately, a different Cl = control silage with formic acid, SBS = unmolassed sugar beet shreds, RB = rolled barley, CP = a commercial product based on molassed sugar beet shreds (75%) and wheat bran (20%), C 2 = control silage without any additives, MSBS = molassed sugar beet shreds, FA = 80% formic acid solution, grFA = granulated formic acid solution, SEM =standard error of means, OM = organic matter, DM = dry matter, DOM = digestible organic matter, W°J != metabolic live weight.NS = non significant type of experiment must be carried out.According to the results of Johnson (1992), a greater increase in animal live weight gain could be expected if the same amounts ofRB were fed separately with un- treated silage rather than as an absorbent.In con- trast, Done and Appleton (1998) found no evidence of better performance in lambs fed a "complete" diet than in those offered silage and supplement as separate components.

Conclusions
In general, there was no difference between SBS, MSB and CP silages in the animal parameters measured.The use of grFA product did not re- duce effluent production but it clearly decreased effluent COD and thereby silage digestibility without affecting growth of lambs.Incorporation of MSBS without formic acid treatment improved silage quality and increased intake, but did not affect lamb growth.Further, it reduced effluent production but increased its COD.Rolled barley did not retain silage effluent, nor was the silage palatable for the lambs.In view of the effects on effluent COD, digestibility and silage intake, it is recommended that SBS be used as absorbent and grFA as silage additive.

Table 1 .
Chemical composition of the absorbents used in

Table 2 .
Effects of absorbent inclusion and formic acid treatment on effluent quantity, composition, dry matter losses, retention rate and chemical oxygen demand.

Table 4 .
Chemical composition and quality of silages in ad libitum feeding period.

Table 5 .
Digestibility and intake of silage (grass silage with absorbent) and of diet (silage plus soya-bean meal) and growth rate of lambs in ad libitum period.