comparison of flail-harvested, precision- chopped and...

Comparison of flail-harvested, precision-chopped and round-bale silages for growing

beef cattle

E. Charmley1† and S. Firth2

1Crops and Livestock Research Centre, Agriculture and Agri-Food Canada, Nappan, Nova Scotia, Canada

2Agrapoint, 92 Webster Street, Kentville, Nova Scotia, Canada

The effects of silage conservation method on silage composition and animal perform-ance were examined in two experiments. In Experiment 1, unwilted, flail-choppedsilages made with or without an additive (sodium nitrite and hexamethylene tetramine)were compared with wilted, round-bale silage. The dry matter (DM) concentration ofround bale silage (460 g/kg) was higher than that of flail silage (214 g/kg) and thisrestricted fermentation and N solublisation. When fed to growing cattle, intake(P<0.01), live-weight (LW) gain (P<0.001) and LW gain to feed ratio (P<0.05) weregreater for round-bale silage than for flail silage. In Experiment 2, flail-harvested silagewas compared with wilted, precision-chopped and round-bale silages conserved eitherwithout or with pre-slicing immediately before baling. The DM concentration of flail,precision-chopped and round-bale silages were 163, 334 and 468 g/kg, respectively.Fermentation in flail silage was more extensive than in precision-chopped and particu-larly round-bale silages, but insoluble-N concentration was unaffected. Round-balesilage was more digestible (P<0.05) than flail or precision-chopped silages. Voluntaryintake was higher for steers fed round-bale silages compared to flail silage (P<0.05),while intake of steers fed precision-chopped silage was intermediate (P>0.05). Steersfed round-bale silages had higher LW gain (1.0 kg/day) than those fed flail (0.7 kg/day)or precision-chopped silage (0.8 kg/day; P<0.05). Efficiency of utilization of DM for LWgain was similar for all silages. Pre-slicing at baling had no effect on animal perform-ance. It is concluded that the increased performance by cattle offered silages made bythe wilted round-bale system was largely due to higher voluntary intake.

Keywords: Cattle; dry matter intake; live-weight gain; silage systems

Irish Journal of Agricultural and Food Research 43: 43–57, 2004

43

†Corresponding author: [email protected]

IntroductionPrecision-chopped ensilage is an effect-ive method of forage conservation butequipment costs frequently result in thisoption not being available to smaller scale livestock farmers, typified inAtlantic Canada by beef producers.Alternative silage systems are availableincluding flail harvesting (Gordon, 1989)and round-bale silage making (Haigh,1990) which require less investment inequipment (Wilkinson, Wadephul andHill, 1996). Research has demon-strated that animal performance fromflail-harvested silage can be as good as, orbetter than, that from wilted, precision-chopped silage (Gordon, 1989). Althoughvoluntary intake tends to be lower for flailand other types of low dry matter (DM)silage, it appears to be offset by increasedefficiency of utilization (O’Kiely, Flynnand Wilson, 1988; Gordon, 1989). The risk of adverse fermentation increases in wetter, flail-harvested silages, especial-ly if the pre-ensiled herbage is low inwater soluble carbohydrates (WSC). Theuse of additives specifically formulatedfor wetter silages may be beneficial undersuch circumstances. Research with round-bale silage, in contrast, has demon-strated increased voluntary intake andanimal performance relative to precision-chopped silage, particularly when forage is baled at a higher DM concen-tration (Charmley et al., 1999). Pre-slicing of forage before round baling is reputed to improve fermentation and animal performance. Little researchis available where these systems have been compared in the same exper-iment. Therefore, two experiments wereconducted over 2 years to compare conservation characteristics and feeding value of silages made using flail, round-bale and precision-chop systems.

Materials and MethodsField operations Two experiments were conducted in suc-cessive years at the Nappan ResearchFarm, Nova Scotia, Canada (45°46' N,64°14' W). In both years, timothy-dominat-ed swards were used. In year 1, second-cutforage (15.3 ha) was harvested from threeadjacent fields. In year 2, first-cut material(9.4 ha) was used from one of the fieldsused in year 1 plus two similar, timothy-dominated fields. Prior to harvesting, fieldswere divided into equal areas and thesewere allocated at random to the variousensiling systems. Thus, within both yearsthe field was the replicate, there beingthree replicates in each year. Harvestingconstraints resulted in all perimeter forage(10 m) being assigned to the flail treatmentin both years. Samples of forage were takenfor determination of chemical compositionfrom all treatments at the time of ensiling,either by sampling loads of silage (flail andprecision-chopped) or sampling the swathimmediately before baling (round-baletreatments). All loads of silage and roundbales were weighed immediately aftertransfer from the field. The harvested DMyield was determined as the weight of for-age corrected for the DM concentrationmeasured in samples taken from loads orfrom the swath immediately before baling.Stubble heights were measured in flail-har-vested and mowed plots with a minimum of20 measurements per cutting method. Experiment 1: Three harvesting methodswere compared in this experiment: flail har-vested without additive (FC), flail harvestedwith additive (FA) and round-bale (RB).Field operations began on 1 August. Flailharvesting using a John Deere crop chop-per 16A (John Deere Corp Dearborne, MI)with a 1.4 m cutting width was initiated firstand once sufficient cleared area was avail-able for equipment manoeuvering, theblocks for round-bale silage were mowed

44 IRISH JOURNAL OF AGRICULTURAL AND FOOD RESEARCH, VOL. 43, NO. 1, 2004

with a rotary disc mower (model FC300G,Kuhn S.A., France) with a 3-m cuttingwidth. Flail-harvested material was collect-ed into 5-t capacity wagons pulled behindthe harvester. Two of these were used alter-nately, to allow for more or less continualoperation of the harvester. Additive wasapplied to half the flail harvested materialby treating alternate loads with a solutioncontaining sodium nitrite (300 g/l) andhexamethylene tetramine (200 g/l; Ultrasil,FSL Bells, UK) at a rate of 2.5 l/t fresh for-age. Additive was applied by gravity feedonto the forage immediately after the flailchamber of the harvester. Forage wasformed into a heap approximately 1.5 mhigh and 3 m wide, on an uncovered con-crete pad using a front end loader. Thismachine was also used for consolidating theheap. Silos were sealed with a double layerof black plastic (0.125 mm thick) which washeld in place with used car tyres. The edgeswere held in place with soil. For round-balesilage, forage was allowed to wilt overnightin the field until the DM reached over 400g/kg. No tedding or raking was performedon the wilted crop. The crop was baled intoapproximately 1.2 m × 1.2 m bales using avariable chamber round baler (model NewHolland 640, New Holland, PN, USA).Bales were transported to a storage areawhere they were wrapped within 2 h of baling using an individual bale wrapper(model UN 7510, Kverneland UnderhangA.S., Norway). Each bale was individuallycovered with white wrap (25 micron) and an approximate two-thirds overlap, thusensuring a minimum of three layers of filmover the forage. Bales were stored on theirside outdoors on an asphalt pad.

All silages were stored for 5 monthsbefore feeding. Silages from the FC andFA treatments were removed from heapsusing a silage block cutter (Siolcut Model150, Alo, Sweden). Silages were givenwithout supplement, to growing steers

accommodated in pens of 8 animals andindividually fed using feed troughsequipped with electronic locking doors(American Calan, Inc., Northwood, NH).Eighteen yearling Hereford steers (initiallive weight (LW) 304 (s.d. 32) kg) wereranked according to LW and blocked into6 groups of 3. Within each block, one steerwas assigned to each treatment. Silageswere offered ad libitum for 98 days, duringwhich time intake was recorded on a dailybasis as the difference between feedoffered and uneaten feed remaining 22 hlater. Rate of gain was determined byweighing steers every week and regressingLW against time. Gain was taken as theslope of the relationship. All animals werecared for in accordance with the guide-lines of the Canadian Council on AnimalCare (CCAC, 1993).Experiment 2: This study employed fourensiling techniques; flail-chopped (FC),precision-chopped (PC), round-bale(URB) and cut round-bale (SRB). Fieldoperations for all treatments began on 12June. Flail harvesting was conducted aspreviously described. Precision-choppedmaterial was mowed using the sameequipment as the round-bale system. Theprecision-chopped and round-bale treat-ments were left in the field for 4 days dueto rainfall (30.5 mm) on 2 days after mow-ing. The PC material was harvested usinga New Holland harvester (model 790;New Holland, PN, USA). The silo for pre-cision-chopped material was made in thesame fashion as the silo for flail silage. Inthis experiment, round baling began at thesame time as the precision-chop harvest-ing, but owing to slower field operations,the DM concentration of round balesilage was higher than that of precision-chopped silage. A fixed chamber roundbaler (model 250RC, Claas, Columbus,IN, USA) was used, fitted with a roto-cutcutting assembly which allowed forage to

CHARMLEY AND FIRTH: COMPARISON OF SILAGE-MAKING METHODS 45

be coarsely sliced immediately before bal-ing. This mechanism was engaged foralternate bales resulting in two round-baletreatments (URB and SRB). Wrappingtook place at the storage area within 2 h ofbaling as previously described.

Samples of harvested forage were takenfrom 10 loads each of flail and precision-chopped treatments and from 5 bales ofeach round-bale treatment. Twenty parti-cles were taken at random from each sam-ple and measured for determination ofmean particle length and size distribution.In round-bale treatments only, 5 bales pertreatment were sampled by coring in order to assess rate of fermentation. Cores were taken at 0, 4, 10, 20 and 70days after ensiling for determination ofchanges in pH, N solubility and concentra-tion of fermentation acids. Cores, 60 cm inlength, were taken using a 2.5-cm dia-meter corer fitted onto an electric drill.After removal of the sample, the hole waspurged with CO2 and sealed with silageplastic repair tape. Density of bales wasestimated from the total weight of baledforage and number of bales per replicate,assuming a bale volume of 1.36 m3.

Following 7 months storage the silageswere fed to 24 crossbred yearling Herefordsteers (initial LW 334 (s.d. 46) kg) in afeeding trial lasting 84 days. Methods wereas described for Experiment 1, except that6 steers were assigned to each of four treat-ments. In addition, six crossbred Herefordsteers weighing approximately 350 kg wereused in a double incomplete latin squaredesign with four diets, three periods andthree animals per square. Cattle were adap-ted to housing in individual pens (3 m × 6m) and collection cages (1 m × 2 m) and tosilage-feeding prior to the trial. At the startof this study cattle were randomly assignedto square and, within square to diet. Eachperiod of the trial was 21 days with totalcollection of faeces taking place over the

last 7 days of each period. Cattle werehoused in individual pens with completefreedom of movement for 14 days and thenconfined by a head yoke within collectioncages, fitted to allow separation and collec-tion of urine and faeces. Water was avail-able at all times and lights were left on 24h/day. Forages were offered ad libitum.

Chemical analysesDry matter in fresh silage was determinedby toluene distillation (Dewar andMcDonald, 1961) and in other material byoven-drying at 50 °C. Organic matter wasdetermined as weight lost upon ashing for5 h at 550 °C. Total N was determined infresh silage by macro-Kjeldahl procedures[7.033–7.037 (AOAC, 1984)]. Insoluble-Nwas determined in fresh silage usingtrichloroacetic acid (TCA; Siddons, Evansand Beever, 1979). Ammonia N was meas-ured as the volatile N fraction in silagejuice distilled on a Kjeltec Autosystem,with N determined in the distillate usingthe Kjeldahl method. Ash-free neutraldetergent fibre (NDF) and acid detergentfibre (ADF), and in vitro digestibility weredetermined by methods described by VanSoest, Robertson and Lewis (1991).Volatile fatty acids and alcohols weredetermined in acid extracts of silage usinggas chromatography (Charmley, Savoieand McQueen, 1997). Lactic acid in acidextracts was determined by the colorimet-ric method of Barker and Summerson(1941). Silage pH was determined on mac-erated samples in distilled water asdescribed by Charmley et al. (1997). Grossenergy (GE) was determined by totalcombustion of fresh silage, with polyethyl-ene as a primer, or dried faeces using anadiabatic bomb calorimeter.

Statistical analysisData from the two feeding trials wereanalysed by standard analysis of variance


procedures of SAS (1995) using a ran-domized block design with treatment andblock as factors. In Experiment 1, pre-planned contrasts were employed to compare the effect of round baling v.flail-harvesting and the effect of additiveaddition to flail-harvested forage. InExperiment 2, the comparisons were flailv. round-bale silage, flail v. precision-chop silage, round bale v. precision chopsilage and unsliced v. pre-sliced roundbale silage. In the digestibility trial,analysis of variance was used to determine the effects of treat-ment, animal and period using SAS(1995).

ResultsField measurementsThe yield of regrowth forage used inExperiment 1 was very low (Table 1). InExperiment 2, where first-cut timothy wasused, the yield was more than twice that

in Experiment 1 (Table 2). In both years,flail harvesting resulted in higher residualstubble heights which probably affectedharvested yield and chemical composi-tion (Tables 1 and 2). Particle length andsize distribution in Experiment 2 weremarkedly different between round-baleand the other harvesting treatments(Table 2). Mean particle length in round-bale silages was over 400 mm and pre-slicing had only a small effect on particlelength and size distribution (Figure 1).Pre-slicing increased the proportion ofparticles less than 500 mm from 0.57 to0.79. Although this resulted in a 0.14 in-crease in wet-bale density this increasewas not significant (Table 2). Precision-chopped silage had the shortest particlelength, with over half less than 50 mm.Flail-chopped silage had a mean particlelength proportionally 0.70 longer and theproportion of particles that were less than50 mm was only 0.18 (Figure 1).


Table 1. Field measurements (s.d.) taken at ensiling – Experiment 1

Flail silage Round-bale

No additive With additive silage

Forage dry matter (DM) at ensiling (g/kg) 209 (7.2) 223 (7.8) 423 (69.4)Harvested yield (t DM/ha) 1.04 1.44 1.25Residual stubble height (mm) 149 (16.1) 149 (16.1) 93 (18.6)Wilting period (h) 0 0 22Additive (l/t DM) 0 11.5 0

Table 2. Field measurements (s.d.) taken at ensiling – Experiment 2

Flail Precision- Round-bale silagesilage chopped silage Control Pre-sliced

Forage dry matter (DM) at ensiling (g/kg) 168 (12.9) 477 (22.4) 585 (64.1) 510 (43.4)Harvested yield (t DM/ha) 2.67 3.44 2.34 2.34Residual stubble height (mm) 170 (8.0) 117 (24.1) 117 (24.1) 117 (24.1)Wilting period (h) 0 85 90 90Mean forage particle length (mm) 98.8 (14.1) 58.4 (10.7) 430 (150.9) 405 (78.8)Wet bale density (kg/m3) – – 348 (68.6) 399 (87.5)

Forages used in both years were of similar chemical composition at ensiling.Concentrations of crude protein (CP),ADF and NDF in Experiment 1 averaged147, 340 and 607 g/kg DM, respectively. InExperiment 2, these concentrations were146, 357 and 629 g/kg DM, respectively.

Silage qualityIn Experiment 1, the DM concentration of direct-cut, flail-harvested silage was similar to the DM concentration of thestanding crop and less than half the DMconcentration of the round-bale silage,

which had been wilted for 22 h (Table 3).As a consequence of this, organic acid con-centration was greater in flail silage thanround-bale silage and the proportion ofinsoluble N was reduced. However, pH was not greatly different. Additive treat-ment of flail silage had relatively smalleffects on silage quality. The proportion ofinsoluble protein in additive-treated flailsilage was intermediate between that of theother silages. The concentration of organicmatter (OM) was less in flail-choppedsilages indicating higher ash concentrationand possible soil contamination.


Figure 1: Effect of conservation method on particle size distribution.

Table 3. Mean (s.d.) chemical composition of silages – Experiment 1 (g/kg dry matter unless otherwise stated)

Flail silage Round-bale silage

No additive With additive

Dry matter (DM) (g/kg) 214 (16.5) 223 (17.9) 460 (10.1)Organic matter 855 (38.1) 885 (13.7) 911 (10.2)Acid detergent fibre 336 (34.0) 327 (46.5) 336 (11.0)Neutral detergent fibre 563 (50.2) 585 (40.5) 614 (22.2)Total N 22.8 (3.65) 23.0 (3.66) 24.2 (6.54)Insoluble N (g/kg total N) 199 (139) 278 (145) 341 (200)Ammonia N (g/kg total N) 211 (66.2) 175 (33.3) 120 (32.0)Lactic acid 16.3 (7.45) 27.4 (11.6) 13.2 (5.98)Lactic acid (g/kg organic acid) 620 (237) 753 (70.1) 754 (175)Acetic acid 7.89 (6.32) 7.14 (1.39) 3.72 (3.11)Butyric acid 0.29 (0.77) 0.22 (0.56) 0.11 (0.27)Total acids 25.7 (8.58) 35.5 (14.1) 16.8 (5.87)pH 5.08 (0.13) 4.58 (0.12) 5.53 (0.31)

In Experiment 2, the DM concentrationof precision-chopped silage was interme-diate between that of the flail silage and thetwo round-bale silages (Table 4). As inExperiment 1, the OM concentration of

flail silage was lower than in other silages.Insoluble-N concentration in flail silage was highly variable, compared to the other silages. However, ensiling methodappeared to have no influence on protein


Table 4. Mean (s.d.) chemical composition of the silages – Experiment 2 (g/kg dry matter unless otherwisestated)

Flail Precision- Round-bale silagesilage chopped silage Control Pre-sliced

Dry matter (DM) (g/kg) 163 (7.4) 334 (35.1) 468 (57.7) 467 (27.8)Organic matter 881 (23.7) 898 (4.0) 916 (4.62) 912 (5.84)Acid detergent fibre 401 (12.1) 390 (18.0) 373 (12.0) 372 (15.7)Neutral detergent fibre 627 (24.9) 631 (12.7) 626 (24.5) 626 (32.1)Total N 23.2 (2.78) 25.2 (1.72) 21.7 (4.99) 23.7 (1.27)Insoluble N (g/kg total N) 411 (124) 378 (72.3) 396 (57.4) 418 (39.8)Ammonia N (g/kg total N) 201 (25.1) 208 (29.7) 171 (18.0) 186 (26.9)Lactic acid nd‡ 52.5 (10.1) 44.6 (12.1) 52.1 (11.5)Lactic acid (g/kg organic acid) nd 759 (36.0) 756 (54.5) 797 (49.5)Acetic acid 8.53 (1.26) 4.68 (0.78) 3.42 (0.68) 3.56 (0.53)Butyric acid 7.36 (0.92) 3.76 (0.66) 2.57 (0.09) 3.04 (0.18)Total acids nd 70.7 (14.39) 50.9 (16.01) 64.9 (13.52)pH 4.84 (0.17) 5.57 (0.39) 4.86 (0.15) 4.93 (0.18)

‡nd = not determined

Figure 2: Change in pH, insoluble N and organic acid concentration in control and pre-sliced round bales during ensiling.

solublisation. In silages where lactic acidwas measured, it accounted (proportionate-ly) for about 0.75 to 0.80 of total organicacids. Compared to Experiment 1, fermen-tation was judged to be more extensive, withhigher concentrations of organic acids. Flailsilage had higher concentrations of aceticand butyric acids, but the concentrations ofthese acids in other silages were similar.

In round-bale silages, the pattern of fer-mentation during ensiling was little affect-ed by pre-slicing (Figure 2). The pH inboth silages had reached 5.0 by 20 daysafter ensiling and insoluble-N concentra-tion had declined from 820 g/kg total N atensiling to just under 650 g/kg on day 20.Pre-slicing had a small effect on organicacid concentration over time, causing ahigher peak in organic acid concentration20 days after ensiling.

Animal performance In Experiment 1, voluntary intake and LWgain were high for all treatments (Table 5).Within the flail silages, steers consumedapproximately 20 g DM/kg LW and addi-tive use had no significant effect onintake. However, intake of round-balesilage was greater, averaging 25 g DM/kgLW (P<0.05). Greater intake of round-bale silage promoted higher LW gain, with steers gaining approximately 1.6 times faster than steers fed flail silage

(P<0.001). Efficiency of utilization wasalso greater for round-bale silage whenconsidered in terms of LW gain per unitintake (P<0.05).

In Experiment 2 (Table 6), DM digest-ibility of round-bale silage was higher thanthat of flail or precision-chopped silage(P<0.05), while gross energy digestibilitywas higher in round bale than precision-chopped silage (P<0.05). However, therewas no difference in digestibility of flailand precision-chopped silage, nor betweenthe two round-bale silages. As in Experi-ment 1, voluntary intake was higher forsteers offered round-bale silages com-pared to flail silage (P<0.05). Voluntaryintake of steers on precision-choppedsilage was intermediate between, and notdifferent to, those given flail silage orround-bale silages. Rates of LW gain were lower in Experiment 2 than inExperiment 1, and steers offered round-bale silages had a higher rate of gain than those on flail silage (P<0.01). TheLW gain of steers offered precision-chopped silage was similar to that of steersgiven flail silage but less than that of steerson round-bale silages (P<0.05). Slicing offorage immediately before baling had noeffect on any aspect of animal perform-ance. In contrast to Experiment 1, efficien-cy of utilization of feed DM for LW gainwas similar for all silages.


Table 5. Effect of ensiling system on animal performance – Experiment 1

Flail silage Round-bale s.e. Significance ofsilage contrasts

No additive With additive Flail v. bale

Dry matter (DM) intake (kg/day) 6.80 7.37 8.90 0.373 **DM intake [g/kg live weight (LW)] 19.7 21.1 24.7 1.34 *LW gain (kg/day) 1.00 0.97 1.61 0.079 ***Gain/intake (g LW/kg DM intake) 149 133 181 12.2 *NRC predicted LW gain (kg/day) 0.70 0.92 1.35 – –

There were no significant additive effects. NRC (1996) was used to predict live-weight gain based on observed dry matter intake, crude protein andneutral detergent fibre.

DiscussionFlail-harvested and round-bale silagesrepresent extreme silage types. The for-mer is directly harvested from the field,without wilting, whilst the latter, at least inCanadian conditions, is extensively wiltedto a DM concentration in excess of 400g/kg. Thus flail-harvested silage is normal-ly characterised by an extensive fermen-tation, while high DM round-bale silagehas a restricted fermentation. Both sys-tems offer cost and labour advantagesover precision-chopped silage for small-scale livestock producers. For example,costs of the equipment (based on pur-chase price at the time of the experi-ments) used for the round-bale and precision-chop systems in these experi-ments were 1.5 and 1.9 times greater thanfor the flail system, respectively. In spiteof the obvious need to compare these sys-tems, the authors found no direct compar-isons in the literature. This was attributedto the more recent advent of round-balesilage, subsequent to the demise in popu-larity of flail-harvesting systems.

Silage compositionThe chemical composition at harvest ofcrops in both experiments was similar, yetanimal performance in the 2 years wasquite different. In Experiment 1, the re-growth timothy was grown under dry con-ditions and the yield of harvested material(1.24 t/ha) was very low and the DM con-centration of the crop at cutting was relat-ively high (>200 g/kg). Based on earlierwork we have conducted in easternCanada, epiphytic lactic acid bacterianumbers would also have been low, prob-ably less that 103 colony forming units oflactic acid bacteria per g of fresh crop(Charmley and McQueen, 1991). Thesefactors, combined with minimal physicaldisruption to the crop during both flailharvesting and especially round balingresulted in silages with very restricted fer-mentations, although ammonia-N valueswere very high. For example, the pH offlail silage ensiled without the additive was over 5 and total acid concentrationwas below 30 g/kg DM. Typically, flail-harvested silage would be expected to


Table 6. Effect of ensiling method on animal performance – Experiment 2

Flail Precision-chop Round-bale silage (RB) s.e. Significance of contrastssilage silage Control Pre-sliced FL v. RB RB v. PC(FL) (PC) (URB) (SRB)

Digestibility coefficientsDry matter (DM) 0.660 0.650 0.740 0.727 0.027 * *Gross energy 0.664 0.623 0.721 0.705 0.030 *

DM intake (kg/day) 5.07 6.62 7.76 7.43 0.78 *DM intake [g/kg live

weight (LW)] 15.0 18.9 21.0 20.4 2.06 *LW gain (kg/day) 0.68 0.76 1.04 0.89 0.079 ** *LW gain (g/kg DM

intake) 148 117 135 121 25.8NRC predicted LW

gain (kg/day) 0.12 0.54 0.83 0.89 – – –

There were no significant FL v. PC or URB v. SRB contrasts. NRC (1996) was used to predict live-weight gain based on observed dry matter intake, gross energy digestibil-ity, crude protein and neutral detergent fibre.

have an extensive fermentation withorganic acid concentration in excess of100 g/kg DM (Gordon, 1986). In Experi-ment 2, the silage was made from primarygrowth timothy and growing conditionsand DM yield were more typical for east-ern Canada. The DM concentration of theflail-harvested crop was lower than inExperiment 1 and fermentation was moreextensive. Unfortunately, analytical prob-lems resulted in the loss of lactic acidanalyses for this treatment. However,based on the concentration of volatilefatty acids and the pH, it is apparent thattotal acid production would have been inthe range of 80 to 100 g/kg DM. Two daysof rainfall, immediately after the flail-harvested material had been ensiled,delayed the harvest of the wilted treat-ments resulting in longer wilting periodsthan in Experiment 1 (90 v. 22 h). Slowdrying rates have been associated withadverse silage fermentation and reducedintake response to wilting by cattle(Wright et al., 2000). The high ammonia-Nand butyric acid concentrations in thesewilted silages suggest that fermentationwas indeed adversely affected.

The low DM concentration, normallyassociated with flail-harvested silages canincrease the likelihood of a clostridial fer-mentation. In addition, the flail action atharvest can lead to soil contamination,and hence clostridial contamination. Thiswas likely in Experiment 1 where OM concentration was lower in flail silage.Consequently, in Experiment 1, an addi-tive specifically designed to inhibitclostridial bacteria was added at ensiling.Under the conditions of Experiment 1, theadditive had little effect on fermentationof the silage, probably because the higherthan expected DM concentration reducedthe activity of Clostridia in the untreatedsilage. This is contrary to European

results (Lingvall and Lättemäe, 1999)where the same levels of sodium nitriteand hexamethylene-tetramide improvedfermentation and reduced ammonia con-centration.

In Experiment 1, CP in all silages washighly soluble, but there was a tendencytoward reduced solubility in the additivetreated silage and particularly in the wilt-ed, round-bale silage. The addition of anadditive containing (proportionately) app-roximately 0.15 N should have increasedtotal-N concentration by 0.25 g/kg DM,with all of this additional N present asammonia N. However, total-N concentra-tion was little influenced and ammonia-Nconcentration actually declined, presum-ably due to alterations in forage-N trans-formations resulting from altered silagefermentation. In Experiment 2, proteinsolubility was similar in all treatments, butgenerally lower than in Experiment 1, pos-sibly because primary growth forage wasused (Tremblay et al., 2000). Both wiltingand additives have been shown to reduceprotein solubility in silages (Charmley,2001) and this effect was observed in Ex-periment 1. The absence of a wilting effecton protein solubility in Experiment 2 canbe attributed to the extended wilting peri-od which allowed for extensive plant pro-teolysis before ensiling (Charmley andVeira, 1991).

In both experiments ammonia-N con-centrations were generally high for allsilages and difficult to explain in light ofthe other fermentation characteristics ofthe silages. Typically, high ammonia-Nconcentration is attributed to deamin-ation by Clostridia. However, Entero-bacteria have also been implicated in theproduction of ammonia from nitrates(McDonald, Henderson and Heron,1991). Butyric acid concentration was relatively high in Experiment 2, suggest-ing some clostridial activity. Butyric acid


concentration was not high in Experiment1, but the dry growing conditions mayhave produced high nitrate concentrationin the crop and favoured the developmentof Enterobacteria over lactic acid bacteriain silage. This combination would havebeen conducive to ammonia productionin the silage.

Voluntary intake In both trials, voluntary DM intake wassome 2 kg/day higher for round-balesilages than flail-harvested silages. Whenthese silages were evaluated using theNRC Nutrient Requirements for BeefCattle (1996) model, estimated intakeswere close to those achieved by cattle fedthe round-bale silages (7 to 8 kg/day). Thissuggests that intake of flail silage waslower than the potential intake based onfibre content and digestibility. Flail-har-vested silage was ensiled without wiltingand at low DM concentration, whereasround-bale silages were extensively wiltedbefore ensiling. Most research has indicat-ed that voluntary intake of wilted silage ishigher than that of unwilted silage (Rohrand Thomas, 1984; Dawson et al., 1999;Wright et al., 2000). No comparisons couldbe found in the literature comparingunwilted, flail-harvested silage and wilted,round-bale silage. The closest comparisonwas one made by Petit et al. (1993). Theyensiled lower DM forage (300 g/kg) usinga self-loading forage wagon and higherDM forage (>500 g/kg) as round bales.Contrary to our work, Petit et al. (1993)found that intake of round-bale silage waslower than that of the silage harvestedusing a forage wagon. Within round-balesilages, voluntary intake normally increas-es with increasing DM concentration(Beaulieau et al., 1993; Dawson et al., 1999; Gordon et al., 1999). Similarly,within long-chopped silage harvested with a forage wagon, voluntary intake

increases with silage DM concentration(O’Kiely, 1992). Thus, we conclude that our results are consistent with expec-tations based on similar studies in the literature.

Several comparisons exist in the litera-ture between flail-harvested and wiltedprecision-chopped silages. Steen (1984)showed that intake was higher for finishingbeef cattle fed wilted precision-choppedsilage compared with flail-harvested silage.Similarly, Gordon (1986) found intake ofprecision-chopped silage by lactating dairycows was about 2 kg/day higher than thatof flail silage. In this study, we foundintake to be similar for precision-choppedand flail-harvested silage.

When comparing the intake of preci-sion-chopped and round-bale silages ofsimilar DM concentration there is no vol-untary intake response to round baling(Nicholson et al., 1991; Petit et al., 1993).However, when the DM concentration of the round-bale silage is higher, then a response in voluntary intake can beexpected (Beaulieu et al., 1993; Charmleyet al., 1999). In Experiment 2 of the pres-ent study, voluntary intake was numericallyhigher for the somewhat drier round-balesilage.

Wright et al. (2000), in a review of experi-ments comparing unwilted and wiltedsilages, concluded that the most importantdeterminants of intake response to wiltingwere the fermentation quality of theunwilted silage and the rate and extent ofwater loss during wilting of the wiltedsilage. In our Experiment 2, the rate ofwilting (44 g water loss/h per kg DM) wasmuch slower than in Experiment 1 (110 gwater loss/h per kg DM), yet the intakeresponse to wilting was similar in bothexperiments (~2 kg/day). This lack ofresponse to the better wilting conditionscan be explained by the fact that inExperiment 1, the unwilted silage was also


of high intake potential, owing to therestricted fermentation. Thus, the absoluteintake of wilted round-bale silage was con-siderably higher in Experiment 1 (~25 g/kgLW) than in Experiment 2 (~20 g/kg LW).

DigestibilityApparent digestibility was only measuredin Experiment 2. In spite of the poor con-ditions during wilting, digestibility washigher in round-bale silages than in flail orprecision-chopped silages. Generally, re-search has shown that increasing the DMconcentration of silage by wilting reducesdigestibility among precision-choppedsilages (Charmley and Thomas, 1987,1989; Yan et al., 1996). Furthermore,Gordon (1986) found the same resultwhen comparing unwilted, flail-harvestedsilage and wilted precision-chopped silage. However, there is some evidencethat digestibility of round-bale silageincreases with DM concentration (Beaulieuet al., 1993) and may be higher than forprecision-chopped silages of similar DMconcentration (Nicholson, Charmley andBush, 1992). Larger particle size couldincrease residency time in the rumen andso increase digestibility relative to veryfinely chopped silages. Charmley et al.(1999) found that intensive conditioningat cutting, causing particle size reduction,reduced digestibility of round-bale silage.When the effect of particle size was exam-ined in Experiment 2, there was no effectof pre-chopping before baling on digest-ibility. However, the difference in particlesize between the two round-bale treat-ments was minimal so an effect ondigestibility was unlikely.

Weight gain We found that LW gain was better forsteers fed round-bale silage than for thosefed flail-harvested silage. The magnitudeof the response was exceptional but the

possibility of compensatory gain wasunlikely since steers had been weaned for at least 3 months and adapted to a pre-cision-chopped silage diet offered toappetite. No direct comparison of flail andround bale silages was found in the litera-ture. However, Petit et al. (1993) conclud-ed that milk production was higher incows offered high DM round-bale silagecompared with lower DM silage coarselychopped in a forage wagon. Live-weightgain was not different between steersoffered flail or precision-chopped silage.Similarly, Gordon (1986, 1989) has sug-gested that output of milk or LW gainfrom flail-harvested, unwilted silage waseither equal to or greater than that fromwilted, precision-chop systems.

Live-weight gain from high DM (>400g/kg) round-bale silage has been found tobe lower (Nicholson et al., 1991), equal to(Charmley et al., 1999) or higher (Beaulieauet al., 1993) than from precision-choppedsilages. However, when the present dataare included, the balance of evidence sug-gests that round-bale silages with a DMconcentration >400 g/kg elicit greaterrates of gain in cattle than precision-chopped silages of moderately lower DMconcentration. A similar conclusion wasreached by Petit et al. (1993) for milk pro-duction in dairy cows.

In general, the LW response to increas-ing DM concentration of silage on LWgain is positive (O’Kiely et al., 1988;Gordon et al., 1999), but when the con-founding effect of harvesting system, inparticular particle size, is introduced, theresponse becomes less clear (Steen, 1985).Some workers have concluded that parti-cle size per se, has little influence on per-formance (Gordon, 1982) while othershave suggested that improved perform-ance results from more finely choppedmaterial (Marsh, 1978). The presentresults from unwilted coarsely-chopped


silage, moderately-wilted, finely-choppedsilage and heavily-wilted unchopped silagesuggest that regardless of particle size,there is a positive relationship betweensilage DM concentration and LW gain.

Utilization of silages Predicted rates of gain based on NRC(1996) and using the actual DM intakesachieved in the experiments showed thatefficiency of DM utilization was higherthan predicted by NRC. This has beennoted before for high-silage diets(Duynisveld and Charmley, 2001) becausethe model underestimates the net energyvalue of forages. However, it was alsoapparent that this discrepancy was greaterin the low-DM flail silages than in thehigher-DM silages. This supports otherresearch suggesting that flail silages areutilized with greater efficiency than wiltedsilages when output is measured as LWgain (Steen, 1985) or milk production(Gordon, 1982).

Debate continues concerning the uti-lization of low-DM unwilted, and high-DM wilted silages, regardless of methodof ensiling. Following earlier studies(Rohr and Thomas, 1984) the consensuswas that higher-DM, wilted silages wereused less efficiently than unwilted silages,largely due to their lower digestibility.When efficiency is measured as LW gainper unit DM intake, this does not accountfor possible differences in digestibility ofsilages or gut fill and body composition ofanimals which can be influenced by treat-ment (Steen, 1984). Charmley and Thomas(1987) showed that efficiency of utiliza-tion of digestible energy was unaffected by wilting and apparent differences infeed/gain ratio could be accounted for by differences in composition of the gain.More recent research involving calorime-try now supports this conclusion (Gordonet al., 1999). While these comparisons

were made among precision-choppedsilages, relatively few papers report oncontrasting ensiling techniques that resultin silages with DM concentrations andnone compare the three systems in Ex-periment 2. However, Petit et al. (1993)compared round-bale silage with lower-DM fine- or coarse-chopped silage. Effi-ciency for milk production was higher forthe round-bale silage than for the coarse-chopped silage, but was not different fromthe fine-chopped silage. This was attrib-uted to a shift in the partitioning of nutri-ents away from body reserves and towardmilk production in cows fed the round-bale silage. Wright et al. (2000) showedthe negative effects on DM intake of pro-longed, wet wilting conditions. It appearsthat poor wilting conditions in Experiment2 may have also eliminated the responsein efficiency of DM utilization for LWgain which was seen in Experiment 1. Thiswas not due to reduced digestibility, whichwas actually higher in round-bale silages,but presumably due to poorer utilizationof digestible energy or a higher energycontent of LW gain (Charmley andThomas, 1987).

ConclusionsThe three systems all yielded silages ofbroadly similar fermentation characteris-tics. Generally voluntary intake and per-formance was satisfactory, regardless ofthe system employed. Potential benefitsassociated with the flail system includelower equipment costs and elimination ofa field wilting period but this ensilage sys-tem resulted in the lowest intakes. How-ever, LW gain and efficiency of utilizationof DM for LW gain were either better thanor similar to that achieved with wilted pre-cision-chopped silage. There was no bene-fit from the use of an additive in flailsilage. The round-bale silage system is


intermediate in equipment costs betweenthe flail and precision-chop systems butvoluntary intake, digestibility and LW gainwere all better for round-bale silage.However, pre-chopping forage at balingdid not influence intake or animal per-formance. It is concluded that animal per-formance is not necessarily directly relatedto the cost and complexity of the silage-making system. Choice of system shouldtherefore be based on the scale and natureof the enterprise and not on expected dif-ferences in animal performance.

AcknowledgementsWe thank T. Gowan, J. Smith, K. Milner, B. Truemanand their support staff for their assistance in the conduct of the research. This research was partiallyfunded by a grant from the Nova Scotia Departmentof Agriculture and Marketing. We also thank FSLBells, UK, for donation of the silage additive.

ReferencesAOAC (Association of Official Analytical Chemists).

1984. ‘Official Methods of Analysis’ (15th edi-tion), AOAC, Arlington, VA, USA.

Barker S.B. and Summerson W.H. 1941. The colori-metric determination of lactic acid in biologicalmaterial. Journal of Biological Chemistry 137:535–554.

Beaulieau, R., Seoane, J.R., Savoie, P., Tremblay, D.,Tremblay, G.F. and Thériault, R. 1993. Effects ofdry-matter content on the nutritive value of indi-vidually wrapped round-bale timothy silage fed tosheep. Canadian Journal of Animal Science 73:343–354.

CCAC (Canadian Council on Animal Care). 1993.‘Guide to the Care and Use of ExperimentalAnimals’, Volume 1, Ottawa, Canada.

Charmley E. 2001. Towards improved silage quality– A review. Canadian Journal of Animal Science81: 157–168.

Charmley, E. and McQueen, R.E. 1991. Factorsinfluencing lactic acid bacteria numbers at har-vest in New Brunswick. Canadian Journal ofAnimal Science 71: 1292–1293.

Charmley E., Savoie P. and McQueen R.E. 1997.Influence of maceration at cutting on lactic acidbacteria populations, silage fermentation and voluntary intake and digestibility of precision-chopped lucerne silage. Grass and Forage Science52: 110–121.

Charmley, E., Savoie, P, McRae, K.B. and Lu, X.1999. Effect of maceration at mowing on silageconservation, voluntary intake, digestibility andgrowth rate of steers fed precision chopped orround bale silages. Canadian Journal of AnimalScience 79: 195–202.

Charmley, E. and Thomas, C. 1987. Wilting ofherbage prior to ensiling: effects on conservationlosses, silage fermentation and growth of beefcattle. Animal Production 45: 191–203.

Charmley, E. and Thomas, C. 1989. Effect of dura-tion of wilting on the conservation of silage andon gains in body components by steers. AnimalProduction 48: 91–98.

Charmley, E. and Veira, D.M. 1991. The effect ofheat-treatment and gamma radiation on the com-position of unwilted and wilted lucerne silages.Grass and Forage Science 46: 381–390.

Dawson, L.E.R., Ferris, C.P., Steen, R.W.J., Gordon,F.J. and Kilpatrick, D.J. 1999. The effects of wilt-ing grass before ensiling on silage intake. Grassand Forage Science 54: 237–247.

Dewar W.A. and McDonald, P. 1961. Determinationof dry matter in silage by distillation with toluene.Journal of the Science of Food and Agriculture 12:790–795.

Duynisveld, J.L. and Charmley, E. 2001. An evalua-tion of models predicting beef cattle performanceusing data from bulls on performance testing.Canadian Journal of Animal Science 81: 581–584.

Gordon, F.J. 1982. The effects of degree of choppinggrass for silage and method of concentrate allo-cation on the performance of dairy cows. Grassand Forage Science 37: 59–65.

Gordon, F.J. 1986. The effect of system of silage har-vesting and feeding on milk production. Grassand Forage Science 41: 209–219.

Gordon, F.J. 1989. Effect of silage additives and wilt-ing on animal performance. In: ‘Recent Advancesin Animal Nutrition’ (ed. W. Haresign and D.J.A.Cole), Butterworths, London, pages 159–173.

Gordon, F.J., Dawson, L.E.R., Ferris, C.P., Steen,R.W.J. and Kilpatrick, D.J. 1999. The influence ofwilting and forage additive type on the energy uti-lization of grass silage by growing cattle. AnimalFeed Science and Technology 79: 15–27.

Haigh, P.M. 1990. The effect of dry matter contenton the preservation of big bale grass silages madeduring the autumn on commercial farms in SouthWales 1983–1987. Grass and Forage Science 45:29–34.

Lingvall, P. and Lättemäe, P. 1999. Influence of hexa-mine and sodium nitrite in combination withsodium benzoate and sodium propionate on fer-mentation and hygienic quality of wilted and long


cut grass silage. Journal of the Science of Food andAgriculture 79: 257–264.

Marsh, R. 1978. A review of the effects of mechan-ical treatment of forages on fermentation in thesilo and on the feeding value of the silages. NewZealand Journal of Experimental Agriculture 6:271–278.

McDonald, P., Henderson, A.R. and Heron, S.J.E.1991. ‘The Biochemistry of Silage’ (2nd edition),Chalcombe Publications, Marlow, UK.

NRC (National Research Council). 1996. ‘NutrientRequirements of Beef Cattle’, National AcademyPress, Washington, DC, USA.

Nicholson, J.W.G., Charmley, E. and Bush, R.S.1992. Effect of moisture level on ensiling charac-teristics of alfalfa in big bales or chopped andcompacted in plastic tubes. Canadian Journal ofAnimal Science 72: 347–357.

Nicholson, J.W.G., McQueen, R.E., Charmley, E.and Bush, R.S. 1991. Forage conservation inround bales or silage bags: effect on ensiling char-acteristics and animal performance. CanadianJournal of Animal Science 71: 1167–1180.

O’Kiely, P. 1992. A note on the performance ofFriesian steers offered unwilted or wilted grasssilage diets from weaning through to slaughter.Irish Journal of Agricultural and Food Research 31:71–75.

O’Kiely, P., Flynn, A.V. and Wilson, R.K. 1988. Acomparison of the chemical composition ofunwilted and wilted grass silage and of the intake,performance, carcass composition and rumenfluid fatty acid concentrations of steers fed thesilages. Irish Journal of Agricultural Research 27:39–50.

Petit, H.V., Tremblay, G.F., Savoie, P., Tremblay, D.and Wauthy, J.M. 1993. Milk yield, intake andblood traits of lactating cows fed grass silage con-served under different harvesting methods.Journal of Dairy Science 76: 1365–1374.

Rohr, K. and Thomas, C. 1984. Intake, digestibilityand animal performance. In: ‘Efficiency of SilageSystems: A Comparison Between Unwilted and

Wilted Silages’ (ed. E. Zimmer and R.J. Wilkins),Landbauforschung Volkenrode, Sonderheft 69:64–70.

SAS. 1995. SAS/STAT User’s Guide, Version 6 (4th

edition), Vol. 2, SAS Institute Inc., Cary, NC, USA. Siddons, R.C., Evans, R.T. and Beever, D.E. 1979.

The effect of formaldehyde treatment beforeensiling on the digestion of wilted grass silage bysheep. British Journal of Nutrition 42: 535–545.

Steen, R.W.J. 1984. The effects of wilting andmechanical treatments of grass prior to ensilingon the performance of beef cattle and beef out-put per hectare. Fifty-Seventh Annual Report ofthe Agricultural Research Institute of NorthernIreland, pages 21–32.

Steen, R.W.J. 1985. The effect of field wilting andmechanical treatment on the feeding value ofgrass silage for beef cattle and on beef output perhectare. Animal Production 41: 281–291.

Tremblay, G.F., Michaud, R., Bélanger, G., McRae,K.B. and Petit, H.V. 2000. In vitro ruminalundegradable proteins of alfalfa cultivars.Canadian Journal of Plant Science 80: 315–325.

Van Soest, P.J., Robertson J.B. and Lewis B.A. 1991.Methods for dietary fiber, neutral detergentfiber, and non-starch polysaccharides in relationto animal nutrition. Journal of Dairy Science 74:3583–3597.

Wilkinson, J.M., Wadephul, F. and Hill, J. 1996.‘Silage in Europe: A Survey of 33 Countries’,Chalcombe Publications, Lincoln, UK.

Wright, D.A., Gordon, F.J., Steen, R.W.J. andKilpatrick, D.C. 2000. Factors influencing theresponse in intake of silage and animal perform-ance after wilting of grass before ensiling: areview. Grass and Forage Science 55: 1–13.

Yan, T., Patterson, D.C., Gordon, F.J. and Porter,M.G. 1996. The effects of wilting of grass prior toensiling on the response to bacterial inoculation.1. Silage fermentation and nutrient utilizationover three harvests. Animal Science 62: 405–417.

Received 7 January 2004


comparison of flail-harvested, precision- chopped and...

Documents