The effect of dietary fat and metabolizable energy ...

The effect of dietary fat and metabolizable energy supply on milk protein concentration of dairy cows J. M. Moorby't, R. J. Dewhurst't, C. Thomas' and S. Marsden2 'Grasslarzd uild Rzin~~irnrzf Sczetlce Department, Scottrsh A~rzculturalCollege, Auchlncrul?~e,Ayr KA6 5HW 'Dnlgefy Agrrc ulture Ltd, 780 Aztec Wesf,Almoudsbury, Brzstol BSI2 4TH

Abstract To inzlesfigate the e$fecf of dietary fat and metabolizable energy ( M E ) on milk protein concentration, an experiment was carried ouf using 12 multiparous early-lacfation Holstein-Friesian dairy cozus. Three diets were offered in a complete Latin-square change-over design, based on ad libitum access to grass silage. One of three concentrates was offered at a rate of 12 kglday, each formulated to supply one of two levels o f ME (12.1 and 13.6 MJlkg dry matter ( D M ) ) atid one of two levels offat (31 and a mean of 88 g acid hydrolysis ether extract per kg DM): low energy, high fat (LEHF); low energy, low fat (LELF); and high energy, high fat (HEHF). The concentration of milk protein was sigtzificanfly higher from animals offered the LELF concentrate (32.5 v. a mean of 31.2 (s.e.d. 0.45) glkg, P < 0.05), because of lower milk yields (31.0 v. a mean of33.4 (s.e.d. 0.63) kglday, P < 0.05). Animals offered the HEHF concentrate produced the highest yields of milk protein but their milk had the lowest concentrations of fat (32.5,34.4 and 31.9 glkgfor LEHF, LELF and HEHF respecfively; s.e.d 1.07; P < 0'05for difference between LELF and HEHF). Silage DM intake was significantly increased by animals offered the LEHF concentrate (9.1, 8.6 and 8.7 (s.e.d. 0.19) kglday, P < 0.05for differences between LEHF and the other two concentrates). Urinary purine derivative excretion, used as an index of microbial protein supply, was highest from animals offered the LELF and HEHF concentrates, zuhich both supplied similar amounts offermentable ME. It is hypothesized that increased d e novo synthesis offatty acids on the low fat diet reduced the availability of glucose for lactose synthesis, leading to reduced milk yields and hence increased milk protein concentrations. Keywords: dairy cows, dietaryfat, metabolizable energy, milk production, milk protein.

Introduction In early lactation, many dairy cows lose condition as energy output in milk exceeds dietary energy intake. With the need to increase the cost efficiency of milk production, fat may be used as a relatively cheap energy source for incorporation into dairy cow lactation rations. However, milk protein concentration has become an ~mportant consideration and although an increase in fat consumption by the dairy cow tends to increase milk yields, it also tends to decrease milk protein concentration (Del'eters and Cant, 1992).

t Present address: Animal Science and Microbiology Department, Institute of Grassland and Environmental Research, Plas Gogerddan, Aberystwyth SY23 3ER.

At high levels of incorporation fat can adversely affect fibre fermentation in the rumen (Coppock and Wilks, 1991). Saponification of oils with calcium or the use of whole oil seeds can reduce this problem by partially 'by-passing' the rumen. In this respect, rumen acetate concentration was increased by increasing the level of rumen protection of supplemental fat (Jenkins and Jenny, 1992) and this was associated with a concomitant-increase in milk yield. At the same time, however, milk protein concentration was seen to decrease slightly (Jenkins and Jenny, 1992). Other workers have found a similar effect: increasing fat supplementation resulted in increased milk yields but a decrease in protein concentration (Drackley and Elliot, 1993). Casper and Schingoethe (1989) proposed that this phenomenon is mediated by a reduction in growth hormone release and an indirect reduction of amino acid

2

Moorby, Dewhurst, Thomas and Marsden

uptake by the mammary gland. Cant et ul (1993), on the other hand, postulated that additional dietary fat decreases mammary blood flow, thereby reducing thc dellvery of nutrients to the mammary gland. Iiecent attempts to reduce the effect of dletary fat on m ~ l kprotun concentrat~onby, for example, the use ot protc~n-tat'hvpass' supplements (Holter et N I , 1993) or high levels of undegradable proteln (Palmqulst et ul 1993) have met wlth only l ~ m ~ t esuccess. d Tntramamlnary nutrlent partitioning, part~cularl y of energv-v~eld~ngsubstrates such as glucose and acetate, can affect the quantlty and quality of m ~ l k produced Fatty aclds absorbed from the d ~ e can t be ~ncorporatedmto mllk fat unchanged (Banks et u l , 1980)-and an increase in the output of long-chain fatty acids in milk is associated with a decrease in the dr ? Z C J ~ Y Jsynthesis of short-chain fatty acids (Faulkner and Pollock, 1989). The extent to which this happens is indicated by the concentration of citric acid in milk (Faulkner and Pollock, 1989), allowing the effects of diet on mammary fatty acid synthesis to be examined and an increased understanding to be gained of the intramammary processes involved in milk production from animals on specific diets. This experiment was designed to investigate the effect of offering two levels of dietary fat at two concentrations of fermentable metabolizable energy (FME) on milk production and composition in dairy cows. Urinary purine derivative excretion was used as an index of microbial protein supply and milk citric acid concentration as an index of mammary fatty acid synthesis.

Within squares, the three treatments were allocated at random to each animal. The data obtained were analysed statistically using analysis of variance with C,LNSI.AI 5 (Lawes Agricultural Trust, 1990). A blocking structure of period X (squarelcow) and a treatment structure of experimental diet were used. For the analysis of urinary purine derivative excretion data, a treatment structure of diet X day X time was used. Because of the non-orthogonal nature of the treatment structure, differences between diets were assessed using a t test. Dietforn~ulationand production The experimental diets ~ 7 e r ebased on ad libitum access to first-cut grass silage. This was supplemented with concentrates offered at a flat rate of 12.0 kg/day which were formulated to provide between them two levels of acid hydrolysis ether extract (AHEE) and two levels of metabolizable energy (ME). Dry-matter (DM) content, crude protein (CP) content, protein degradability and the ratio of starch to digestible crude fibre (a component of the food compounder's formulation matrix) were all formulated to be similar across the three concentrates. The composition of each of the three concentrates, low energy, high fat (LEHF), low energy, low fat (LELF) and high energy, high fat (HEHF) is given in Table 1. The lower ME level was intended to be moderate in terms of the requirements for a dairy cow in post-peak lactation. The concentrates consisted of a relatively low quality carbohydrate energy source plus added fat (mainly

Material and methods Aninlals und their m a r ~ u g e n ~ e n f Twelve multiparous Holstein-Friesian cows, at weeks 8 to 10 of lactation at the start of the experiment, were drawn from the Scottish Agricultural College Auchincruive herd. They were housed in a metabolism unit in individual stalls fitted with de Boer yokes and were milked i n situ using a vacuum line and bucket units at about 05.30 h and 15.30 h. Milk yields were recorded at each milking by weighing. Experirnerztal desigrz The experiment was a complete change-over design based on four 3 X 3 Latin squares. Each experimental period was divided into adaptation and collection periods of 3 weeks and 1 week in length respectively. The mean milk yields from the 7 days prior to the 'xperiment were used to allocate animals to Latin squares, with the three lowest yielding cows assigned to square one, the next three to square two and so on to the three highest yielders in square four.

Table 1 Sumniary of the, lngredlrnt rorrlpos~tlorlof the three cxpcruncntal concentrates (glkljfuesh zi~el~lrt) Concentrate LEHF Barley Wheat Wheatteed Rice bran Molassed sugar-beet pulp High protein maize gluten meal 00-Rapeseed meal Sunflower seed meal Field beans Toasted soya hulls High protein soya-bean meal Fat Palm oil Molasses Minerals and vitamins

LELF HEHF

3

Milk protein and dietary fat of dairy cows palm 011) (LEHF), a hlgher quahty carbohydrate energy source w ~ t hvery little added fat (LELF) or the LELF energ) sources plus the LEHF added fat (HEFIF) Therefore, 111 addltlon to two contrast~ng energy levels and two contrasting rates of fat inclusron, the three concentrates offered two contrasting FME levels The log~cal fourth concentrate that would have allowed a 2 X 2 factorial invest~gat~on would have been one conta~n~ng h~gh energy and low fat This was not possible, however, because an Increase In ME beyond that achieved on n the LELF concentrate was not achievable w ~ t h ~the bounds of the other formulation cr~ter~a

A~zirnulfeeding Cows were offered fresh grass silage ad libiturn daily at approximately 09.30 h. This was done by offering proportionately about 0.1 more silage than the previous day's intake and by topping up individual animal's food bins during the day if necessary. The concentrate part of the diet was offered in two equal portions of 6.0 kg at each milking. The silage offered during period 1 was a first-cut grass silage prepared with a formic acid silage additive (Add-F, BP Nutrition (UK) Ltd. Northwich, Cheshire). A new clamp was opened at the end of period l and a first-cut grass silage that was prepared using a bacterial inoculant (EcoSyl, ICI Bio Products, Billingham, Cleveland) was offered during periods 2 and 3.

Sample collection and analysls During the last 10 days of each experimental period, silage intake was measured by weighing out the silage offered and weighing back the refusals the follow~ngmorning. Small samples (approx. 200 g) of the silage offered were collected daily and bulked for analysis. Similarly, approximately 200 g of silage refusals were collected from each animal and bulked over the 10 days for DM analysis. Silage samples were frozen immediately after collection and stored at -20°C until analysed. Concentrate samples were bulked over each experimental period and stored at -20°C until analysed. Spot urine samples were taken by vulva1 stimulation at about 10.30h and 14.30h on each of 2 days consecutively at the end of each collection period. The samples were immediately diluted 1 in 20 (75 p1 urine in 1.5ml) with O.lmol/l ammonium dihydrogen orthophosphate solution to avoid the precipitation of uric acid from undiluted urine when frozen and thawed. The diluent also contained 0.1 mol /l allopurinol as an internal standard. The samples were either analysed immediately for creatinine and the purine derivatives uric acid and

allantoin, or were frozen upright and stored at -20°C until analysed. Sample5 of faeces were collected from each cow at the same t~mesas unne samples, tak~ngcare to avoid contamlnatlon The samples were fro7en Immediately and were stored at -20°C unt11 being drled at 60°C for storage before later analvs15 tor lndlgestible acld-detergent fibre content (Penning and Johnson, 1983) Milk yields were recorded daily throughout the experiment. Milk samples were taken at four consecutive milkings, starting with an afternoon milking, and were preserved using Lactab milk preservative tablets (Thompson and Capper Ltd, Runcorn, Cheshire, UK) and storage at 4'C. Samples from the first two milkings of the four were also taken for analysis of milk CP, casein, non-protein nitrogen and urea; subsamples of this were frozen and stored at -20°C for later analysis of milk minerals. Methods of analyses of food, faeces, urine and milk were carried out as described by Moorby ' t al. (1996).

Results The mean composition of the silages offered during the experiment is presented in Table 2 and the composition of the concentrates offered in Table 3. Silage DM changed between experimental periods

Table 2 Composrfrorr of d a x c offered throughout the ~rperz~ncvlt (inearz of thrce qamples, c a ~ hhulked over 10 days, ~rn1uc.iI I I glkg d r y matter ( D M i uule5s otherzozqrz stated) Mean Dry matter (glkg) Organic matter Crude protein Metabolizable energy (MJ/ kg DM) Rumen degradable protein Undegradable protein Neutral-detergent fibre Acid-detergent fibre Water-soluble carbohydrates Ether extract Acid hydrolysis ether extract In vitvo organic matter digestibility ( g / k g OM) D-value NH,-N (g/ kg total N ) pH Calcium Phosphorus Magnesium Potassium Sodium

254 932 182 11.5 155 27 435 261 40 36 56 784 718 89 3.7 6.2 3.4

2.5 20.8 4.0

s

d

49-6 0.6 9.5 0.06 8-4 1.2 14-2 9.5 24.3 6.4 6.3 4.9 3-8 30.9 0.06 1.40 0.67 0.60 2.71 1.29

4

Moorby, Dewhurst, Thomas and Marsden

Table 3 Composit~onof the experlrnental conceiztrate portzons of the dlet (values In glkg dry rnatter ( D M ) unlt.55 othfru~zscstated) Concentrate LEHF LELF HEHF Dry matter (g/kg) Organic matter Crude protein Metabolizable energy (E3) (MJ/kg DM) Fermentable metabolizable energyt (MJ 1kg DM) Ether extract Acid hydrolysis ether extract Starch Water-soluble carbohydrates Acid detergent fibre In vitro organic matter digestibility (g/ kg OM) Calcium Phosphorus Magnesium Potassium Sodium

858 864 191

857 918 185

860 913 182

12.1

12.1

13.6

9.1 84.5 92.0 128 105 128

11.1 18.7 30.6 238 97.4 139

10.8 77.8 84.2 240 93.4 128

698 16.7 9.6 4.6 17.1 4.3

811 10.5 6.1 3.4 14.4 3.2

779 14.3 7.0 3.5 12.7 3.1

t Estimated: FME = ME - 0.033 X AHEE.

(202, 259 and 301 g DM per kg for periods 1, 2 and 3 respectively), although the analysis of the DM was relatively constant (e.g. 171, 187 and 188 g CP per kg DM, and predicted energy contents of 11.4, 11.5 and 11.5 MJ ME per kg DM). No differences in concentrate composition were seen between samples from the different experimental periods, which is as

expected since concentrates were produced in a single batch. The mean daily intakes of silage DM, total food DM, CP, ME, estimated FME and acid hydrolysis ether extract (AHEE) are given in Table 4. Food FME concentration was estimated from the ME and AHEE contents of the concentrate (concentrate FME = ME 0.033 X AHEE) and silagc ME (silage FME = 0.71 X ME), assuming additivity (Agricultural and Food Research Council (AFRC, 1992)). Because of differences in the FME densities of the concentrate portions of the diet, the ratios of effective rumen degradable protein (ERDP) to FME of the diets consumed differed between treatments (14.3, 12.5 and 12.6 g ERDP per MJ FME). -

There was a significant increase in the DM intake of animals offered the LEHF concentrate due to an increase in the silage intake. Whole tract apparent digestibility of dietary organic matter (Table 4) was calculated from the change in concentration of indigestible acid-detergent fibre between food and faeces; diet digestibility was significantly lower in animals offered the LEHF concentrate (P < 0.001), although numerically the difference was small. Milk production and composition were significantly affected by dietary treatment (Table 5). The concentrations of milk solids were significantly higher from animals offered the LELF concentrate. Milk yields, however, were also lowest from animals offered this concentrate. No significant differences due to treatment were seen in the ratios of protein/ fat, proteinllactose or fat/lactose. Similarly, there

Table 4 Effect ofconcentrate treatment on mean dally zntake of dry matter ( D M ) and of crude protein (CP), tnefabolizabl~rnergy ( M E ) and aczd hydrolysis ether extract (AHEE), and estzmated itztukes of effective rumen degradable protezn (ERDP), dzgestlblt undegradcd protezn (DUP), andfermentable metabolizable energy (FME) (AFRC, 1992) (whole t r a ~ apparent t orgnnzc nzattei (OM) dlgcstzbil~tyand urrnary purlne derivatzue excretion expressed ln relatlon to urznary creatlnzne ~on~errtratlon) Concentrate Sigmficancet LEHF (1)

LELF (2)

HbHF (3)

9.1 19.4 3.6 2.4 0.62 229 168 1.5 0.79 3.24

8.6 18.9 3.5 2.3 0.48 223 184 0.8 0.81 3.51

8.7 19.0 3.5 2.3 0.48 240 182 l .4 0.82 3.51

S

ed

1-2

13

2-3

p

Silage DM (kg/ day) Total DM (kg/ day) CP (&/day) ERDP (kg/ day) DUP (kglday) ME (MJ/ day) FME (M1 / day) AHEE (kg/ day) Whole tract apparent digestibility of OM (gig) Allantoin + uric acid / creatinine (mol 1mol)

0.19 0.19

0.003 0.104

* S

*"" *

X**

t Significance of difference of effects between concentrate treatments; 1-2 signifies difference between treatments LEHF and LELF, etc.

Milk protein and dietary fat of dairy cows Table 5 Effect of dzetary treatme~zton mrlk yield and composition, and on yzelds of mllk components Concentrate Agmficancet LEHF (1)

LELF (2)

HEHF (3)

s

ed

1-2

1-3

2-3

Milk yield (kg/ day) Crude protein (g/ kg) True protein (g / kg) Casein (g / kg) Whey$ (gikg) Non-urea NPN (g/ kg) Urea (glkg) Fat (g/&) Lactose (g/ kg) Citric acid (g/ kg) Crude protein yield (giday) True protein yield (glday) Casein yield (glday) Whey yield (glday) Non-urea NPN y~eld(g / day) Urea yield (g/ day) Fat yield (glday) Lactose (glday) Citric acid yield @/day)

t Significance of difference of effects between concentrate treatments; 1-2 signifies difference between treatments LEHF and LELF, etc. $ Whey calculated as true protein - casein.

was no effect of diet on casein as a proportion of true protein, with a grand mean of 0.82. The concentration and yield of citric acid in milk (Table 5) increased significantly in response to additional dietary fat (LEHF v. LELF concentrates). Milk concentrations of potassium, sodium, chlorine, calcium and phosphorus are presented in Table 6. Despite significant differences between the effects of diets LELF and HEHF on K and Na concentrations, the ratio of K to Na in milk was not affected by Table 6 Eflect of dietary treatment concentrations (values in @kg)

On

mean milk mineral

Concentrate Significancet LEHF LELF HEHF (1) (2) (3) s.e.d. 1-2 1-3 2-3 Sodium Potassium Chlorine Calcium Phosphorus

0.40 1.64 0.94 1.12 0.94

0.39 1.61 0.96 1.13 1.00

0.41 1.68 0.98 1.11 0.93

0.011 0.031 0.031 0.025 0,021

* * *

*

t Significance of difference of effects between concentrate treatments; 1-2 signifies difference between treatments LEHF and LELF, etc.

dietary treatment (means 4.2, 4.2, and 4.1 (s.e.d. 0.11) g/g, for diets LEHF, LELF, and HEHF respectively). Likewise, the lactose/Cl ratios were not significantly affected by dietary treatment (means 51.7, 51.4 and 49.2 (s.e.d. 1.86) g/g, despite significant dietary effects on milk lactose concentrations for all three diets. The rate of excretion of the purine derivatives allantoin and uric acid, expressed in spot samples of urine as a ratio to the creatinine concentration (Table 4). was similar for animals offered the LELF and HEHF concentrates. These diets had similar dietary FME contents; animals offeredthe LEHF concentrate, which were supplied with approximately 15 MJ FME per day less than animals offered the other concentrates, had significantly lower rates of purine derivative excretion. The excretion of both allantoin and uric acid differed with time and day of sampling (Table 7), although the changes were in opposite directions on the 2 days of collection such that the combined purine derivative excretion data was not significantly affected by day of sampling. The gross efficiency of dietary protein utilization for milk protein production (i.e. milk protein output/ CP intake; Table 8) was significantly different between the high and low ME concentrate diets. More

Moorby, Dewhurst, Thomas and Marsden

6

:\!C'

2.76

I!C, ,\Cl/ C \'L1

'3.4 l 0.40

0-27 3.00 I 1-9

0.067

7.96 0-47 3.38 9.8

(1.03l 0.08 1 (i.h4

'3.81 II

iii f i ~ i111i1h Table 8 C;I.IIV i , f t i ~ i e ' i ~ i - ~[ i~~,c~/ i i r i//i r i ~ / ~ , II~I/I:I~/I~I~I p r i ~ t ~/I~.~I(/LI~ iii ~IOII f11iilk ~III/~,III ( ! i ~ t / ~ i ~ / / i ~[ i~~ri~i t~i ~/ iii/iil,~,, i, , i i i ;;l;:) C'otice~itr Signitic,incet I.CIik (l) C'r~~de protein .l-l.lle pl.otciri C,lscin

I.EI.I; (7)

I IEtiF; (3) s.c.11.

0.271 0.280

0-2')1

0.257

0-27.7

0.0009

0-211 0.21h 0.223

0.0037

0.263

0.0040

1-2

1-3

**X

2-3 X

**

**

t S i ~ i i i t i w n c co f d~ttescnceot effc.cts h c t ~ v e e nconcentrate treatuients; 1-2 s i g l ~ r f i e s ciifferencc, I.EIIF '111dLELF. ctc.

between t r ~ ~ t t r n e n t s

m ~ l kproteln wa5 produced per umt d~etaryCP consumed at the hlgher dens~tyof concentrate ME, particularly \$hen cornparlng the two hlgh fat concentrates

Discussion Three concentrates were offered wh~cliallowed the cornparlwn ot three comb~nat~ons of factors (a) the effect ot fat content at s~mllarME dens~ties(LELF 11 LEF-IF), (b) the effect ot ME dens~tyat s ~ m ~ l afat r content5 (LEFIF o HEHF), and (c) the effect of ME dens~tyat sim~larFME dens~ties(LELF v HEHF) The difference5 In concentrate fat and ME In (a) and (b) respect~velpwere associated wlth a d~fferenceIn potential FME, and the d~fferenceIn ME In (c) was mediated bv a d~fferencein fat content All three concentrate5 had s~mllarCP contents, and were formulated to conta~n s ~ m ~ l arat105 r of starch to dlge5tlble crude flbre 111 practice, the concentrates conta~neds ~ m ~ l levels ar of acld-detergent f ~ b r e(and mater-soluble carbohydrates) but d~tferedIn thelr contents ot starch, meaning that the mqor dltference5 In ME supply fol each concentrate were effected bv thelr starch and fat contents

3 11 0.24 3.47

I4.il

0.067 0.03 1 0.08 1 0-h4

?*X

S**

*X*

X**

**X

***

Although the compos~t~on of the three concentrates d~ffered quite markedly, as formulated, the consumption of 51lage was allowed to vary freely S~lageIntake was hlghest In an~malsoftered the LEHF concentrate, possibly as a result of its lower starch content (Thomas, 1987), although numerically the differences were small However, because of this small Increase In s~lageIntake, anlmals offered the LEHF concentrate consumed more CP (and some 200 g digestible undegraded proteln per day more) than anlmals offered the other concentrates, and yet the y~eldsof mllk proteln from those ammals were no d~fferentfrom the others The efficiency of use of food protein for milk protein production was significantly less from animals offered both low ME diets than those offered the high ME diet. On diets differing only in silage quality, with a constant concentrate regime, the efficiency of food protein utilization for milk production has been found to cover a considerable range (0.24 to 0.32; Dewhurst r t al., 1996), indicating the potential of a number of factors to influence this. Flowever, one theory is that the utilization of amino acids for gluconeogenesis may have been reduced on the high ME diet (Lees t,t al., 1990), leading to an increased availability for milk protein production. This would also help to explain the difference in milk protein production between the two high FME diets since both dietary CP supply and microbial protein capture were similar for these two diets. The lncrease in s~lageIntake by animals offered the LEHF concentrate was not cnough to compensate for the d~fferencebetween concentrate FME densities the ERDPIFME rat10 ot the dlet of these anlmals was therefore h~gherthan that of the other dlets because the ERDP intake was slmllar by an~malson all three dlets The dletary ERDPIFME rat10 1s an important factor for mlcrob~alprotein synthesis from rumen degradable d~etaryCP For lactating d a ~ r ycows, an ERDPIFME rat10 of about 11 g/MJ is recommended

Milk protein and dietary fat of dairy cows (AFRC, 1992), which was exceeded by all the diets offered in this study. In this study, the urinary excretion of purine derivatives was used as a simple index of microbial protein synthesis (Moorby et al., 1996). Purine derivative excretion from animals offered the two higher FME concentrates, LELF and HEHF, was significantly higher than that of animals offered the lower FME concentrate. The ERDPIFME ratio of the LEHF concentrate diet was almost 2 g / MJ higher than the other diets, indicating that the supply of FME in that diet was limiting for microbial protein production. Thus, a lack of effective nitrogen capture by the rumen microbial population may have contributed to the lower gross efficiency of milk protein production on this diet.

Mzlk productzon and compositlnn Milk protein concentration was significantly affected by dietary treatment and the results suggest that this was controlled by a combination of protein supply to the mammary tissue and milk volume. The daily yields of milk protein were lowest from animals offered the LELF concentrate although the large differences in milk protein concentration were brought about by a combination of protein production and milk volume. This is in agreement with many other studies in which increases in dietary fat content have decreased milk protein concentration but not decreased protein yields (e.g. Drackley and Elliot, 1993; Holter et al., 1993; Palmquist et al., 1993). An increase in the supply of dietary fat led to a decrease in the concentration of milk CP. This was apparently due to significant increases in milk yields (and more specifically the volume of water produced by the animals) since there was no significant difference in the yield of milk protein between the two low ME diets. Increasing the ME supply at the high fat level did not significantly affect either milk yields or milk protein concentrations. The effect of diet on the fatty acid content of milk fat is generally well characterized (Baer, 1991). Milk fatty acids are derived either from the blood, which in turn may be obtained from the diet (DePeters et al., 1987 and 1989; Cant et al., 1993), or from de novo synthesis by the mammary gland. Addition of palmitic acid to dairy cow diets can increase the palmitic acid content of milk fat and reduce the de novo synthesis of fatty acids (Banks et a l , 1980). The concentration of citrate in milk is an index of de noao fatty acid synthesis (Faulkner and Pollock, 1989) the lower its concentration in milk, the greater the rate of fatty acid synthesis by the mammary gland and in this study the output of citrate in milk was significantly higher on the two high fat diets than on the low fat diet, indicating a decrease In de novo fatty

-

7

acid synthesis. Glucose is used by the mammary gland for the production of fats - not for the incorporation of carbon but for nicotinamide adenine dinucleotide phosphate (NADP) reduction and aglycerol-P formati~n (Forsberg ct al., 1985). The incorporation of preformed fatty acids into milk fat is energetically efficient since it reduces the need for both NADPH and a-glycerol-P units, because less NADPH is required for fatty acid chain elongation and because 1 g of milk fat with a high proportion of long-chain fatty acids contains fewer molecules than 1 g of fat with a greater proportion of short-chain fatty acids. If glucose is spared from the process of fatty acid synthesis, it is available for other purposes in the mammary gland. In this study, as in previous studies (Banks et al., 1980; Cant et al., 1993; Del'eters et al., 1987 and 1989), the increase in supply of dietary fatty acids apparently reduced the level of mammary fatty acid synthesis so that dietary fatty acids were incorporated into milk fat in preference to the production of new ones. At the same time as de rrovo fatty acid synthesis was decreased, milk lactose yields increased: The daily production of milk was increased by the addition of fat to the diet, as is expected with increased lactose yields. Lactose concentrations, however, were significantly reduced by the addition of fat in the diet; this finding has also been reported by other groups (DePeters et al., 1987 and 1989). The reduced lactose concentration between diets LELF and HEHF was apparently balanced by increased concentrations of potassium and sodium. The increased lactose yields on the high fat diets, and in particular on the HEHF diet suggests that more glucose was available for lactose production on these diets as less glucose was used for fatty acid synthesis.

Conclusions In this experiment, supplementary dietary fat reduced the concentrations of milk solids. The concentration of protein, like that of fat and lactose, was reduced in the milk from animals offered high fat concentrates, despite significant increases in daily yields of protein and lactose on the high ME concentrate. The reductions in milk solids concentrations were brought about mainly through significant increases in yields of water as the de nova synthesis of fatty acids for milk fat was apparently reduced and lactose production was increased, drawing more water into milk and diluting the solids. Dietary fat supplied little or no FME, and this was observed in the lower rates of urinary purine derivative excretion from animals fed the low ME, high fat concentrate. Animals offered the high ME, high fat concentrate yielded more milk protein than animals offered the low ME, low fat (but equal FME)

8

Moorby, Dewhurst, T h o m a s a n d M a r s d e n

concentrate, indicating that amino acids may have been spared from gluconeogenesis by the extra supply of ME. It is therefore concluded that the concentration of protein in milk depends not only on the supply of precursors for milk protein production, but also o n the supply of precursors for fat and lactose production which will ultimately determine milk yields.

References Agricultural and Food Research Council. 1992. Technical Committee on Responses to Nutrients. Report no. 9. Nutritive requirements of ruminant animals: protein. Nutrition Abstracts and Reviews, S~riesB 62: 788-835. Baer, R. J. 1991. Alteration of the fatty acid content of milk fat. lourllal of Food Protection 54: 383-386. Banks, W., Clapperton, J. L. and Kelly, M. E. 1980. Effect of oil-enriched diets on the milk yield and composition, and on the composition and physical properties of the milk fat of dairy cows receiving a basal ration of grass silage. /ourrzal of Dairy Research 47: 277-285. Cant, J. P., DePeters, E. J. and Baldwin, R. L. 1993. Mammary uptake of energy metabolites in dairy cows fed fat and its relationship to milk protein depression. Journal of Dairy Science 76: 2254-2265. Casper, D. P. and Schingoethe, D. J. 1989. Model to describe and alleviate milk protein depression in early lactation dairy cows fed a high fat diet. Journal of Dairy Science 72: 3327-3335. Coppock, C. E. and Wilks, D. L. 1991. Supplemental fat in high-energy rations for lactating cows: effects on intake, digestion, milk yield, and composition. lournal of Animal Science 69: 3826-3837. DePeters, E. J. and Cant, J. P. 1992. Nutritional factors influencing the nitrogen composition of bovine milk - a review. /ournal ($Dairy Scierlce 75: 2043-2070. DePeters, E. J., Taylor, S. J. and Baldwin, R. L. 1989. Effect of dietary fat in isocaloric rations on the nitrogen content of milk from Holstein cows. \ouri~al of Dairy Scienct. 72: 2949-2957. DePeters, E. J., Taylor, S. J., Finley, C. M. and Famula, T. R. 1987. Dietary fat and nitrogen composition of milk from lactating cows. [ournal of Daivy Scienc~70: 1192-1201. Dewhurst, R. J., Mitton, A. M., Offer, N. W. and Thomas, C. 1996. Effects of the composition of grass silages on milk

production and nitrogen utilization by dairy cows. Animal Scieizc~62: 25-34. Drackley, J. K. and Elliot, J. P. 1993. Milk composition, ruminal characteristics, and nutrient utilization in dairy cows fed partially hydrogenated tallow. [ournal o f Dairy Science 76: 183-196. Faulkner, A. and Pollock, H. T. 1989 Changes in the concentration of metabolites m mllk from cows fed on dlets supplemented with soybean oil or fatty acids /ournal of Dairy Rrscarch 56: 179-183 Forsberg, N. E., Baldwin, R. L. and Smith, N. E. 1985. Roles of glucose and its interaction with acetate in maintenance and biosynthesis in bovine mammary tissue. ]ournal c!fDairy Science 68: 2544-2549. Holter, J. B., Hayes, H. H., Kierstead, N. and Whitehouse, J. 1993. Protein-fat bypass supplement for lactating dairy cows. /ournu/ of Dairy Science 76: 1342-1352. Jenkins, T. C. and Jenny, B. F. 1992. Nutrient digestion and lactation performance of dairy cows fed combinations of prilled fat and canola oil. 1ournal of Dairy Science 75: 796-803. Lawes Agricultural Trust. 1990. Gei~stat5 rgerence manual. Clarendon Press, Oxford. Lees, J. A., Oldham, J. D., Haresign, W. and Garnsworthy, P. C. 1990. The effect of patterns of rumen fermentation on the response by dairy cows to dietary protein concentration. British Iournal ofNutrition 63: 177.186. Moorby, J. M., Dewhurst, R. J., Thomas, C. and Marsden, S. 1996. The influence of dietary energy source and dietary protein level on milk protein concentration from dairy cows. Animal Sclmce 63: 1-10. Palmquist, D. L., Weisbjerg, M. R. and Hvelplund, T. 1993. Ruminal, intestinal and total digestibilities of nutrients in cows fed diets high in fat and undegradable protein. Journal of Dairy Science 76: 1353-1364. Penning, P. D. and Johnson, R. H. 1983. The use of internal markers to estimate herbage digestibility and intake. 2. Indigestible acid detergent fibre. Journal of Agricultural Sciet~ce,Cambr~dge100: 133-738. Thomas, C. 1987. Factors affecting substitution rates in dairy cows on silage based rations. In Recent advances in 1987 (ed. W. Haresign), pp. 205-218. animal nutrition Butterworths. London. -

(Rcce1verl5March 1997-Acc~pted 20 January 1998)

The effect of diet and housing on the development of sole haemorrhages, white line haemorrhages and heel erosions in Holstein heifers C. T. Livesey', T. Harringtonzt, A. M. Johnston" S. A. May3and J. A. Metcalf'Vpter~nary1,aborator~~s Agrr~cy, Addlt~~toile, Surrey KT75 3 N B ' A D A S Rrtdil) t

Quantity

Ingredient

Quantity

(ka , L,

fresh weight ddy)

Ingredient Grass silage Sodium hydroxide treated straw Brewers' grains Concentrate (see l'ablc 2)

Rolled wheat Molassed sugar-beet teed Molasses Extracted soya-bean meal Fish meal Mineral and vitamin supplement

4-10t 3.5 10 1.5

t The grass s~lagecontent was adjusted according tu dry matter content and energy concentrat~onbased on s ~ l a g e analys~s

Ai11rnal5 Holste~n he~fers had been reared Forty con;entionally at a growth rate of about 0.7 kg/day and calved down at approximately 2 years of age between October 1994 and January 1995, with a mean live welght of 545 (s.d. 38.7) kg after calving. During the summer of 1994 the heifers were kept at grass. From late August they were allowed free access to a total mixed ration (TMR) whilst still at summer grazing, due to a shortage of grass. The TMR (Table 1) comprised grass silage, sodiumhydroxide treated straw, brewers' grains and a small quantity of home mixed concentrate (Table 2). The same TMR was offered after the heifers had been housed until calving.

l;xper~'mrntaldeslgtl The experiment had a 2 X 2 factor~aldes~gn The prlmary treatments were housing in cubicles (C) or straw yards (S) and the secondary treatments were two lactat~ondiets designated high concentrate (H) and low concentrate (L): The treatment groups were designated: cubicles, high concentrate diet (CH); cubicles, low concentrate diet (CL); straw yards, high concentrate diet (SW); and straw yards, low concentrate diet (SL).

h10 100 I00

130 25

75

were weighed before the last trimester of pregnancv, in Julv 1994 and were then blocked by slre, wlthln sire by lifetime we~ghtgaln, w ~ t h ~11ven we~ghtgain by ascending value (Veerkamp ct a l , 1995) and anlmals within each block were allocated to one of the four treatment groups (Table 3)

The

Tveatnzents All heifers were housed in cubicle yards for training in September. Then approximately 6 weeks before predicted calving dates they were housed in cubicles or straw yards according to their allocated treatment. The cubicle yard used before calving was separate from the post-partum accommodation but the cubicle beds, bedding and passageways were similar to the post-partum cubicle yards. Heifers in cubicle yards were moved to a straw yard 3 to 5 days before calving and to their post-partum cubicle yard 1 to 3 days alter calving. The cubicles used after calving had butyl-rubber mats and shavings as bedding. Cubicle beds were 2.3 m long by 1.2 m clear between divisions, with a 75 mm fall from front to rear and a 0.20m rear step. The cubicle divisions were cantilevered free-standing galvanized steel tubing, which was also used to create a lunge space at the front of the cubicle with a headrail. The passageways between the cubicles were 2.2 m wide and were scraped four times daily by an automatic scraper. Shavings were refreshed daily onto the lower onethird of the cow beds. The single straw yard used

Table 3 Gtji~c>tic index and lifefirrle z c ~ e i ~ hgains t usrd in blockirig the ht7ifi.r~ prior to allucatiorr f o treatrrzerrf. Tilne of allociiticir~ to h ~ ~ ~ itreat,rrcrrt ilg relatioe to ciil?liizg dafc I S also gi.r)rn.All data preserzfcd (as mean arid standard deaiationsfur each trc,atnrrllt Blocking parameter

Treatmentt

(E)

s.d.

Lifetime live-weight gain (kdday)

CH

35 38 35 34

9.8 8.9 5.3 6.2

0.69 0.68 0.69 0.70

PIN

CL SH SL

t C = cubicles; S = straw yards; H = high concentrate; 1,

=

low concentrate.

.lime . of allocation s.d.

to housing (days pre-calving)

0.05 1 0.039 0.043 0-072

35 36 39 41

s.d 3.1 0.8 0 .>'.-l

1.6

Foot abnormalities in dairy heifers before calving to accommodate heifers allocated to straw yards was also used as the calving yard for heifers allocattd to cubicles. This yard was deep litter straw over a concrete floor and heifers in this yard had no access to concrete. One to 3 days after calving heifers allocated to the straw yard treatment groups were moved to one of two identical straw yards which had a 0.25 m step into a concrete feeding passage and were allocated at a rate of 6.2 m* per cow. One big bale of straw (approx. 256 kg) was used to bed u p each group of 20 heifers daily and yards were cleaned out monthly. The floors were concrete with a reinforcing mesh to reduce thermal cracking and had been mechanically grooved in a scluare pattern at 100 mm centres. The feeding passages were all 4-3 m wide and were scraped twice daily by the automatic scraper. After calving all four groups were in the same building and consequently the distance walked to the milking parlour was similar for both housing treatments. The dietary treatments commenced at calving. The diets contained proportionately 0.60 (H) and 0.30 (L) concentrate on a dry-matter basis respectively, mixed with the same grass silage given ad libiturn as a TMR. The concentrate used throughout was home-mlxed from rolled wheat, sugar-beet feed, molasses, extracted soya-bean meal, fish meal and a vitamin and mineral supplement. The proportions of ingredients in the concentrate are given in Table 2. The silage and pre-mix were analysed for dry matter (Ministry of Agriculture, Fisheries and Food (MAFF), 1985), ash (Her Majesty's Stationery Office (HMSO), 1982), crude protein (HMSO, 1982), amylase-treated neutral-detergent fibre (NDF) (MAFF, 1993), oil (MAFF, 1993), sugar (HMSO, 1982), starch (HMSO, 1982), water-soluble carbohydrates (MAFF, 1993), ammonia N (MAFF, 1985), p H (MAFF, 1985), neutral cellulase with gammanase digestibility (NCCD) (MAFF, 1993), acid-detergent fibre (ADF) (Clancy and Wilson, 1966), total fermentation acids, acetic and butyric acids (Fussell and McCalley,1987) as detailed in Table 4. Metaboli~ableenergy was determined from organic matter digestibility which was predicted by near infrared spectroscopy (Agricultural and Food Research Council, 1993). The concentrate pre-mix and silage were both sampled once a week and bulked into six samples each representing 1 month. The analytical results (Table 4) are the mean of the six bulk samples for silage and concentrate. Starting at housing in September, heifers were footbathed every 2 weeks in an antibiotic solution of Lincospectin (0.6 g/]; Upjohn Ltd) or Erythrocin

11

Table 4 Cowlpositiorr of grass silu,ye ilrrd corlccrrtrilte pre-rrrir 1,ffired izfter cnloing ( 0.05)

Table 10 c;rou)! irltakc~sof Hie lucti~tiorrtotill rrrixc2d r i i t i o ~ ~ik,y s ill!/ rrruttc,r pc'r kc+r per du!/J girleir by ii~tzt,kofexperirilcvt .

.

l reatment group

Cubicles C ~ n c e n t r ~level ~te

High

Low

Straw yards High

Low

Week of exper~ment 1 2 3 3 5

6 7 8 9 10 11 12 13 14 15

lameness. Vermunt (1992) has used this term to describe a range of abnormalities in the feet including haemorrhages into horn and heel erosions. These were considered to be the result of mild laminitis which could have been caused by toxaemic infections, nutritional problems and traumatic underfoot conditions such as rough concrete. Vermunt (1992) suggested that this represented an early stage in the development of foot disease. This study suggested that SLS consists of several different conditions which were individually expressed in response to different stress factors. The independent effects on sole haemorrhages and heel erosions could have been associated with common factors such as introduction to housing, pregnancy, parturition and changes in metabolism associated with lactation.

Discussion

The size of the standard errors associated with lesion scores in Tables 7 to 9 suggests that the genetic, environmental and nutritional history of individual animals may predispose heifers to foot problems. These factors may be anatomical (foot and limb conformation) or physiological (metabolic response to diet) and are likely to be wholly or partly influenced by the genotype. If such factors can be identified it would be possible to select dairy animals for low susceptibility to lameness, as is already being attempted in Germany (Distl, 1996). With hindsight the sensitivity of the experiment would have been increased by using pre-partum lesion prevalence or lesion score as a blocking factor. However, heifers were not uniformly affected by all three types of lesion and especially in the case of heel erosion the prevalence was very low before calving. Therefore, pre-partum occurrence of each lesion type would need to have been considered as a separate blocking factor and this would have been complicated further if lesion score was taken into consideration.

The term 'subclinical' laminitis was first proposed by Peterse (1979) to describe haemorrhages into hoof horn which were not necessarily associated with

Laminitis causes haemorrhagic lesions and abnormal horn formation (Vermunt, 1992)but lesions included

Ofhcv abnovi~ralifics Abnormalities apart from lesions associated with SLS were uncommon. There was only one mild lesion of digital dermatitis observed throughout the study and abrasion or contusion of the limbs was not recognized as a significant problem for any of the groups. In spite of paring of all the feet at each examination to restore normal conformation the soles of the feet of many heifers had returned to a flat or relatively flat conformation by the next examination.

Foot abnormalities in dairy heifers in SLS may not always be caused by laminitis. In this study there was evidence that laminitis was not responsible for the lesions observed. If laminitis had been the common factor responsible for heel erosions and haemorrhages in the sole and white line, the lesions would have been expressed with a common relationship to each other. In contrast, these data show that the lesions develop independently. In addition, if the pre-calving white line haemorrhages had originated in the laminar corium they would have persisted at the bearing surface of the foot for at least the first half of lactation because the growth rate of the wall has been shown to be between 0.5 and 1 cmlmonth for a wide range of situations (Vermunt and Greenough, 1994). The alleviation of white line haemorrhage scores for heifers housed in straw yards within 12 weeks suggests these haemorrhages must have originated in the corium of the sole and not the laminar corium. Laminitis resulting from concentrate overload in the diet would have been most likely to occur when the low concentrate pre-calving diet was changed to the lactation diet. The horn produced in the distal region of the laminar corium around calving would have been visible at the bearing surface of the feet by 12 weeks post partum. Therefore, the failure of the high concentrate diet to induce white line haemorrhages in heifers in straw yards is unlikely to be an artefact caused by the delay in appearance of haemorrhages originating in the laminar corium at the bearing surface of the feet. Consequently there was no evidence that the change to the high concentrate diet induced acute laminitis at least for heifers in straw yards. The group food intake records do not take account of straw bedding eaten by heifers in straw yards. The average intakes are within the expected normal range for the type of animals, diets and level of milk production. The addition to and removal of heifers from the experimental groups as they calved and completed the study accounts for some of the observed variation in average food intakes. Sole regrowth between foot examinations suggests that the sole was replaced approximately every 6 weeks under the conditions of growth, wear and foot paring in this experiment and the growth rate of sole was at least 5 mmlmonth, which is consistent with other published estimates of wall growth rate (Vermunt and Greenough, 1994; Offer et al., 1997). The haemorrhages observed in the horn of the sole and the white line would have been incorporated into the horn as it was synthesized. This suggests that the sole haemorrhages observed before calving were caused before the study commenced, when heifers were at summer grazing. Sole haemorrhages

15

recorded 6 weeks post partz~rn occurred in keratin synthesized at or soon after calving and haemorrhages recorded 12 weeks post parturn occurred at least 6 weeks after calving. Sole haemorrhage scores increased in all groups after calving suggesting the occurrence of this lesion was independent of the treatments. However lesion scores were higher for heifers housed in cubicles and heifers given the high concentrate diet, suggesting these treatments exacerbated sole haemorrhages. The positive interaction between cubicle housing and the high concentrate diet after calving, manifested by a particularly high sole haemorrhage score and marked increase in white line lesion score for group CH at 6 weeks post partum, was no longer present by 12 weeks post parturn, suggesting that other factors, which were only present at calving, may have contributed to this interaction. White line haemorrhages were alleviated and sole haemorrhages were less severe for heifers in straw yards. Other authors have reported the association between housing and lameness (Arkins, 1981; Baggott and Russell, 1981), the apparent benefit of straw bedding (Bazeley and Pinsent, 1984), the importance of cubicle design (Leonard et al., 1994) and time spent lying down (Colam-Ainsworth et al., 1989). In this experiment the additional stress experienced by cubicle housed heifers may have been imposed by spending less time lying down than heifers housed in straw yards, or when entering and leaving the cubicles, or by standing with hind feet in the cubicle passage. However, these factors were not studied in this experiment. Heel erosion appears to be comprised of two distinct lesions: pitting of the keratin, presumably caused by corrosive environmental conditions, and underrunning of the heel, which is probably associated with abnormal horn formation. These two conditions may be unrelated but it is possible that chances " in keratin svnthesis could adverselv affect keratin quality and predispose the horn to corrosive damage. The lesion scoring system combined surface pitting and underrunning within the same score. The increase in heel erosion scores in all groups during early lactation (Table 9) suggested that the development of this lesion was independent of the imposed treatments but the significantly higher lesion scores experienced by heifers in straw yards suggested this environment exacerbated the development of heel erosions. It is possible that either walking in deep straw or the contamination of the bedded area with dung and urine contributed to this lesion.

16

Livesey, Harrington, Johnston, May and Metcalf

Locomotion scores obtained during this study suggested that the high concentrate diet was associated with a higher incidence of lameness. The white line haemorrhages and heel erosions encountered in this experiment did not apparently cause lameness or discomfort since the high concentrate diet did not increase lesion scores for these two conditions. Sole haemorrhage lesion scores were increased by housing in cubicles and by the high concentrate diet but there was no direct correlation between sole haemorrhage score and locomotion score for individual heifers. Therefore, the cause of lameness could not be ascribed directly to the visible lesions, although the closest relationship was with sole haemorrhage. It is likely that cows bearing excessive weight on their soles are more likely to induce haemorrhages into sole horn and are probably in discomfort because of inappropriate loading of the sole.

Conclusions The results of this study suggest that the term 'subclinical laminitis syndrome' can be a misnomer when used to describe haemorrhagic lesions in hoof horn. White line haemorrhages were alleviated by housing in straw yards and exacerbated by cubicle housing. Heel erosion and sole haemorrhages occurred in early lactation independent of the treatments but heel erosion was exacerbated by straw yard housing and sole haemorrhages were exacerbated by both cubicle housing and the high concentrate diet. Even moderately severe heel erosions or haemorrhagic lesions in the sole and white line did not appear to cause lameness, since the incidence of lameness was increased by the high concentrate diet with no effects of housing treatments.

Acknowledgements We wish to thank the following: A. Breeze and the staff at ADAS Bridgets who took part in this study, C. A. Collins for assistance with experimental design and statistical analysis, R. D. Hancock, R. W. Blowey and S. B. Drew for their advice and encouragement, and Mrs L. P. Cooper for typing the manuscript. The work was funded by the Ministry of Agriculture, Fisheries and Food (MAFF).

References Agricultural and Food Research Council. 1993. Enerm und proteln requlrernents of ruminants. An adv~sory manual prepared by the AFRC Technical Committee on Responses to Nutrients. CAB International, Wallingford, Oxfordshire. Arkins, S. 1981. Lameness in dairy cows. Irish Vetrri~zary \ourr~ul35: 135-140. Baggott, D. G. and Russell, A. M. 1981. Lameness in cattle. British Veterinary Journal 137: 113-132.

Bazeley, K. and Pinsent, P. J. N. 1984. I'reliminary observations on a series of outbreaks of acute laminit~sin Rccoril 115: 619-622. dairy cattle. V~,tc.rii~ar!l Clancy, M. J. and Wilson, R. K. 1966. Development and application of a new chemical method tor predicting the digestibility and intake of herbage samples. Procprclirlgs c!f the 10th ir~tc~r~lntioirnl grtassllind congress, pp. 445-453. Colam-Ainsworth, P., Lunn, G. A., Thomas, R. C. and Eddy, R. G. 1989. Behaviour of cows in cubicles and its possible relationship with laminitis in replacement dairy heifers. Vrtivinary Record 125: 573-575. Distl, 0 . 1996 Improvements In the health of cattle through the genet~cselect~onof tra~tsrelated to the feet and legs Tlernrztllche Unlsthau 51: 331-340 Esslemont, R. J. and Peeler, E. J. 1993. The scope for raising margins in dairy herds by improving fertility and health. British Veterinary lour~inl149: 537-547. Fussell, R. J. and McCalley, D. V. 1987. Determination of volatile fatty acids (C2-C5) and lactic acid silage by gas chromatography. Atzalyst 112: 1213-1216. Greenough, P. R. and Vermunt, J. J. 1994. In search of an epidemiologic approach to investigating bovine lameness problems. Eighth interizational syrnposiunl on disorders of the ruminant digit (ed. I-'. R. Greenough), Bailfi Caiinda, pp,l86196. Her Majesty's Stationery Office. 1982. The ,ked~ng stuffs regulations. HMSO, London. Leonard, F. C., O'Connell, J. and O'Farrell, K. 1994. Effect of different housing conditions on behaviour and foot lesions in Friesian heifers. Veterinary Record 134: 490-494. Manson, F. J. and Leaver, J. D. 1988. The Influence of dietary protein intake and of hoof trimming on lameness in dairy cattle. Aiiirnal Production 47: 191-199. Ministry of Agriculture, Fisheries and Food. 1985 The anulysls of agrzcultural materials Agr~culturalDevelopment and Advisory Serv~ce, reference book no 427, MAFF, London Ministry of Agriculture, Fisheries and Food. 1993 Pred~ctlonof cizergy values of compound feedzng stuffs f o ~farrr~ ainrnals MAFF, London Offer, J. E., Logue, D. N. and Roberts, D. J. 1997. The effect of protein source on lameness and solear lesion formation in dairy cattle. A~iinlalScience 65: 143-149. Peterse, D. J. 1979 Nutrltron as a poss~blefactor 111 the pathogenes~sof ulcer of the sole m cattle Tljdschr~ft71oor Drergr?ieeskunde 104: 966-970 Veerkamp, R. F., Hill, W. G., Stott, A. W., Brotherstone, S. and Simm, G. 1995. Selection for longevity and yield in dairy cows using transmitting abilities for type and yield. Animal Scirrzcr 61: 189-1 97. Vermunt, J. J. 1992 "Subcl~n~cal" lammit~sIn d a ~ r ycattle New Zcaland Vrtc.tcrlnary/ourilal40: 133-138 Vermunt, J. J. and Greenough, P. R. 1994 Predrspos~ng factors of l a m ~ n ~m t ~cattle s Brltlslz Vrterlnary [ournal 150: 151-164

(Recerrlrd l 6 August 1997-Accepted

1 1 Fcbruary 1998)

18

Aston, Fisher, McAllan, Dhanoa and Dewhurst

Table 1 iorriy1~1iitioiiof tlzc coricc2~~tratc's ikgllfiesh i ( ~ ( ~ l x l ~ t )

Cc~ncenlratet

Iliglccl~ent l3arlt.1 Wheat Wheatfved M Y ~ ,gluten ~IL~ F1511 niedl HLII' \ova-bean meal Fat supplement$ M1neral5

243 121 46 36 110 390 17 7

390 245 98 98 (2

18 '31 15

t In th15and other I able\ H = h ~ g hcrude proteln (338 g CP per kg DM), L = low CP (156 g per kg DM) and M, the mcdlum CP concentrate was an equal part? mlxture of H and 1 f Megalac Volac Ltd, Royston, HCrtfordshlrc

concentration whilst milk fat concentration tended to decline. Responses to the additional energy supplied by concentrates are limited by reduced silage intakes and increased partitioning to body fat. Residual effects are a feature of responses to additional protein that have received renewed interest in dry cows, whilst a 'low milk protein syndrome' which develops in early lactation and is difficult to rectify by short-term dietary changes has been postulated (Moorby et al., 1996). Aston et al. (1995) showed that, even in established lactations, low milk protein concentrations could be completely rectified by increasing concentrate allocations. This does not preclude the possibility of a lack of response in other situations, perhaps when opportunities to increase silage intake are restricted by low quality or management factors. The objective of the work reported here was to investigate the short- and long-term response of cows to the level of protein supplied in a fixed low allocation of concentrates with a good grass silage available ad llbitunl.

through '3 6-mm die. Minerals wcre added during mixing to meet the requirements of lactating cows given by the Agricultural Research Council (1965). Degradability of the protein in the concentrates estimated using the principles and values given in Agricultural and Food Research Council (1992) was 0.73 and 0.62 ( g / g ) for L and H concentrates respectively. The silage offered to the cows was a 1 : 1 mixture of first and second cuts of late maturing cultivars of perennial ryegrass (Loliutrr perennc.) harvested on 19 to 21 May and 29 June to 1 July 1992 respectively. Crass was wilted for 2 h, harvested using a precision-chop machine, ensiled in roofed bunkers and covered with plastic sheets weighted with straw bales. Add-Safe (BP Nutrition Ltd; a mixture containing formic acid, ammonium formate, propionic acid and ammonium propionate) was applied at 5 I / t grass. Five dietary treatments were imposed based on grass silage supplied ad libitunr together with concentrate given at a flat rate of 5 kg/day. The treatments were designated LL, LH, MM, HL and HH according to the type of concentrate given. Three groups of 22, 11 and 22 cows received L, M and H concentrates respectively during weeks 4 to 12. At the end of week 12 the cows given L and H were assigned at random to receive either L or H concentrate in weeks 13 to 21 with an abrupt change-over of concentrate type where appropriate. From the beginning of week 13 there were 11 replicates of the five treatments. Three strategies provided a similar average amount of concentrate CP per day during weeks 4 to 21, a flat rate, treatment MM and two stepped rates treatments LH and HL.

Material and methods

Anlnrals and the~rrtzanagcrner~t Mrlk production r~xprrrmrr~t Flfty-f~ve HolsteinFriesian cows in t h e ~ rsecond and later lactatlons were selected for the experiment Thew mean calving date was 12 November (k 24 days) 1992 Allocat~on to treatment was at random from within blocks of five second or subsequent parity cows wlth similar dates of calv~ng

Diets ~lndtr~~ltmerlts Dietary treatments were based on concentrates given with grass silage. Three concentrates were used in the experiment; two concentrates were prepared with similar concentrations of metabolizable energy (ME) and either 156 (L) or 338 (H) g CP per kg DM, the third concentrate (M) comprised equal proportions of L and H. The CP content of the concentrate was increased by soya-bean and fish meals in the ratio 3 : 1 (Table 1). The ingredients for the concentrates were mixed, ground in a hammer mill fitted with a 3.2-mm screen and pelleted

From 3 weeks before their expected dates of calving until turn-out to pasture the cows wcre housed in a cubicle building equipped with Amcal individual food access doors (Amcal, USA). Prior to calving 2 kg fresh weight of rolled barley and 7 kg DM of grass silage were offered daily. For 2 weeks after calving all cows were offered the same diet of grass silage ad libitunl and 3 kg of M concentrate increased in two steps so that 5 kg/day was given in a covariance period during days 8 to 14 of lactation. Change-over to experimental treatment was made during week 3.

Protein in supplements to grass silage for dairy cows

19

Table 2 L'llt~ii~iral corrrpositlorr 17fthe siliz~runii corrcerltvi~tt~s ig/k,y dry r~rattt,v( D M ) iorlc~sssh?tc>il) Colicen tr,?te

H

Silage

I -

Mean 'loluene DM (g/kg trcsh wcight) Oven UM (as analysed g / k g fresh weight) Ash Acid-detergent fibre Ac~d-detergentinsoluble nitrogen Neutral-detergent fibre Water-soluble carbohydrate Starch Oil Crude protein Insoluble nitrogen Gross energy (MJ1kg DM) Metabolizable energy (MJ/ kg DM)t D-value it7 oh~o pH Fermentation acids Lactic Acetic Propionic Butyric Ammonia-nitrogen (g/ kg total N)

s.e.

252

3.23

83 311 0.73 499 23

0.8 2.6 0-029 3.8 0-2

tracc 45 162 16.3 20.0 12.3 0.72 3.72 115 15.9 1.7 2.3

91

I .h 0.31 0.43

Mean

880 '10

63 0.67 171 84 277 32 338 49.3 18.8 13.0

s.e.

Mean

S.?

1-7 1.O 2-1 0.050 3.3 2.4 7.3

873 78 62 0.68 200 56 454 28 156 23.0 17.9 12.8

2.3

0.9 3.1 0.82 0.03

I -H

2.2 0.053 3.7 1 43 5.4

1.8 2.6 0.28 0.04

0.014 3.3 0.75 0.13 0.45 3.4

t Measured in zli7~~with sheep at the maintenance level of feeding (ME as digested energy X 0.815).

Milk yields were recorded twice daily in a herringbone-parlour equipped with an automated gravimetric system and samples of milk for analysis were taken from two consecutive milkings each week. Live weights and condition scores were recorded weekly. The recording and sampling procedures, methods for chemical analyses and in viz10 measurements of apparent digestibility and ME concentration in silage and concentrates were described by Aston et al. (1994a and b). The concentration of urea in the milk was determined by the method of Oltner and Sjaunja (1982).

Apparent dlgestrhrlity und nztrogeiz bulutzte exprrrn~enf The apparent digest~bility and nitrogen (N) p a r t ~ t ~ o n ~ of n g dlets contaln~ng L, M and H concentrates (5 kg fresh welght) together w ~ t hgrass silage ad llbrtunz were determined In a separate change-over des~gn exper~ment w ~ t h four th~rd par~tycows, three diets and four perlods The cows were housed In indiv~dualstalls and the exper~ment commenced In weeks 12 to 14 of lactat~on Two weeks were allowed for adaptation to the diets and food consumption was measured over 6 days consecutively. Faecal and urine outputs were collected over a 6-day period which lagged 1 day

behind the measurement of intake. One day before collection commenced a separator mounting (Velcro, Selectus Ltd) was attached around the vulva and anus with contact adhesive (Evo Stik, Evode Ltd). At 09.00 h the following day a urine collector fitted with 1.5 m of flexible hose was placed over the vulva and fastened to the separator mounting. Urine was retained in a 25-1 plastic container positioned in a drainage channel behind the cow. To collect faeces a 350-mm wide flexible plastic chute was secured by four elasticated straps attached to a girth band on the cow and supported with two straps fastened to a tubular framework at the rear of the stall. Faecal output was directed by the chute into a container suspended in the drainage channel. A heel bar prevented the cow stepping backwards. Urine was preserved during collection by adding sufficient acid (350 g H,SO, per 1) into the containers to maintain pH between 2 and 3. Individual faecal and urine outputs were collected, weighed and mixed at 09.00 h each day; proportionally 0.05 by weight was retained. Daily subsamples of food, faeces and urine were stored at -20°C. At the end of each collection period bulked samples of food, faeces and urine from individual cows were mixed and

Aston, Fisher, McAllan, Dhanoa and Dewhurst

20

Table 3 L)ari!/ iritak~5r f d r y ~ r r a t t ~jnh/l) r arid ili~yestill/eei~rrg!/( D E ) h!/ ci1ic>sxir1e11 i4 ci~riceritrnt~~ iii\~}~/yirl,y l.ih (l.), 2-17 (,U) or ,338 (Mg c,rudt, proteiri pc'r Ag DM togc,thr,v ivith gmss sila,4y ad libiturn Treatments

DM mta ke (kg) Week5 4 to 12 S~lage Total Week5 I? to 2 1 51ldge rota1 Weeks 4 to 21 Silage Total DE intake (MJ) Weeks 4 to 12 Weeks 13 to 21 Weeks 4 to 21

LL

l>

LH

HI

MM

12 5 16 8

M

HH

13 7 18 0

I ?9 18 2

5 tJ d

**X

X*

12 8 17 5

l? 5 177

12 1 164

17 5 178

0 47 0 47

12 4 16 7

12 6 16 8

1 17 18 0

128 17 1

13 h 18 1

0 44 0 44

246

t In t h ~ and s other Tables, week5 4 to 12 d f

255 248 = 49,

267 260 264

Sfatlsfzcal analysls The significance of treatment effects was determined by analysis of variance. Food intakes, the yields and composition of milk solids, live-weight change and condition scores were adjusted by covariance using the mean values measured over 7 days in week 2 of lactation. The results of the apparent digestibility and N balance study were analysed using REML (residual maximum likelihood; Genstat 5 Committee, 1987) with cows and periods treated as random effects and treatments as fixed effects.

X-

**

** *X

268 239 250

week5 13 to 21 and 4 to 21 d t

subsampled. Anaiytical methods were described by Aston et al. (199413).

S~gn~t~cance

0 39 0 '39

11 Y l6 2

237 244

t

266 269 = 47

unless stated otherw~se

on the digestibil~tyexperiment, though the relative differences in forage proportion of DM intakc (0.744, 0.764 and 0.761 for L, M and H respectively) in the production experiment were maintained in the digestion experiment (0.718, 0.746 and 0.746 respectively). Intakes of digestible energy (DE), using values determined in the digestion experiment with lactating cows, are also included. Silage intake responses are illustrated in Figure 1. Effects of corzcei~trafeCP leorzl. Increasing concentrate CP from 671 g/day to 1067 g/day (treatment LL compared with MM) increased the intake of dietary

Results The results of chemical analysis (mean of 16 silage samples and eight samples of each concentrate) are given in Table 2. Both the silage and concentrates were of consistent chemical composition during the course of the experiment; the CV for silage toluene DM, N and lactic acid contents were 0.05, 0.04 and 0.11 respectively; CV of 0.05 were observed for concentrate CP and neutral-detergent fibre (NDF) contents. A well preserved grass silage with 0.85 of fermentation acids in the form of lactic acid was used in the experiment. The major chemical difference between the H and L concentrates was a substitution of CP (+l82 g/kg DM) for starch (-177 g/kg DM); there were only small changes in oil, ash and watersoluble carbohydrate contents.

Voluntary sllage intake The results of analysis of variance for voluntary intake of silage as well as total intakes are given in Table 3. Silage intake was slightly lower for the cows

Weeks after calving Figure 1 Silage toluene dry-matter intake ( k g l d a y ) for cows offered one of three concentrates varying in crude protein concentration (156 (L), 247 (M) and 338 (H) g l k g DM) in lactation weeks 4 to 12 and 13 to 21 (not covarianceLL; LH; +)( MM; adjusted). (A) (---W---)HL;(+)HH.

Protein in supplements to grass silage for dairy cows

21

Table 4 Dally yit~ldqof nidk (kg) irrrii rrrrlk .olrrl\ i y ) cow5 yrzleii 5 ky: concentriztr srrpplyrnx 156 ( L ) 247 ( M ) 07 138 ( H ) g rrudr proterri prr A,? DM togctlrer rrlrth yr~zsssrlagt ad llblturn I leatmcnts Weeks 4 to 12 Mllk Fat Proterl~ Lactose Week5 13 to 71 Mllk Fat I'rote~n Lactose Week5 4 to 21 M~lk Fat Proteln Lactose

L1

I

LH

26 h 1001 739 1228

MM

HT

29 6 1128 830 1360

H

HH

29 7 1103 866 1353

se

d

0 95 34 3 23 2 43 1

21 7 845 606 992

25 9 1019 792 1178

24 2 976 708 1107

21 6 877 633 999

25 3 957 762 1124

1 26 40 0 28 5 57 0

24 3 924 674 1118

26 0 1010 764 1195

26 9 1052 769 1233

25 4 991 748 1172

27 8 1030 816 1243

115 38 5 26 7 52 4

CP by 570 g/day and raised (P < 0.01) intakes of silage, DM and DE in all periods. The second increment of 400g CP per day (treatment HH compared with MM), predictably, increased the intake of dietary CP but had no effects on the intakes of silage, DM or DE.

Effects of CP distribution. Intakes of silage, total DM and DE were greater (P < 0.05) when a similar amount of concentrate CP was supplied at a flat rate

Sign~f~cance

*X

*** **X

** ** S**

***

** ** X>

***

(strategy MM) compared with stepped rates (strategies LH and HL) during weeks 4 to 21. Stepping up the amount of CP given in the concentrate in week 13 (LL v. LH) produced positive intake responses (P < 0.05) of 0.9 kg silage DM and 18 MJ DE per day, stepping down the amount of CP (HH v. HL) reduced intakes of silage DM and DE by 1.4 kg (P < 0.07) and 27 MJ DE per day respectively.

Milk production, milk composition and live-weight and condition score change Effects of concentrate CP level. Milk (P < 0.05), fat (P < 0.01), protein (P < 0-001) and lactose (P < 0.05) yields were greater when cows received treatment MM compared with LL during weeks 4 to 21 (Table 4). Giving the second increment of CP in the concentrate (treatments MM v. HH) had no additional effects on the yields of milk, fat and lactose but tended to increase milk protein yield. Covariance adjusted milk production responses to CP in the concentrate are illustrated in Figure 2.

Milk composition results are given in Table 5. Milk fat concentration tended to be higher for treatment MM than for LL or HH in all periods of the experiment. In the first period milk protein concentration was greater when cows received treatment HH compared with LL or MM, however the effect did not achieve significance. Increasing the supply of concentrate CP produced a linear response in milk protein concentration during weeks 13 to 21 Weeks after calving (P < 0.01) and overall (P = 0.08); the temporal pattern of response of milk protein concentration to Figure 2 Covariance-adjusted milk yield (kgiday) for cows concentrate CP is shown in Figure 3. Although there offered one of three col~centratesvarying in crude protein appeared to be a reduction in lactose concentration concentration (156 (L), 247 (M) and 338 (H) g / k g DM) in with diet HH (P = 0.055) this was small and is lactation weeks 4 to 12 and 1 3 to 21. (A) LL; (---+---) LH; ( -aa) MM; ( - - W ) HL; (A) probably not biologically meaningful. In all periods HH. of the experiment the relationship between

22

Aston, Fisher, McAllan, Dhanoa and Dewhurst

Table 5 C ' O I I I ~ O S~f ~ ~the I Ornilkfiorn II tuuls Xrvell 5 hg tonter~tratt~ i u ) ~ p l y l r ~756 g ( 1 ) 247 ( M )or 338 (11) ,y ~ r u d cprotrzrr7 pt7r hg L>M togrthrr uvth grass sllngcx ad l ~ b ~ t u m Treatmcnts Weeks 4 to 12 Fat (g/kg) I'rote~n(g/ kg) Lactose (g/ kg) Urea (mmoli 1) Weeks 13 to 21 Fat (g / kg) Protein (g/ kg) Lactose (g/ kg) Urea (mmol/ l) Weeks 4 to 21 Fat (glkg) Protein (g / kg) Lactose (g/ kg) Urea (mmol 11)

IL

L

LH

37.6 27.9 46.3 5-5

MM

HL

38.3 28.2 45.9 6.5

t1

HH

37-5 29.2 45.5 7.7

ed

51gnlflcance

1.09 0.69 0.40 0.29

***

5

38.7 28.0 45.7 5.2

39.4 30.9 45.5 7.5

40.6 29.4 45.7 6.3

40.8 29.1 46.1 5.3

38-8 30.6 44.6 7.3

1.S5 0.72 0.56 0.3 1

*X*

37.8 27.8 45.9 5.4

38.9 29.6 46.0 6.5

39.4 28.8 45.8 6.4

39.4 29.2 46.0 6.7

38.0 29.9 44.9 7.4

1.34 0.82 0.51 0.28

**S

concentrate CP and milk urea concentration was linear and highly significant. Live weight increased throughout the experiment on treatment H H whereas on LL and MM weight increased in the second period only (Table 6; Figure 4). Between weeks 4 to 21 there was little change in the weight of cows given LL compared with weight gains by those receiving MM (P < 0.05) and HH

(P < 0.01). Condition-score change was consistently greater with treatment H H than with treatments LL or MM but the effects did not achieve statistical significance.

Effects of pattern of distribution of concentrate CP. Varying the pattern of distribution of CP in a fixed amount of concentrate (strategies LH, MM and HL) did not influence the production of milk and milk solids or milk composition. In the second period there was a consistent, though non-significant, carryover effect on milk protein concentration from giving a higher level of CP (LL v. HL) in weeks 4 to 12. Increasing the supply of CP in the concentrate in the second period (treatment LL v. LH) increased daily milk (P < 0.01), fat and protein (P < 0.001) and lactose (P < 0.01) yields by 4.2 kg, 174 g, 186 g and 186 g respectively and protein concentration (P < 0.01) by 2.9 g / kg. Stepping down CP level (treatment H H 71. HL), produced negative responses of 3.7 kg and 80 g (P < 0-05), 129 g (P < 0.001), 125 g (P < 0.05) and 1.5 g / kg respectively.

Weeks after calving Figure 3 Covariance-adjusted milk protein concentration (g/ kg) for cows offered one of three concentrates varying in crude protein concentration (156 (L), 247 (M) and 338 (H) g l k g DM) in lactation weeks 4 to 12 and 1 3 to 21. (-)LL;f--4---)LH;(&)MM;(---W---)IIL; )-( HH.

During weeks 4 to 21 strategies LH (P < 0.05) and MM produced higher weight gains than HL. Cows lost weight (P < 0.05) and mobilized body reserves gained in the first period when the level of concentrate CP was reduced (treatments HL 71. H H ) in weeks 13 to 21. Following a small weight loss in early lactation there was a non-significant increased weight gain of 209g/day (LH v. LL) when the amount of concentrate CP was raised. Conditionscore change was not significantly affected by treatment but did reflect the differences in liveweight change measured for strategies LH and HL and for treatments HL 71. HH.

Protein in supplements to grass silage for dairy cows Table 6 C / i n ~ ~ , yiir~ ,isi i ~ r7ilcigilt [zrld co~lilitioilscore I J couls ~ [ w ~ ~ t rpi^ i r ~I;!: DM to::c2thc,r ic~itllgvuss sllqyc ad libit~un

gii~t'rr

5 /iS

C I I I I C E I I ~ I . I I / YS I I ~ ~ ~ I / L / I1:ih IIS

23 ( L ) , 247 ( M , (71. .3.3S ( i f )

crri,ic --

Treatments

Live weight week 2 (kg) Live-weight changc (g/cilly) Weeks 1to 12 Weeks i3 to Zi Weeks 4 to 21 Condition score week 2 Condition-score change Wecks 4 to 12 Weeks 13 to 21 Weeks 4 to 21

[>Id

I.

LH

MM

HI

H

HH

b.t%.d

Sigriili~a11ct'

577 2 1 l 01 7

310 188 2.3

0.02 0.0 1 0.05

0.01 0.14

compared w ~ t hthe second increment M ~ l kN output was ra~sed(P < 0 01) by glvlng M or H compared w ~ t hL concentrate but glvlng H compared with M had no effect The efficiency of convert~ngfood N lnto milk N plus retained N was not s~gnificantly different between treatments but tended to decline as concentrate CP Increased

Discussion Apparerlt dryestzbll~tyatld r~ltrogcnpart~t~ot~rrrg Responses to CP are often confounded w ~ t h responses to energy which result from effects on voluntary intakes and/or digestibilities. The results of the digestibility experiment with cows, confirm the success of producing diets which were of very similar apparent digestibility whilst varying in CP content from approximately 150 to 190 g / k g DM; the negative effect on apparent digestibility often Weeks after calving associated with having less than 180 g CP per kg DM Figure 4 Live weights (kg) for cows offered onc of three (Oldham and Smith, 1982) was very small in the concentrates varying in crude protein conce~itration(156 current work. The lower forage: concentrate ratio (L), 247 (M) and 338 (H) g l k g DM) in lactation weeks 4 to with the low protein concentrate should have LL; 12 and 13 to 21 (not covariance adjusted). (--C-) increased the apparent digestibility. Although total (---a---) LtI; (A) MM; (---I---) HL; (A) digestibilities were not significantly reduced by the HH. lowest concentrate CP treatment. an increased forage voluntary intake made up a major component of the Apparcn t digestibility utzd rlitrogtv iltilizution response to protein; the forage intake response to Silage and total DM intakes increased (P < 0.001) protein remains to be characterized and incorporated when M or H was offered compared with L into rationing schemes. As concentrate CP increased concentrate but, as in the production experiment, the efficiency of conversion of dietary N to milk N there was no intake response from H compared with ( g / @ tended to decline: 0-225, 0.215 and 0,187 M (Table 7). The apparent digestibility of DM, (digestion experiment), 0.265, 0.246 and 0.231 organic matter and gross energy was not (production experiment). On average, the efficiency significantly affected by additional CP in the of conversion of dietary N to milk N was lower than concentrate, whereas the apparent digestibility of in the series of experiments reported by Dewhurst tzt acid-detergent fibre (P < 0.01), NDF (P < 0.05) and N ul. (1996) and this may reflect the lower levels of (P < 0.01) increased. Intakes of N and apparently concentrate supplementation used in the current digested N and the output of urinary N were work. The range of nitrogen efficiencies (g milk-N increased (P representative breeds from each of the major breed groups. With this in mind, the total productivity of crossbred and straightbred breeds of African, European and Indian origins is being measured over a range of breeding, rearing and finishing environments. The focus of this paper is on one component of production viz.growth rate. It presents results for live weights at birth, weaning and at 18 months of age for some straightbred and first cross animals from the different breed groups, provides

estimates of heterosis for growth for some of the crosses, ranks the sire breeds for genetic potential for growth and discusses potential uses of the sire breeds in crossbreeding programmes.

Material and methods Breeds and animals The study was conducted at the National Cattle Breeding Station, 'Belmont', near Rockhampton, Queensland, on three crops of straightbred and crossbred calves born in 1991, 1992 and 1993. The zebu breeds (Bos iizdicus) used were the Brahman (B) and the Boran (Bo), the sanga breed (African Ros taurus) was the Tuli (Tu), and the European breeds (Bos taurus) were the Charolais (Ch) and the Belmont Adaptaur (HS), a synthetic breed of nominally '/, Hereford (H) X l/, Shorthorn (S) derivation. The Belmont Red (AX), a Bos taurus synthetic breed of nominally G Africander I/qH %S, and the Belmont BX (BX), a synthetic of nominally '/,B 1/,H '/,S, were also used. The same H and S cows were used in the formation of the HS, AX and BX lines, each of which has been closed and inter se mated since 1953. The B, AX, BX and HS lines have been selected since the late 1960s primarily for high growth rates on pasture at Belmont. Details of selection in the lines are available elsewhere (Frisch, 1981; Mackinnon et al., 1991). There had not been any selection within the Boran and Tuli in Australia. Most of the B animals were direct samples from commercial seed-stock herds or were the progeny of animals sampled from those herds. The Ch was sampled as semen from commercial seed-stock herds. At the end of 1990, 1991 and 1992, females from the AX, B, BX and HS lines were allocated at random within age and previous lactation status, to different sire breeds to produce straightbreds and crossbreds according to the design shown in Table 1. Except during the 10-week breeding season, all of the cows grazed as a single herd. At the end of each breeding season the herd was divided for logistical reasons on the basis of sex of the calves. From then on, calves of the same sex remained together and were treated alike. Up to weaning, none of the calves or their dams was treated to control endo- or ecto-parasites and none was offered supplementary food. All animals were vaccinated against calf-hood diseases. All of the male calves remained entire until the end of the study. Drought prevailed for the latter half of 1994 and from 9 to 12 months of age the 1994 crop of males was offered a supplement of a grain-urea-molasses block at the rate of approximately 2 kg per head per week.

Beef breeds of African, European and Indian origins - 1 Their heifer contemporaries concurrently on irrigated pasture.

were

reared

Srrrs The 10 AX, 18 B and 10 BX sires were selected for high estimated breeding value (EBV) for 600-day live weight. The 15 HS sires were selected primarily for high 550-day live weight and high resistance to cattle ticks (Booph~lus microplus). Within each of these breeds, the same sires were used to produce straiehtbred and crossbred calves. The 10 Bo. 10 Tu and six Ch sires were selected at random. At least two bulls of every breed were joined in successive years. The prime purpose of the study was to estimate breed rather than sire effects. Consequently the number of sires per breed was maximized rather than the number of progeny per sire. U

Llve weights Calves born on weekdays were individually identified and weighed within 24 h of birth. The interval for calves born on weekends was up to 48 h. All of the calves were weighed at weaning, which occurred on a single day when the mean age of the calves was 180 days and at approximately 18 months of age. Treatments At a mean age of 8 months, the calves from each crop were allocated at random within sex, breed, sire and age, and age and previous lactation status of the dam, to two treatments. From then to 17 months, the 'treated' animals were treated every 3 weeks with anthelmintic (Nilverm Injectable, Pitman-Moore Australia Limited) to control gastrointestinal nematodes ('worms' - primarily Haemonchus, Oesophagostomurn, Trichostrongylus and Cooperia spp.). They were also inspected for cattle ticks (Boophilus microplus) and if any were found on any animals, all of the treated animals were dipped in acaricide (Tactic, Hoechst, Australia). The other half of the crop was the untreated 'control'. A few control animals from the 1992 crop were treated to prevent possible death from parasites. These animals and any that were showing signs of poisoning from Lantana (Lantana camara) were excluded from the analyses. Statistical analysis Live weights of the calves at birth, weaning and 18 months, were analysed using least-squares procedures (Statistical Analysis Systems Institute, 1992). Each model included the effects of years, genotype and sex of the calf, combined age and previous lactation status of the dam (called here AGEPLS), the interactions between main effects, and residual. Only first-order interactions were included in the final model as all higher order interactions were non-significant (P > 0.05). Covariance was used

29

to adjust live weights for differences between animals in their dates of birth. The mean squares for main effects were tested against the residual mean square unless there was a significant interaction involving the main effect. In the latter case, the main effect was tested against the interaction. For the analysis of live weights at 18 months, the effect of treatment to control parasites and its interaction with genotype of calf were included in the model. Live weights at 18 months were estimated from birth weights and daily gains from birth to the age at final weighing. Preliminary analysis of 18 months live weight had shown that sex X treatment and breed X sex X treatment effects were not statistically significant (P > 0.40). Sex differences in 18-month live weight were therefore estimated from the data for both treatments combined. Data for treatment differences and the parasite burdens that induced those differences are presented in part two of this series (Frisch and O'Neill, 1998). B, Bo and Tu were used as common sire breeds on HS, AX, BX and B dams. This subset of data was used to estimate the effect on live weight at different ages of sire breeds, dam breeds, their interaction, and calf sex and its interaction with sire and dam breeds. The average effect of dam breed was calculated across the three common sire breeds. Comparisons of the seven sire breeds were valid only across B dams. Purebred Bo and Tu calves were born on 'Belmont' to maiden heifers that were reared as a separate herd. Their birth weights, corrected using the constants for AGEPLS calculated from the other breeds, were used to estimate heterosis for birth weights for the Bo and Tu crossbreds. Prior to weaning, the purebreds were transferred to another location. Heterosis for live weights at weaning and 18 months have therefore not been calculated. For the AX, B and HS, the linear functions of the means for live weights of each sex at each age were used to estimate heterosis as the deviation of the F, from the mean of the parents. Straightbred Ch were not included in the comparisons and heterosis effects were not estimated. The approximate differences required for statistical significance between breedsex means and for statistical significance from zero were then calculated using Studentized range (Snedecor and Cochran, 1980).

Results Birth weights The least-squares estimates for birth weights are shown in Table 1. For each dam breed, mean birth weights of calves by Tu sires were similar to those of their straightbred contemporaries and consistently lower than those of crossbred calves sired by B or Bo

Frisch and O'Neill

30

Table 1 l.r'ixst-si[l~iircscstrlrrritc2sof hirtlr ioc2i~1its(r~icairsi s.c.) f i r liotz ci~lilt'~ (,f i.ilr.h srZs itr c,nclr gc,rrotypc. (110. of nllinrals gii7tvi rrr ~lll~l~llt/l~~sl~s~ S ~ r breedt e

Dam breed tiS

Sex

t1S

M

34.5 0.68 (44) 31 . 1 f 0.60 (52) 32.8 f 0.47

F AX

Brced mean M F

BX

Breed mecm M

F R

Breed mean M F

Breed mean Regression o f calf birth weight on day o f birth b, = 0.020 f 0.005 M

AX

RX

R

Ro

+

37-5f0-92 (24) 35-8f 0.73 (38) 36.7 f 0.61 4 1 i- 0.73 (36) 36.0 f 0.73 (37) 38.8 f 0.53 36.6 f 0.85 (26) 31.0 f 0.86 (26) 33.8 f 0.60 32.0 f 0.53 (72) 28.3 f 0.54 (72) 30.1 k 0.39

36.8 f 1 14 (15) 33.4 i0.98 (20) 35.1 f 0.76 39.4 0.93 (22) 34.4 f 0.84 (29) 36.9 f 0.63 35.6 f 0.92 (25) 32.2 f 0.84 (27) 33.9 f 0.63 31.7 f 1.25 (17) 29.0 f 1.06 (18) 30.3 f 0.86

+

+

36.9 0.78 (32) 35.2 f 0.82 (31 36.1 f 0.57

32.7 0.78 (32) 30.5 f 0.78 (32) 31.6 f 0.55 31.3 f 0.92 32.5 f 0.75 30.2 f 0.78 (26) (35) (32) 30.9 f 0.74 31.1 0.80 28.8 0.97 (37) (30) (20) 31.0 f 0.62 31.9 f 0.56 29.5 0.63

+

+ +

F Sire breed meant

t Breed codes: HS = Belmont Adaptaur; AX

For straightbreds and crossbreds, the birth weights of calves from the different dam breeds ranked AX > HS > BX z B with the difference between each dam breed being statistically significant (P < 0-01).

33.8 f 0.95 (21) 31.9 0.85 (28) 32.9 f 0.65 35.5 f 0.76 (34) 34.7 f 0.76 (36) 35-1 f 0.55 32.4 f 0.87 (26) 30.1 f 0.83 (28) 31.2 f 0.6 I 29.7 f 0.75 34.6 f 0.98 (34) (22) 28.5 f 1.21 35.3 f 0.90 (15) (24) 29.1 f 0.72 35.0 0.68

+

+

Dam breed meant 36.1 f 0.61

(60) 33.9 f 0.54 (86) 34.9 0.43 38.9 f 0.51 (92) 34.9 f 0.49 (102) 36.9 f 0.38 34.9 f 0.54 (77) 31.1 0.52 (81) 33.0 f 0.39 31 .l f 0.49 (123) 28.6 f 0.54 (105) 29.9 f 0.40

+

+

37.0 f 0.41 35.9 f 0.55 32.8 f 0.46 (158) (79) (115) 32.9 0.39 32.1 f 0.51 31.4 f 0.49 (173) (94) (107)

+

+

34.9 f 0.30 34.1 0.40 32.1

= Belmont Red; BX Charolais. $ Calculated within common sire and dam breeds only.

bulls. The difference between sexes was also consistently less for Tu-sired than for B- and Bo-sired calves. Birth weights of B- and Bo-sired male calves from HS. AX and BX dams were consistently heavier than those of their straightbred contemporaries. The corresponding females were significantly heavier only from HS dams. B-sired calves from HS and AX dams were consistently heavier than Bo-sired contemvoraries. Within B dams, Ch-sired calves were s~gnificantlyheavier than all other calves.

+

Ch

lu

=

+ 0.36

Belmont BX; B = Brahman; Bo

=

Boran; Tu

=

Tuli; Ch

=

Birth weights (kg) adjusted for AGEPLS for the purebred Bo and Tu males were 31.2 (s.e. 0.61) (10 calves) and 32.6 (s.e. 2.29) (8) respectively and for the corresponding females were 27.6 (s.e. 0.90) (13) and 29.3 (s.e. 1.45) (11) respectively.

Ht.trrosis for hirth u~r~ighfs The estimates of heterosis for birth weights are shown in Table 2. For each dam breed (except B) and sex, heterosis generally exceeded 13% for B-sired calves and 11% for Bo-sired calves and was generally less than 6% for Tu-sired calves. Estimates within B dams were generally negative. Within HS and B dams, heterosis for females consistently exceeded that of male contemporaries

Beef breeds of African, European and Indian origins - 1

31

Table 2 L~tlmatr5of h c t e ~ o s ~(kg, \ ' X ) ) fur hlrth i l l e ~ g h of t ~ Ilve calves ofeach 5c.x 1n facl1 ,yc,~~otypet

Sire breed

Heter0515

Dam breed HS

Dam breed AX

M

M

F

F

Dam breed BX

Dam breed B

M

M

E

F

Dam breed B 51re breed Heterosls

M

F

*,** Means are significantly different from zero * P < 0.05, ** P < 0.01. Differences between means of >0.8 kg are significant (P < 0.05). t For breed codes see Table 1.

while for all dam breeds, the heterosis for Tu-sired females consistently exceeded that of Tu-sired males. For males born to B dams, most estimates

of heterosis were significantly negative and for females, generally not significantly different from zero.

Table 3 Lrast square5 ~ s t l m a t e sof weanl~zgwt~glzts(meat15 f S e j fur calves of each \ex parenthebe5 j

111

cnch ge~rotype(no of a~ilnzalsglveir In

S ~ r breedt e Dam breed

Sex

HS

HS

M

157.8 f 3.0 (43) 143.7 f 2-8 (50) 150.7 f 2.1

F

AX

Breed mean M

Breed mean M F

B

Breed mean M F

Breed mean Regression of calf weaning weight on day of birth h, = 4.754 f 0.228 M

BX

185.8 f 3.5 (31) 178.5f 3.7 (29) 182.2 f 2.6

F BX

AX

185.4 f 3.8 (26) 176.5f 3.5 (31 181.0 f 2.6 197.0 f 4.1 205.9 f 3.3 190.9 3.3 (25) (35) (35) 184.5 f 3.3 191.2 f 3.3 179.0 4.5 (36) (28) (18) 190.8 f 2.8 198.6 f 2.5 185.0 f 2.8

F Sire breed mean$

t For breed codes see Table 1. f Calculated within common sire and dam breeds only.

+ +

B

Bo

173.6 f 3.8 (26) 174.8 f 3.2 (37) 174.2 f 2.6 204.6 f 3.2 (36) 186.6 f 3.2 (37) 195.6 f 2.3 199.4 3.7 (27) 172.5 f 3.8 (26) 186.0 f 2.6 191.6 f 2.3 (75) 174.4 f 2.5 (65) 183.0 f 1.8

172.2 5.3 (14) 162.3f 4.5 (18) 167.2 f 3.5 193.2 f 4.0 (23) 177.9 f 3-7 (29) 185.5 f 2.7 199.9 f 4.1 (24) 180.7 f 3-7 (27) 190.3 f 2.8 188.6 f 5.5 (17) 174.4 4.8 (17) 181.5 f 3.8

+

+

+

lu

171.6 f 4.4 (19) 160.1 f 3.8 (27) 165.9 f 2.9 190.6 f 3-6 (30) 178.5 3.4 (35) 184.5f 2.5 200.0 f 3.9 (25) 180.4 f 3.8 (26) 190.2 f 2.7 194.5f 3.3 205.6 f 4.6 (33) (20) 185.3 f 5.4 213.2 f 4.2 (14) (21 189.9 f 3.2 209.4 f 3.2

+

192.3 f 1.7 188.5 12.3 189.2 f 2.0 (160) (78) (107) 177.0 f 1.7 173.8 f 2.2 176.1 f 2.1 (165) (91) (102)

+

Ch

184.7 1.3 181.2 f 1.7 182.6 f 1.6

Dam breed meant 172.5 f 2.6 (59) 165.7 f 2.3 (82) 169.1f 1.8 196.1 f 2.2 (89) 181.0 f 2.1 (101) 188.6 f 1.6 199.8f 2.3 (76) 177.9f 2.2 (79) 188.8 f 1.6 191.6 2.0 (121) 178.0 2.3 (96) 184.8 l .7

+ + +

32

Frisch and O'Neill

Wearling weights Table 3 shows the least-squares estimates for weaning weights. Over all genotypes, B X Ch calves were the heaviest, proportionately 0.38 heavier than the lightest genotype, the HS. Differences between straightbred AX, BX and B were small. Progeny from HS dams were significantly lighter than their contemporaries from other dams. Within B dams, Ch-sired calves were significantly heavier than all other calves, AX-sired calves were significantly heavier than all but the Ch-sired calves, HS- and Tusired calves were of similar weights and heavier than calves sired by BX, Bo and B bulls. Crossbred males by B, Bo and Tu sires were significantly heavier than straightbred males within HS, AX and BX dams but crossbred female contemporaries were consistently heavier only within HS dams. Over all dams, differences between the three sire breeds were small. However, B-sired calves were heavier than Bo- and Tu-sired calves from HS and AX dams, lighter than Bo- and Tu-sired calves from BX dams, and lighter than Tu-sired calves from B dams. B X HS calves were significantly heavier than HS X B calves but the differences between reciprocal AX and B (AX/B) and between reciprocal BX and B (BX/B) crosses were small and not statistically significant.

Bo and Tu sires were generally small and not significant. Within B dams there were significant differences between sire breeds such that over both sexes, Ch > AX > HS > BX, Bo, Tu > B with a 15% (49 kg) weight range between the Ch- and B-sired progeny. However, although Ch-sired males were heavier than AX-sired males, the difference was small and not significant. Of the straightbreds, HS calves were significantly lighter than all other calves, B and AX males were of similar weights and both were significantly heavier than the BX males. The small difference between B and BX females was not significant. Both were lighter than AX females although only the B was significantly so. Over the three common sire breeds, there were significant differences between dam breeds with AX > BX, B > HS and a 6% (22 kg) range between the mean live weights of progeny born to AX and HS dams.

Heferosisfor 18-month weight Table 6 shows estimates of heterosis for live weight at 18 months. Within sex, heterosis was generally significantly highest for HS/B genotypes and significantly lowest for BX/B genotypes and was significant for all genotypes except BX X B females. Although there were significant differences between reciprocal crosses, there was no consistent pattern as to which reciprocal had the higher estimate.

Heterosis for weaning weights Table 4 shows estimates of heterosis for weaning weights for reciprocal crosses. For calves from B dams, heterosis was significantly highest for HSsired progeny and significantly lowest for BX-sired progeny of each sex. Heterosis did not differ significantly between sexes within any sire breed. However, this pattern was not repeated for calves by B sires. There was significant positive heterosis for both sexes only within AX dams. Heterosis was not consistent between reciprocal crosses.

Live weights at 18 months Table 5 shows live weights at 18 months of age. For each genotype, males were consistently heavier than females with the difference varying within genotypes from about 24 kg for the BX to about 55 kg for the BX X B. Within HS, AX and BX dams, all crossbred males except AX X Tu were significantly heavier than their corresponding straightbred male contemporaries. However, the corresponding crossbred females were consistently heavier than their straightbred contemporaries only within HS dams. Within HS and AX dams, B-sired progeny were the heaviest and Tu-sired progeny were the lightest but within B and BX dams, the differences between the progeny of B,

Discussion Birth weights For each dam breed, Tu-sired progeny were consistently lighter than B-sired or Bo-sired progeny and for male calves, consistently lighter than the respective straightbred males (Table 1). Similar rankings for the same sire breeds have been reported from the USA (Cundiff e f al., 1995; Baker, 1996; Chase et al., 1996). Relatively low birth weights and low heterosis identifies the Tu as a potential sire breed for use in those situations where increased birth weight of crossbreds is undesirable. Cundiff et al. (1995) and Rowan and Josey (1995), who used some of the Bo and Tu bulls used in the current study, reported that the incidence of dystocia corresponded closely with the ranking of the sire breed for calf birth weight. Suppression of birth weights of calves from B dams has long been recognized (Cartwright, 1973). An additional phenomenon may be operating for Tusired calves from B dams. Birth weight of female B X Tu was similar to that of corresponding B females. However, the B X Tu males were proportionately 0-07 lighter than B males. Plasse et al. (1995) in summarizing results from crosses between Brahman

33

Beef breeds of African, European and Indian origins - 1 Table 4 Estiiilatcs r,fhctc~rosis(kg, % ) f o r rcieai~ii~g ?or~ixlztsof calves ofeach sex in each reciprocal cross ~ c r l o t y p c t Sire breed B

Dam breed B Sire breed

Units

M

F

Dam brccd

Units

M

F

kg

23.3*** 13.3 17.2'* 9.1 2.4 1.3

25.4"* 16.0 14.7"" 8.3 3.5 2.0

HS

kg 'X,

AX

kg '26 kg

1.1 4.6 15.9"* 8.4 10.9"' 5.8

15.7** 9.9 10.1** 5.7 -3.0 -1.7

I is

'X)

kg

AX

%B

kg '%)

BX

BX

*' Means are significantly different from zero (P < 0.01). Differences between means of >4.7 kg are significant (P < 0.05). t For breed codes see Table 1.

and several South American Criollo breeds reported a similar phenomenon for birth weights of FIB X Criollo. Their birth weights were similar to or up to proportionately 0.08 below those of straightbred B contemporaries. The reasons for this phenomenon and for the similarity of response of the Criollo

breeds and the Tuli are unclear. However, the latter may indicate that the sanga and Criollo breeds are more closely related than their recent origins would suggest. The relatively greater difference between birth weights of the reciprocal B/AX crosses compared with the other reciprocal crosses may be

Table 5 Least-~quare~ mtzrnates of llve welghts at 18 months of age for males and females of each genotype (no of anlmals gzven in parentheses) S ~ r breedt e Dam breed HS

Sex

HS

M

297.8 f 4.5 (42) 264.7 f 4.4 (44) 281.2 f 3.2

F AX

Breed mean M

Breed mean M

F B

Breed mean M F

Breed mean Regression of 18 months live weight on day of birth h, =4.116f0.034 M

BX

350.6 f 5.3 (29) 320.9 1 5 . 3 (30) 335.8 1 3 . 8

F BX

AX

336.9 f 5.8 (24) 312.8 5.3 (29) 324.8 f 4.0 369.9 l 6.0 385.8 f 5.0 352.9 1 4 . 8 (23) (33) (35) 332.3 k 5.0 357.0 1 5 . 7 320.7 f 6.7 (34) (25) ('8) 351.1 3.9 368.9 13.9 336.8 f 4.2

+

+

F Sire breed meant

t For breed codes see Table 1. $ Calculated within common sire and dam breeds only.

B

Bo

357.3 f 6.0 (23) 324.7 f 4.8 (35) 341.0 f 3.9 392.6 f 4.8 (36) 339.7 f 4.8 (36) 366.2 f 3.5 367.5 f 5.6 (26) 312.0 5.7 (25) 339.7 4.0 353.2 1 3 . 5 (71 310.3 3.9 (55) 331.7 2.7

345.3 17.6 (14) 312.6 f 6.9 (17) 328.9 15.2 368-1 f 6.0 (23) 320.5 5.5 (28) 344.3 24.1 362.9 f 5.9 (24) 319.9 f 5.5 (27) 341.4 f 4.1 359.3 f 6.8 (18) 315.0 f 7.1 (16) 337.1 5.0

+ + + +

+

+

Tu

+ +

320.6 6.5 (19) 293.0 5.6 (27) 306.8 f 4.3 345.1 f 5.1 (32) 317.4 l 5.0 (34) 331.3 f 3.7 365.4 1 5 . 7 (25) 312.5 f 6.1 (22) 339.0 f 4.2 358.5 5.0 388.4 f 6.5 (33) (20) 322.0 7.7 372.0 f 6.3 (14) (21) 340.2 f 4.6 380.2 f 4.6

+ +

367.2 1 2 . 5 358.0 f 3.2 348.0 f 2.8 (156) (79) (109) 321.7 2.4 316.8 3.1 312.1 f 3.0 (151) (88) (97)

+

Ch

+

344.4 f 1.8 337.4 k 2.4 330.1 f 2.2

Dam breed meant

Frisch and O'Neill

34

associated wlth the same phenomenon The Afr~candercomponent of the AX when ~nteract~ng with the B uterine environment may have resulted un a relatlvelv lower birth welght of the F,B X AX calves than would bc prcd~ctedfrom the b~rthwe~ghtof the straightbred AX For both sexes, there was no significant difference in hcterosis for Bo- or for B-sired calves from HS dams. However, heterosis for B-sired calves from AX dams was significantly higher than that for Bo-sired calves. For both dam breeds, heterosis was consistently least for Tu-sired calves (Table 2). As a general principle, this indicates that the genetic distances between the Brahman and the HS and between the Bo and the HS are similar. Likewise, the Bo and AX are more closely related than the B and AX and the Tu is more closely related than either of the zebu breeds to the two taurine breeds. Similarly, the lack of significant heterosis for Bo-sired calves from B dams suggests that the B and Bo are relatively closely related. These results are not unexpected given the likely evolutionary history of the breeds (Frisch et al., 1997). Wearung zuelghts Within each dam breed, weaning weights of crossbreds (with the exception of B X Bo calves) were consistently higher than those of the corresponding straightbreds with the advantage ranging from 1% for Tu-sired calves from AX dams to 16% for B-sired calves from HS dams. Breed and heterosis effects contributed to the crossbred advantage. Within B and AX dams, straightbred calves were the progeny of bulls selected for high EBV for 600-day live weight. Nevertheless, an additional 0-08 proportional increase in weaning weight was achieved through the effects of heterosis. Within B dams, the greatest increase in weaning weight (proportionately 0.14) was achieved by crossing to

the Charolais. Smaller increases were achieved by crossing to the other, smaller, taurine breeds. The differences between these F,s in weaning weights are a direct reflexion of differences in mature sizes of the sire breeds (discussed later). Within HS and AX dams, breed effects can be expected to favour B sires over Bo and Tu sires. Weaning weights ranked accordingly (Table 3). In addition, heterosis effects can be expected to be greater for B and Bo than for Tu. Within BX dams, heterosis effects were likely to have been similar for Bo and Tu sires and least for B sires. Weaning weights then favoured Bo- and Tu-sired calves. Within B dams, heterosis effects were zero for Bsired calves and could be expected to be less for Bo(indicus X indicus) than for Tu-sired (indicus X taurus) calves. In this case, weaning weights favoured Tusired calves. Nevertheless, the effect of heterosis between Bo and B was sufficient to compensate for the lower mature size of the Bo. Despite the expected effects of heterosis and the marked morphological differences between the Bo and Tu, the weaning weights of their calves within HS, within AX and within BX dams were essentially the same. The effects of differences in heterosis created by the different breed combinations were however fully manifest at 18 months (Table 5). Cundiff et al. (1995) have reported a similar ranking for weaning weights of B-, Bo- and Tu-sired calves from Hereford and Angus females. However, in the temperate environment of Nebraska, the HerefordAngus reciprocal cross calves had higher weaning weights than the Bo- or Tu-sired calves and the advantage of the B-sired calves over the Bo- or Tusired calves was greater than in the present study. For the HS X B males, the estimate of heterosis for weaning weight was not significantly different from zero which. when the estimate for the reciwrocal cross is considered, indicates that realized grokth of

Table 6 Estirizatrs ofhrtevosis (kg, %i for live 70c.ights at 18 rnonths ofage for inales andfemales in each reciprocal cross gcnofyprt

Dam breed B

Sire breed B

Sire breed

Units

M

F

Dam breed

Units

M

F

HS

kg

44.4** 13.6 33.9**

44.8**

HS

kg

37.2** 12.9

kg

31.8+" 9.8 40.7,"

2.6, between dam breeds of >2.9 and between sire breeds of >2.7 kg are significant (P < 0.05).

Frisch and O'Neill

42

gained signihcantly more than control HS but the corresponding treated animals had similar gains to one another Within AX dams, gains of treated progenv of B and of Bo sires were slm~larto one another but only those of the B-sired progeny were significantly higher than those ot treated progeny by AX and Tu sires In the controls, gains of progeny ranked B > Bo > AX, Tu w ~ t hthe differences between the slre breeds bang signihcant Cams of the control progeny of the B slres were slmilar to those of treated progeny of AX and Tu sires Within BX dams, the gains of treated progeny of each of the sire breeds were similar to one another. However, in the controls, gains were significantly higher for the progeny of B sires than for the progeny of BX and Tu sires. Gains of control progeny of B0 sires were intermediate between those of B and BX sires and not significantly different (P > 0.05) from either. Within B dams, gains of treated progeny ranked Ch > AX > HS, Tu > BX, Bo, B. However, the differences between Tu, Bo and BX were not significant. Gains of the Ch- and AX-sired controls were significantly higher than those of all other controls and gains of B controls were significantly lower than those of the HS- and BX-sired controls. Over the three common sire breeds, gains of treated progeny from AX and HS dams were slmilar to one another and significantly higher than those of B and BX dam< Differences between gains of control progeny were small although the difference between HS and BX dams was s~gnificant

Response fo triatinent. Over all breeds, response to treatment was highest for HS and AX X Tu and

lowest for B X BX, B, B X Bo and BX X B Over all dam breeds, response to treatment was least for Bs~redprogeny and greatest for TU-s~redprogeny These s ~ r ed~fterences were generallv consident within each of the dam breeds Over the common slre breeds, responses were least for progeny of B dams and greatest for progeny of AX dams Responses of R- and Bo-sired progeny from each dam breed were consistently significantly lower, while responses of Tu-sired progeny (with the exception of HS X Tu), were consistently significantly higher, than those of the corresponding straightbreds. For the straightbreds, responses were significantly highest for HS, significantly lowest for B and significantly higher for AX than for BX. The difference in response of reciprocal cross genotypes was significant for B/AX (B X AX and AX X B) and B/BX but the differences were small and not consistently in favour of either the B sire or the B dam. Over the four dam breeds, the ranking of responses of B-sired progeny corresponded with the rank~ngof responses of the straightbreds. With a single exception in each case, the same was true for Bo- and Tu-sired progeny. The response of B X Ch was significantly greater than that of all other progeny from B dams and there was no significant difference between the response of B X AX and B X Tu progeny.

Het~rosis. Table 2 shows estimates of heterosis for live-weight gains over the treatment period. For HS/ B and AX/B genotypes there was significant positive heterosis for gains in both treatments but in the B/ BX genotypes, heterosis was significant only for gains of control animals. In each genotype, heterosis for control animals was consistently higher than that for treated animals.

Table 2 Ili~teroslsfol-li7~e-wc;~yht gaiils qf treatid and c.oiltrol airirrralsfor ci7c.h rcciprocul L.ross ge~lotypet

Sire breed K

Dam breed B S~rcbrced FiS

Un~ts

Treated

Control

Dam breed

19.0**

HS

'%)

11.1** 8.5 16.1** 11.7

kg

1.9

(X)

1.4

kg %

AX

BX

kg

Treated

Contlol

kg

179-

13.7

24-l** 21.3

AX

(X, kg

12.8**

17.9**

9.3

BX

'70 kg

16.8

16.7** 13.6 8.2"* 6.6

Un~ts

'X,

1.6 1 .2

14.6 4-8"

3.9

",*" Means are significantly different from zero * P < 0.05, ** I' < 0.01. Differences between means >3.9 kg are significant (P< 0.05).

t For breed codes we Table 1

Beef breeds of African, European and Indian origins - 2 Table 3 Log iiiiii i~rithiriet,c.irrzarr i i s . c . ) trck i.ollirt i p c ai~irr1111 ~ pcJr day) for each gc2nutype (710.ofar~lninls~iilc'iiirr pavcrrtlrosc~i) Sire breed t -.

Dam breed

HS

AX

BX

B

Bo

'Tu

Dam breed mean$

Ch

-.-

HS

Log Arith

AX

1.0s Arith

BX

Log Arith

B

Lug Arith

Sire breed meant Log Arith

+

1.629 0.050 42.6 (66)

1-24110.063 1-171i 0.082 1-45310.069 17-4 14-8 28.4 (45) (2.5) (38) 1 635 i 0 . 0 6 0 1.260 0.066 1.360 i 0.066 1.594 0.060 43.1 18.2 22.9 39.2 (45) (49) (37) (48) 1.460 f 0.066 1.191 f 0.064 1-246 10.066 1.433 10~06.5 28-8 15.5 17.6 27.1 (44) (37) (37) (39) 1-372 10.069 1.270 i 0 . 0 6 0 1.296 i 0.067 1.170 0.042 1.258 0.094 1.424 i 0.070 1-545 i O.OX4 23.6 18.6 19.8 14.8 18.1 26.5 35 1 (37) (40) (38) (89) (22) (33) (301 1.245 i 0.031 1.275 0.038 1.503 l0.07i 17.6 18.8 31-8 (220) (121) (158)

+

+

+

+

+

1.347 ir 0.010

-77.2 (108) 1.419 t 0.038 26 7 (1 '14) 1 321 C 0.077 20-9 ! 113) i -777 L 0.03h 18.L) (IW)

p -

t For breed codes sec Table I. $ Calculc~tedwithin common sire and dam breeds onlv.

Tick counts Table 3 summarizes results of tick counts. Over all genotypes, tick counts were significantly higher for straightbred AX and for HS than for all other genotypes except AX X Tu. Tick counts of Tu-sired progeny were significantly lower (HS), similar to (AX and BX), or significantly higher than (B) those of corresponding straightbreds. Tick counts for Chsired progeny were significantly higher than those of all other progeny from B dams except Tu-sired progeny, while tick counts of progeny of HS and Tu sires were significantly higher than those of straightbred B. Over the three common sire breeds, tick counts of progeny of AX dams were significantly higher than those of progeny of B and BX dams. Over the four dam breeds, tick counts of B- and of Bo-sired progeny were similar and significantly lower than those of Tu-sired progeny and the respective straightbred progeny (other than B). Table 4 shows heterosis for mean tick counts. Heterosis was positive for all genotypes and generally significantly higher for HS/B and AX/B genotypes than for BX/B genotypes. W o r i ePggg ~ ~ co~irlfs

Table 5 summarizes results from worm egg counts. Over all genotypes, worm egg counts were highest for HS X Tu and lowest for B and in every dam breed, worm egg counts were consistently lowest for B-sired progeny, intermed~atefor Bo-sired progeny and generally highest for Tu-sired progeny. For progeny t r o m B dams, worm egg counts of AX- and

Table 4 Hrtc,rosrs f i ~ ri n m n tick courlts for each rrciproc.111cross ,ym?otl/pet Dam breed B Sire breed

HS AX

RX

Units Hcterosis ticks 'X, ticks 'X ticks 'X

5.1" 17.8 10.4** 35.9 2.0 9.2

Sire breed B Dam breed

HS AX

Units

t-leterosis

ticks 'X, ticks

11.3*" 39.4 lO.8** 37.2

'%I

BX

ticks %

6.3" 29.8

*,** Means are significantly different from zero * 1' i0.05, ** P < 0.01. Differences between means >4.3 ticks are significant (I-' < 0.05). t For breed codes see Table 1.

of Tu-sired progeny were higher, but not always significantly so, than the progeny of the other sire breeds. The differences between progeny of other sire breeds were small and not significant. Over the common sire breeds, worm egg counts of progeny of B dams were significantly lower than those of the other dam breeds while the differences between the progeny of AX and BX dams approached significance (P < 0.08). Within each reciprocal cross the difference between worm egg counts was small and not significant. Table 6 shows estimates of heterosis for worm egg counts. There was significant positive heterosis of

Frisch and O'Neill

44 Table 5 Log and nrlthmetic mean (i4

)

ulorm c : y ~count for eaclz grrlotype (110 ofar~~n~lzls gluc2nIn piirc~~lthe~e~) Sire brcedt

Dam breed

t IS

HS

Log Arith

AX

Log Arith

RX

Log Arith

B

Img Arith

Sire breed meant

BX

AX

B

KO

Tu

Ch

Dam breed meant

2.66 f 0.045 455 (69)

2.39 i 0.052 2.58 f 0.064 2.72 i 0.057 2.57 f 0.037 245 38 1 522 372 (45) (27) (37) (109) 2.63 f 0.053 2.48 k 0.048 2.62 k 0.055 2.68 f 0.049 2.59 f 0,034 428 302 412 479 389 (44) (58) (40) (53) (151) 2.59 f 0.052 2.45 k 0.054 2.50 f 0.053 2.62 0.051 2.53 2 0.034 387 279 31 7 421 339 (45) (41 (42) (43) (126) 2.41 0.056 2.53 f 0.052 2.38 f 0.055 2.34 f 0.039 2.42 f 0.071 2.48 f 0.056 2.42 0.058 2.43 k 0.034 255 342 240 22 1 260 302 266 269 (38) (47) (39) (101) (23) (38) (36) (162) 2.43 f 0.030 2.53 f 0.035 2.63 f 0.032 269 339 42 7 (245) (132) (171)

+

+

+

Log Arith

t For breed codes see Table 1. $ Calculated within common sire and dam breeds only.

Table 6 Heterosis for mean worm egg courlts before treatnlerlt began for each reciprocal cross genotypet Dam breed B bire breed HS

Units

Heterosis

worm eggs

83" 24.6 -17 -5.2 64* 21

.-

OL

AX BX

worm eggs YO worm eggs Yo

-

-

-

Dam breed

Sire breed B Units Heterosis

HS

worm eggs

AX

worm eggs 76 worm eggs (%

93**

-

77.5 -.

U

BX

23 7.1 25 8.2

* Means are significantly different from zero, P < 0.05. Differences between means >h2 eggs are significant ( P < 0.05). t For breed codes see Table 1.

similar magnitude for both B/HS reciprocals, no significant heterosis for either B/AX reciprocals, and significant positive heterosis for the B X BX.

Discussion Live-weight gains In the treated group, differences between genotypes in live-weight gains are mainly a reflexion of differences in growth potentials of the parental breeds and heterosis generated by each cross. Estimates for each are presented in part one of this series (Frisch and O'Neill, 1998). In the present study it is the effect that ticks and worms have on the expression of growth potential and heterosis for growth and the consequent changes in the ranking of

the genotypes for growth rates that are being considered. Both ticks and worms affected growth of every genotype. The correlation across'breeds between worm egg counts and response to treatment was 0.78 (P 10.01) and between tick counts and response to treatment, 0-61 (P 0.01). However, neither total parasite burdens nor relative burdens of ticks and worms can be expected to remain constant either between years o; regions. Any changes can be expected to change not only absolute growth rates but also the rankings of the breeds for growth. The extreme case is demonstrated by the B and HS straightbreds. When treated, there was very little difference in their growth rates, but when untreated, growth rate of the B was only slightly depressed but exceeded that of the HS by proportionately 0.26. Likewise, as tick challenge increased, growth rates of B- and Bo-sired progeny could be expected to maintain the same relativity but as worm challenge increased, the difference in gain in favour of B-sired progeny could be expected to increase. The effects of parasites shown in Table 1 must therefore be interpreted, not as absolute values, but as comparative values for the different genotypes exposed to low to moderate burdens of both ticks and worms. Given the range of breeds and environments, it is impractical empirically test all scenarios. The need is then to be able to predict the ranking of different genotypes as parasite burdens alter. At higher levels of challenge, which frequently occur in tropical regions, correspondingly greater responses to treatment occur. An example is

Beef breeds of African, European and Indian origins - 2 provided by a previous study (Frisch and Vercoe, 1984), in which the responses of B, BX and HS to treatment to control ticks and worms were each about four times those in the present study. Thus, responses increase multiplicativ;ly, not additively, as parasite burdens increase (see also Wharton et al., 1970). Had four times the parasite challenge occurred in the present study, it could be expected that there would be an upward shift in the ranking of the more resistant breeds. Predicted gains for control B would then be 132 - (5.6 X 4) = 110 kg (from Table 1). Similarly, gains for the control B X Bo, and B X HS would then be 115 and 105 kg respectively compared with 97, 91 and 77 kg for the less parasite resistant B X AX, B X Tu and B X Ch respectively. Thus, although the control B X Ch had top ranking for gains in the present study, at high levels of tick challenge their ranking can be expected to be below In general, at that o f t h e more resistant high levels of parasite challenge, the B could be expected to have markedly higher gains than any of the other straightbreds, and B-sired progeny from dam breeds other than B could be expected to have higher growth rates than Bo- or Tu-sired progeny from those same dam breeds. Similarly, it could be expected that crossbreds by B-, Bo- or Tu-sires from HS dams would have higher growth rates than straightbred HS, B- and Bo-sired crossbreds from AX, BX and B dams would have higher growth rates than the corresponding straightbreds, and Tu-sired progeny from AX, BX and B dams would have lower growth rates than the corresponding straightbreds. The particular genotype, whether straightbred or crossbred, that will have the highest growth is therefore very dependent on both the growth potential (Frisch and O'Neill, 1998) and resistance of the breed, and the level of parasite challenge. The responses of the HS X Tu and B X BX are anomalous and caution should be exercised if using them. Based on tick counts (Table 3) and worm egg counts (Table 5) and comparisons with AX X Tu and BX X B respectively, the responses of the HS X Tu and B X BX are expected to be about 27 and 7 kg respectively. Although the gains of all genotypes were depressed, at the low levels of parasite burdens experienced the response to treatment of those genotypes with the highest resistance was insufficient to warrant the cost of treatment. For the B, for example, the response was equivalent to 6.7 kg/year or about $AU7 per animal per year (1998 prices) while the cost of each treatment was about $AU4. Any cost of treatment to control parasites would therefore have to be below $7 annually to break even. For the HS, the genotype with the greatest response, the corresponding value was $36. However, treatment of the HS (128.5 kg) would then only achieve the same result as could be achieved by the B without any treatment (126.3 kg).

45

Similar methods could be used to determine the most profitable genotype for any given parasite challenge. Heterosis for live-weight gains (Table 2) arises from heterosis for growth potential and for resistance to environmental stresses (Frisch, 1987). Growth potential of the F, approaches that of the parent with the higher growth potential while resistance approaches that of the parent with the higher resistance. The consistently higher heterosis for gains of the controls is a reflexion of the high levels of heterosis for resistance to parasites. Environmental stresses other than ticks and worms also had differential effects on growth rates of F,s and straightbreds (unpublished data) and may account for the relatively high heterosis for the treated animals. The growth rate advantage of the F, over the less resistant parent can therefore be expected to increase as parasite challenge increases. The resistance of the various Fls was consistently lower than or similar to that of the more resistant parent suggesting that at high levels of parasite challenge, only those F,s from breeds of high resistance can be expected to maintain a growth rate advantage, albeit small, over the more resistant parent. Any additional advantage arising from crossing resistant breeds must therefore come from effects on components of production other than growth. These are currently being assessed. Ticks The mean 3-weekly tick count averaged over all controls over the 10-month treatment period was 4.3 (s.e. 1.0) ticks per animal per day. This low infestation presumably resulted from conditions that were unfavourable for the establishment of higher burdens. However, despite low infestations, the regression of live-weight gains on tick counts was b = -0.42 (s.e. 0.113) kg per tick per 10 months or about 0.5 kg per tick per year for each genotype. Accordingly, genotypes have been ranked for resistance on the basis of tick counts.

The straightbreds can be compared directly and ranked B > BX > HS, AX in order of decreasing resistance. However, the Bo, Tu and Ch can only be ranked from tick counts of their crossbred progeny. Because of possible differences in heterosis for tick resistance between different crosses (Table 4), sire breeds should be compared for resistance only within crosses likely to generate the same amount of heterosis. Thus, the two indicine breeds, the B and Bo were compared within HS dams, the two sanga breeds, the AX and Tu, and the two European breeds, the HS and Ch, were compared within B dams, and the Bo and BX were compared within B dams. The BX is likely to be equally unrelated to the Bo and Tu and it is therefore used to compare them.

46

Frisch and O'Neill

Mort~detalled rt,asons tor uslng thcsc compar~sons are presented in part one of t h ~ sscrles (Fr~schand O'Ne~ll,1998) O n the basis of tick counts for TTS X B and HS X Bo, the R and Bo have similar resistance, ;l conclusion supported by the similarity of tick counts for the progeny of both sire breeds irrespective of dam breed. There was no significant difference between tick counts of B X AX and B X Tu or between AX and AX X Tu, suggesting that the AX and Tu have similar resistance. This is further supported by the similarity of live-weight responses of the B X AX and B X Tu. Comparisons of tick counts of B X HS and R X Ch indicate that the HS has higher resistance than the Ch. This is supported by the difference in response to treatment in favour of the B X HS, which had similar worm egg counts to the B X Ch (Table 5). Comparisons of tick counts for BX X Bo and BX X Tu suggest that the Tu has lower resistance to ticks than does the Bo. This is supported by the consistently higher tick counts for Tu-sired progeny compared with Bo-sired progeny regardless of dam breed. On the basis of these comparisons, the breeds have been ranked in order of decreasing resistance to ticks as B, Bo > BX > AX, HS, Tu > Ch. However, it should be noted that some HS animals are carriers of a major gene for tick resistance (Frisch, 1994; Kerr et al., 1994) and are therefore not tvvical of other breeds of ,, European origins. The comparatively high resistance of the FIB X HS is therefore unlikely to be repeated if lowly resistant Hereford X Shorthorns (or any other lowly resistant breed of European origins) were crossed to B dams. An indicator of the likely result is provided by the tick count for the FIB X Ch which was significantly higher than that of the FIB X HS and over twice that of the straightbred B. Lemos cf ul. (1985) have also reported that the mating of lowly resistant sire breeds to other indicine dam breeds produced progeny that had markedly lower resistance than the straightbred indicine breed. In many regions tick challenge is far higher than that experienced in the present study. 'The question is then whether there is any advantage in live-weight gains if crossbreds ralher than straightbred B were used. Since it is impractical to answer the question empirically, prediction must be used. From values for live-weight gains of treated animals (Table l), tick counts (Table 3) and the effect of ticks on live-weight gain (-0.5 kg per tick per year), it is possible to calculate, for example, that when the tick count on B controls was 50 per day, the live-weight gains of contemporary B X Ch controls would be the same as that of the control B, i.e. whenever the B carried 50 ticks per day the advantage

lies with the B. Likewise, when the B carried 15, 23, 38, 42 or 102 ticks per day, their gains would equal those of the B X BX, B X Tu, R X HS, B X Ro and B X AX respectively. Thus, where tick infestations are low to moderate (c20 ticks per day on B controls), live-weight gains of these crossbreds will exceed those of the R. Values could be calculated for the other genotypes used in the study and the relative suitabilities of those genotypes assessed for particular regions.

Worn~s Unlike the situation for ticks, the effect of worm egg counts on growth was not consistent across breeds. Comparisons of the worm egg counts of the strai~htbreds indicate that the R had the hiehest " resistance to worms and that any difference in resistance between the HS, AX and BX was small. However, the within breed regressions of live weight at the start of the treatment and mean worm egg count were -0.015 (s.e. 0.005) and -0.021 (s.e. 0.008) kg per egg for the HS and AX respectively but 0.000 (s.e. 0.009) and 0.015 (s.e. 0.008) kg per egg for the BX and B respectively. Thus, the effect of a similar worm burden will be greatest for HS and AX and least for B. It has long been recognized that equal worm egg counts are not always reflected in equal effects on live-weight gains. Turner and Short (1972) reported that the BX had similar worm egg counts to, but markedly lower live-weight responses to treatment than, the HS and AX. Likewise, B animals failed to show live-weight responses to drenching when worm egg counts were low (Frisch and Vercoe, 1984; Frisch, 1987). Thus, at least where B is involved, the rankinn of breeds for resistance must consider both worm egg counts and responses to treatment. O n this basis, the ranking of the straightbreds for resistance to worms in order of decreasing resistance is B > BX > AX, HS. This ranking is consistent with that established previously (Turner and Short, 1972; Frisch and Vercoe, 1984; Frisch, 1987). C,

'3

The same dam breeds as those used to rank the sire breeds for tick resistance can be used to rank them for worm resistance. On the basis of comparisons of worm egg counts and live-weight gain responses of the HS X B and HS X Bo, the B has higher resistance to worms than does the Bo. This conclusion is sunvorted bv the consistent difference in both worm L I egg count and response to treatment in favour of Bsired progeny from the other dam breeds. Comparisons of B X AX and B X Tu indicate that the AX and Tu have similar resistance to worms. This conclusion is supported by the comparisons of worm egg counts and responses of the AX and AX X Tu. Comparisons of worm egg counts of the B X HS and B X Ch suggest that the HS and Ch have similar resistance to worms. The greater response of the B X

Beef breeds of African, European and Indian origins - 2 Ch compared with the B X HS is likely to have arisen from differences in tick resistance (Table 3). Comparisons of EX X Bo and BX X TLIindicate that the Bo has higher resistance to worms than does the Tu. This is supported by the consistently higher worm egg counts and generally higher responses to treatment for Tu-sired compared with Ro-sired progeny. On the basis of these comparisons, the breeds are ranked in order of decreasing resistance to worms as B > Bo, BX > AX, Ch, HS, Tu. Worm burdens for all genotypes were low to moderate. In other situations, worm challenge is far higher. The question is then whether there are advantages in live-weight gains if crossbreds rather than straightbred B were used. However, the relationship between worm egg counts and liveweight gains differed between genotypes. This precluded valid calculations of egg counts at which gains of B and any given straightbred would equal one another. Cornpared with the B, the Bo had moderate resistance and the AX and Tu had low resistance to the worm species present in the current study. Any advantages associated with using the African breeds either as crossbreds or straightbreds, can therefore be expected to decrease as challenge from worms increases.

Implications for crossbreeding None of the breeds in the present study was outstanding for both high resistance and high growth rates at all levels of challenge from ticks and worms. While the Bo could match the B for resistance to ticks, none of the breeds could equal the B for resistance to both ticks and worms. However, at the levels of parasite challenge experienced, growth rates of many crossbreds exceeded that of the B without the need for chemical control of parasites. At the highest levels of parasite challenge, only B X BX and B X Bo could match the growth rates of the straightbred B. However, the advantage in growth rate was small and, as for any F,, the magnitude of advantages in other components of production will therefore have a major influence on any decisions relating to crossing of these breeds. The advantages of particular crossbreds over straightbreds provides scope for improving productivity through appropriately combining breeds for particular environments. In tropical regions of high parasite challenge and in the absence of other equally resistant breeds, the high resistance of the B (and by inference, other Indian zebus), and the high heterosis the B generates when crossed to taurine breeds, will ensure that the B remains an essential component of any crossbreeding programme. Where the B is already established, potentially the greatest increase in absolute growth rates could be obtained by crossing to Ch (or by inference, to any other large

47

breed of European origins) F-Io.vvever, because ot comparative lack of res~stanceof the B X Ch, the cross has limited apphcat~on 111 reglons o f h ~ g h paraslte challenge unless those parasites arc controlled bv chem~cal or other mean5 In 5~1th reglons, the AX, H5 and TLIoffer advantages over the Ch both from the standpoint of ease of u5e of bulls and In terms ot eff~c~c.ncv of growth of their F , progeny In the present study, only straightbreds and F,s h a w been considered. However, the F, is only the first step in any systematic crossbreeding programme and the question of which breeds arc best suited for the prodbction of subsecluent generations must be addressed. Consider the case where the F, was an F, B X HS. Interbreeding (to form a BX) will result in a loss of heterosis and hence both growth (Table 1)and resistance (Tables 3 and 5) will decline relative to the F,. Back-crossing to the B (i.e. BX X B) will maintain high resistance and slightly increase growth rates, but effects on other components of production may preclude its use. Back-crossing to the HS (or other breeds of European origins) will reduce resistance and growth rates though may produce advantages in other components of production. Crossing to an AX, Bo or Tu or combinations of these (e.g. AX X Bo) or combinations with other breeds of European origins (e.g. Ch X Bo) can be expected to maintain high levels of heterosis and have much less effect on resistance and growth than back-crossing to the HS (or other breeds of European origins). However, the use of crossbred bulls has merit only if those crossbreds are also more productive than the straightbreds from which they were formed. For the AX X 80, therc are advantages in both resistance to ticks (Table 3) and growth irrespective of the level of challenge from ticks and worms (Table 1).The merits of other crosses in terms of resistance to ticks and worms and growth rates at various levels of challenge could be predicted by reference to the relevant tables. While systematic crossbreeding betw-een the more resistant breeds would maintain moderate to high levels of resistance and growth in the presence of ticks and worms, it is not always a feasible system to use. Multibreed synthetics offer a potential solution but if 'resistant' and 'susceptible' breeds are used to form the synthetic, it can be expected that, as for the BX, at least half of the F, heterosis for resistance will be lost on interbreeding and growth in the presence of ticks and worms can be expected to decline relative to the F, (Table 1). The loss of resistance could be avoided if only resistant breeds were used. While the AX, HS, Tu and Bo have high resistance relative to the Ch, they are far less resistant than the B for resistance to one or both parasites. For

48

Frisch and O'Neill

crossbreed~ngto reach ~ t sfull potential In tropical areas of high t ~ c kand worm challenge, there 1s a need to h a v e access to other tropicallv adapted breeds that not only have high resistance to these yaras~tesbut are also unrelated to the Tndlan zebus There 1s a need to determine whether such breeds exist

Acknowledgements We wish to thank G. Halford, J. Quilty, J. Davies, 1. Gray and R. Holmes for care and supervision of the animals and assistance with collectioii of data. Special thanks arc due to A. Day who not only assisted with the above tasks but who also diligently prepared and counted all worm egg samples. Mrs D. Fleetwood and MS M. Matheson are thanked for preparation of the manuscript. A major part of the funding for the project was provided by the Meat Research Corporation through ownership of 'Belmont' and the cattle, and through provision of a grant that enabled the study to proceed.

References Frisch, J. E. 1987 I'hys~olog~calreasons for heter0515 In growth of Bo5 ~ndrcu\ X BOS taurus ]ournal of Agrzculturnl Sclence, Cambrldge 109: 213-230 Frisch, J. E. 1994. Identification of a major gene for resistance to cattle ticks. Proceedings of t h e f f t h world co?rgress orr genetics applied to livestock production, Guelph, zjol. 20, pp. 293-295. Frisch, J. E. and O'Neill, C. J. 1998. Comparative evaluation of beef cattle breeds of African, European and Indian origins. 1. Live-weights and heterosis at birth, weaning and 18 months. Anirrlal Science 67: 27-38.

Frisch, J. E. and Vercoe, J. E. 1984. An analysis of growth of different cattle genotypes reared in different environments. /o~rrrrizlofAgriirilturr71 Sric~irc-c),Carrrhrrii,yc~103: 137153. Kerr, R., Frisch, J. E. and Kinghorn, B. P. 1994. Evidence for a rn'ljor gene for tick resistance in cattle. Procc,rdirrg.; of tlzc f f t h iclorlci coizgrcss oli g~rrt~ticsnpplirrf to lii~ustock productiorr, Crrelpii, rlo/. 20, pp. 265-268. Lemos, A. M., Teodoro, R. L., Olivera, G. P. and Madalena, F. E. 1985. Comparative performance of six Holstein-Friesian X Guzera grades in Brazil. 3. Burdens of Boophilrrs rrricropl~rs under field conditions. Aurrnal Productiorr 41: 187-191. Roberts, F. H. S. and O'Sullivan, P. J. 1950. Methods tor egg counts and larval cultures for strongyles infecting the gastrointestinal tract of cattle. Australial~ [~ziriral 1!f Agricultural Rrscilrch 1: 99-102. Snedecor, G. W. and Cochran, W. G. 1980. Statistical ~nethods,ser~entheditioir. lowa State University Press, Ames, Iowa. Statistical Analysis Systems Institute. 1992. S A S user's ~ u i d estatistics, , version 6 edition. SAS First Inc., Cary, N.C. Turner, H. G. and Short, A. J. 1972. Effects of field infestations of gastrointestinal helminths and of the cattle tick (Boophilrrs microplus) on growth of three breeds of cattle. Alrstralinn Iournnl of A~riculturnlResearch 33: 177-193. Wharton, R. H., Utech, K. B. W. and Turner, H. G. 1970. Resistance to the cattle tick, Boopl~ilusmicroplus, in a herd of Australian Illawarra Shorthorn cattle: its assessment and heritability. Australiarr ]ourr?al of Agricultural Researdr 21: 163-181.

Effect of undernutrition and refeeding on digestion in Bos taurus and Bos indicus in a tropical environment P. Grimaud", D. Richard', A. Kanwk', C. Durieri a n d M. Doreau' 'Centre Interi~atlonalLIPRrchrrchc-D6~~cIop~~eiiient sur I'Elcvag~pn zone Subhurnldc, 01 BP 454 Robo Dloulasso Burklrla Faso 'Centrc de Coopcratron lnternat~onaleen Re~herclzeAgronomzque pour /c Developpern~ntlDipartementElevagr t,t Medecznr Vetc;rrnalre tJnpays Troptcaux, BP 5035,34032 Moiltpelller, Fraizcc 'Inst~tutNntlorzal de la Recherche Agronoinzque, Laboratozre d~ Blon~Arre,Route de St Cyr, 78026 Versa~lles,Fratzcr Vi~stttutNatlonal de la Recherche Agrotzornlqu~,Luboratolr~Sous Nutrltlorl deq Rurnlnants, The11 63122 S t Genes Champarzelle, France

Abstract The effect of undevfEeding and refeeding on digestion was studied in Bos taurus and Bos indicus cows. Eight nonlactating cows, four B. taurus and four B. indicus (live weight 156 kg and 207 kg respectizlely) were first given a forage-based diet at a level above energy maintenance requirements for 4 weeks (3.65 and 4.66 kg dry matter ( D M ) per day respectively for B. taurus and B. indicus). They were then restricted at a low level of intake for 2 months (1.83 and 2.33 kg D M per day respectively for B. taurus and B. indicus) and finally refed at the first level for 2 months. Digestion measurements were made before the underfeeding period, at 3 and 8 weeks of underfeeding and at 3 and 8 weeks of refeeding. O r p n i c matter apparent digestibility decreased with underfeeding and increased with refeeding (0,637, 0.591, 0,652, 0.692 and 0.669 i n B. taurus and 0,674, 0.560, 0.580, 0.698 and 0,692 i n B. indicus, respectively 1 week before, 3 and 8 weeks after underfeeding, and 3 and 8 weeks after refeeding). This lower apparent digestibility at low level of intake was not expected either by ruminal particle retention time, which increased when intake decreased, or by measurements of microbial activity: DM degradability measured in situ and ruminal particle size did not vary with level of intake. A n effect of the length of underfeeding and refeeding was seen: the apparent digestibility tended to increase after several weeks of undernutrition and was higher affer refeeding than before underfeeding. N o difference was observed between the two genotypes.

Keywords: cattle, refeeding, rumen drgestion, tropical climate, underfeeding, zebu.

Introduction Ruminant females are often subject to periodic food restriction. These periods where body reserves are mobilized are generally followed by phases of refeeding and lipid storage. In temperate countries, underfeeding can be due to an economic strategy of the farmer to minimize cost of feeding. It can also be the physiological situation in high-producing dairy cows. The lack of roughage in many tropical countries due to unfavourable climatic conditions also induces reductions in food intake. When food intake is very low, reproduction performances are Ilmited: fertility rates of cows on Sahelo-Sudanian rangelands are less than 50% (Galina and Arthur, 1989). To assess the nutritional deficit, whatever the amount of food restriction, it is essential to evaluate

the nutritive value of roughage and diet. The knowledge of their digestibility and the modifications of this digestibility when intake changes, permit an appropriate food supplementation to be proposed. Restricting food intake usually results in increased organic matter (OM) digestibility (Schneider and Flatt, 1975; Galyean and Owens, 1991). However, recent bibliographical reviews show that this relation is inconsistent and that the negative relationship between intake and digestibility is not always observed for intakes lower than maintenance level (Chilliard et al., 1995). In particular, the effect of duration of underfeeding and/or refeeding on digestibility, the between-genotypes differences in

50

Grimaud, Richard, KanwC, Durier and Doreau

digestive capacities and the effect of underfeeding on digestion after subsequent refeeding, are mainly uncertain. These factors have been analysed in a n experiment conducted with the purpose of determining whether cows of tropical breeds, which undergo a 2-month period of low energy food intake followed by a 2-month period of high energy intake, are able to show adaptations of digestion when faced with such modifications. This experiment aimed to answer several questions: (1) does a difference in digestive processes occur between maintenance and low intakes? (2) does an adaptation to undernutrition exist during the period of low intake? (3) does an adaptation to refeeding exist when cows return to maintenance intake? (4) is the digestive efficiency at maintenance intake similar before and after a period of undernutrition? This experiment was carried out with Bos tuurzrs and Bos indicus cattle, to compare the response of digestive events to undernutrition for these two genotypes.

Material and methods Animals, experin~entaldesign and diet Eight non-lactating non-pregnant adult cows, four Peuhl zebus (B. indicus) and four Baoules (B. taurus), were used as experimental animals. The experiment took place in a tropical environment in Burkina Faso, at Bobo-Dioulasso. This area is classified as a subhumid zone with four distinct seasons, dry-cool, dryhot, wet-cool and wet-hot. The annual mean rainfall is about 1100 mm, with rains falling from June through October. Mean minimum and maximum temperatures range from 17" to 23OC and from 33" to 37°C respectively. This experiment was carried out between December 1995 and April 1996, during the dry seasons, in a tsetse fly-proof shed where air was continuously renewed. Cows were penned in individual boxes. Their body weight at the beginning of the experiment was 207 (s.e. 18) kg and 156 (s.e. 23) kg for B. indicus and B. taurus, respectively. Three of each genotype were fitted with permanent ruminal cannulas made of polyamide and polyvinyl chloride (Synthesia, Nogent sur Marne, France); surgery took place more than 10 weeks before initiation of the experiment and was performed under general anaesthesia (Xylazine, Bayer, Germany). The cows received antibiotic treatment (Streptapen, Avotec, France) for each of 5 days after surgery. Throughout the experiment the cows received a forage-based diet (ratio forage/concentrate equal to 79/21) of constant composition at two different levels. They were first adapted to the diet at ad libitu~nintake for 3 weeks. Dry-matter (DM) intake was on average 95 and 92 g / k g metabolic weight (Mo7" respectively for B. irzdicus and B. taurus. Food was then offered for 8 weeks in restricted amounts

(proportionately 0.1 less than ad lihitlliil consumption for each animal). Mean daily intake was 4.66 and 3-65 kg DM respectively for B. irldicus and R. tnurrfs. This was defined as the high level (HL) of intake. The level of intake was then decreased by half from one day to the next (2.33 and 1.83 kg IIM daily, respectively for B. indiczts and R. tili~rus)and cows remained for 8 weeks at this low level (LL). At the end of this phase, cows were refed for 3 days until they achieved level HL. This last phase lasted 8 weeks. Five experimental periods were defined: HI, at the end of the first high intake period; L1 and L2 after 3 and 8 weeks of undernutrition; H2 and H3 after 3 and 8 weeks of refeeding at high intake. Duration of each period was 12 days. The diet was formulated so that decreasing food intake implied a shortage in energy but not in protein. Animal requirements were calculated according to INRA recommendations (Institut National de la Recherche Agronomique, 1989):300 kJ and 3.25 g/kg M07"or net energy and intestinally digestible protein, respectively. The diet covered 1.2 and 0.6 of net energy requirements at levels HL and LL and was rich enough in crude protein (CP) so that the supply of intestinally digestible protein was 1.8 and 0.9 of requirements for levels HL and LL. A mineral supplement was offered throughout the experiment so that total intake of minerals covered requirements. Diets were based on a mixture of rice straw, cottonseed meal and sugar-cane molasses. Concentrate was given at 07.30 h; rice straw was offered twice a day, just after the concentrate and at 15.00 h. Water was offered ad lihitum in buckets. Composition of the diet and net energy density determined from foodstuff composition according to food tables of Institut National de la Recherche Agronomique (1989) are given in Table 1 .

Measurements and clzetnical anaiyses Cows were weighed at the end of each of the five periods. Faeces were collected for h days (1 to h). Table 1 Cl~erri~cnl conrpo~it~oil nild rzct c,ncJrgyd~>rl\rtyo f 11 diet gl?leri to co~c~s nt hrgli 1rrlc.l (MC)nr~dlow lcvel ILL) of rritizke HL and L L t

Chemical coinpositio~i(g/ kg DM) Organic matter Crude protein Neutral-detergenL fibre Acid-detergent fibre Calcium Phosphorus Net energy density ( M J l kg DM)

909 87 h22

387 3.78 2.48 1.43

t Composition (g/kg): rice straw, 790.0; cottonseed meal, 140.0; cane molasses, 70.0.

Digestion in tropical breeds of cows After they were homogenized and weighed, a 0 07 allquot from each dally faecal collect~onwas dried to estlmate the total faecal DM production Food samples were composited withln each penod The DM content of toods and faeces was determined by drying at 80°C for 48 h and the ash content was determined bv a s h ~ n gsamples at 550°C for 6 h Neutral-detergent fibre (NDF) and acid-detergent fibre (ADF) were analysed according to Goering and Van Soest (1970) in dried foods and faeces Hem~cellulose was considered as the difference between NDF and ADF Nitrogen was determ~ned by the Kleldahl method on fresh foods and faeces Retention time in the digestive tract was measured on all the animals using chromium-mordanted forage as a solid marker: after giving 50 g of Crmordanted rice straw on day 4 at 07.00 h, 26 samples of rectal contents were collected by hand 2, 5, 8, 11, 14, 17, 20, 23, 26, 29, 33, 37, 41, 48, 53, 57, 61, 72, 82, 96, 106, 120, 130, 144, 154 and 168 h later. To calculate ruminal liquid turn-over, a pulse dose of polyethylene glycol (PEG) (40 g in 150 m1 of water) was infused into the rumen of cannulated animals on day 4 at 08.00 h and determined in 10 samples taken via ruminal cannula 1, 4, 7, 10, 13, 16, 19, 22, 25 and 28 h after the infusion. I'EG was measured by turbidimetry (Hydgn, 1955); Cr was extracted using nitric acid according to Blincoe et al. (1987) then determined by atomic emission spectrophotometry. Ruminal liquid was sampled via the ruminal cannulas using a tube placed in the ventral sac, at 07.00 (just before the meal), 09.00 and 12.00 h on days 1 and 2. The p H was immediately measured using a combination electrode after filtration through cheesecloth. A pooled sample was constituted and preserved with a solution of 1% (vol/vol) ophosphoric acid (1 ml in 10 m1 rumen liquid). On this sample, volatile fatty acids (VFA) concentration and composition were analysed by gas liquid chromatography using 4-methylvaleric acid as an internal standard. At the end of each period (day 12 at 11.00 h), the rumen of the six cows was manually emptied and weighed; a representative sample of 1 kg was dried at 80°C for 48 h for DM determination. Another 1 kg sample was used to determine the particle size of the rumen content using an analytical sieve shaker fitted with six sieves (100 pm; 200 pm; 400 pm; 800 pm; 2 mm and 4 mm). The rest of the contents were then returned to the rumen. 111 slt~r DM disavvearance of the straw was determined using polyester bags with heat-sealed edges (53 pm pore size, Internal dimensions 5 X 10 cm; Ankom, Fairport, NY). Triplicate (for each of I

5l

three B ~tzdrtus)and duplicate (for each of three R tullru5) bags with approximately 3 g (DM basis) of straw ground through a 08-mm screen were introduced ~ n t othe rumen on day 4 at 09 00 h and t e dthe rumen for 3, h, 12, 24, 48, 96 were ~ i i c ~ ~ b a In and 120 h After removal, bags were washed under tap water then rlnsed and dried at 80°C for 48 h to determme DM residue5 Blood samples were collected at the caudal vein on day 5 at 08.00 h in a heparinized vacutainer. After centrifugation (3000 X g at 4OC for 15 min) plasma was stored frozen (-20°C) until analysed using a multianalyser (Elan, Merck-Clevenot, Nogent-surMarne, France) and commercial kits for glucose (Merckotest; Merck, Nogent-sur-Marne, France), urea (Boehringer Mannheim, Meylan, France), nonesterified fatty acids (NEFA-C test, Wako, Biolyon, Lyon, France), and 3-hydroxybutyrate (Barnouin c,t al., 1986). Culculufiorzs and statist~calarlulyses Ruminal dilution rate of the liquid phase was calculated by the exponential decrease of PEG concentrations. The ruminal and total tract particle retention times were calculated from the Cr excretion curve by modelling according to Dhanoa et al. (1985). The kinetics of DM disappearance trz situ were adjusted to the exponential model: D(t) = a + b (1 cc'), where D ( t ) is the percentage of disappearance from the bag for the time t, of the rapidly (a) and slowly (b) degradable components, and c is the rate of degradation of b. Degradability D was calculated as a + bc / (c + k) where k is the particle passage rate. Theoretical degradability was calculated using the same k value (the overall mean obtained by Dhanoa's method, i.e. 0.0223) for all animals and periods. Effective degradability was calculated by taking k as the effective passage rate for each animal and each period, obtained by Dhanoa's method. Mean particle size was first determined using the formula: I: P,S',, where P, is the percentage of particles retained on sieve i (S,) and S', is the mean size of particles retained on sieve i. The size is determined for i = 200 pm to 2 mm as (S, - S, ,)/2; mean size of particles retained on the 4 m m sieve is taken as 6 mm; mean size of particles passing through the 100 pm sieve is taken as 50 pm. Medium particle size was determined by the method of Waldo et ul. (1971) by plotting the cumulative percentages of retained particles against the logarithm of sizes S', of sieves. +

1

Statistical analysis was performed by analysis of variance with the GLM procedures of Statistical Analysis Systems Institute (1987), according to the

52

Grimaud, Richard, Kanwe', Durier and Doreau

model l,,= p + g , + a + p, + gp,, + where Y was Results The chemical composition of foods did not vary the dependent vana&e, p was the overall mean, g, was the effect of genotype, B talrrlrs or B. l n ~ l r t u s( 1 = throughout the experiment and the climatic 1 or 2), a,,,, was the effect of animal nested w ~ t h ~ n variations (mean temperature, hygrometry) varied in a narrow range, so that it has been concluded that genotype ')j= 1 to 4), p, the effect of period (k = 1 to the differences observed for digestive events could 5), gpik was the interaction between genotype and be attributed to the effects of intake changes. period, and c,,,, was the error. The effect of genotype was tested taking the animal effect as the error. A p p a r e n f digcsfibilify Orthogonal contrasts were used to determine: l) effect of intake (periods 1 , 4 and 5 v.periods 2 and 3), Apparent digestibility of DM, OM, NDF, ADF and 2) adaptation to undernutrition (period 2 v. period 3), hemicellulose (Table 2) was significantly lower at 3) adaptation to refeeding (period 4 zl. period 5), 4) low intake than at high intake, for B. taurus and B. irzdicus. Apparent digestibility increased between effect of a period of undernutrition on digestion at periods L1 and L2, showing an adaptation to level HL (period 1 v. periods 4 and 5). All these contrasts are independent. When the interaction undernutrition (significant AU contrast, except for ADF). Apparent digestibility was higher after the between genotype and period was significant, undernutrition period than before so that a orthogonal contrasts were determined separately for significant HL contrast was observed, except for B. taurus and B. indicus in a second statistical analysis, not detailed in the tables. Level of hemicellulose. No adaptation to refeeding occurred significance was declared at P < 0.05. (non-significant AR contrast).

Table 2 Dry inatter ( D M ) and net ener,yy ( N E ) intake per kg Mi1'brzd apparent digestibility of DM, organic nlatter (OM), neutraldetergeiztfibre (NDFI, acid-detergeiztfibre (ADF) aizd (NDF-ADF) in cows on high and low levels ofiirtake

Periodt Item DM intake (g/ day per kg Mfl75) Bos taurus Bos indicns NE intake (MJIday per kg M'175) Bos taurus Bos indicus Apparent digestibility DM Bos taurus Bos indicus OM Bos taurus Bos ilzdiczrs NDF Bus tazlrus Bos indicus

H1

L1

L2

H2

H3

82.7 85.4

43.1 44.7

43.4 44.5

83.5 89.3

83.1 88.6

s.e.3

Statistical significance5

0.35 0.36

0.18 0.19

0.18 0.19

0.36 0.38

0.35 0.38

0.595 0,637

0.533 0.493

0.601 0,521

0.651 0.659

0,627 0.654

1.50

G X P**,I"", AU**, HL*

0.637 0.674

0.591 0-560

0.652

0.692 0.698

0.669 0.692

1.34

0.580

G X P**, I**, A US,HL**

0.652 0.674

0.612 0.596

0,685 0.595

0.699 0.716

0.683 0.696

1-50

G X P*", I**, AU*, HL*

0.652 0.669

0.594 0.609

0.672 0.594

0.723 0.707

0.681 0.697

1.83

I**, HL*

0.653 0.681

0.638 0.576

0.704 0.597

0.662 0.717

0.685 0.695

1.88

G X P*", I*", AU*

ADF Bos taurus Bos indicus

NDF-ADF Bos taurus Bos indicus

t HI, H2, H3: high level of intake before undernutrition, after 3 weeks and after 8 weeks of refeeding, respectively; L1, L2: low level of intake after 3 weeks and 8 weeks of undernutrition, respectively. $ Standard error of trealent means = -\/ (residual mean squarelir), with ,I = 4. 5 G X P, genotype X period interaction, I, level of intake, (orthogonal contrasts, HI, H2, H3 v. L1, L2), AU: adaptation to undernutrition (orthogonal contrasts, L1 71. L2), HL: consequences of a period of undernutrition on digestion at high level of intake (orthogonal contrasts, H1 v . H2, H3).

Digestion in tropical breeds of cows

53

Table 3 Ri~rrrrrrizlpool sizr: ilry rruzttrr (DM), rc~izter.,nrrd total zcvi,~htsil~rllDM cor~terrt;vurr~irzizllicyu~ddil~rtiorrrntc ilrrd urcurr ru~tziirizl rtzttvitiorr titncz; 1rreui7 total trurt rcti~rztioiitinze; irz cozcis or1 kiiylz n~zif1 0 i ~levels c!f irztnkf, Periodt

I""

0.576 0.550

0.534 0,611

0.592 0.629

0.614 0.537

0.045

G X P*

0.312 0.373

0.237 0.467

0.106 0.194

0.141 0.258

0.138

l**

0.367 0.438

0.452 0.693

0.487 0.789

0.508 0.620

0.143

I*, AU*

0.201 0.203

0.262 0.343

0.221 0.389

0.308 0.172

0.076

G X P*"

interaction, I, level o f intake, (orthogonal contrasts, HI, f-12, H3 71. L1, L2), AU: adaptation to undernutrition (orthogonal contrasts, L1 71. L2, HL: consequences of a period of undernutrition on digestion at high level of intake (orthogonal contrasts, H1 a. H2, H3).

56

Grimaud, Richard, KanwC, Durier and Doreau

genotypes was observed for any metabolite; a significant interaction between genotypes and period was noted for glucose and urea.

Discussion Genc~rale f i c f ofiilfuke or1 digesfion The decrease in energy intake induced a decrease in OM and cell wall apparent digestibility. This low digestibility results in a lower VFA concentration, accompanied by a higher pH at low intake. Based on data from experiments mostly carried out at intakes higher than or equal to maintenance (review of Tyrrell and Moe, 1975), an increase in digestibility with reduction in level of intake was expected, mainly due to a higher mean retention time of particles in the rumen. However, the absence of a change in digestibility at low intake has been observed by some of the authors who worked at levels of intake lower than maintenance (McCraham, 1964 and Keenan et al., 1969, in sheep and Agabriel et al., 1995 in cattle) and Gingins et al. (1980) even found the lowest digestibility at the lowest intake. Until now, no measurement of microbial activity or retention time in the rumen has been made to explain this absence of a change or a drop in digestibility when intake was decreased to a very low level. In this experiment, we observed paradoxically a longer particle retention time at the low levels of intake. The absence of an expected increase in digestibility in spite of this longer particle retention time is probably due to the high value for retention time at high levels of intake, so that the efficiency of digestion is unlikely to be increased by a longer retention time. Such a statement had been made by Glenn et al. (1989), Varga and Prigge (1982) and Lugirtbuhl et nl. (1994). This is consistent with the low increase in the proportion of degraded rice straw after 48 h of presence in the rumen at both high and low levels of intake. It could be supposed that variations in ruminal particle size with intake are involved. Generally, ruminal particle size decreases when intake decreases (Okine and Mathison, 1991); however an increase in duodenal particle size has been observcd by Doreau et al. (1986) at low intakes, due to disturbances in rumination. Such a result has not been observcd in this experiment. As neither a difference in particle retention time in the rumen nor a difference in particle size explain the variations in apparent digestibility, microbial activity is probably involved. With forage-based diets, in situ DM degradability either did not vary (Aitchison et al., 1986) or increased (Kabr6 et al., 1995) when intake decreased and digestibility increased. Kabre et al. (1994) showed that such an increase was due to a

higher en~yniaticactivity of cellulolytic bacteria. 111 our experiment, a slight dccrease in theoretical degradability is observed. A limitation of microbial activity may occur. The low magnitude of the decrease in theoretical degradability compared with the decrease in apparent digestibility could be due to a lack of sensitivity of the irt situ method to appreciate differences in fibrolytic activity (Nozic're and Michalet-Doreau, 1997). Moreover, the decrease in protozoa when intake is reduced (Grimaud and Doreau, 1995) is not totally taken into account by the it7 situ method. Ruminal water content, very high at both levels o f intake, can also be unfaa7ourable to the microbial activity, due to a physical change in ruminal medium. of duration o$ ~ ~ ~ d e r t l u t r i t i o~ lI l I C refeeding ~ or1 modificatior~sof d i g e s t i o ~ After the initial drop in apparent digestibility subsequent to undernutrition, apparent digestibility increased. This is consistent with previous results (Crimaud and Doreau, 1995) in which a decrease in digestibility was observed after 1 week of undernutrition, the initial digestibility being restored after 5 weeks of undernutrition then constant until 18 weeks of undernutrition. On the contrary, Ortigues and Vermorel (1996) observed a trend to a decrease in digestibility between 3 and 6 weeks of undernutrition. In the present experiment, no explanation can be found, since from L1 to L2 particle retention time and in srfu degradability decreased, both factors being associated with a decrease in apparent digestibility.

Ef?cf

A higher digestibility after an undernutrition period than before has been observed in sheep by Thornton et al. (1979) and by Perrier and Doreau (1995). Although data are lacking, this effect may be associated with an increase in efficiency of utilization of food during compensatory growth in bovine (O'Donovan, 1984). However in the present experiment, this increase in efficiency cannot be related to variations in particle retention time, microbial activity or particle size.

Cornparisorl between B. indicus nrzd B. taurus No difference in apparent digestibility between genotypes was found in our trial, where animals were fed with a high crude fibre content diet. This is consistent with the results of Duckworth (1946) who observed a higher digestibility in B. tatlrus only for diets low in fibre. Such an observation agrees with the results of Kennedy (1982), who has shown that for forage diets, OM apparent digestibility did not differ significantly between Brahman crossbred and Hereford steers, whereas ruminal OM apparent digestibility was higher for Brahman crossbred. This latter difference was attributed to a longer rumen

Digestion in tropical breeds of cows retention of less rapidly digestible dietary components in H. ir~dicus.On the contrary, Phillips et al. (1960) reported a trend to a shorter mean ruminal particle retention time for B. indicus than for R. tnurus. In this experiment, no difference was noted at either level of energy intake in both rumen and total tract retention times of the two genotypes, although several authors reported a smaller total digestive tract size of R. indicus, which is considered to be one of the characteristics responsible for the apparently higher degree of heat tolerance exhibited by the zebu (review of Schneider and Flatt, 1975). It may be concluded that, for diets high in fibre, digestive efficiency in both B. indicus and R. taurus is similar. Conclusion

This experiment has shown a clear decrease in apparent digestibility when energy intake was decreased to a level lower than maintenance, in both B. taurus and B. indicus cattle. Literature data have shown that the rble of particle retention time in digestion efficiency was not important at low intakes. However the drop in apparent digestibility observed in this trial has not been explained by two characteristics closely related to the microbial activity: in situ degradability and size of particles. Other factors which influence microbial activity such as r61e of protozoa, limiting nutritional factors, physical conditions in the rumen, should be taken into account in further experiments.

Acknowledgements The authors wish to acknowledge the assistance of S. de la Rocque and T. Lefran~oisfor help in surgery, C. Vogel for skill in making of sampling devices, G. Kam Sami, N. Nacro, B. Djiry and B. Saldou for care of the animals and help with sampling, H. Sanou, L. Sidibe, J. Chabrot and R. Lefaivre for chemical analyses. This experiment was supported by the program 'Action incitative programm6e ADELE-H' financed by INRA.

References Agabriel, J., Petit, M., Ollier, A. and Chilliard, Y. 1995. Effects of underfeeding on body reserves variations and on energy efficiency of lactating Charolais cows. Ai~nales dc Zootechnie 44: (supplement) 317. Aitchison, E. M., Gill, M. and Osbourn, D. F. 1986. The effect of supplementation with maize starch and level of intake of perennial ryegrass (Loliurn perenne CV.Endura) hay on the removal of digesta from the rumen of sheep. British Journal ofNiltritiorl 56: 477-486. Barnouin, J., El Idilbi, N., Chilliard, Y., Chacornac, J. P. and Lefaivre, R. 1986. Microdosage automatise sans deproteinisation du 3-hydroxybutyrate plasmatique chez les bovins. Anizales de Recherches Vit6rinaires 17: 129-139. Blincoe, C., Theisen, M. 0. and Stoddard-Gilbert, K. 1987. Sample oxidation procedures for the determination of

57

chromium and nickel in biological material. Corrrrrr~~rricntioils irz Soil Scic~ilicailil Plant Analysis 18: 687-697. Chilliard, Y., Doreau, M., Bocquier, F. and Lobley, G. E. 1995. Digestive and metabolic adaptations of ruminants to variations in food supply. In Rec~irl lierlrlopineizts irz tht, rlutritiorr ofl~erbioor.cs(ed. M. Journet, E. Grenet, M. H. Farce, M. Thericz and C. Dcmarquilly). IJroc.eedirrgs of tlze fourth iilterrlatiorzal syrrlposirrni or1 the iz~rtritiorl of h~~rhioorc~s, pp. 329-360. I N R A Editions, Paris. Dhanoa, M. S., Siddons, R. C., France, J. and Gale, D. L. 1985 A mult~compartmental model to describe marker excret~onpatterns m rum~nant faeces Hrrt~sh /ourilcal of Nutritron 53: 663-671 Doreau, M., Lomri, A. I. and Adingra, A. 1986. Influence d'un faible niveau d'ingestion sur la digestion et Ic comportement alimentairc chez la vache recevant un regime trPs digestible. R(,productiorr, Nutrition, Developrrleizt 26: 329-330. Duckworth, J. 1946. A statistical comparison of the influence of crude fibre on the digestibility of roughage by Bos indicus (Zebu) and Bos taurus cattle. Tropical Agriculture 23: 4-8. Galina, C. S. and Arthur, G. H. 1989. Review of cattle reproduction in the tropics. 2. Parturition and calving intervals. Ailirrlal Breedir~gsAbstracts 57: 679-686. Galyean, M. L. and Owens, F. N. 1991. Effects of diet composition and level of feed intake on site and extent of digestion in ruminants. 111 Physiological aspects of digestioiz and rrzetabolisin in ruiwinants (ed T. Tsuda, Y. Sasaki, R. Kawashima). Proceedings of seventh symposium on rutninant physiology, pp. 483-514. Academic Press, San Diego. Gingins, M., Bickel, H. and Schiirch, A. 1980. Efficiency of energy utilization in undernourished and realimented sheep. Livestock Production Scieizce 7: 465-471. Glenn, B. P., Varga, G. A., Huntington, G. B. and Waldo, D. R. 1989. Duodenal nutrient flow and digestibility in Holstein steers fed formaldehyde- and formic acid-treated alfalfa or orchardgrass silage at two intakes. /ournal of Aninlal Science 67: 513-528. Goering, H. K. and Van Soest, P. J. 1970. Forrzge fiber analysis (apparatus, reagents, procedures, and some applications). Agricultural handbook no. 379. ARS-USDA, Washington, DC. Grimaud, P. and Doreau, M. 1995 Effect of extended underfeedmg on digest~on and nitrogen balance In nonlactating cows louriral of Aiz~malSc~ence73: 211-219 Hydhn, S. 1955. A turbidimetric method for the determination of higher polyethylene glycols in biological materials. Kungliga Lantbrukshogskolarzs Arznalcr 22: 139-145. Institut National de la Recherche Agronomique. 1989 Rurnznant nutrltroi~ retorr~rnended allozuance4 a t ~ dfeed tableq John Libbey Eurotext, Paris KabrC, P., Doreau, M. and Michalet-Doreau, B. 1995 Effects of underfeedlng and of fish meal supplementation on forage digest~onm sheep Jouri~alof Agricultural Sclence, Cambrldcye124: 119-127 KabrC, P., Martin, C. and Michalet-Doreau, B. 1994 ~ erumen s sol~d-adherentmicroorganlrms Enzyme a c t ~ v ~ t of In chron~callyunderfed ewes lournal of the Scieiice of Food arld Axr~culturt~ 65: 423-428

58

Grimaud, Richard, Kanw6, Durier and Doreau

Keenan, D. M., McManus, W. R. and Freer, M. 1969. Changes in the body composition and efficiency of mature sheep during loss and regain of live weight. /oiirrri~l of A~rici~ltrrral Screlrcl., Cn~rthri~i