Genetic Relationships among Body Condition Score, Body Weight, Milk Yield, and Fertility in Dairy Cows

J. Dairy Sci. 86:2193–2204 American Dairy Science Association, 2003.

Genetic Relationships among Body Condition Score, Body Weight,Milk Yield, and Fertility in Dairy Cows

D. P. Berry,*† F. Buckley,* P. Dillon,* R. D. Evans,* M. Rath,† and R. F. Veerkamp‡*Dairy Production Department, Teagasc,Moorepark Production Research Centre, Fermoy, Co. Cork, Ireland†Department of Animal Science, Faculty of Agriculture,University College Dublin, Belfield, Dublin 4, Ireland‡Institute for Animal Science and Health (ID-Lelystad),P.O. Box 65, 8200 AB Lelystad, The Netherlands

ABSTRACT

Genetic (co)variances between body condition score(BCS), body weight (BW), milk production, and fertility-related traits were estimated. The data analyzed in-cluded 8591 multiparous Holstein-Friesian cows withrecords for BCS, BW, milk production, and/or fertilityfrom 78 seasonal calving grass-based farms throughoutsouthern Ireland. Of the cows included in the analysis,4402 had repeated records across the 2 yr of the study.Genetic correlations between level of BCS at differentstages of lactation and total lactation milk productionwere negative (−0.51 to −0.14). Genetic correlations be-tween BW at different stages of lactation and total lac-tation milk production were all close to zero but becamepositive (0.01 to 0.39) after adjusting BW for differencesin BCS. Body condition score at different stages of lacta-tion correlated favorably with improved fertility; ge-netic correlations between BCS and pregnant 63 d afterthe start of breeding season ranged from 0.29 to 0.42.Both BW at different stages of lactation and milk pro-duction tended to exhibit negative genetic correlationswith pregnant to first service and pregnant 63 d afterthe start of the breeding season and positive geneticcorrelations with number of services and the intervalfrom first service to conception. Selection indexes inves-tigated illustrate the possibility of continued selectionfor increased milk production without any deleteriouseffects on fertility or average BCS, albeit, genetic meritfor milk production would increase at a slower rate.(Key words: body weight, body condition score, fertil-ity, selection index)

Abbreviation key: AVGBCS = average body conditionscore; AVGBW = average body weight; Cumfat240 =cumulative fat yield to d 240 of lactation; Cummilk120,

Received October 17, 2002.Accepted December 19, 2002.Corresponding author: Donagh P. Berry; e-mail: dberry@

moorepark.teagasc.ie.

2193

Cummilk180, Cummilk240 = cumulative milk yieldto d 120, 180, and 240 of lactation, respectively; Cum-prot240 = cumulative protein yield to d 240 of lactation;CVg = coefficient of genetic variation; DairyMIS =Dairy Management Information System; FSCO = firstservice to conception interval; HUK = Holstein UnitedKingdom; IFS = interval to first service; NS = numberof services per cow; PR63 = pregnant 63 d after thestart of the breeding season; PRFS = pregnant tofirst service.

INTRODUCTION

Until recently breeding programs worldwide withinthe Holstein-Friesian breed have been based almostentirely on increased milk production per cow. Littleor no emphasis was placed on ancillary traits relatingto health and reproduction efficiency. It has now beenrecognized that selection in dairy cattle solely for highmilk production is generally accompanied by reducedfertility (Hoekstra et al., 1994; Grosshans et al., 1997;Royal et al., 2000; Roxstrom et al., 2001; Evans et al.,2002; Royal et al., 2002a) and reduced health (Eman-uelson et al., 1988; Pryce et al., 1998). It is for thisreason that most countries have begun to include traits,other than those associated with milk production, intheir selection indexes (Philipsson et al., 1994; Visscheret al., 1994; Heringstad et al., 2000; Veerkamp et al.,2002). Although progress towards increased milk pro-duction may be reduced, these selection indexes suggestthat better overall economic efficiency will be obtainedwhen functional nonproduction traits are included inselection objectives. Many factors however, hinder theinclusion of fertility and health traits in a selectionobjective, most notably the lack of available data andtheir low heritability (Hoekstra et al., 1994; Grosshanset al., 1997; Dechow et al., 2001; Veerkamp et al., 2001).However, Philipsson (1981) suggested that ample addi-tive genetic variation exist for fertility traits to warranttheir inclusion in breeding objectives. Nevertheless, in-terest is accruing in indicator traits that 1) can be more

BERRY ET AL.2194

easily recorded, 2) can be measured early in life, and3) possess a coheritability that is larger than the herita-bility of the fertility/health traits. Potentially interest-ing indicator traits include BCS and BW. Body condi-tion score and BW both have been shown to exhibitmoderate heritabilities (Gallo et al., 2001; Pryce et al.,2001; Veerkamp et al., 2001; Berry et al., 2002; Royalet al., 2002b). The heritabilities for BCS change andBW change tend to be lower (Pryce et al., 2001; Berryet al., 2002; Dechow et al., 2002).

Most estimates of genetic (co)variances between BCS,BW, and fertility traits have been derived from datasetscontaining small numbers of animals with repeatedBCS/BW observations (Veerkamp et al., 2000; Pryce etal., 2001) or large numbers of animals with a singleBCS/BW observation per animal (Moore et al., 1992;Veerkamp et al., 2001; Royal et al., 2002b). These stud-ies indicated negative genetic correlation between BCSand fertility-related interval traits (Dechow et al., 2001;Pryce et al., 2001; Veerkamp et al., 2001; Royal et al.,2002b). Dechow et al. (2002) reported that cows thatare genetically inclined to lose more BCS in early lacta-tion will have a prolonged calving to first service inter-val. Many studies (Darwash et al., 1997; Royal et al.,2000) have used milk progesterone assays as indicatorsof the commencement of luteal activity. Using thismethodology, Royal et al. (2002b) reported a significantgenetic relationship between commencement of lutealactivity and BCS; they suggested that each unit in-crease in BCS (scale 1 to 9) would bring forward thecommencement of luteal activity by on average approxi-mately 6 d. Veerkamp et al. (2001) reported that BCSin early lactation showed a stronger positive geneticcorrelation with first-service conception rate and 56 dnonreturn rate to first service than BCS in later lac-tation.

Few studies have estimated genetic correlations be-tween BW and fertility traits. Moore et al. (1992) esti-mated a low but favorable genetic correlation (−0.07)between BW at calving and days open in first-lactationHolstein cows. A similar genetic correlation was ob-served by Veerkamp et al. (2000) on first lactation ani-mals using milk progesterone assays as an indicator ofthe commencement of luteal activity. This correlationwas strongest for BW at wk 15 (−0.54). Veerkamp etal. (2000) also reported a strong negative genetic corre-lation (−0.80) between BW change in the first 15 wk oflactation and commencement of luteal activity. How-ever, neither of the above studies adjusted BW for dif-ferences in BCS prior to estimating the genetic correla-tions between BW with fertility despite BCS being re-ported to explain 12 to 45% of the genetic variationin BW (Veerkamp and Brotherstone, 1997; Berry etal., 2002).

Journal of Dairy Science Vol. 86, No. 6, 2003

The objective of this study was to estimate genetic(co)variances between BCS, BCS change, BW, BWchange, milk production, and fertility traits on a largenumber of animals with several BCS, BW, and milktest-day observations per animal and assess the subse-quent effects of alternative selection objectives on thegenetic response for each trait.

MATERIALS AND METHODS

This study, carried out over 2 yr (1999 and 2000),was comprised of 78 spring-calving dairy herds (74 com-mercial and four research herds) in the south of Irelandwith a potential 8928 spring-calving cows available forinclusion in the data set. Herd size ranged from 30 to240 cows. All herds were incorporated into the DairyManagement Information System (DairyMIS) run byMoorepark (Crosse, 1986). The DairyMIS is a recorder-based computerized system collecting detailed stock,farm inputs, production, and reproduction informationon a monthly basis.

Pedigree Information

Of the cows available 48% were herd book registeredwith Holstein United Kingdom (HUK). Four genera-tions of ancestry on the paternal and maternal sidewere identified for 92 and 66% of these cows, respec-tively. For the remaining 52% of the cows not registeredwith HUK, sire and maternal grand sire were obtainedfrom DairyMIS. For these herds the paternal ancestryand maternal grand sire ancestry was provided by HUKto the same level as for the pedigree cows. The propor-tion of North American Holstein-Friesian genetics foreach sire/maternal grand sire present in the data setwere calculated as outlined by Berry et al. (2002).

Within the edited data set, there were 818 differentsires with daughters. The number of daughters per sireranged from 1 to 415, and the average was 10.1 daugh-ters per sire. In total 6340 of the cows with records hadidentified maternal grandsires. The additive genetic re-lationship matrix included 20,613 animals.

BCS, BW and Milk Production Traits

A total of 7250 cows from 66 herds had recordedspring calving dates and greater than three BCS or BWrecords or both. Body condition score was measured ona scale of 1 (thin) to 5 (fat) with increments of 0.25(Lowman, 1976). Body condition score and BW at d 5,60, 120, 180, and 240 (BCS only) were estimated foreach cow using a smoothing spline as outlined by Berryet al. (2002). Average BCS (AVGBCS) and average BW(AVGBW) were calculated as the mean of these esti-

BODY CONDITION SCORE, WEIGHT, AND FERTILITY 2195

mated test day records, in which the animal had allrecords present; this minimized bias by ensuring allcows had the same number of BCS/BW records evenlydispersed throughout the lactation. Body conditionscore change and BW change were calculated as thedifference between the traits of interest (where animalshad an estimate for both test days). Cows with missingvalues for any variable were coded as such but remainedin the analyses for the rest of the variables.

A total of 8541 cows from 78 herds had recordedspring calving dates and milk records. Test-day recordsfor each cow were obtained from the Irish Dairy Re-cording Cooperative. Three cumulative milk yield traitsto d 120, 180, and 240 (Cummilk120, Cummilk180,Cummilk240) of lactation, fat yield to d 240 of lactation(Cumfat240), and protein yield to d 240 of lactation(Cumprot240) were also derived fitting a smoothingspline for each cow (Berry et al., 2002).

Reproductive Traits

Five fertility variables similar to those used interna-tionally (Grosshans et al., 1997; Veerkamp et al., 2001;Evans et al. 2002) were calculated: interval to first ser-vice (IFS), first service to conception interval (FSCO),pregnant to first service (PRFS), number of servicesper cow (NS), and pregnant 63 d after the start of thebreeding season (PR63). The start of the breeding sea-son for each herd was defined as the first service daterecorded in that herd; start of breeding dates wereavailable for both years of the study. In total 8315 cowshad identified first-service records. In Ireland mostfarmers use AI for the first 6 wk of the breeding seasonand natural mating thereafter. Ninety-two percent offarmers observed cows more than twice daily for estrusduring the breeding season, while 99% of farmers usedtail paint or a vasectomized bull or both as an aid toestrus detection. This facilitated all services to be accu-rately recorded. Beginning 40 to 50 d after the start ofthe breeding season all herds were visited on threeor four occasions, at approximately 40-d intervals, toperform pregnancy diagnosis by transrectal ultrasoundimaging (Aloka 210D * II, 7.5 MH3). Cows that wereinseminated at least 28 d and not observed in estrusagain after insemination were scanned to confirm preg-nancy. Subsequently, all cows in the study were deter-mined to be pregnant or not by rectal palpation at least56 d after the end of the defined breeding season.

Data Analysis

A multivariate analysis for all 26 traits simultane-ously was not computationally feasible. For this reason,a series of bivariate analyses were carried out in AS-


REML (Gilmour et al., 2002). Herd-season groups wereformed. Season was defined as month of calving. Herd-season groups with less than four cows had their recordsmoved into an adjoining season group from the sameherd to facilitate a more accurate estimate. Holsteinpercentage of the cows in the present study varied from0 to 75%. Average Holstein percentage was 53%. How-ever, the maximum Holstein percentage a cow on thestudy may have is 75% as the maternal granddam isassumed to have 0% Holstein genes. The latter wasassumed because most maternal granddams, and theirproportion of Holstein-Friesian genes were unknown,in addition the base population in Ireland prior to themideighties was predominantly British Friesian. Leastsquares means for all traits were calculated using AS-REML for cows with 0 and 75% Holstein genes. Thenumber of cows with 0 and 75% Holstein genes were426 and 1457, respectively. The output also suppliedthe standard error of the difference, which was used tocalculate the significance of the Holstein effect at the5% level (1.96 * standard error of the difference).

The average length of the breeding season was 15wk across the 78 herds in both years of the study. Thisranged from 9 to 25 wk for individual herds. Senatoreet al. (1996) showed that PRFS was positively relatedto the number of ovulations prior to first service. Thenumber of services per cow will be influenced by thelength of the breeding season in each herd. For thesereasons, a quadratic polynomial regression for both thenumber of days between calving and the start of thebreeding season and between calving and the end ofthe breeding season were added to the model whenanalyzing the fertility traits.

The following linear animal model was used for theunivariate and bivariate analysis of all fertility re-lated traits:

Yijkpq = µp + HYSj + lk + ∑2

t=1

btHolt + ∑2

m=1

bmSOBm

+ ∑2

n=1

bnEOBn + ai + PEi + eijkpq

where

Yijkpq = observation for trait p on animal i,µp = overall mean for trait p,

HYSj = fixed effect of herd by year by month ofcalving interaction (j = 574),

lk = fixed effect of lactation number (k = 1,2, 3, 4+),

= fixed effect of a quadratic polynomial∑2

t=1

btHolt

regression for the percentage of NorthAmerican Holstein-Friesian genes,

BERRY ET AL.2196


m=1

bmSOBm

regression for the number of days be-tween calving and the start of the breed-ing season,


n=1

bnEOBn

regression for the number of days be-tween calving and the end of the breed-ing season,

ai = random additive genetic effect,PEi = random permanent environmental ef-

fect for each cow, andeijkpq = random residual term.

The model applied to the nonfertility traits was thesame except the quadratic regression on calving to thestart of the breeding season and the quadratic regres-sion on calving to the end of the breeding season werenot included. Additive genetic, permanent environmen-tal, and residual covariances between traits were esti-mated from the bivariate analysis.

Following the analyses, the genetic correlations esti-mated between all 26 traits were incorporated into a26 × 26 matrix. Due to the large number of bivariateanalysis carried out, some of the eigenvalues of thecorrelations matrix were negative and were thereforemade positive, and the correlation matrix recalculatedusing the eigenfunctions (Hill and Thompson, 1978).The matrix was made positive definite to facilitate sub-sequent selection index calculations. In this new posi-tive definite matrix, 77% of the correlations hadchanged by less than 0.05, and 95% of the correlationschanged by less than 0.10. The standard errors, how-ever, were not adjusted since the change in correlationswere so small and are thus likely to have little effecton the standard errors of the correlations. In the presentstudy, estimates of those genetic correlations are pre-sented.

Adjustment of a trait (e.g., trait 1) for differences inanother trait (e.g., trait 2) in the present study wasachieved by including trait 2 as a covariate for trait 1in the model and the parameters subsequentlyreestimated.

Selection Index Methodology

The effects of four alternative selection objectives onresponses per trait were studied using selection indextheory (Hazel, 1943). Traits of interest were milk yield,fat yield, protein yield, AVGBW, AVGBCS, IFS, andPR63. The four alternative selection objectives were 1)selection for increased milk production based on theyield index (shown below); 2) selection on the yield indexwith an economic weight of −20% on AVGBW (relativeto protein yield in genetic SD terms); 3) selection on


the yield index with an economic weight on AVGBCSto restrict change in PR63; 4) selection on the yieldindex with an economic weight on PR63 to obtain nochange in PR63. All four selection objectives assumedthat milk, fat, and protein yields were available from100 half-sib daughter groups. For selection objective 2,it was assumed that daughters also had records forAVGBW, while selection objective 3 assumed measure-ments on all daughters for AVGBCS, as well as theproduction records. Selection objective 4 was derivedas if records were available for PR63 on all daughters,as well as the production records.

The yield index was defined as milk yield + fat yield+ protein yield with chosen economic values consistentwith those currently adopted in the economic breedingindex for Ireland (Veerkamp et al., 2002); €−0.076,€0.90, and €5.70 for milk yield, fat yield, and proteinyield, respectively. The accuracy of the breeding valuefor PR63 was also calculated over different-sized half-sib daughter groups with three alternative selectionindexes (a) PR63, (b) AVGBCS, and (c) PR63 +AVGBCS; the breeding objective was PR63. The geneticand phenotypic parameters used for the calculation ofthe optimal index weights in all index calculations werethose observed in the present study. The vector of opti-mal index weights (b) was calculated for each of theobjectives as b = P−1Ga where P−1 = the inverse ofthe phenotypic (co)variance matrix of the traits in theselection index, G = the genetic covariance matrix be-tween traits in the selection goal and the selection in-dex, and a = the vector containing the economic valuesfor the goal traits.

The genetic change per trait from selection on each

of the four objectives was calculated as Rj = i*b′*Gj

σ iwhere Rj = a vector with the genetic change per traitj, i = selection intensity (in the present study this wasassumed to equal one), b′ = transpose of the vectorcontaining the index weights, Gj = the jth column of aG matrix containing the genetic covariances betweenthe trait j and the index traits; and σi = the standarddeviation of the index used. The expected response fromselection using the objectives is illustrated in the pres-ent study as genetic gain following one cycle of selectionwith a standardized selection intensity of one.

RESULTS

The number of observations, least squares means,genetic standard deviations, heritabilities, permanentenvironmental effects, and coefficients of genetic varia-tion are summarized for the production and fertilitytraits in Tables 1 and 2, respectively. The sum of IFS


Table 1. Number of observations, least squares means, genetic SD (σg), heritability (h2), permanent environ-mental effect (c2), their SE, and a coefficient of genetic variation (CVg) for each of the BCS, BW, and milkproduction traits.

Trait1 No Obs Mean σg h2 c2 CVg

CS5 9840 3.08 0.20 0.29 (0.041) 0.06 (0.037) 6.4CS60 10,061 2.92 0.19 0.43 (0.048) 0.19 (0.044) 6.6CS120 9557 2.93 0.18 0.43 (0.050) 0.20 (0.046) 6.3CS180 9735 2.94 0.19 0.41 (0.047) 0.12 (0.043) 6.6CS240 9307 2.96 0.20 0.34 (0.044) 0.04 (0.039) 6.7AVGBCS 8523 2.97 0.19 0.58 (0.056) 0.13 (0.052) 6.5CS60-5 9750 −0.16 0.08 0.07 (0.020) 0.06 (0.025)CS240-5 8953 −0.11 0.11 0.07 (0.016) 0.00 (0.000)BW5 9575 544.7 33.0 0.39 (0.047) 0.25 (0.044) 6.1BW60 9828 546.2 32.8 0.53 (0.053) 0.29 (0.051) 6.0BW120 9357 562.9 34.3 0.57 (0.055) 0.27 (0.053) 6.1BW180 9513 576.3 32.2 0.45 (0.050) 0.29 (0.047) 5.6AVGBW 8725 558.3 33.9 0.60 (0.057) 0.28 (0.055) 6.1BW60-5 9496 0.5 8.1 0.06 (0.017) 0.06 (0.024)BW180-5 9133 33.8 9.1 0.05 (0.016) 0.12 (0.024)Milk60 12,372 28.4 1.8 0.21 (0.034) 0.33 (0.032) 6.2Cummilk120 12,263 3160 195 0.24 (0.036) 0.42 (0.034) 6.2Cummilk180 11,945 4326 278 0.26 (0.038) 0.44 (0.036) 6.4Cummilk240 10,909 5218 352 0.28 (0.041) 0.44 (0.039) 6.7Cumfat240 8847 202 18.4 0.44 (0.050) 0.23 (0.047) 9.1Cumprot240 8847 180 11.9 0.32 (0.045) 0.38 (0.043) 6.6

1CS5, CS60, CS120, CS180, CS240 = BCS on d 5, 60, 120, 180, and 240 of lactation, respectively; AVGBCS= average BCS; CS60-5, CS240-5 = change in BCS between the respective test days; BW5, BW60, BW120,BW180 = body weight on d 5, 60, 120, and 180 of lactation, respectively; AVGBW = average BW; BW60-5,BW180-5 = change in body weight between the respective test days; Milk60, = milk test day yield on d 60of lactation, Cummilk120, Cummilk180, Cummilk240 = cumulative milk yields to d 120, 180, and 240 oflactation, respectively; Cumfat240 = cumulative fat yield to d 240 of lactation; Cumprot240 = cumulativeprotein yield to d 240 of lactation.

and FSCO equate to the number of days open (90 d),which corresponds to a calving interval of 373 d (theaverage gestation length was 283 d). Heritabilities forBCS and BW were larger than those reported for milkyield. Heritability estimates for the fertility traits wereall less than 0.03, yet a considerable genetic variationexisted for some of these traits; FSCO and PRFS bothshowed a coefficient of genetic variation (CVg) ofgreater than 10%, while IFS exhibited the lowest CVg(2.4%). The CVg for BCS, BW, and milk productionwere all less than 7% with the exception of Cumfat240,which was 9.1%. No permanent environmental vari-ance existed for IFS and PR63 with the univariate anal-ysis; hence, this component was not included in subse-quent bivariate analysis for either of the two traits.

Table 2. Number of observations, least squares means, genetic SD (σg), heritability (h2), permanent environ-mental effect (c2), their SE, and a coefficient of genetic variation (CVg) for each of the fertility traits.

Trait1 No Obs Mean σg h2 c2 CVg (%)

IFS (d) 12,262 72.8 1.72 0.02 (0.008) 0.000 (0.000) 2.4FSCO (d) 10,661 16.7 2.71 0.01 (0.008) 0.001 (0.021) 16.2PRFS (%) 12,262 0.49 0.051 0.01 (0.007) 0.025 (0.018) 10.5NS (Number) 12,262 1.8 0.15 0.02 (0.010) 0.002 (0.020) 8.7PR63 (%) 12,262 0.70 0.060 0.02 (0.008) 0.000 (0.000) 8.5

1IFS = Interval to first service; FSCO = first service to conception interval; PRFS = pregnant to firstservice; NS = number of services per cow; PR63 = pregnant 63 d after the start of the breeding season.


Milk yield, fat yield, and protein yield were signifi-cantly (P < 0.05) higher for animals with 75% Holsteingenes (5252 kg, 202 kg, and 180 kg for Cummilk240,Cumfat240, and Cumprot240, respectively) over ani-mals with no Holstein genes (4957 kg, 191 kg and 171kg for Cummilk240, Cumfat240, and Cumprot240, re-spectively). However, Holstein percentage had no sig-nificant effect on any of the fertility traits, even thoughanimals with 75% Holstein genes tended to have longerIFS (73.2 d vs. 72.7 d), longer FSCO (17.1 d vs. 16.8 d),lower PRFS (48% vs. 50%), greater NS (1.80 vs. 1.78),and poorer PR63 (70% vs. 72%) than animals with noHolstein genes.

Genetic correlations among BCS and BW at the samedays in milk varied from 0.37 to 0.47. Table 3 shows

BERRY ET AL.2198

Table 3. Genetic correlations (and their SE) between milk productionwith BCS and BCS change.

Trait1 Cummilk240 Cumfat240 Cumprot240

CS5 −0.33 −0.14 −0.27(0.10) (0.10) (0.10)

CS60 −0.49 −0.32 −0.46(0.08) (0.08) (0.08)

CS120 −0.51 −0.33 −0.48(0.08) (0.08) (0.08)

CS180 −0.50 −0.35 −0.45(0.08) (0.08) (0.08)

CS240 −0.50 −0.35 −0.44(0.08) (0.09) (0.09)

CS60-5 −0.27 −0.43 −0.45(0.15) (0.14) (0.14)

CS240-5 −0.47 −0.45 −0.53(0.07) (0.07) (0.07)

AVGBCS −0.46 −0.30 −0.43(0.08) (0.08) (0.08)

1CS5, CS60, CS120, CS180, CS240 = BCS on d 5, 60, 120, 180,and 240 of lactation, respectively; AVGBCS = average BCS; CS60-5,CS240-5 = change in BCS between the respective test days; Cum-milk240 = cumulative milk yield to d 240 of lactation; Cumfat240 =cumulative fat yield to d 240 of lactation; Cumprot240 = cumulativeprotein yield to d 240 of lactation.

the genetic correlations between BCS and BCS changeat different stages of lactation with milk production.All correlations were negative and ranged from −0.53 to−0.14 and were stronger than those previously reportedfrom a subset of the data (Berry et al., 2002).

The genetic correlations between BW with milk pro-duction were all close to zero (−0.07 to 0.09). However,after adjusting BW for differences in BCS, all geneticcorrelations between BW and milk production becameslightly to moderately positive (0.01 to 0.39).

The genetic correlations between BCS, BW, and milkproduction with the five fertility traits are summarizedin Tables 4, 5, and 6, respectively. Irrespective of thestage of lactation, BCS was positively correlated withPRFS and PR63 and negatively correlated with IFS andNS. Body condition score was genetically uncorrelated

Table 4. Genetic correlations (and their SE) between BCS and BCS change with fertility traits.

Trait1 CS5 CS60 CS120 CS180 CS240 CS60-5 CS240-5 AVGBCS

IFS −0.33 −0.38 −0.33 −0.44 −0.42 0.06 −0.06 −0.37(0.10) (0.08) (0.09) (0.09) (0.09) (0.18) (0.17) (0.08)

FSCO −0.15 0.07 0.00 0.03 −0.07 0.23 0.07 0.02(0.23) (0.17) (0.18) (0.18) (0.22) (0.27) (0.29) (0.17)

PRFS 0.46 0.28 0.34 0.26 0.24 −0.22 −0.20 0.34(0.21) (0.24) (0.24) (0.25) (0.26) (0.30) (0.26) (0.24)

NS −0.52 −0.39 −0.40 −0.33 −0.35 0.17 0.17 −0.42(0.11) (0.11) (0.12) (0.13) (0.14) (0.23) (0.19) (0.11)

PR63 0.41 0.29 0.37 0.39 0.42 −0.19 0.00 0.35(0.15) (0.13) (0.13) (0.13) (0.14) (0.23) (0.22) (0.13)

1CS5, CS60, CS120, CS180, CS240 = BCS on d 5, 60, 120, 180, and 240 of lactation, respectively; AVGBCS= average BCS; CS60-5, CS240-5 = change in BCS between the respective test days; IFS = interval to firstservice; FSCO = first service to conception interval; PRFS = pregnant to first service; NS = number of servicesper cow; PR63 = pregnant 63 d after the start of the breeding season.


with FSCO. Body weight throughout lactation consis-tently exhibited negative genetic correlations with IFS,PRFS, and PR63, while positive genetic correlationswere evident between BW with both FSCO and NS.Adjustment of BW for differences in milk yield had verylittle effect on the genetic correlations between BW andfertility with the exception of BW in mid- to late lacta-tion, which became more negatively correlated withIFS.

The direction of the correlations between milk pro-duction with the fertility traits were similar to thoseobserved between BW with the fertility traits (i.e., nega-tive correlations between milk production with PRFSand PR63 and positive correlations between milk pro-duction with FSCO and NS). However, unlike BW, totallactation milk production showed no correlations withIFS (−0.09 to 0.04). The directions of the correlationswere similar irrespective of whether Cummilk240,Cumfat240, or Cumprot240 was applied. The pheno-typic correlations between BCS, BW, and milk produc-tion with the fertility traits were all low (−0.12 to 0.09)and are not presented here.

Selection Indexes

The phenotypic and genetic correlations used to cal-culate the optimal index weights are shown in Table 7,while the expected responses from selection on the fouralternative selection objectives are summarized in Ta-ble 8. Selection on the yield index alone is predictedto reduce AVGBCS but have little effect on AVGBW,suggesting that the reduction in BCS is compensatedby an increase in body size. Interval to first service isexpected to decrease, as is PR63 following selection onthe yield index alone.

The expected reduction in AVGBCS and AVGBW asa consequence of selection for increased milk productionwas intensified when a relative economic weight of


Table 5. Genetic correlations (and their SE) between BW and BW change with fertility traits.

Trait1 BW5 BW60 BW120 BW180 BW60-5 BW180-5 AVGBW

IFS −0.28 −0.19 −0.22 −0.29 0.28 0.10 −0.25(0.11) (0.09) (0.09) (0.10) (0.17) (0.19) (0.09)

FSCO 0.32 0.33 0.34 0.37 0.08 0.20 0.37(0.19) (0.17) (0.17) (0.19) (0.29) (0.31) (0.17)

PRFS −0.11 −0.23 −0.26 −0.26 −0.26 −0.38 −0.22(0.25) (0.23) (0.23) (0.23) (0.28) (0.27) (0.23)

NS 0.07 0.14 0.19 0.23 0.26 0.30 0.15(0.15) (0.13) (0.13) (0.14) (0.22) (0.24) (0.13)

PR63 −0.18 −0.23 −0.22 −0.21 −0.23 −0.16 −0.24(0.13) (0.12) (0.12) (0.13) (0.23) (0.25) (0.11)

1BW5, BW60, BW120, BW180 = Body weight on d 5, 60, 120, and 180 of lactation, respectively; AVGBW= average BW; BW60-5, BW180-5 = change in body weight between the respective test days; IFS = intervalto first service; FSCO = first service to conception interval; PRFS = pregnant to first service; NS = numberof services per cow; PR63 = pregnant 63 d after the start of the breeding season.

−20% was applied to AVGBW despite little effect on theresponse in milk production. However, the expecteddecrease in IFS from selection on the yield index alonewas reduced when BW was selected against, whilePR63 continued to deteriorate.

An economic weight of 45% (relative to protein yieldin genetic SD terms) on BCS was required within theselection objective so as to have no effect on PR63. Thisselection objective reduced the expected response inprotein and milk yield by 19 and 50%, respectively, overa selection objective based on the yield index alone.

To restrict the expected response in PR63 to zero—assuming only the three milk production traits andPR63 are in the index—it was necessary to apply aneconomic weight of 36% to PR63 in the selection objec-tive (relative to protein yield in genetic SD terms). Thisimplies that similar relative economic weighting in ge-netic standard deviation terms would have to be appliedto either BCS or PR63 to restrict change in PR63 tozero. This relative weight is similar to that applied tomilk yield in the index assuming a progeny group sizeof 100 half-sib daughters. Expected responses in both

Table 6. Genetic correlations (and their SE) between milk production with fertility traits.

Trait1 Milk60 Cummilk240 Cumfat240 Cumprot240

IFS 0.04 −0.01 −0.08 −0.09(0.11) (0.10) (0.09) (0.10)

FSCO 0.33 0.31 0.20 0.27(0.23) (0.21) (0.19) (0.20)

PRFS −0.18 −0.29 −0.31 −0.37(0.26) (0.25) (0.24) (0.22)

NS 0.44 0.46 0.24 0.37(0.12) (0.18) (0.18) (0.13)

PR63 −0.18 −0.22 −0.31 −0.17(0.15) (0.14) (0.13) (0.15)

1Milk60 = milk test day yield on d 60 of lactation, Cummilk240 = cumulative milk yield to d 240 oflactation; Cumfat240 = fat yield to d 240 of lactation; Cumprot240 = protein yield to d 240 of lactation; IFS= interval to first service; FSCO = first service to conception interval; PRFS = pregnant to first service; NS= number of services per cow; PR63 = pregnant 63 d after the start of the breeding season.


AVGBW and AVGBCS were negative using this selec-tion index.

The inclusion of only PR63 in the selection indexserved as a better predictor (based on the accuracy ofthe selection index as a representation of the selectiongoal) of PR63 in the selection objective than the inclu-sion of only AVGBCS in the index, when the numberof daughters with records was greater than 18 daugh-ters per sire (Figure 1). However, a combined index ofPR63 and BCS consistently produced a more accurateestimate of the breeding value for PR63 up to 100daughters per sire than an index with only PR63.

DISCUSSION

The objective of this study was to estimate the geneticparameters for BCS, BW, milk production, and fertility-related traits and to estimate the expected response toselection from alternative selection objectives. Resultsfrom this study suggest that a favorable, moderate tostrong genetic relationship exists between BCS withfertility, while increased BW or milk production tended

BERRY ET AL.2200

Figure 1. Accuracy of the breeding value for pregnant 63 d afterthe start of the breeding season (PR63) as a function of progenygroup size for daughters that have records for either PR63 (�), bodycondition score (�), or PR63+body condition score (×).

to reduce PRFS and PR63 and increase the NS andFSCO. The animals observed across the 74 commercialherds in the present study were an accurate representa-tion of the Irish national dairy herd; thus, genetic pa-rameters estimated in the present study could apply innational genetic evaluations.

Variance Components

Heritability estimates from the present study arevery similar to most other international studies, whichlooked at BCS (Veerkamp and Brotherstone, 1997;Gallo et al., 2001; Pryce et al., 2001), BCS change (Pryceet al., 2001; Dechow et al., 2002), BW (Veerkamp andBrotherstone, 1997; Veerkamp et al., 2000; Sønder-gaard et al., 2002), milk yield (Veerkamp and Broth-erstone, 1997; Pryce et al., 1998; Dechow et al., 2001),and fertility traits (Grosshans et al., 1997; Pryce etal., 1998). However, Royal et al. (2002a) reported largeheritability estimates for some endocrine fertility pa-rameters and stressed their potential for inclusion ina selection index to improve fertility. The lack of a per-manent environmental effect for IFS and PR63 may bedue to the seasonal calving system adopted in Ireland(i.e., animals with short IFS, which conceive to thatservice in 1 yr, will have long IFS in the following yearto maintain the seasonal calving pattern, and cows,which do not become pregnant 63 d after the start ofthe breeding season, are more likely to be culled). Thesubstantial CVg for all the fertility traits with the possi-ble exception of IFS were in very close agreement withthe CVg estimated for comparable fertility traits withina similar seasonal calving system adopted in NewZealand (Grosshans et al., 1997). This confirms thatdirect selection for fertility may prove beneficial (Phil-


ipsson, 1981). Similar to most other studies (Hoekstraet al., 1994; Roxstrom et al., 2001; Evans et al., 2002)higher genetic merit for milk production was associatedwith a longer FSCO, a greater NS, poorer PRFS, andfinally lower PR63. Genetic correlations between milkproduction and IFS were all close to zero (−0.09 to 0.04)although the standard errors were large (0.09 to 0.11).These correlations are weaker than those reported byprevious authors (Hoekstra et al., 1994; Roxstrom etal., 2001) between IFS and milk yield (0.27 to 0.44).

Genetic Correlations with BCS

The negative genetic correlations between BCS withmilk, fat, and protein yield (−0.51 to −0.14) agrees withprevious studies (Pryce et al., 2001; Veerkamp et al.,2001). There was a tendency for BCS measured in earlylactation to give the weakest correlations with milkproduction, which is consistent with results from Veer-kamp et al. (2001) using random regression models.Based on these genetic correlations, a cow with a supe-rior genetic merit for Cummilk240 of 1000-kg milk, willhave a 0.25 BCS unit lower average BCS than a cowwith a genetic merit of 0 kg for milk. Genetic correla-tions between BCS change in early lactation and milkproduction (−0.45 to −0.27) agree with those from De-chow et al. (2002); the discrepancy in the signs of thecorrelations may be explained by the different traitdefinitions of the BCS loss traits. Therefore, if selectionfor higher milk production alone continues, the geneticmerit for BCS will reduce continuously, and manage-ment practices may have to be altered to compensatefor the deleterious genetic effects.

A biological interpretation of the negative genetic cor-relation between BCS and milk production is the appar-ent relationship between BCS with energy balance andtissue mobilization (Pryce and Løvendahl, 1999). Be-cause body tissue may be used in part to fuel milkproduction, a moderate to strong antagonistic geneticcorrelation between BCS and milk production is there-fore expected.

The favorable genetic correlations between BCS andfertility observed in the present study agree with previ-ous phenotypic (Domecq et al., 1997) and genetic stud-ies (Dechow et al., 2001; Pryce et al., 2001; Veerkamp etal., 2001). Adjusting these relationships for phenotypicmilk yield had no effect on the direction of the correla-tions between BCS with IFS, PRFS, NS, and PR63;correlations involving BCS with FSCO were all closeto zero. Thus, regardless of yield, cows with low BCSwill exhibit poorer fertility, suggesting that genes asso-ciated with body tissue mobilization may have pleio-trophic effects or be closely linked to genes controllingfertility in animals. Royal et al. (2002b) suggested that


Table 7. Phenotypic (above the diagonal) and genetic (below the diagonal) correlations used in the calculationof the optimal index weights.

Traits1 Protein yield Fat yield Milk yield AVGBW AVGBCS IFS PR63

Protein yield 0.73 0.88 0.20 −0.11 −0.01 −0.01Fat yield 0.62 0.62 0.14 −0.12 −0.01 −0.02Mik yield 0.73 0.29 0.14 −0.18 0.01 −0.04AVGBW −0.03 0.03 −0.01 0.44 −0.07 0.01AVGBCS −0.43 −0.30 −0.46 0.44 −0.11 0.09IFS −0.09 −0.08 −0.01 −0.25 −0.37 −0.21PR63 −0.17 −0.31 −0.22 −0.24 0.35 −0.18

1Protein yield = cumulative protein yield to day 240 of lactation; fat yield = cumulative fat yield to day240 of lactation; milk yield = cumulative milk yield to day 240 of lactation; AVGBW = average BW; AVGBCS= average BCS; IFS = interval to first service interval; PR63 = pregnant 63 d after the start of the breedingseason.

the pathways in which these correlations express them-selves could be through either 1) hormones controllingintermediary metabolism having a direct effect on ovar-ian function, or 2) reproductive hormones, which regu-late ovarian function having a direct effect on interme-diary metabolism. Thus, low BCS resulting from selec-tive breeding for higher milk yield may alter circulatinghormonal levels of growth hormone and corticosteroids,while simultaneously reducing circulating levels of in-sulin and insulin-like growth factor-1, both of whichmay have deleterious effects on subsequent fertility.Similarly, low BCS as a consequence of selection forhigher yield may alter the level of circulating reproduc-tive hormones such as gonadotrophins resulting in pos-sible implications for subsequent fertility (Royal et al.,2002b). However, evidence is still not clear, and severalalternative pathways are possible.

Genetic correlations, involving BCS change and fer-tility, expressed large standard errors and therefore

Table 8. Effect of four different selection objectives on individual responses per trait. The response isexpressed assuming a standardized selection differential of one.

Selection objective1

Traits2 Objective 1 Objective 2 Objective 3 Objective 4

Protein yield (kg) 10.3 10.1 8.3 9.9Fat yield (kg) 14.2 13.7 12.3 12.5Milk yield (kg) 147 144 74 125AVGBW (kg) −0.7 −8.2 6.0 −1.9AVGBCS (BCS units) −0.06 −0.08 0.03 −0.05IFS (d) −0.20 −0.10 −0.49 −0.24PR63 (%) −0.009 −0.006 0.000 0.000

1Objective 1: selection for increased milk production based on a yield index (1.0 protein yield + 0.2 fatyield − 0.4 milk yield) with records on daughters for milk production; Objective 2: selection on the yieldindex with a weight of −20% on BW (relative to protein yield in genetic SD terms) in the selection objectivewith production records and BW records on daughters; Objective 3: selection on the yield index with anoptimum economic value on BCS to obtain no change in PR63 and with production and BCS records availableon daughters; Objective 4: selection on the yield index with an optimum economic value on PR63 to obtainno change in PR63 and with records on daughters for milk production and PR63.

2Protein yield = cumulative protein yield to day 240 of lactation; fat yield = cumulative fat yield to day240 of lactation; milk yield = cumulative milk yield to day 240 of lactation; AVGBW = average BW; AVGBCS= average BCS; IFS = interval to first service interval; PR63 = pregnant 63 d after the start of the breedingseason.


make interpretation of the results difficult. Neverthe-less, the genetic correlation between change in BCSbetween d 5 and 60 and NS was similar to that reportedby Dechow et al. (2002) on similar traits in second lacta-tion animals. Based on the genetic parameters reportedin the present study between BCS and fertility, an in-crease in genetic merit for average BCS of 1.00 BCSunit will reduce the IFS by 3 d, increase PRFS by 9percentage units, reduce the NS by 0.32, and increasethe PR63 by 11 percentage units.

Genetic Correlations with BW

The near zero genetic correlations between BW andmilk production signify that selection for milk yieldhas a negligible genetic effect on the BW of an animal.However, when BW was adjusted for differences inBCS, all genetic correlations between BW and milkproduction became positive; correlations between

BERRY ET AL.2202

AVGBW and the milk production traits ranged from0.15 to 0.39. These correlations agree more closely withcorrelations estimated between measures of size andmilk production; Brotherstone (1994) estimated geneticcorrelations between stature and milk production ofbetween 0.16 and 0.25. Therefore, although selectionfor increased milk production may increase cow size,the reduction in BCS associated with such an increasein milk production will ultimately result in a negligiblegenetic effect of milk production on BW.

Genetic correlations, estimated between BW and thefertility traits, indicated that although geneticallyheavier cows are served sooner, they require more ser-vices and have a longer interval from first service toconception, which may be due in part to their inferiorPRFS. All these factors combine to give lower PR63.However, the genetic correlations involving PRFS, NS,and PR63 did possess large standard errors althoughthe direction of the correlations is consistent acrossthe different BW test-day records. The correlations alsoagree with a long-term selection experiment conductedin Minnesota, where cows selected for high body sizetended to require more services per conception thancows selected for low body size (Hansen et al., 1999).The negative correlations between BW and IFS are con-sistent with those reported by Veerkamp et al. (2000)between BW and commencement of luteal activity be-fore and after adjusting for genetic merit in milk pro-duction. Following the adjustment of BW for differencesin milk yield, the genetic correlations between adjustedBW and IFS were stronger for BW in midlactation thanfor BW in early lactation, which also agrees with Veer-kamp et al. (2000). Moore et al. (1992) reported aslightly negative genetic correlation (−0.07) betweenBW at calving and days open in Holsteins, which hasbeen shown to be a similar trait to IFS (Jansen, 1985;Grosshans et al., 1997).

Although a considerable proportion of the variationin BW is due to BCS (Veerkamp and Brotherstone,1997; Berry et al., 2002), the difference in signs of thegenetic correlations between BW with PRFS, NS, andPR63 to the genetic correlations observed between BCSwith PRFS, NS, and PR63 indicate that factors otherthan BCS also contribute to the genetic variation inBW. This was substantiated when BW was adjustedfor differences in BCS, and the genetic correlations be-tween BW and fertility were reestimated. The signsof the correlations remained the same although thestrength of the correlations increased (with the excep-tion of the correlations between BW and IFS, whichbecame slightly weaker).

Selection Indexes

It is important to bear in mind that the expectedresponses from selection are influenced by the parame-


ters used in calculating the optimal index weights. Thestandard errors of some of the genetic correlations esti-mated were large; therefore, care should be taken ininterpreting the magnitude in selection responses.Lindhe and Philipsson (1998) reported large effects onexpected genetic gains per trait when genetic correla-tions were assumed between the traits in the index asopposed to zero genetic correlations. In agreement withprevious studies, selection for milk yield alone is likelyto reduce BCS (Pryce et al., 2001; Veerkamp et al., 2001;Pryce et al., 2002) and pregnancy rates (Grosshans etal., 1997). It is therefore imperative that selection in-dexes incorporate other nonproduction traits to mini-mize the antagonistic effect on correlated traits thatmay prove to have economical or ethical implicationsin the future.

Selection for reduced BW as a means of reducingmaintenance costs has been adopted in the breedingobjectives of some countries (Visscher et al., 1994; Mont-gomerie, 2002). In a national progeny-testing programthe number of BW records available is likely to be low.However, the genetic correlations between AVGBW andBW at different stages of lactation were all greater than0.95; this indicates that one BW measurement maysuffice for inclusion in a national selection objective.The heritability for individual BW test-day recordswere lower than for AVGBW, which implies that ahigher weight on BW might be required to achieve thesame genetic gain, when only one BW record is availablenationally. Similar conclusions are applicable to BCSmeasured only once in a national progeny-testingprogram.

In the present study, applying a negative weight toBW in the selection objective reduced the favorabletrend in IFS over selection on the yield index alone,while BCS was also predicted to decline further. Thelarger decline in BCS from selecting for reduced BW isdue to the moderate genetic correlations between BCSand BW found in the present study. Because body tissuereserves act as a biological buffer making up the deficitin energy required for milk synthesis, a continued re-duction in genetic merit for BCS may in time becomethe major limiting factor in response to selection forincreased milk production. Nevertheless, genetic corre-lations estimated in the present study between produc-tion and BCS were less than an absolute value of 1,suggesting that simultaneous selection for increasedmilk production while increasing or maintaining BCSat its current level is possible.

The ability to select for increased milk productionwithout a corresponding deterioration in BCS was high-lighted when an economic weighting of 45% relative toprotein yield was applied to BCS. Although a reductionin BCS of −0.06 BCS units from selection on the produc-


tion index appears small, the inclusion of BCS in theselection objective with a positive economic value em-phasizes the role that BCS mobilization plays in achiev-ing high milk yield. Substituting the genetic parame-ters for AVGBCS with those for BCS on d 5 of lactationhad very little effect on the economic weighting neces-sary to achieve similar results; the economic weightingapplied to BCS relative to protein yield was increasedfrom 45 to 47%.

The slight negative effect that the first two selectionobjectives had on PR63 may indicate the need for theinclusion of a fertility-related trait in the selection ob-jective. PR63 would be a possible fertility trait for inclu-sion because it measures the ability of an animal toconceive early in the breeding season—a criteria ofgreat importance in seasonal calving systems. In thepresent study, PR63 had one of the largest heritabilitiesof the fertility traits investigated and also had a largeCVg (8.5%). However, PR63 is not a routinely availablemeasure for the Irish dairy cow population. Neverthe-less, the current economic index in Ireland includessurvivability (Veerkamp et al., 2002), which is likely tobe correlated with PR63, since animals not pregnanthave a greater chance of being culled; infertility ac-counted for 50 and 46% of the total culling reasons onDairyMIS farms for 1999 and 2000, respectively. Thefinal selection objective illustrates that it is possible torestrict the expected change in PR63 to zero by includ-ing PR63 in the selection objective with an economicweight of 36% relative to protein yield. This objectivewas achieved at the expense of a slight reduction (4%)in the expected response in protein yield over selectionfor increased production alone and a continued declinein both BCS and BW. Selection objective 4 is more effi-cient than selection objective 3, when records are avail-able on 100 daughters per sire, since the expected re-sponse in milk production is greater in the former. How-ever, with smaller progeny group sizes BCS is a betterindicator of PR63 than PR63 itself (Figure 1).

CONCLUSIONS

It can be concluded from the present study that al-though the heritability for fertility traits is low, thereis sufficient genetic variation present to allow directgenetic improvement in fertility with large progenygroup sizes. However, there is a cost associated withlarge progeny-testing schemes. Also, small progenygroup sizes may lead to inaccurate breeding values forfertility traits, reducing the overall response to selec-tion if included in a selection objective. It is for thesereasons that indicator traits are of interest for breedingvalue estimation for fertility. In the present study thecoheritability of BCS and BW with fertility was larger


than the heritability for most of the individual fertilitytraits, signifying that with small progeny group sizesfaster rates of genetic improvement may result if indi-rect selection for fertility is practiced. The continualselection for increased milk production with no accounttaken of BCS will result in lower genetic levels for BCSthroughout lactation, which will have deleterious ef-fects on fertility, since BCS was shown to be favorablycorrelated with improved fertility. Body weight wasnegatively correlated with PRFS and PR63, while itwas positively correlated with NS and FSCO. The inves-tigation of different selection objectives showed the pos-sibility of selecting for increased milk production, whilesimultaneously maintaining PR63 at its current level,by including either BCS or PR63 in the selection objec-tive with positive economic weightings.

ACKNOWLEDGMENTS

The authors wish to acknowledge with gratitude Al-lied Irish Bank, the AI managers Association, the Hol-stein-Friesian Society of Great Britain and Ireland,Dairy Levy Farmer Funds, and EU Structural Funds(FEOGA) in financing the research program. The tech-nical assistance of D. Cliffe, T. Condon, and J. Keneally,and the guidance of Dorian Garrick in the initial stagesof the study is also acknowledged.

REFERENCES

Berry, D. P., F. Buckley, P. Dillon, R. D. Evans, M. Rath, and R. F.Veerkamp. 2002. Genetic parameters for level and change of bodycondition score and body weight in dairy cows. J. Dairy. Sci.85:2030–2039.

Brotherstone, S. 1994. Genetic and phenotypic correlations betweenlinear type traits and production traits in Holstein-Friesian dairycattle. Anim. Prod. 59:183–188.

Crosse, S. 1986. Dairy herd management information system—DAIRYMIS II. Pages 325–361 in Part 2: Animal Health and Ma-chine Milking. Moorepark 25th Anniversary Publ.

Darwash, A. O., G. E. Lamming, and J. A. Woolliams. 1997. Estima-tion of genetic variation in the interval for calving to postpartumovulation of dairy cows. J. Dairy Sci. 80:1227–1234.

Dechow, C. D., G. W. Rogers, and J. S. Clay. 2001. Heritabilities andcorrelations among body condition scores, production traits, andreproductive performance. J. Dairy Sci. 84:266–275.

Dechow, C. D., G. W. Rogers, and J. S. Clay. 2002. Heritability andcorrelations among body condition score loss, body condition score,production and reproductive performance. J. Dairy Sci.85:3062–3070.

Domecq, J. J., A. L. Skidmore, J. W. Lloyd, and J. B. Kaneene. 1997.Relationship between body condition scores and conception atfirst artificial insemination in a large dairy herd of high yieldingHolstein cows. J. Dairy Sci. 80:113–120.

Emanuelson, U., B. Danell, and J. Philipsson. 1988. Genetic parame-ters for clinical mastitis, somatic cell counts, and milk productionestimated by multiple-trait restricted maximum likelyhood. J.Dairy Sci. 71:467–476.

Evans, R. D., F. Buckley, P. Dillon, and R. F. Veerkamp. 2002. Geneticparameters for production and fertility in spring-calving Irishdairy cattle. Irish J. Agric. Food Res. 41:43–54.

BERRY ET AL.2204

Gallo, L., P. Carnier, M. Cassandro, R. Dal Zotto, and G. Bittante.2001. Test-day genetic analysis of condition score and hearthgirth in Holstein-Friesian Cows. J. Dairy Sci. 84:2321–2326.

Gilmour, A. R., B. R. Cullis, S. J. Welham, and R. Thompson. 2002.ASREML Reference Manual. New South Wales Agriculture, Or-ange Agricultural Institute, Orange, NSW, Australia.

Grosshans, T., Z. Z. Xu, L. J. Burton, D. L. Johnson, and K. L. Macmil-lan. 1997. Performance and genetic parameters for fertility ofseasonal dairy cows in New Zealand. Livest. Prod. Sci. 51:41–51.

Hansen, L. B., J. B. Cole, G. D. Marx, and A. J. Seykora. 1999.Productive life and reasons for disposal of Holstein cows selectedfor large versus small body size. J. Dairy Sci. 82:795–801.

Hazel, L. N. 1943. The genetic basis for constructing selection indexes.Genetics. 28:476–490.

Heringstad, B., G. Klemetsdal, and J. Ruane. 2000. Selection formastitis resistance in dairy cattle: A review with focus on thesituation in the Nordic countries. Livest. Prod. Sci. 64:95–106.

Hill, W. G., and R. Thompson. 1978. Probabilities of non-positivedefinite between-group or genetic covariance matrices. Biometrics34:429–439.

Hoekstra, J., A. W. van der Lugt, J. H. J. van der Werf, and W.Ouweltjes. 1994. Genetic and phenotypic parameters for milkproduction and fertility in upgraded dairy cattle. Livest. Prod.Sci. 40:225–232.

Jansen, J. 1985. Genetic aspects of fertility in dairy cattle based onanalysis of A.I. data—a review with emphasis on areas for furtherresearch. Livest. Prod. Sci. 12:1–12.

Lindhe, B., and J. Philipsson. 1998. Genetic correlations betweenproduction with disease resistance and fertility in dairy cattleand consequences for total merit selection. Acta Agric. Scand.48:216–221.

Lowman, B. G., N. Scott, and S. Somerville. 1976. Condition scoringof cattle. Bull. of the East Scotland College of Agriculture, No.6. rev. ed.

Montgomerie, B. 2002. Subject: Cow fertility and breeding objectives.http://www.aeu.org.nz/ext7.html. Accessed May 29, 2002.

Moore, R. K., B. W. Kennedy, L. R. Schaeffer, and J. E. Moxley.1992. Relationships between age and body weight at calving, feedintake, production, days open, and selection indexes in Ayrshiresand Holsteins. J. Dairy Sci. 75:294–306.

Philipsson, J. 1981. Genetic aspects of female fertility in dairy cattle.Livest. Prod. Sci. 8:307–319.

Philipsson, J., G. Banos, and T. Arnason. 1994. Present and futureuses of selection index methodology in dairy cattle. J. Dairy Sci.77:3252–3261.

Pryce, J. E., M. P. Coffey, S. H. Brotherstone, and J. A. Woolliams.2002. Genetic relationships between calving interval and bodycondition score conditional on milk yield. J. Dairy Sci.85:1590–1595.

Pryce, J. E., M. P. Coffey, and G. Simm. 2001. The relationshipbetween body condition score and reproductive performance. J.Dairy Sci. 84:1508–1515.


Pryce, J. E., R. J. Esslemont, R. Thompson, R. F. Veerkamp, M. A.Kossaibati, and G. Simm. 1998. Estimation of genetic parametersusing health, fertility and production data from a managementrecording system for dairy cattle. Anim. Sci. 66:577–584.

Pryce, J. E., and P. Løvendahl. 1999. Options to reduce vulnerabilityto metabolic stress by genetic selection. Pages 119–128 in Meta-bolic Stress in Dairy Cows. J. D. Oldham, G. Simm, A. F. Groen,B. L. Nielsen, J. E. Pryce, and T. L. J. Lawrence, ed. Br. Soc.Anim. Sci. Occasional Publ. No. 24.

Roxstrom, A., E. Strandberg, B. Berglund, U. Emanuelson, and J.Philipsson. 2001. Genetic and environmental correlations amongfemale fertility traits and milk production in different parities ofSwedish red and white dairy cattle. Acta Agric. Scand. 51:7–14.

Royal, M. D., A. O. Darwash, A. P. F. Flint, R. Webb, J. A. Woolliams,and G. E. Lamming. 2000. Declining fertility in dairy cattle:Changes in traditional and endocrine parameters of fertility.Anim. Sci. 70:487–501.

Royal, M. D., A. P. F. Flint, and J. A. Woolliams. 2002a. Geneticand phenotypic relationships among endocrine and traditionalfertility traits and production traits in Holstein-Friesian dairycows. J. Dairy Sci. 85:958–967.

Royal, M. D., J. E. Pryce, J. A. Woolliams, and A. P. F. Flint. 2002b.The genetic relationship between commencement of luteal activ-ity and calving interval, body condition score, production, andlinear type traits in Holstein-Friesian dairy cattle. J. Dairy Sci.85:3071–3080.

Senatore, E. M., W. R. Butler, and P. A. Oltanacu. 1996. Relationshipsbetween energy balance and post-partum ovarian activity andfertility in first lactation dairy cows. Anim. Sci. 62:17–23.

Søndergaard, E., M. K. Sørensen, I. L. Mao, J. Jensen. 2002. Geneticparameters of production, feed intake, body weight, body composi-tion, and udder health in lactating dairy cows. Livest. Prod. Sci.77:23–34.

Veerkamp, R. F., and S. Brotherstone. 1997. Genetic correlationsbetween linear type traits, food intake, live weight and conditionscore in Holstein-Friesian dairy cattle. Anim. Sci. 64:385–392.

Veerkamp, R. F., P. Dillon, E. Kelly, A. R. Cromie, and A. F. Groen.2002. Dairy cattle breeding objectives combining yield, survivaland calving interval for pasture-based systems in Ireland underdifferent milk quota scenarios. Livest. Prod. Sci. 76:137–151.

Veerkamp, R. F., E. P. C. Koenen, and G. De Jong. 2001. Geneticcorrelations among body condition score, yield, and fertility infirst-parity cows estimated by random regression models. J. DairySci. 84:2327–2335.

Veerkamp, R. F., J. K. Oldenbroek, H. J. Van Der Gaast, and J. H.J. Van Der Werf. 2000. Genetic correlations between days untilstart of luteal activity and milk yield, energy balance, and liveweights. J. Dairy Sci. 83:577–583.

Visscher, P. M., P. J. Bowman, and M. E. Goddard. 1994. Breedingobjectives for pasture based dairy production systems. Livest.Prod. Sci. 40:123–137.

Genetic Relationships among Body Condition Score, Body Weight, Milk Yield, and Fertility in Dairy Cows

Documents