Acta agriculturae Slovenica, 121/4, 1–10, Ljubljana 2025 doi:10.14720/aas.2025.121.4.13600 Original research article / izvirni znanstveni članek Genetic parameters for growth traits in the Slovenian beef cattle popu- lation Barbara LUŠTREK 1, 2, Milena KOVAČ 1, Klemen POTOČNIK 1 Received May 18, 2023; accepted November 23, 2025. Delo je prispelo 18. maja 2023, sprejeto 23. novembra 2025. 1 University of Ljubljana, Biotechnical Faculty, Department of Animal Science, Ljubljana, Slovenia 2 Corresponding author, e-mail: barbara.lustrek@bf.uni-lj.si Genetic parameters for growth traits in the Slovenian beef cattle population Abstract: This study compared genetic parameter es- timates and correlations between estimated breeding values (EBVs) obtained using four single-trait animal models differ- ing in the definition of the common herd-time environmental effect. Analyses were based on field performance records for birth weight (BW), weight at 90 days (W90), and weight at 210 days (W210) in Charolais and Limousin calves born in Slo- venian suckler herds between 1990 and 2017. Model variants defined contemporary groups by herd only (H), or by herd combined with one-year (HY), two-year (HY2), or five-year (HY5) time periods. Heritability estimates varied across mod- els: for BW, direct h² ranged from 0.25 to 0.39, for W90 from 0.11 to 0.37, and for W210 from 0.34 to 0.57. Correspond- ing maternal h² ranged from 0.10 to 0.13 (BW), 0.12 to 0.18 (W90), and 0.20 to 0.29 (W210). The proportion of variance due to common herd-time effects ranged from 0.25 to 0.44, and EBV correlations between models ranged from 0.61 to 0.95. The HY2 model provided the most balanced partition- ing of phenotypic variance across traits and herds, providing the most practical solution for national genetic evaluations under the current data structure. Key words: beef cattle, Charolais, Limousin, suckler cows, growth traits, genetic evaluation, heritability, models, model comparison, Slovenia Ocena genetskih parametrov za lastnosti rasti v populaciji slovenskega mesnega goveda Izvleček: V raziskavi smo primerjali ocene genetskih parametrov in korelacije med plemenskimi vrednostmi (PV), ocenjenimi s štirimi različnimi eno-lastnostnimi modeli živa- li, ki so se razlikovali v definiciji skupnega okoljskega vpliva črede skozi čas. Analize so temeljile na podatkih terenskih meritev za telesno maso ob rojstvu (BW), pri 90 dneh (W90) in pri 210 dneh starosti (W210) telet pasem šarole in limuzin, rojenih v slovenskih čredah krav dojilj med letoma 1990 in 2017. Primerjalne skupine so bile definirane kot: samo čre- da (H), ali čreda v kombinaciji z enoletnim (HY), dvoletnim (HY2) oziroma petletnim (HY5) časovnim obdobjem. Ocen- jeni dednostni deleži so bili: za BW 0,25–0,39 (direktni) in 0,10–0,13 (maternalni); za W90 0,11–0,37 in 0,12–0,18; za W210 0,34–0,57 in 0,20–0,29. Delež variance zaradi skupne- ga okoljskega vpliva črede je znašal 0,25–0,44, korelacije med PV različnih modelov pa 0,61–0,95. Model HY2 je zagoto- vil najbolj uravnoteženo razdelitev fenotipske variance med vplive in se je izkazal kot najustreznejša izbira za uporabo v nacionalnem genetskem vrednotenju pri obstoječi strukturi podatkov. Ključne besede: mesno govedo, šarole, limuzin, dojilje, lastnosti rasti, genetsko vrednotenje, heritabiliteta, modeli, primerjava modelov, Slovenija Acta agriculturae Slovenica, 121/4 – 20252 B. LUŠTREK et al. 1 INTRODUCTION Charolais (CHA) and Limousin (LIM) are the most important specialised beef breeds in Slovenia, where suckler herds are managed extensively, primar- ily relying on grazing during the vegetation season (e.g., Krupová et al. (2025)). The main commercial products of these herds are weaned calves, with birth weight (BW), weight at 90 days (W90; corresponding to the onset of grazing), and weaning weight at 210 days (W210; end of grazing) routinely recorded for national genetic evaluations. Among these, W90 and W210 are particularly relevant for assessing both the calf ’s genet- ic potential for growth and the dam’s maternal ability, while W210 directly affects breeder income (Simčič and Čepon, 2007; Madsen et al., 2025). Phenotypic expres- sion of calf growth is influenced by both genetic and en- vironmental sources of variation. The genetic contribu- tion is typically partitioned into direct additive genetic effects and maternal additive genetic effects, the latter acting through the dam’s genotype as well as the ma- ternal environment she provides (Willham, 1972; Koch, 1971; Assan, 2025; Daneshi et al., 2025). These maternal effects are particularly pronounced in suckler systems, where calves rely entirely on their dams for milk (e.g., Davies & Denholm, 2025; Gellatly et al., 2025). Environmental influences on calf growth traits in- clude both permanent and temporary effects, such as climatic and nutritional variation, as well as measure- ment error, which reduces precision in genetic studies (Falconer, 1989). In the context of Slovenian extensive herds, important sources of shared environmental vari- ation also arise from common herd environment and year of birth. The year effect typically captures annual fluctuations in climate and vegetation, affecting milk production and thus early growth. The common herd environment includes management practices such as feeding, health care, calving protocols, culling strate- gies, and grazing routines, which are shared by all ani- mals within a herd (Troxel & Simon, 2007; Hasan et al., 2024; Edwards et al., 2025). To account for both spatial and temporal variation in management and environ- mental conditions, many genetic evaluation models use a herd-year interaction effect as an environmental grouping factor (Robinson, 1991; Phocas & Laloë, 2004; Schenkel et al., 2024). Such grouping enables more ac- curate separation of environmental and genetic effects, thereby improving the estimation of breeding values (EBVs). The objective of this study was to assess how dif- ferent specifications of the common environmental ef- fect – defined as herd, herd-year, herd-2 years, or herd-5 years (collectively referred to as herd(-time)) – influ- ence the estimation of genetic parameters and breeding values for early growth traits in Slovenian beef cattle. Specifically, we evaluated how these alternative tempo- ral definitions of the common herd effect affect vari- ance partitioning and EBV rankings. 2 MATERIALS AND METHODS Genetic parameters and EBVs were estimated for BW, W90, and W210 using field test records from the Slovenian national routine genetic evaluation scheme for beef cattle. The analysed dataset included Charolais (CHA) and Limousin (LIM) calves of both sexes, born between 1990 and 2017 and reared in suckler herds un- der extensive conditions. All three traits were recorded in accordance with ICAR guidelines (International agreement of recording practices, 2018). The data struc- ture was comparable across breeds and sexes. Calvings predominantly occurred in late winter and spring, with animals typically kept on pasture from May to October. The parity effect was categorised into four groups based on parity number and calving per- formance, reflecting current national evaluation prac- tice: first- and second-parity cows were assigned to separate classes, cows in their third to ninth parity were assigned to class 3, and those in their tenth or higher parity to class 4. The number of records per parity class was as follows: for BW, 993 (class 1), 859 (class 2), 2529 (class 3), and 159 (class 4); for W90, 737, 609, 1880, and 112; and for W210, 949, 792, 2447, and 157, respectively. The data originated from 57 herds of varying sizes. While three herds had more than 500 animals, most contained fewer than 100 animals. The breeding pro- gramme initially involved the largest herd, with others joining progressively over time, resulting in an increase in average herd size during the study period. Due to the predominance of small herds, a single sire was used in 60.4–62.0% of herds per year, and two sires in up to 19.0% of herds (Figure 1). Larger herds typically used artificial insemination and multiple sires, thereby con- tributing most to genetic connectedness across herds. Genetic ties between herds were assessed based on ped- igree links and sire usage, confirming sufficient related- ness to support across-herd evaluation. Although phenotypic data were available for ani- mals born between 1990 and 2017, analyses of sire dis- tribution (Figure 1) were limited to 1995–2017 to reflect stable recording practices. Pedigree information used for the genetic evaluation of all three traits spanned five generations of ancestors (Table 1), including some ani- mals originating from other breeds. Adjusted weights for W90 and W210 were calcu- Acta agriculturae Slovenica, 121/4 – 2025 3 Genetic parameters for growth traits in the Slovenian beef cattle population lated by interpolation to the exact ages of 90 and 210 days, in accordance with ICAR standards. Throughout this paper, W90 and W210 refer to these pre-adjusted weights. The number of records available for analysis was similar for BW (4540) and W210 (4345), as both traits are mandatory in routine recording, while few- er observations were available for the optional W90 (3338) (Table 2). Birth weight was assumed to corre- spond to age 0 days, as calves are weighed within 24 hours of birth. The average trait values were 43.8 kg for BW, 142.3 kg for W90, and 257.9 kg for W210. 2.1 MODEL STRUCTURE Four alternative single-trait animal models were tested to analyse sources of variation in each trait. These included the model currently used in the nation- Figure 1: Distribution of sires across herds (birth years are denoted by sequential numbers: 1 = 1995, …, 23 = 2017) N Breed LIM CHA Other Total Animals 3060 3126 151 6337 Sires 265 273 10 548 Dams 781 847 141 1769 Table 1: Pedigree structure N: number; LIM: Limousin; CHA: Charolais; Other: other breeds Trait Number of records Breed Weight (kg) Age (days) LIM CHA x SD x SD T 4540 2333 2207 43.8 6.6 0 0 BW M 2208 1126 1082 45.2 6.7 F 2332 1207 1125 42.5 6.5 T 3338 1846 1492 142.3 11.9 87.8 9.4 W90 M 1653 875 778 145.7 12.1 88.2 9.4 F 1685 971 714 139.1 11.8 87.5 9.4 T 4345 2181 2164 257.9 16.0 204.0 14.3 W210 M 2165 1082 1083 266.5 16.3 205.6 14.3 F 2180 1099 1081 249.4 15.8 202.5 14.2 Table 2: Data structure and descriptive statistic BW: birth weight; W90: weight at 90 days; W210: weight at 210 days; T: total; M: male; F: female; x: mean; SD: standard deviation Acta agriculturae Slovenica, 121/4 – 20254 B. LUŠTREK et al. fect, direct additive genetic effect (a), maternal additive genetic effect (m), and residual error (e). Depending on the model variant, the random common environmental effect was defined either as a single herd effect (hm) or a herd-time interaction effect defined for each calendar year (hym), two- year period (hy2m), or five-year period (hy5m). In the H model, a single group was defined per herd over the entire period, without any time division. The HY5 and HY2 models defined the herd–year interaction using five-year and two-year periods, respectively. The most detailed structure was ap- plied in the HY model, where a sepa- rate group was defined for each calen- dar year. The HY model corresponds to the model used for BW in the na- tional genetic evaluation, while the H model is used for W90 and W210 (Ocena plemenskih vrednosti, 2018). The response variable (yijklmn) repre- sented the observed trait value (BW, W90, or W210, in kg) for each animal. The distributions of BW, W90, and W210 were approximately normal. The (co)variance structure included direct and maternal additive genetic effects, their covariance, com- mon herd-time effects (depending on the model), and residual error. Genetic effects were assumed to follow a multivariate normal distribution with variances pro- portional to the additive relationship matrix, while en- vironmental and residual effects were assumed uncor- related and homoscedastic. 2.2 CONTEMPORARY GROUP STRUCTURE As expected, increasing the fragmentation of the time component resulted in smaller contemporary groups. The median group size clearly reflected this pattern, ranging from 5 to 27 animals per herd for BW, 4 to 16.5 for W90, and 5 to 22 for W210 (Table 4). This pattern indicates that most herds were small, with only a few large herds forming the upper range of group sizes. The number of contem- porary groups increased, while their median size decreased, as the tempo- ral definition of the herd effect became more detailed. The impact was most pro- nounced in smaller herds, whereas larger herds contributed disproportionately to the overall data structure in all model variants. This high variability in group size reflects the heterogeneous and unbalanced structure of the Slovenian beef popula- tion, particularly in terms of herd size distribution. Model Model equation HY yijklmn = μijklmn + Bi + Sj + Lk + Pl + hym + aijklmn + mijklmn + eijklmn HY2 hy2m HY5 hy5m H hm Table 3: Model description HY: herd-year; HY2: herd-2 years; HY5: herd-5 years; H: herd al evaluation. The overall model structure was identical across variants, differing only in the definition of the common herd(-time) environmental effect. All models included the same fixed effects: breed, sex, year of birth, and parity class. Random effects included the direct ad- ditive genetic effect, the maternal additive genetic ef- fect, and the residual. In addition, a random common herd(-time) effect was included to account for shared environmental influences. The statistical model included both fixed and ran- dom effects, with the main distinction among model variants being the definition of the common environ- mental grouping (Table 3). Fixed effects were breed (Bi; i = 1, 2), sex (Sj; j = 1, 2), year of birth (Lk; k = 1995, …, 2017), and parity class (Pl; l = 1–4). Random effects included the common herd(-time) environmental ef- Trait Model N Groups Number of animals within group Median Min Max BW HY 427 5 1 77 HY2 254 9 1 142 HY5 142 13 1 291 H 54 27 4 965 W90 HY 345 4 1 64 HY2 220 6 1 127 HY5 122 10 1 267 H 50 16.5 4 850 W210 HY 422 5 1 76 HY2 264 7 1 125 HY5 149 12 1 276 H 57 22 4 985 Table 4: Number and size of contemporary groups BW: birth weight; W90: weight at 90 days; W210: weight at 210 days; N: number; Min: minimum; Max: maximum Acta agriculturae Slovenica, 121/4 – 2025 5 Genetic parameters for growth traits in the Slovenian beef cattle population 2.3 ESTIMATION PROCEDURES (Co)variance components were estimated for each trait using the restricted maximum likelihood (REML) method implemented in the VCE-6 software package (Kovač & Groeneveld, 2008). Based on these estimates, EBVs were obtained using best linear unbiased predic- tion (BLUP) under a single-trait model. The same pedi- gree file was used for all models to ensure comparabil- ity. To assess the impact of different definitions of the common herd(-time) effect on sire ranking, Pear- son correlation coefficients between sire EBVs were calculated. Only sires with EBV accuracy ≥ 0.30 were included in the comparison. Descriptive statistics and correlation analyses were performed using SAS/STAT software (SAS 9.4 Institute Inc., Cary, NC, USA). 3 RESULTS AND DISCUSSION This study aimed to evaluate the impact of model- ling the common herd(-time) environment on the esti- mation of genetic parameters and breeding values for early growth traits in Slovenian Charolais and Limousin calves. Using national field data, we analysed three eco- nomically relevant traits; birth weight (BW), weight at 90 days (W90), and weight at 210 days (W210), under four model variants differing in the temporal definition of the common herd environment (herd only, herd- year, herd-2 years, and herd-5 years). The analysis re- vealed that the choice of herd(-time) grouping notably affected the partitioning of phenotypic variance and EBV ranking. Among the tested models, the two-year herd-time interaction (HY2) provided the best balance between precision and interpretability of genetic pa- rameters. To explore how different model structures influ- ence genetic evaluation, we compared estimates of ad- ditive and maternal variances, common environmental effects, and residual components across traits. Addition- ally, correlations between estimated breeding values (EBVs) from different models were analysed to assess their robustness under alternative grouping schemes. Variance component estimates for each trait and model combination are presented in Table 5. The esti- mates are reported in the original measurement units (kg²) and include direct additive genetic variance (σ²ₐ), maternal additive genetic variance (σ²m), com- mon herd(-time) environmental variance (σ²c), and residual variance (σ²ₑ). Where available, standard er- rors (SEs) are shown in parentheses. Heritabilities and genetic correlations are summarised in Table 6, includ- ing direct heritability (h²ₐ), maternal heritability (h²ₘ), the proportion of variance explained by the common herd(-time) effect (c²), residual proportion (e²), and di- rect–maternal additive genetic correlation (rₐₘ). 3.1 BIRTH WEIGHT (BW) Phenotypic variance (σ²ₚ) for BW was similar across all models, ranging from 27.5 to 32.3 kg², with Trait Model σ²ₚ σ²c σ²ₐ σ²ₘ σ²e BW HY 27.5 6.8 (1.0) 10.8 (1.6) 3.5 (0.8) 11.5 (0.9) HY2 30.5 10.4 (1.5) 8.5 (1.4) 3.0 (0.7) 13.0 (0.8) HY5 32.2 12.4 (1.9) 7.9 (1.3) 3.1 (0.7) 13.5 (0.8) H 32.3 12.3 (2.6) 9.2 (1.3) 3.9 (0.7) 13.4 (0.8) W90 HY 604.8 264.9 (31.0) 66.7 (27.1) 70.1 (17.4) 266.1 (17.3) HY2 567.7 224.8 (30.4) 127.5 (33.2) 78.9 (18.0) 259.0 (19.3) HY5 561.1 220.8 (37.0) 155.5 (37.1) 90.8 (18.6) 261.3 (21.0) H 558.9 205.1 (47.9) 205.9 (38.0) 99.2 (18.2) 250.5 (21.3) W210 HY 1752.0 734.6 (86.2) 600.7 (109.7) 409.5 (58.1) 574.1 (55.4) HY2 1861.9 809.6 (100.7) 636.0 (112.5) 376.6 (56.7) 616.9 (60.4) HY5 1750.1 709.7 (113.3) 806.7 (112.0) 410.7 (54.7) 590.5 (55.0) H 1611.9 527.4 (114.8) 919.0 (106.6) 466.7 (58.0) 601.8 (57.1) Table 5: Estimates of variances for body weight at birth (BW), weight at age 90 (W90) and 210 (W210) days (in kg²). Standard errors are shown in parentheses. σ²ₚ: phenotypic variance; σ²ₐ: direct additive genetic variance; σ²ₘ: maternal additive genetic variance; σ²c: variance for the common herd envi- ronment; σ²e: residual variance; HY: herd-year; HY2: herd-2 years; HY5: herd-5 years; H: herd Acta agriculturae Slovenica, 121/4 – 20256 B. LUŠTREK et al. the lowest value in the HY model (Table 5). Slight vari- ation in the estimated variance components, and thus in phenotypic variance, reflects differences in model structure, particularly in the specification of the com- mon herd(-time) environmental effect. Direct additive genetic variance (σ²ₐ) declined when the herd-only model (H) was replaced with longer herd-time interac- tion groupings (HY5), but increased again with finer fragmentation (HY2 and HY). Maternal additive ge- netic variance (σ²ₘ) was lowest in HY2 and HY5, and slightly higher in HY and H models. The proportion of phenotypic variance attributed to the common herd(-time) effect (c²) decreased substantially in the HY model (25%) compared to 38% in the HY5 and H models. Residual variance (σ²ₑ) was relatively stable, contributing 41%–43% across models (Table 6). Estimates of direct heritability (h²ₐ) ranged from 0.25 (HY5) to 0.39 (HY), and maternal heritability (h²ₘ) from 0.10 to 0.13 (Table 6). The highest h²ₐ in HY may be inflated due to small group sizes and low sire over- lap, leading to confounding of environmental effects with genetic variance. Broader groupings (H, HY5) likely produced more accurate separation of genetic and environmental components. Genetic correlations between direct and maternal additive effects (rₐₘ) were moderately negative, ranging from –0.42 (HY) to –0.54 (H) (Table 6). Stronger antagonism in H and HY5 sug- gests clearer distinction of direct and maternal contri- butions when contemporary groups are broader. There is limited literature on the genetic evalua- tion of traits in suckler populations, particularly un- der extensive systems. However, several studies pro- vide estimates for BW in similar breeds. For LIM, Meyer (1992) reported h²ₐ = 0.22, h²ₘ = 0.05, and rₐₘ = –0.16, while for CHA, reported values were h²ₐ = 0.42, h²ₘ = 0.17, and rₐₘ = –0.39. These results are broadly comparable to our estimates, especially those from the HY5 and HY models, suggesting that the different de- grees of herd-time fragmentation can yield estimates similar to published benchmarks when appropriately structured. In our data, the HY5 model for LIM yielded heritability close to Meyer’s estimates, while the HY model for CHA showed strong agreement with values for that breed. These consistencies suggest that the defi- nition of contemporary group plays a pivotal role in aligning genetic parameter estimates with known breed characteristics. In comparison, Crews et al. (2004) reported h²ₐ = 0.53, h²ₘ = 0.16, and σ²ₑ = 9.6 kg² for BW in Ca- nadian Charolais cattle, which are generally higher than our estimates. Genetic correlations in H and HY5 models were closest to their results. Differences may reflect larger dataset, more consistent management, and single-herd structure. Čepon et al. (2008, 2009) re- ported higher h²ₐ (0.62–0.74) and lower h²ₘ and rₐₘ, but these were based on test station data without common herd(-time) effects, which likely reduced environmen- tal variance. In summary, the results for birth weight reveal that direct heritability was moderate (0.25–0.39), and maternal heritability was low (0.10–0.13). The choice of herd(-time) grouping substantially affected variance partitioning, with HY inflating genetic variance and H/HY5 better balancing environmental sources. Trait Model c² h²ₐ h²ₘ e² rₐₘ BW HY 0.25 (0.03) 0.39 (0.05) 0.13 (0.03) 0.42 (0.04) −0.42 (0.11) HY2 0.34 (0.04) 0.28 (0.05) 0.10 (0.02) 0.43 (0.04) −0.43 (0.12) HY5 0.38 (0.04) 0.25 (0.04) 0.10 (0.02) 0.42 (0.04) −0.48 (0.11) H 0.38 (0.05) 0.28 (0.04) 0.12 (0.02) 0.41 (0.04) −0.54 (0.08) W90 HY 0.44 (0.03) 0.11 (0.04) 0.12 (0.03) 0.44 (0.04) −0.46 (0.17) HY2 0.40 (0.04) 0.22 (0.05) 0.14 (0.03) 0.46 (0.05) −0.61 (0.10) HY5 0.39 (0.04) 0.28 (0.06) 0.16 (0.03) 0.47 (0.06) −0.70 (0.08) H 0.37 (0.06) 0.37 (0.07) 0.18 (0.03) 0.45 (0.06) −0.71 (0.07) W210 HY 0.42 (0.04) 0.34 (0.06) 0.23 (0.03) 0.33 (0.04) −0.57 (0.08) HY2 0.43 (0.04) 0.34 (0.06) 0.20 (0.02) 0.33 (0.04) −0.59 (0.07) HY5 0.41 (0.04) 0.46 (0.06) 0.23 (0.02) 0.34 (0.05) −0.67 (0.06) H 0.33 (0.05) 0.57 (0.06) 0.29 (0.03) 0.37 (0.05) −0.69 (0.05) Table 6: Estimates of genetic parameters and variance component proportions. Standard errors are shown in parentheses. c²: proportion of common herd in environmental variance; h²ₐ: direct heritability; h²ₘ: maternal heritability; e²: proportion of residual variance; rₐₘ: direct-maternal (additive) genetic correlation; HY: herd-year; HY2: herd-2 years; HY5: herd-5 years; H: herd Acta agriculturae Slovenica, 121/4 – 2025 7 Genetic parameters for growth traits in the Slovenian beef cattle population 3.2 WEIGHT AT 90 DAYS (W90) For W90, phenotypic variance (σ²ₚ) ranged be- tween 558.9–604.8 kg² across models (Table 5). Genetic variances (σ²ₐ, σ²ₘ) decreased as herd-time groups be- came more fragmented. In the HY model, σ²ₐ dropped markedly (66.7 kg²), likely due to inadequate sire over- lap across small herds, confounding genetic with en- vironmental effects. The maternal component (σ²ₘ) was less variable but still showed the highest value in the H model (99.2 kg²). Common herd(-time) vari- ance (σ²c) increased with fragmentation: from 205.1 kg² (H) to 264.9 kg² (HY), contributing 37% to 44% of σ²ₚ. Residual variance remained high across all models (44%–47%) (Table 6). Direct heritability (h²ₐ) ranged from 0.11 (HY) to 0.37 (H). Maternal heritability (h²ₘ) was highest in H (0.18) and lowest in HY (0.12). The genetic correlation rₐₘ became more negative with fragmentation (Table 6), from –0.46 (HY) to –0.71 (H), indicating stronger antagonism where genetic effects were better separated. Only a few studies include W90 specifically. Čepon et al. (2008, 2009) reported h²ₐ = 0.23–0.33, h²ₘ = 0.12, and rₐₘ = –0.61 for W90 in Slovenian CHA herds un- der test-station conditions, which differ from our more field-based population structure. Our estimates in HY2 and HY5 are relatively consistent with these findings, supporting their applicability in structured genetic evaluations. Ulutaş et al. (2001) reported rₐₘ = –0.46 for 100-day weights in Welsh Black suckler cattle, which aligns closely with our HY model estimate (rₐₘ = –0.46). These values are consistent with Lee (2002), who sum- marised genetic antagonism between direct and ma- ternal additive effects in beef breeds, with rₐₘ ranging from –0.21 in Gelbvieh to –0.32 in Simmental. The more extreme values observed in our study may reflect higher variability in management conditions, maternal dependence of early calf growth, and the explicit mod- elling of both direct and maternal additive effects in fragmented herd structures. The results for W90 indicate that direct heritability was the lowest among traits (0.11–0.37), while mater- nal heritability was moderate. Increased fragmentation inflated c² and reduced h²ₐ. The H and HY2 models provided the most balanced partitioning. Negative rₐₘ values indicate strong antagonism. 3.3 WEIGHT AT 210 DAYS (W210) Phenotypic variance for W210 ranged from 1611.9 kg² (H) to 1861.9 kg² (HY2) (Table 5). Di- rect additive genetic variance (σ²ₐ) was highest in H (919.0 kg²) and decreased with fragmentation, espe- cially in HY2 (636.0 kg²) and HY (600.7 kg²). Mater- nal additive variance (σ²ₘ) followed a similar pattern, with H having the highest value (466.7 kg²). Common herd(-time) variance (σ²c) ranged from 527.4 kg² (H) to 809.6 kg² (HY2), accounting for 33% to 43% of pheno- typic variance (Table 6). Residual variance was lowest for this trait (33%–37%). Direct heritability (h²ₐ) was highest for this trait, ranging from 0.34–0.57, with the highest in H. Maternal heritability also reached 0.29 in H, compared to 0.20–0.23 in other models (Table 6). The genetic correlation rₐₘ was again most negative in H (–0.69) and least negative in HY (–0.57). Compared to the literature, Crews et al. (2004) re- ported h²ₐ = 0.22, h²ₘ = 0.10, and σ²ₑ = 500.2 kg² in Canadian CHA cattle, all lower than our W210 esti- mates, particularly for residual variances. Similarly, Meyer (1992) summarised literature values for wean- ing weight in LIM cattle, reporting h²ₐ = 0.16 and h²ₘ = 0.15, both lower than our findings. Čepon et al. (2008, 2009) reported h²ₐ = 0.29, h²ₘ = 0.12, and rₐₘ = –0.30, all lower than our findings except for the genetic correlation, suggesting that our broader data- set and inclusion of herd effects enabled better variance separation. Meyer (1997) reported rₐₘ ranging from –0.65 to –0.30 for LIM weaning weight, matching our range of rₐₘ from –0.69 to –0.57. These results reinforce the importance of modelling both direct and maternal components, particularly in extensive systems where the dam‘s influence on calf growth is substantial. In Czech CHA populations, Vostrý et al. (2007) found σ²ₐ = 71.3–167.2 kg², σ²ₘ = 27.2–76.8 kg², and rₐₘ = –0.15–0.42 for weaning weight, generally low- er than our values. Their higher residual variance (σ²e = 658–690 kg²) may reflect differences in data com- pleteness or environmental noise. In contrast, our high- er additive and maternal variance, and lower residuals, suggest successful partitioning in our chosen models (HY2/HY5). Kennedy and Henderson (1975) reported σ²c = 222–238 kg² (25–28%) for Hereford and 178–331 kg² (25–41%) for Aberdeen Angus at weaning, while other estimates ranged from 0–8%. Our estimates of σ²c = 527–810 kg² (33–43%) for W210 are higher and may reflect greater heterogeneity in Slovenian herds, smaller group sizes, or stronger maternal dependence. These results highlight the substantial influence of com- mon herd(-time) effects, particularly in fragmented, ex- tensive systems. W210 stood out as the trait with the highest to- tal genetic contribution. The h²ₐ and h²ₘ were highest among all traits, and c² was also substantial. Models H Acta agriculturae Slovenica, 121/4 – 20258 B. LUŠTREK et al. and HY5 provided the best balance of variance com- ponents. Phenotypic correlations (r) between body weights at different ages vary among beef-type breeds and production systems. In Bali and Nguni cattle, both reared primarily for beef, low to moderate correlations were reported between birth and weaning weights (r = 0.10–0.34), while correlations between weaning and later weights were substantially higher (r = 0.90) (Assan, 2006; Gunawan and Sari, 2012). Similar pat- terns were described in European beef breeds (Krupa et al., 2005) and are supported by a meta-analysis dem- onstrating that prenatal and early postnatal nutrition influence both calf birth and weaning weights (Barce- los et al., 2022). Collectively, these findings indicate that birth weight has only a limited association with wean- ing weight, whereas weights measured at later develop- mental stages are more strongly interrelated and better reflect cumulative growth potential. 3.4 ESTIMATED BREEDING VALUE (EBV) COR- RELATIONS BETWEEN MODELS Table 7 presents Pearson correlations between sire EBVs across models. Only sires with EBV accuracy ≥ 0.3 were included. Correlation coefficients were high- est for BW and W210 (up to 0.95) and lowest for W90 (as low as 0.61). The highest agreement was between HY and HY2 (0.89–0.95), indicating that adding an ex- tra year did not substantially change the genetic evalu- ation. The lowest correlations were between H and HY (0.61–0.80), where environmental grouping structure differed most. These patterns indicate that EBV rank- ings were sensitive to the definition of the common herd(-time) effect. Greater fragmentation (HY) led to lower EBV correlations with broader groupings (H). 3.5 LIMITATIONS AND FUTURE PERSPECTIVES Across all traits, genetic correlations between di- rect and maternal additive effects became progressive- ly less negative with increasing fragmentation of the herd(-time) grouping. The largest shift occurred for W90 in the HY model, consistent with patterns ob- served in other suckler systems. While such correlations reflect genetic associations, they do not imply causality. Calf growth is influenced by both its own genetic po- tential and the genetic merit of the dam, particularly in early development. These effects are known to be nega- tively correlated due to physiological and evolutionary trade-offs (Willham, 1972). As shown by the literature, negative rₐₘ values are common and relevant for breeding decisions, particu- larly in suckler herds where the maternal contribution is substantial and cannot be ignored. The inclusion of both direct and maternal additive genetic effects in models is therefore essential for unbiased genetic evaluations. Although the distinction between mater- nal additive genetic and environmental effects was ac- knowledged, they were not separated in this study due to small herd sizes and convergence challenges. Com- bining them resulted in more stable model fitting. In- terpretation of comparative genetic parameters should consider that, with the exception of Simčič and Čepon (2007) and Čepon et al. (2008, 2009), most referenced studies were conducted on populations considerably larger than the Slovenian cattle population, potentially limiting direct comparability due to differences in pop- ulation size effects on parameter estimation precision. Future research using larger, more balanced datasets from similarly-sized populations would help address both the limitaions in parameter estimation precision and enable better separation of maternal genetic and environmental effects. Direct heritability was highest for W210 and lowest for W90, while maternal heritabil- ity was similar across traits, with a slight peak at W210. W90 had the highest environmental (c² and e²) propor- tions, making it most sensitive to herd(-time) grouping. Although season of calving could theoretically ac- count for additional environmental variation, its inclu- sion in the model was not expected to substantially im- prove model fit, as more than 95% of calvings occurred within the same season. In addition, introducing season into the model would further fragment the contempo- Trait N Model HY5 HY2 HY BW 113 H 0.85 0.81 0.80 HY 0.87 0.95 HY2 0.92 W90 87 H 0.72 0.65 0.61 HY 0.80 0.89 HY2 0.87 W210 109 H 0.87 0.77 0.75 HY 0.80 0.89 HY2 0.89 Table 7: Correlations among estimated breeding values for BW, W90 and W210 between models (sires with EBV ac- curacy ≥ 0.3) BW: birth weight; W90: weight at 90 days; W210: weight at 210 days; HY: herd-year; HY2: herd-2 years; HY5: herd-5 years; H: herd; N: number of estimated breeding values Acta agriculturae Slovenica, 121/4 – 2025 9 Genetic parameters for growth traits in the Slovenian beef cattle population rary groups, which in the context of our dataset was not considered an optimal approach. Therefore, season was not included as a model effect, since its potential impact on the results was expected to be negligible. This study underscores the importance of con- temporary group definition in genetic evaluation, particularly in small, extensively managed herds. Al- though limited by the absence of SNP data, the use of single-trait models, and potentially inconsistent meas- urement timing, the findings highlight the influence of herd(-time) fragmentation on variance estimates and EBV rankings. Incorporating SNP genotypes and genomic prediction models (e.g., GBLUP or ssGBLUP) in future research would enable more accurate parti- tioning of genetic and environmental effects, improve prediction accuracy, and support early genomic selec- tion, especially when opportunities for extensive prog- eny testing across herds are limited. 4 CONCLUSIONS The common herd environment significantly af- fected genetic parameter estimates for early growth traits. Among the models tested, the two-year herd- time interaction (HY2) provided the most balanced variance partitioning and stable EBV correlations. Ex- cessive fragmentation (e.g., by year) distorted results, while broader intervals (HY5) performed similarly to HY2. Future evaluations should adapt contemporary group definitions to herd structure and data size. In- cluding genomic information is recommended to fur- ther improve the accuracy of genetic evaluations in small, extensively managed beef cattle populations. 5 REFERENCES Assan, N. (2006). A phenotypic study of growth traits in in- digenous Nguni cattle of Zimbabwe. African Journal of Agricultural Research, 1(2), 24–26. Assan, N. (2025). Direct–maternal genetic covariance in ruminants: Magnitude, direction, and implications for growth trait improvement. International Journal of Inno- vative Science and Research Technology, 10(8), 494–505. https://doi.org/10.38124/ijisrt/25aug022 Barcelos, S. d. S., Nascimento, K. B., Silva, T. E. d., Mez- zomo, R., Alves, K. S., de Souza Duarte, M., Gionbelli, M. P. (2022). The effects of prenatal diet on calf perfor- mance and perspectives for fetal programming studies: A meta-analytical investigation. Animals, 12(16), 2145. https://doi.org/10.3390/ani12162145 Crews, D. H., Lowerison, J., M., Caron, N., Kemp, R. A. (2004). Genetic parameters among growth and carcass traits of Canadian Charolais cattle. Canadian Journal of Animal Science, 84(4), 589–597. https://doi.org/10.4141/A04-019 Čepon, M., Simčič, M., Malovrh, Š. (2008). Genetic param- eters for growth in Charolais calves. Acta Agricultu- rae Slovenica, 92(2), 111–117. https://doi.org/10.14720/ aas.2008.92.2.15143 Čepon, M., Simčič, M., Malovrh, Š. (2009). Estimation of ge- netic parameters for body weight in Charolais calves in Slovenia. Italian Journal of Animal Science, 8(4), 735–742. https://doi.org/10.4081/ijas.2009.735 Daneshi, M., Borowicz, P. P., Montgomery, V., Entzie, Y. L., Sy- ring, J. G., King, L., … Caton, J. S. (2025). Effects of ma- ternal nutrition and one-carbon metabolite supplementa- tion on fetal jejunal morphology and hexose transporter expression in beef cattle. Veterinary Sciences, 12(9), 884. https://doi.org/10.3390/vetsci12090884 Davies, R., Denholm, K. (2025). A holistic approach to colos- trum management: Enrichment of maternal colostrum combined with extended colostrum feeding as con- trol measures for bovine rotavirus‐associated neonatal calf diarrhoea. Veterinary Record Case Reports, e70214. https://doi.org/10.1002/vrc2.70214 Edwards, A. K., Vedovatto, M., Gurie, J., Marcon, H., Scaglia, G., Fannin, J. M., … Gentry, G. T. (2025). 150 Effect of adopting multiple management technologies on cattle performance in a small grazing beef cow-calf system. Journal of Animal Science, 103(Supplement_2), 75–76. https://doi.org/10.1093/jas/skaf170.088 Falconer, D. S. (1989). Introduction to quantitative genetics. UK/New York: Longmans Green/John Wiley & Sons. Gellatly, D., Lei, Y., Neale, A., Smith, L., Edgar, E., Bloomfield, B., … Thompson, S. (2025). Impacts of maternal bovine appeasing substance administered at weaning on behav- ioral and physiological adaptation of beef heifers to the feedlot. Animals, 15(19), 2788. https://doi.org/10.3390/ ani15192788 Gunawan, A. and Sari, R. (2012). Estimates of Genetic and Phenotypic Trends of Growth Traits in Bali Cattle. Me- dia Peternakan, 35(2), 85. https://doi.org/10.5398/med- pet.2012.35.2.85 Hasan, F. M., Lomax, S., Thomson, P., Islam, M. R., Chlingary- an, A., Clark, C. (2024). The impact of rainfall on beef cat- tle growth across diverse climate zones. Animal, 101336. https://doi.org/10.1016/j.animal.2024.101336 International Agreement of Recording Practices. (2018). ICAR recording guidelines. ICAR. Retrieved from https://www.icar.org/index.php/icar-recording-guidelines/ Kennedy, B. W., Henderson, C. R. (1975). Components of variance of growth traits among Hereford and Aberdeen Angus calves. Canadian Journal of Animal Science, 55(4), 493–502. https://doi.org/10.4141/cjas75-061 Koch, R. M. (1971). The role of maternal effects in animal breeding: VI. Maternal effects in beef cattle. Journal of Animal Science, 35(6), 1316–1323. https://doi.org/10.2527/ jas1972.3561316x Kovač, M., Groeneveld, E. (2008). VCE-6 user’s guide and Ref- erence Manual Version 6. Ljubljana: University of Ljublja- na, Biotechnical Faculty, Department of Animal Science. Krupa, E., Oravcová, M., Polák, P., Huba, J., Krupová, Z. (2005). Acta agriculturae Slovenica, 121/4 – 202510 B. LUŠTREK et al. Factors affecting growth traits of beef cattle breeds raised in Slovakia. Czech Journal of Animal Science, 50(1), 14–21. https://doi.org/10.17221/3990-CJAS Krupová, Z., Krupa, E., Brzáková, M., Veselá, Z., Ma- lát, K. (2025). Beef breeding systems and preferences for breeding objective traits. Animals, 15(15), 2175. https://doi.org/10.3390/ani15152175 Lee, C. (2002). On the negative estimates of direct and mater- nal genetic correlation – A review. Asian-Aust. Journal of Animal Science, 15(8), 1222–1226. https://doi.org/10.5713/ ajas.2002.1222 Meyer, K. (1992). Variance components due to direct and maternal effects for growth traits of Australian beef cattle. Livestock Production Science, 31(3–4), 179–204. https://doi.org/10.1016/0301-6226(92)90017-X Meyer, K. (1997). Estimates of genetic parameters for weaning weight of beef cattle accounting for direct-maternal envi- ronmental covariances. Livestock Production Science, 52(3), 187–199. https://doi.org/10.1016/S0301-6226(97)00144-9 Ocena plemenskih vrednosti. (2018). Retrieved from https://rodica.bf.uni-lj.si/web/gov/1802/eval/_main_ eval_6.html Phocas, F., Laloë, D. (2004). Genetic parameters for birth and weaning traits in French specialized beef cattle breeds. Livestock Production Science, 89(2–3), 121–128. https://doi.org/10.1016/j.livprodsci.2004.02.007 Robinson, G. K. (1991). That BLUP is a good thing: The esti- mation of random effects. Statistical Science, 6(1), 15–32. https://doi.org/10.1214/ss/1177011926 Schenkel, F., Miller, S., Wilton, J. (2004). Herd of origin effect on weight gain of station-tested beef bulls. Livestock Pro- duction Science, 86(1–3), 93–103. https://doi.org/10.1016/ S0301-6226(03)00143-X Simčič, M., Čepon, M. (2007). Factors affecting actual and ad- justed 90-day, 205-day and 365-day weight of Charolais calves. Poljoprivreda, 13(1), 111–114. Retrieved from https://hrcak.srce.hr/16067 Troxel, T. R., Simon, K. (2007). Best management practices for small beef cow-calf herds. Little Rock: University of Arkansas, Agriculture and Natural resources, Arkan- sas Cooperative Extension Service. Retrieved from https://www.uaex.edu/publications/PDF/FSA-3117.pdf Ulutaş, Z., Dewi, I. A., Saatçi, M. (2001). Six different models for estimation of genetic parameters and breeding values for pre-weaning and post-weaning weights in suckler cat- tle. Turkish Journal of Veterinary & Animal Sciences, 25(2), 79–184. Vostrý, L., Přibyl, J., Veselá, Z., Jakubec, V. (2007). Selec- tion of a suitable data set and model for the estima- tion of genetic parameters of the weaning weight in beef cattle. Archives Animal Breeding, 50(6), 562–574. https://doi.org/10.5194/aab-50-562-2007 Willham, R. L. (1972). The role of maternal effects in ani- mal breeding: III. Biometrical aspects of maternal effects in animals. Journal of Animal Science, 35(6),1288–1293. https://doi.org/10.2527/jas1972.3561288x