Associations between weaning and post-weaning body measurements in Limousin cattle and their potential use for early preselection Authors: Márton János Demény, Dóra Lili Brassó, Csaba Rácz, Márton Szűcs, János Tőzsér Published: 2026 ABSTRACT Highlights•Body weight and mass index showed the greatest variability in Limousin young bulls.•Body parameters and weaning weight were moderately and strongly correlated.•Body parameters and weight were strongly correlated at 13–14 months of age.•Withers and rump height showed moderate correlation between the two time points.•Pin and hip width showed weak and moderate correlations between the two time points. Introduction The measurement of body parameters in cattle, including the Limousin breed, is a crucial aspect of livestock management that impacts animal performance, productivity, and overall herd health. Numerous studies have demonstrated strong relationships between body measurements, conformation traits and growth or productive performance, highlighting their value in breeding and management decisions (Abreu et al., 2018; Dominguez-Castaño et al., 2024; Koenen & Groen, 1998). Morphological traits can effectively differentiate cattle breeds, reflecting genetic diversity and environmental adaptation (Traoré et al., 2016). Body weight and body size parameters recorded at weaning are reliable indicators of growth and are widely used to evaluate breeding efficiency and heritability in cattle populations (Nurgiartiningsih et al., 2022). In addition, the application of genetic markers associated with body measurement traits plays a key role in determining cattle conformation and growth traits (Gao et al., 2019). Several studies have shown that body dimensions can be used to accurately estimate live weight, providing practical benefits for nutritional planning and economic decision-making in beef production systems (Haq et al., 2020; Paputungan et al., 2013). The use of random regression models, in addition to single-trait models, has enabled researchers to pinpoint functional candidate genes involved in body weight variations (Ribeiro et al., 2023). Comprehensive body measurements enable informed management decisions that support animal health and farm profitability, while also providing a basis for exploring the genetic background of growth traits. The significance of measuring weight and body parameters at various ages in beef cattle to fit growth models describing their development has been demonstrated in the international literature (Crispim et al., 2015; Ribeiro et al., 2023; Yang et al., 2024). In addition, automated, non-contact methods based on image analysis of linear body dimensions may facilitate the identification of early morphometric indicators to support selection and management decisions (Gritsenko et al., 2023; Ruchay et al., 2020). Despite the recognised importance of morphological traits, body weight, and linear body measurements in improving beef cattle production and supporting selection decisions, limited information is available on the consistency and repeatability of these traits across different growth stages in Limousin cattle. In particular, there is a lack of research investigating the relationship between early-life body measurements and later development and how this can be used effectively for early preselection in practical breeding programmes. Key body measurements, including withers height, rump height, back length, shoulder width, hip width and pin width, are widely used indicators of physiological and genetic development and are considered useful for describing growth patterns and selection in beef cattle. Therefore, the primary objective of this study was to evaluate whether body measurements recorded at weaning are associated with corresponding measurements at 13–14 months of age in Limousin cattle. The secondary objective was to compare weaning-age body measurements between bull calves retained for breeding (‘Breeding-approved’ bulls) and those removed from breeding selection (‘Removed’ bulls). Furthermore, our aim was to identify which linear traits show sufficient phenotypic stability, considered to be practical preselection indicators. We hypothesised that significant associations exist between measurements taken at the two time points, particularly for withers and rump height, reflecting their more uniform development. Moreover, we expected stronger correlations and greater homogeneity (CV% below 10%) of body measurements at the later growth stage, and that calves selected for breeding would perform better in body weight and at least one body size parameter than the population average. Since the study is observational and based on routinely collected farm data, the results should be interpreted as evidence of association rather than validated prediction. Materials and methods Ethical approval Ethical clearance was granted by the Institutional Animal Welfare and Use Committee of Széchenyi István University Faculty of Agricultural and Food Sciences (clearance number: SZE-AKMK/MAB/012/2024). Study population and the measured parameters The study used data collated from 13 Limousin breeding farms over a period of approximately 3 years (January 2022 – February 2025). The breeding farms were all located in the same region of Hungary, in Transdanubia. In accordance with strict rules governing breeding farms, calves were raised using similar technology until weaning. The practice of cattle farming is based on the activity of grazing. Within the rotational grazing system, farmers strive to extend the grazing period to its maximum extent. Consequently, cattle typically graze on pasture from April until October. During the winter months, the animals are fed fodder available to farmers, most commonly grass and alfalfa hay, haylage, silage, or a mixture of these. Calves are reared alongside their mothers and are provided with pasture grass or hay as a supplement. In calf nurseries, they are given calf starter feed until weaning. Following weaning, the majority of farms provide grass hay and calf starter feed. However, on certain farms, silage or haylage is also made available to the animals at this age in minimal quantities for the purpose of familiarisation. In the later stages of rearing and breeding heifers, the role of haylage and silage in the animal's diet increases, right up to the time the animals are transferred to the breeding herd. The available nutrient-composition ranges of the major feeds are summarised in Table 1.Table 1Nutrient composition of major feeds offered to calves on participating farms.ParameterConcentrated feedaAlfalfa haybGrass haybCorn silagebHaylageb,cDry matter (% as-fed)88.00–90.2392.6–93.084.3–93.125.5–46.052.4–66.3Crude protein (%)15.35–20.12NA5.7–11.93.8–8.313.7–17.9NEm (MJ/kg)7.01–7.16NA4.76–5.897.86–8.164.65–5.81NEg (MJ/kg)4.80–4.91NA2.42–3.455.20–5.462.31–3.38Starch (%)NANA2.4–3.832.3–41.72.1NDF (%)NANA60.5–70.337.6–43.257.2Crude fat (%)1.72–4.77NA1.08–2.052.64–3.511.59–3.24Crude fiber (%)3.54–11.11NA32.65NA35.71Ash (%)9.05–10.40NA3.55–8.761.06–3.185.21–9.86Calcium (Ca) (%)0.730–2.1000.500–1.3700.264–1.4200.140–0.6170.380–1.690Phosphorus (P) (%)0.435–0.6900.149–0.5160.081–0.2600.110–0.3380.160–0.280Magnesium (Mg) (%)0.166–0.3500.124–0.6110.092–0.2790.150–0.3220.160–0.210Copper (Cu) (mg/kg)10.00–48.223.09–13.102.10–7.855.00–7.216.66–12.00Zinc (Zn) (mg/kg)40.00–229.5617.0–41.814.0–20.016.0–50.018.0–26.6Manganese (Mn) (mg/kg)40.00–199.5225.0–59.420.7–139.040.0–43.027.0–66.5Selenium (Se) (mg/kg)0.300–2.2280.05–0.26<0.04–0.190.050.12Values are reported as minimum–maximum ranges; NA = data not available.NEₘ = net energy for maintenance; NEg = net energy for gain; NDF = neutral detergent fibre.aConcentrated feed comprises commercially manufactured and farm-mixed concentrate feeds offered to calves; nutrient values are expressed on an as-fed basis.bNutrient values for roughage feeds are expressed on a dry matter basis, except dry matter, which is expressed as a percentage of the as-fed material.cHaylage comprises grass haylage, triticale–pea–vetch haylage, alfalfa haylage, and barley haylage.Available feed-composition records from 2022 to 2025 were pooled across years.Feed-composition records were available from eight farms, which together accounted for 748 of the 817 animals in the repeated-measurement dataset and 829 of the 928 animals in the bull-classification dataset. Table 1Nutrient composition of major feeds offered to calves on participating farms.ParameterConcentrated feedaAlfalfa haybGrass haybCorn silagebHaylageb,cDry matter (% as-fed)88.00–90.2392.6–93.084.3–93.125.5–46.052.4–66.3Crude protein (%)15.35–20.12NA5.7–11.93.8–8.313.7–17.9NEm (MJ/kg)7.01–7.16NA4.76–5.897.86–8.164.65–5.81NEg (MJ/kg)4.80–4.91NA2.42–3.455.20–5.462.31–3.38Starch (%)NANA2.4–3.832.3–41.72.1NDF (%)NANA60.5–70.337.6–43.257.2Crude fat (%)1.72–4.77NA1.08–2.052.64–3.511.59–3.24Crude fiber (%)3.54–11.11NA32.65NA35.71Ash (%)9.05–10.40NA3.55–8.761.06–3.185.21–9.86Calcium (Ca) (%)0.730–2.1000.500–1.3700.264–1.4200.140–0.6170.380–1.690Phosphorus (P) (%)0.435–0.6900.149–0.5160.081–0.2600.110–0.3380.160–0.280Magnesium (Mg) (%)0.166–0.3500.124–0.6110.092–0.2790.150–0.3220.160–0.210Copper (Cu) (mg/kg)10.00–48.223.09–13.102.10–7.855.00–7.216.66–12.00Zinc (Zn) (mg/kg)40.00–229.5617.0–41.814.0–20.016.0–50.018.0–26.6Manganese (Mn) (mg/kg)40.00–199.5225.0–59.420.7–139.040.0–43.027.0–66.5Selenium (Se) (mg/kg)0.300–2.2280.05–0.26<0.04–0.190.050.12Values are reported as minimum–maximum ranges; NA = data not available.NEₘ = net energy for maintenance; NEg = net energy for gain; NDF = neutral detergent fibre.aConcentrated feed comprises commercially manufactured and farm-mixed concentrate feeds offered to calves; nutrient values are expressed on an as-fed basis.bNutrient values for roughage feeds are expressed on a dry matter basis, except dry matter, which is expressed as a percentage of the as-fed material.cHaylage comprises grass haylage, triticale–pea–vetch haylage, alfalfa haylage, and barley haylage.Available feed-composition records from 2022 to 2025 were pooled across years.Feed-composition records were available from eight farms, which together accounted for 748 of the 817 animals in the repeated-measurement dataset and 829 of the 928 animals in the bull-classification dataset. Nutrient composition of major feeds offered to calves on participating farms. Values are reported as minimum–maximum ranges; NA = data not available. NEₘ = net energy for maintenance; NEg = net energy for gain; NDF = neutral detergent fibre. Concentrated feed comprises commercially manufactured and farm-mixed concentrate feeds offered to calves; nutrient values are expressed on an as-fed basis. Nutrient values for roughage feeds are expressed on a dry matter basis, except dry matter, which is expressed as a percentage of the as-fed material. Haylage comprises grass haylage, triticale–pea–vetch haylage, alfalfa haylage, and barley haylage. Available feed-composition records from 2022 to 2025 were pooled across years. Feed-composition records were available from eight farms, which together accounted for 748 of the 817 animals in the repeated-measurement dataset and 829 of the 928 animals in the bull-classification dataset. In all breeding herds, housing takes place in deep-litter barns constructed from wood or metal. These barns feature an exercise area that is open on one or more sides and a feeding passage. On smaller farms, the process of weaning is typically executed once or twice annually, coinciding with the animals' transition to and from pastures. On larger farms, the process of weaning is typically conducted every 2–3 months, on average, when the calves reach 7 months of age. This is facilitated by the practice of year-round breeding. Subsequent to weaning, once the calves have acclimatised (typically after 4–5 days, but no later than 10 days, after weaning), weight and body measurements are taken. The technology for keeping bulls participating in the self-performance test is the most clearly defined. In this test, the bulls must achieve a minimum performance level in groups of two to five animals. They are kept in deep litter or open pens and fed a diet of concentrate feed, hay and silage in order to qualify as breeding bulls. The test period generally lasts 150–180 days and must be completed before the animals reach a maximum age of 450 days. Guidelines on the husbandry and feeding technologies, registration and qualification of bulls, and data collection are set out in the breeding programme of the Association of Hungarian Limousin and Blonde d'Aquitaine Breeders, which is available in Hungarian on the association’s website (www.limousin.hu). Climatic conditions showed the following tendencies. In 2022, across all analysed farms, the annual temperature ranged from −9.51 °C to 37.30 °C, with a mean of 11.91 °C. The mean annual humidity was 71.50%, and the total annual precipitation was 571.59 mm. In 2023, the annual temperature ranged from −9.17 °C to 35.43 °C, with a mean of 12.23 °C. The mean annual humidity was 71.73%, and the total annual precipitation was 866.05 mm. In 2024, the annual temperature ranged from −9.26 °C to 36.98 °C, with a mean of 12.93 °C. The mean annual humidity was 73.50%, and the total annual precipitation was 592.61 mm. In 2025, the annual temperature ranged from −8.66 °C to 37.23 °C, with a mean of 11.78 °C. The mean annual humidity was 73.62%, and the total annual precipitation was 513.85 mm. The Hungarian Meteorology Service provided climatic data. A total of 817 breeding cattle, including bulls (n = 469) and heifers (n = 348), were examined for body size data. Each animal went through two body measurements, first time at weaning (at an average age of 230 days) and second time at 13–14 months of age (at an average age of 424 days) in accordance with the breeding regulations of the Association of Hungarian Limousin and Blonde d'Aquitaine Breeders. Fig. 1 illustrates the process of measurements and selection of weaned calves.Fig. 1The process of measuring body dimensions and selecting calves following weaning. The process of measuring body dimensions and selecting calves following weaning. The study design had two distinct datasets: 817 cattle of both sexes were measured at both time points, and a separate dataset of 928 weaned bull calves was classified as breeding-approved (n = 469) or not approved and subsequently sold or allocated to fattening (n = 459). Subsequent evaluations focused on comparing the body size data of bull calves at weaning (n = 928). Following weaning and body measurements, the animals entered the breeding selection process. Bulls that successfully completed the own performance test and were ultimately approved for breeding were classified as ‘Breeding-approved’ bulls (n = 469). The remaining bulls, which were removed from the breeding selection process before final breeding approval, were classified as ‘Removed’ bulls (n = 459). These animals were subsequently sold or slaughtered, typically after reaching approximately 300 kg, although some attained body weights of 600–700 kg before slaughter. Table 2 illustrates a detailed summary of number of animals involved in the analyses each year, per measurement category, per farm, and per sex.Table 2Number of animals involved in the study regarding year, measurement category, farm, and sex.YearMeasurement categoryFarmSex2022at weaning1 – 99Bull – 1803 – 10Heifer – 1194 – 245 – 676 – 427 – 169 – 210 – 511 – 1412 – 213 – 182022at 13 to 14 months1 – 36Bull – 764 – 10Heifer – 405 – 256 – 77 – 1611 – 212 – 213 – 182023at weaning1 – 62Bull – 2032 – 11Heifer – 1424 – 175 – 806 – 788 – 639 – 1510 – 1011 – 92023at 13 to 14 months1 – 103Bull – 2012 – 11Heifer – 1153 – 104 – 235 – 876 – 608 – 39 – 210 – 511 – 122024at weaning4 – 5Bull – 865 – 75Heifer – 876 – 388 – 552024at 13 to 14 months1 – 22Bull – 1694 – 13Heifer – 1715 – 816 – 898 – 1019 – 1510 – 1011 – 92025at 13 to 14 months5 – 29Bull – 236 – 2Heifer – 228 – 14Numbers after the hyphens in column “Farm” denote the number of animals on that farm within the given year and measurement category. Table 2Number of animals involved in the study regarding year, measurement category, farm, and sex.YearMeasurement categoryFarmSex2022at weaning1 – 99Bull – 1803 – 10Heifer – 1194 – 245 – 676 – 427 – 169 – 210 – 511 – 1412 – 213 – 182022at 13 to 14 months1 – 36Bull – 764 – 10Heifer – 405 – 256 – 77 – 1611 – 212 – 213 – 182023at weaning1 – 62Bull – 2032 – 11Heifer – 1424 – 175 – 806 – 788 – 639 – 1510 – 1011 – 92023at 13 to 14 months1 – 103Bull – 2012 – 11Heifer – 1153 – 104 – 235 – 876 – 608 – 39 – 210 – 511 – 122024at weaning4 – 5Bull – 865 – 75Heifer – 876 – 388 – 552024at 13 to 14 months1 – 22Bull – 1694 – 13Heifer – 1715 – 816 – 898 – 1019 – 1510 – 1011 – 92025at 13 to 14 months5 – 29Bull – 236 – 2Heifer – 228 – 14Numbers after the hyphens in column “Farm” denote the number of animals on that farm within the given year and measurement category. Number of animals involved in the study regarding year, measurement category, farm, and sex. Numbers after the hyphens in column “Farm” denote the number of animals on that farm within the given year and measurement category. The ‘Breeding-approved’ and ‘Removed’ bull groups originated from the same farms, with each farm including animals from both groups. Within each farm, housing and general management conditions before weaning were comparable between the two eventual groups, and they received the same preweaning feeding regime. Following selection and at the start of the self-performance test, the Breeding-approved bull group received a concentrate formulated for breeding animals, whereas the Removed bull group received a concentrate formulated for growth. However, only body measurements recorded at weaning were included in the present analysis. The animals were weighed using certified livestock scales. Body measurements were obtained either in a purpose-built measuring frame or in a commercially manufactured frame with adjustable sides. In all cases, the animals were positioned on a level surface and restrained to prevent turning or moving more than approximately 0.5 m forward or backwards. When required, due to the animal's temperament, measurements were taken after the animal had been restrained with a neck catch. Whenever possible, animals were allowed to assume a natural standing posture. All measurements were performed by the same specialist, who was trained by experts from the Limousin and Blonde d'Aquitaine Breeders' Association and subsequently completed additional professional training in France. At the time of the study, the specialist had experience measuring >5000 animals. Duplicate measurements within the same measurement session of the same animals were not feasible. Repeated handling and measurement would have imposed a substantial logistical and labour burden on the breeders. The body measurement points and the measuring equipment used are shown in Fig. 2.Fig. 2Illustration of the key body measurement points on the cattle’s body, including height, width and length measurements, together with the measuring equipment (vertical bars indicate points measured with a measuring stick, while arrows demonstrate tape measurements). Illustration of the key body measurement points on the cattle’s body, including height, width and length measurements, together with the measuring equipment (vertical bars indicate points measured with a measuring stick, while arrows demonstrate tape measurements). In addition to measuring weight and body parameters at the time of measurement, body mass index (BMI) and overgrowth index (OI) were also calculated as follows.BMI=bodyweight(kg)*100withersweight(WH)OI=rumpheight(RH;cm)*100withersheight(WH;cm) In the formulae, RH denotes the term “rump height", and WH denotes the term "withers height". Created with BioRender.com. First, age-corrected values were calculated by adjusting the raw measurements to the mean age of the corresponding sex and measurement group, to evaluate the potential effect of age differences at the time of recording. The same age-correction procedure was applied to all body measurements. For each trait, a separate linear regression model was fitted with the recorded trait as the dependent variable and age (days) as the independent variable. The regression coefficient (b) obtained from the fitted model was then used to adjust the individual measurements to the mean age of the corresponding group according to the following equation:Tc=Ti±b(AGavg−Agi),where Tc is the age-corrected value of the trait, Ti is the recorded value of the trait for the individual, b is the regression coefficient describing the relationship between the trait and age, AGavg is the mean age of the corresponding sex and measurement group (days), and AGi is the age of the individual (days). The regression coefficient was estimated separately for each body measurement, sex, and measurement group, whereas farm- and year-specific regression coefficients were not fitted. Considering the relatively narrow age range within each measurement period, a linear relationship between age and body measurements was assumed to provide an appropriate approximation of age-related changes over the studied intervals. To evaluate whether the available sample size provided sufficient precision for the body measurements, the maximum allowable measurement error was estimated using the following equation:N>(λ×SDE)2 where N is the required sample size; λ is the standard normal deviate corresponding to the selected confidence level (λ = 2.57624; α = 0.01, P = 0.99); SD is the standard deviation of the evaluated trait; and E is the predefined allowable measurement error (cm). As an example, for the withers height of the bulls, the mean value was 112.25 cm, and the corresponding standard deviation was 4.649 cm. Substituting different allowable error values into the equation indicated that a sample size of at least 36, 64, and 143 animals was required to achieve allowable errors of 2.0, 1.5, and 1.0 cm, respectively. The same calculation was performed separately for each body measurement and each evaluation group. The estimated maximum allowable error values for all body measurements and groups are presented in Table 3.Table 3Estimated maximum acceptable error levels for body measurements based on the available sample sizes.Estimated maximum acceptable error (cm)NWHRHBLSWHWPWBulls1st measurement469110.50.50.50.52nd measurement4691110.50.50.5Heifers1st measurement3481110.50.50.52nd measurement34811110.50.5Breeding-approved and Removed bull groups9280.50.50.50.50.50.5 Table 3Estimated maximum acceptable error levels for body measurements based on the available sample sizes.Estimated maximum acceptable error (cm)NWHRHBLSWHWPWBulls1st measurement469110.50.50.50.52nd measurement4691110.50.50.5Heifers1st measurement3481110.50.50.52nd measurement34811110.50.5Breeding-approved and Removed bull groups9280.50.50.50.50.50.5 Estimated maximum acceptable error levels for body measurements based on the available sample sizes. Based on these calculations, the available sample sizes were sufficient to achieve a maximum allowable measurement error of 1.0 cm for all evaluated body measurements. For hip width and pin width, the available sample sizes also supported a higher level of precision, corresponding to a maximum allowable error of 0.5 cm in all groups. These allowable error values were utilised to evaluate whether the observed differences exceeded the anticipated measurement uncertainty. Correlation levels between body weight and body parameters were determined according to Szabó et al. (2004), as follows:•r | ≤ 0.4: weak correlation.•0.4 < | r | ≤ 0.7: moderate correlation.•0.7 < | r | ≤ 0.9: strong correlation.•0.9 < | r |: very strong correlation.rAVG was defined as the average correlation coefficient for all parameters (all body measurement traits) for data recorded at weaning and at 13–14 months of age. •r | ≤ 0.4: weak correlation. r | ≤ 0.4: weak correlation. •0.4 < | r | ≤ 0.7: moderate correlation. 0.4 < | r | ≤ 0.7: moderate correlation. •0.7 < | r | ≤ 0.9: strong correlation. 0.7 < | r | ≤ 0.9: strong correlation. •0.9 < | r |: very strong correlation. 0.9 < | r |: very strong correlation. Statistical analysis The data were evaluated using the Statistical Package for Social Sciences software (SPSS 24.0). Descriptive statistics, involving mean and standard deviation, were conducted to compare the age-adjusted and raw data of weaning bulls, including various body measurement parameters, age, and weight at measurement. The distributions of the analysed variables were assessed using Shapiro–Wilk tests, skewness and kurtosis statistics, histograms with superimposed normal curves, Q–Q plots, and box plots. The Shapiro–Wilk tests indicated statistically significant departures from exact normality (Supplementary Table S2). However, because formal normality tests can be sensitive to minor deviations in large samples (n = 817), the Shapiro–Wilk results were interpreted together with skewness, kurtosis, and graphical diagnostics (Ghasemi & Zahediasl, 2012; Kim, 2013). The distributions of most linear body measurements were approximately symmetric, although more pronounced departures from normality were observed for some age and derived-index variables (Supplementary Figure S1). Based on the combined assessment of the formal tests and graphical diagnostics, the distributions of the linear body measurements were considered sufficiently close to normal for the planned parametric analyses. Parametric methods, including correlation, analysis of variance, and linear regression, have also been applied to continuous morphometric traits in cattle and buffalo populations (Adinata et al., 2023; Getaneh et al., 2020; Vohra et al., 2015). Nevertheless, the selection of Pearson correlation in the present study was based primarily on the quantitative nature of the variables, the objective of quantifying linear associations, and graphical assessments of the distributions and pairwise relationships. The measurement records were obtained from the database of the Association of Hungarian Limousin and Blonde d’Aquitaine Breeders, where data-entry and recording errors had been checked before the records were accepted into the database. The repeated-measurement analysis included cattle with records available at both measurement times (n = 817), with no missing values for the variables included in this analysis. The bull-classification dataset comprised 928 animals. No data imputation was performed. One Removed bull had an incomplete record and was therefore excluded from the complete-case General Linear Model (GLM) analysis, resulting in a sample size of 927 animals for that analysis. No observations were excluded solely on the basis of their statistical identification as outliers. Paired-samples t-test was applied for the comparison of means of age-adjusted and raw data. For the paired-samples t-test, Cohen’s d effect size was applied according to the following categories (Cohen, 1988) (Table 4):Table 4Cohen’s d effect sizes and d-values for the paired-samples t-test.Effect sized-valueSmalld≈0.2Mediumd≈0.5Larged≈0.8 Table 4Cohen’s d effect sizes and d-values for the paired-samples t-test.Effect sized-valueSmalld≈0.2Mediumd≈0.5Larged≈0.8 Cohen’s d effect sizes and d-values for the paired-samples t-test. Independent-samples t-tests were used to compare body measurements and related traits between the Breeding-approved and Removed bull groups. Cohen’s d and its 95% confidence interval were calculated to quantify the magnitude and precision of the differences between the two groups. Effect sizes were calculated with the pooled standard deviation. To evaluate the effects of farm, year, and sex on the recorded traits, separate general linear model analyses were performed for the measurements obtained at weaning and at 13–14 months of age (n = 817). Farm (13 levels), year (2022–2025), and sex (bull and heifer) were included simultaneously as fixed main effects without interaction terms. Type III sums of squares were used, and effect sizes were expressed as partial eta squared (ηp²). For the separate bull-classification dataset, an additional GLM was performed to evaluate the effects of farm (12 levels), year (2022–2025), and final classification (‘Breeding-approved’ or ‘Removed’) on weaning-age traits. These factors were included simultaneously as fixed main effects without interaction terms, using Type III sums of squares. The complete-case analysis included 927 bulls (469 Breeding-approved and 458 Removed); one Removed bull was excluded because of incomplete data. Sex was not included because all animals were male. Age was analysed as a dependent trait rather than as a covariate. Effect sizes were expressed as partial eta squared (ηp²). Correlations between body measurement parameters were analysed using Pearson bivariate correlation. Scatterplots of the primary repeated-measurement relationships were inspected for non-linearity and potentially influential outlying observations; no marked non-linear patterns or isolated observations suggesting substantial distortion of the relationships were identified. Spearman’s rank correlation was considered but not selected as the primary method because the analysed body measurements were quantitative rather than ordinal variables, and the objective was to quantify linear associations between measurements. Linear regression models were fitted between the first and second measurements of each body dimension. In the analysis, the independent variable was the data of the first measurement, while the dependent variable was the results of the second measurement. Pearson correlation coefficients and unstandardized regression slopes were reported with their 95% confidence intervals. Regression outputs also included the standard errors of the slopes, intercepts, and standard errors of the estimate. A significance level of P < 0.05 was adopted as the threshold for statistical significance. No adjustment for multiple testing was applied because the correlation analyses were exploratory and involved biologically related, non-independent traits. Accordingly, the reported P-values are nominal and were interpreted together with the magnitude, direction, biological plausibility, and consistency of the observed associations. Results Descriptive statistics of the age-adjusted and raw data of weaning bulls For individual body measurements and calculated indices, the correlations between corrected and raw data ranged from r = 0.847 to r = 0.999 (P < 0.001) (Table 5). For both heifers and bulls, we obtained similar or even stronger correlations for the results measured at both measurement times. The paired-samples t-test did not indicate statistically significant differences (P > 0.05) between the raw and corrected data. Since there was a strong positive correlation between the two databases and the differences were not statistically significant, the raw data were used for further analysis to facilitate the interpretation of results. CV% for most parameters were below 10%, indicating low variability in the data. Body weight, corrected body weight and body mass index showed slightly higher, moderate CV values.Table 5Descriptive statistics (mean, standard deviation and coefficient of variation) of the age-adjusted and raw data of weaning bulls (n = 469) regarding age, weight and body measurement parameters.Body measurement parametersMeanStd. DeviationCV%First age at measurement (days)227.9421.0879.25Body weight (kg)307.0441.43413.49Corrected body weight (kg)307.0434.77111.32Withers height (WH) (cm)112.254.6464.14Corrected WH (cm)112.254.2013.74Rump height (RH) (cm)121.574.7063.87Corrected TH (cm)121.574.2113.46Back length (BL) (cm)61.803.6965.98Corrected BL (cm)61.803.5015.67Shoulder width (SW) (cm)32.732.6057.96Corrected SW (cm)32.732.5217.70Hip width (HW) (cm)37.972.4316.40Corrected HW (cm)37.972.2966.05Pin width (PW) (cm)12.521.0228.16Corrected PW (cm)12.530.9987.97Body mass index (kg*100/WH)272.8129.70210.89Corrected body mass index273.0925.2729.25Overgrowth index (TH*100/WH)108.321.2211.13Corrected overgrowth index108.321.2201.13Corrected weight and body measurement parameters were adjusted to an age of 228 days.Std. deviation: standard deviation.CV%: Coefficient of variation. Table 5Descriptive statistics (mean, standard deviation and coefficient of variation) of the age-adjusted and raw data of weaning bulls (n = 469) regarding age, weight and body measurement parameters.Body measurement parametersMeanStd. DeviationCV%First age at measurement (days)227.9421.0879.25Body weight (kg)307.0441.43413.49Corrected body weight (kg)307.0434.77111.32Withers height (WH) (cm)112.254.6464.14Corrected WH (cm)112.254.2013.74Rump height (RH) (cm)121.574.7063.87Corrected TH (cm)121.574.2113.46Back length (BL) (cm)61.803.6965.98Corrected BL (cm)61.803.5015.67Shoulder width (SW) (cm)32.732.6057.96Corrected SW (cm)32.732.5217.70Hip width (HW) (cm)37.972.4316.40Corrected HW (cm)37.972.2966.05Pin width (PW) (cm)12.521.0228.16Corrected PW (cm)12.530.9987.97Body mass index (kg*100/WH)272.8129.70210.89Corrected body mass index273.0925.2729.25Overgrowth index (TH*100/WH)108.321.2211.13Corrected overgrowth index108.321.2201.13Corrected weight and body measurement parameters were adjusted to an age of 228 days.Std. deviation: standard deviation.CV%: Coefficient of variation. Descriptive statistics (mean, standard deviation and coefficient of variation) of the age-adjusted and raw data of weaning bulls (n = 469) regarding age, weight and body measurement parameters. Corrected weight and body measurement parameters were adjusted to an age of 228 days. Std. deviation: standard deviation. CV%: Coefficient of variation. Effects of farm, year, and sex on the examined traits The GLM analyses showed that farm was significantly associated with all examined traits at both measurement times, with partial eta-squared values ranging from 0.135 to 0.311 at weaning and from 0.080 to 0.369 at 13–14 months of age (Supplementary Tables S1a and S1b). At weaning, the largest farm-associated effect sizes were observed for body weight (ηp² = 0.311) and withers height (ηp² = 0.290), whereas values for the remaining body measurements ranged from 0.236 to 0.277. Year generally showed smaller and trait-dependent effects, with the largest effect observed for hip width at both measurement times (ηp² = 0.127 and 0.132, respectively). Sex-associated effect sizes were relatively small at weaning, with the largest effect observed for withers height (ηp² = 0.100), but became considerably larger by the second measurement, particularly for body weight (ηp² = 0.724), body mass index (ηp² = 0.718), and shoulder width (ηp² = 0.532). Correlations between body parameters Results on the correlation between body parameters measured at weaning (at an average age of 230 days) are shown in Table 6.a. There was a moderate (SW1: r = 0.592) and strong (WH1: r = 0.773) correlation between body parameters and weight at the time of weaning. The exception to this was pin width, where a weak correlation was observed (PW1: r = 0.293). It is worth mentioning that there was a weak negative correlation (r = −0.166) between weaning weight and overgrowth index (n = 817, P ≤ 0.001). Hip width (HW1) was moderately correlated with withers height (WH1), rump height (RH1), back length (BL1), and shoulder width (SW1) (r = 0.419–0.694). The average correlation coefficient between the traits was rAVG = 0.581 (n = 817, P ≤ 0.001). Pin width (PW1) showed a weak positive correlation with all traits, with the weakest correlation in the case of SW1 (r = 0.096) and the strongest in the case of WH1 (r = 0.293). The overgrowth index showed weak negative correlations with values between r = −0.001 (P = 0.976) and r = −0.354 (P ≤ 0.001).Table 6.aCorrelations between body parameters and their 95% confidence intervals measured at weaning (at an average age of 230 days) (n = 817).Body parametersWeight1WH1RH1BL1SW1HW1PW1BMI1OI1Weight110.770⁎⁎0.773**0.652**0.592**0.656**0.293**0.975**−0.166**WH10.741–0.797#10.964**0.575**0.494**0.558**0.293**0.611**−0.354**RH10.744–0.8000.959–0.96910.566**0.515**0.594**0.283**0.627**−0.094**BL10.611–0.6900.527–0.6190.517–0.61110.430**0.419**0.232**0.612**−0.167**SW10.545–0.6340.440–0.5440.462–0.5630.372–0.48410.694**0.096**0.568**−0.039HW10.615–0.6930.509–0.6040.548–0.6370.361–0.4740.657–0.72810.199**0.623**−0.001PW10.229–0.3540.229–0.3550.219–0.3450.166–0.2960.027–0.1630.132–0.26410.263**−0.104**BMI10.971–0.9780.566–0.6530.583–0.6670.567–0.6530.520–0.6130.579–0.6630.198–0.3261−0.087*OI1−0.018–0.119−0.082–0.055−0.040–0.097−0.144-(−0.008)−0.082–0.0560.047–0.1820.033–0.169−0.005–0.1321Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index.⁎⁎Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values.⁎Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values.#95% confidence intervals for each correlation (under the diagonal). Table 6.aCorrelations between body parameters and their 95% confidence intervals measured at weaning (at an average age of 230 days) (n = 817).Body parametersWeight1WH1RH1BL1SW1HW1PW1BMI1OI1Weight110.770⁎⁎0.773**0.652**0.592**0.656**0.293**0.975**−0.166**WH10.741–0.797#10.964**0.575**0.494**0.558**0.293**0.611**−0.354**RH10.744–0.8000.959–0.96910.566**0.515**0.594**0.283**0.627**−0.094**BL10.611–0.6900.527–0.6190.517–0.61110.430**0.419**0.232**0.612**−0.167**SW10.545–0.6340.440–0.5440.462–0.5630.372–0.48410.694**0.096**0.568**−0.039HW10.615–0.6930.509–0.6040.548–0.6370.361–0.4740.657–0.72810.199**0.623**−0.001PW10.229–0.3540.229–0.3550.219–0.3450.166–0.2960.027–0.1630.132–0.26410.263**−0.104**BMI10.971–0.9780.566–0.6530.583–0.6670.567–0.6530.520–0.6130.579–0.6630.198–0.3261−0.087*OI1−0.018–0.119−0.082–0.055−0.040–0.097−0.144-(−0.008)−0.082–0.0560.047–0.1820.033–0.169−0.005–0.1321Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index.⁎⁎Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values.⁎Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values.#95% confidence intervals for each correlation (under the diagonal). Correlations between body parameters and their 95% confidence intervals measured at weaning (at an average age of 230 days) (n = 817). Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index. Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values. Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values. 95% confidence intervals for each correlation (under the diagonal). Results on the correlation between body parameters measured at 13–14 months of age (average age of 424 days) are shown in Table 6.b. There was a strong positive correlation between body measurements taken at 13–14 months of age and body weight (BL2: r = 0.726, SW2: r = 0.812) (n = 817, P ≤ 0.001). The exception in this case was PW2, where a weak negative correlation (r = −0.009) was observed but was not statistically substantiated (P > 0.05). A weak negative correlation (r = −0.192) was also observed between body weight and the overgrowth index (n = 817, P ≤ 0.001). The WH2, RH2, BL2, SW2, and HW2 traits were moderately correlated with each other, but overall, the correlations were stronger than those observed in the weaning results. The lowest value (moderate correlation; r = 0.424) was between RH2 and HW2, while the highest (very strong correlation; r = 0.958) was between WH2 and RH2. The average correlation coefficient between the traits was rAVG= 0.640 (n = 817, P ≤ 0.001). Pin width (PW2) showed a weak correlation with all traits, with the weakest in the case of SW2 (r = −0.093) and the strongest in the case of RH2 (r = 0.165). The overgrowth index also showed weak negative correlations, ranging from r = −0.251 to r = −0.027 (n = 817, P ≤ 0.001); however, not all associations were statistically significant. The BMI indicated the highest (very strong) correlation with body weight (r = 0.989).Table 6.bCorrelations between body parameters and their 95% confidence intervals measured at an average age of 424 days (n = 817).Body parameterWeight2WH2RH2BL2SW2HW2PW2BMI2OI2Weight210.761⁎⁎0.729**0.726**0.812**0.762**−0.0090.989**−0.192**WH20.731–0.789#10.958**0.705**0.546**0.613**0.123**0.660**−0.246**RH20.695–0.7590.952–0.96410.683**0.490**0.424**0.165**0.632**0.041BL20.691–0.7570.669–0.7380.645–0.71810.551**0.551**0.108**0.682**−0.148**SW20.787–0.8340.496–0.5930.436–0.5400.502–0.59710.674**−0.093**0.821**−0.251**HW20.732–0.7890.568–0.6540.580–0.6640.502–0.5970.634–0.71010.122**0.753**−0.027PW2−0.077–0.0600.055–0.1900.098–0.2310.040–0.176−0.160-(−0.024)0.053–0.1891−0.0310.132**BMI20.987–0.9900.620–0.6970.589–0.6710.644–0.7170.797–0.8420.721–0.781−0.099–0.0381−0.168**OI2−0.257-(−0.125)−0.309-(−0.180)−0.028–0.109−0.215-(−0.081)−0.314-(−0.185)−0.096–0.0410.064–0.199−0.233-(−0.100)1Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index.⁎⁎Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values.⁎Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values.#95% confidence intervals for each correlation (under the diagonal). Table 6.bCorrelations between body parameters and their 95% confidence intervals measured at an average age of 424 days (n = 817).Body parameterWeight2WH2RH2BL2SW2HW2PW2BMI2OI2Weight210.761⁎⁎0.729**0.726**0.812**0.762**−0.0090.989**−0.192**WH20.731–0.789#10.958**0.705**0.546**0.613**0.123**0.660**−0.246**RH20.695–0.7590.952–0.96410.683**0.490**0.424**0.165**0.632**0.041BL20.691–0.7570.669–0.7380.645–0.71810.551**0.551**0.108**0.682**−0.148**SW20.787–0.8340.496–0.5930.436–0.5400.502–0.59710.674**−0.093**0.821**−0.251**HW20.732–0.7890.568–0.6540.580–0.6640.502–0.5970.634–0.71010.122**0.753**−0.027PW2−0.077–0.0600.055–0.1900.098–0.2310.040–0.176−0.160-(−0.024)0.053–0.1891−0.0310.132**BMI20.987–0.9900.620–0.6970.589–0.6710.644–0.7170.797–0.8420.721–0.781−0.099–0.0381−0.168**OI2−0.257-(−0.125)−0.309-(−0.180)−0.028–0.109−0.215-(−0.081)−0.314-(−0.185)−0.096–0.0410.064–0.199−0.233-(−0.100)1Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index.⁎⁎Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values.⁎Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values.#95% confidence intervals for each correlation (under the diagonal). Correlations between body parameters and their 95% confidence intervals measured at an average age of 424 days (n = 817). Weight: body weight (kg), WH: withers height (cm), RH: rump height (cm), BL: back length (cm), SW: shoulder width (cm), HW: hip width (cm), PW: pin width (cm), BMI: body mass index, OI: overgrowth index. Correlation is significant at the 0.01 level (P ≤ 0.01) based on unadjusted P-values. Correlation is significant at the 0.05 level (P ≤ 0.05) based on unadjusted P-values. 95% confidence intervals for each correlation (under the diagonal). Correlations between the parameters of repeated measurements The results indicated a weaker correlation between body measurements taken at weaning age (rAVG = 0.581) than at 13–14 months of age (rAVG = 0.640, n = 817). With regard to withers height (WH), a positive, moderate correlation was found between the two measurement time points (Fig. 3.a), with a standard error of 3.95 cm (n = 817, r = 0.618, r2 = 0.382, SE = 3.952; P < 0.001). In the case of rump height (RH), a moderate positive correlation was also observed between the two measurement time points (Fig. 3.b) with a standard error of 4.08 cm (n = 817, r = 0.631, r2 = 0.398, SE = 4.079; P < 0.001).Fig. 3a and 3.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) withers height (WH) and rump height (RH) measurements. Points indicate the results of the first (WH1; RH1) and second (WH2; RH2) measurements. Relationship between data is illustrated by the linear regression line. a and 3.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) withers height (WH) and rump height (RH) measurements. Points indicate the results of the first (WH1; RH1) and second (WH2; RH2) measurements. Relationship between data is illustrated by the linear regression line. In the case of back length (BL), a moderate positive correlation was found between the two measurement time points with a standard error of 4.15 cm (n = 817, r = 0.415, r2 = 0.172, SE = 4.151; P < 0.001) (Fig. 4.a). In the case of shoulder width (SW), a weak positive correlation was detected between the two measurement time points with a standard error of 3.96 cm (n = 817, r = 0.267, r2 = 0.071, SE = 3.966; P < 0.001) (Fig. 4.b).Fig. 4a and 4.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) back length (BL) and shoulder width (SW) measurements. Points indicate the results of the first (BL1; SW1) and second (BL2; SW2) measurements. Relationship between data is illustrated by the linear regression line. a and 4.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) back length (BL) and shoulder width (SW) measurements. Points indicate the results of the first (BL1; SW1) and second (BL2; SW2) measurements. Relationship between data is illustrated by the linear regression line. In the case of hip width (HW), a moderate positive correlation was found between the two measurement time points with a standard error of 3.08 cm (n = 817, r = 0.433, r2 = 0.188, SE = 3.079; P < 0.001) (Fig. 5.a). In the case of pin width (PW), a weak positive correlation was detected between the two measurement time points with a standard error of 0.95 cm (n = 817, r = 0.264, r2 = 0.070, SE= 0.956; P < 0.001) (Fig. 5.b). No similar studies on correlations in pin width by age have been reported in the international literature for comparison.Fig. 5a and 5.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) hip width (HW) and pin width (PW) measurements. Points indicate the results of the first (HW1; PW1) and second (HW2; PW2) measurements. Relationship between data is illustrated by the linear regression line. a and 5.b Correlations between the first (at an average age of 230 days) and second (at an average age of 424 days) hip width (HW) and pin width (PW) measurements. Points indicate the results of the first (HW1; PW1) and second (HW2; PW2) measurements. Relationship between data is illustrated by the linear regression line. The results of Pearson correlation and linear regression are presented in Table 7.Table 7Pearson correlations and linear regression parameters between body measurements recorded at weaning and at 13–14 months of age in Limousin cattle (n = 817).Body measurement parameterr (95% CI)r²Slope, B (SE)(95% CI)InterceptSE of estimate (cm)P-valueWithers height (WH)0.618(0.573–0.658)0.3820.669 (0.030)(0.610–0.727)51.9133.952<0.001Rump height (RH)0.631(0.588–0.670)0.3980.700 (0.030)(0.640–0.759)51.5124.079<0.001Back length (BL)0.415(0.357–0.470)0.1720.461 (0.035)(0.392–0.531)43.6214.151<0.001Shoulder width (SW)0.267(0.202–0.330)0.0710.385 (0.049)(0.290–0.480)26.5683.966<0.001Hip width (HW)0.433(0.376–0.487)0.1880.592 (0.043)(0.508–0.677)23.7573.079<0.001Pin width (PW)0.264(0.199–0.327)0.0700.262 (0.033)(0.196–0.327)11.2200.956<0.001CI: Confidence interval. Table 7Pearson correlations and linear regression parameters between body measurements recorded at weaning and at 13–14 months of age in Limousin cattle (n = 817).Body measurement parameterr (95% CI)r²Slope, B (SE)(95% CI)InterceptSE of estimate (cm)P-valueWithers height (WH)0.618(0.573–0.658)0.3820.669 (0.030)(0.610–0.727)51.9133.952<0.001Rump height (RH)0.631(0.588–0.670)0.3980.700 (0.030)(0.640–0.759)51.5124.079<0.001Back length (BL)0.415(0.357–0.470)0.1720.461 (0.035)(0.392–0.531)43.6214.151<0.001Shoulder width (SW)0.267(0.202–0.330)0.0710.385 (0.049)(0.290–0.480)26.5683.966<0.001Hip width (HW)0.433(0.376–0.487)0.1880.592 (0.043)(0.508–0.677)23.7573.079<0.001Pin width (PW)0.264(0.199–0.327)0.0700.262 (0.033)(0.196–0.327)11.2200.956<0.001CI: Confidence interval. Pearson correlations and linear regression parameters between body measurements recorded at weaning and at 13–14 months of age in Limousin cattle (n = 817). CI: Confidence interval. Measurement 1, recorded at weaning (average age: 230 days), was used as the independent variable, and measurement 2, recorded at 13–14 months of age (average age: 424 days), as the dependent variable. Values in parentheses below ‘r’ are 95% confidence intervals for Pearson correlation coefficient. ‘B’ denotes the unstandardized regression slope; values in the first parentheses following ‘B’ are its standard errors, and values in the second parentheses are its 95% confidence intervals. ‘r²’ denotes the squared Pearson correlation coefficient. SE of estimate denotes the residual standard error of the regression model. All P-values are nominal. Correlations between corresponding measurements were positive for all traits, ranging from r = 0.264 for pin width to r = 0.631 for rump height (Table 7). The corresponding r² values ranged from 0.070 to 0.398. Regression slopes were also positive: a 1-cm difference in a body measurement at weaning was associated with an estimated 0.262–0.700-cm difference in the corresponding measurement at 13–14 months of age, depending on the trait. All slope estimates were statistically significant, and their 95% confidence intervals excluded zero (P < 0.001). Comparison of breeding-approved and removed bulls at weaning age The weight and body size data of the bull calves at weaning were compared according to their later classification as Breeding-approved or Removed bulls following the breeding selection process. The study included 928 bull calves, of which 469 were ultimately classified as Breeding-approved bulls and 459 as Removed bulls based on the outcome of the breeding selection process. As shown in Table 8, the Removed bulls group had higher standard deviation values, even though the Breeding-approved bulls group had higher average values, except for pin width. There was no difference between the ages of the two groups (P = 0.122) as supported by the results of the t-test (Table 9.a). Based on the results, we found that, with the exception of the overgrowth index and pin width, the performance of Breeding-approved bulls exceeded that of Removed bulls in all parameters (P < 0.05) according to the results of t-test (Table 9.a). The greatest difference was found in measured weight, where Breeding-approved bulls showed 37.233 kg higher (P < 0.001 – Table 9.a) body weight than Removed bulls. In terms of body size, there were slight differences between the groups, ranging from 1.168 to 3.162 cm, with higher values observed for the Breeding-approved bulls group in most body parameters. In the case of PW, although the difference of 0.159 cm was statistically significant (Table 9.a), practical experience and the estimated allowable measurement error suggest that it is negligible.Table 8Comparison of body measurement parameters between Breeding-approved bulls (n = 469) and Removed bulls (n = 459).Body measurement parametersClassificationMeanStd. DeviationCV%Age at first measurement (days)Breeding-approved227.94a21.0879.3Removed223.28a60.98327.3Body weight (kg)Breeding-approved307.04b41.43413.5Removed269.81a47.24017.5Withers height (WH) (cm)Breeding-approved112.25b4.6464.1Removed109.10a5.8245.3Rump height (RH) (cm)Breeding-approved121.57b4.7063.9Removed118.41a5.8604.9Back length (BL) (cm)Breeding-approved61.80b3.6966.0Removed58.64a4.2107.2Shoulder width (SW) (cm)Breeding-approved32.73b2.6058.0Removed30.72a3.53411.5Hip width (HW) (cm)Breeding-approved37.97b2.4316.4Removed36.80a3.3409.1Pin width (PW) (cm)Breeding-approved12.52b1.0228.2Removed12.37a1.0228.3Body mass index (kg*100/WH)Breeding-approved272.81b29.70210.9Removed245.60a36.00514.7Overgrowth index (TH*100/WH)Breeding-approved108.32a1.2211.1Removed108.56b1.5671.4a,bDifferent letter within a row of each body measurement parameter between ‘Breeding-approved’ and ‘Removed’ indicates a significant difference (P < 0.05).⁎Std. deviation: standard deviation.CV% Coefficient of variation.Table 9.aMean statistical differences (Levene's Test for Equality of Variances and t-test for Equality of Means) in body measurement parameters between the Breeding-approved and Removed bull groups.Body measurement parametersLevene's Test for Equality of Variancest-test for Equality of MeansFSig.tdfSignificance Two-Sided PMean DifferenceStd. Error DifferenceAge at first measurement (days)7.5800.0061.549563.9000.1224.6613.008Body weight (kg)6.9100.00912.746903.587<0.00137.2332.921Withers height (WH) (cm)20.0350.0019.106874.244<0.0013.1540.346Rump height (RH) (cm)17.1610.0019.051876.745<0.0013.1620.349Back length (BL) (cm)5.4690.02012.113905.381<0.0013.1530.260Shoulder width (SW) (cm)45.5770.0019.869841.725<0.0012.0150.204Hip width (HW) (cm)27.3400.0016.082836.193<0.0011.1680.192Pin width (PW) (cm)0.0030.9562.362925.5700.0180.1590.067Body mass index (kg*100/WH)9.3500.00212.545886.431<0.00127.2142.169Overgrowth index (TH*100/WH)4.0810.044−2.679865.1790.008−0.2470.092⁎df: degrees of freedom; Std. Error Difference: standard error of the difference. Table 8Comparison of body measurement parameters between Breeding-approved bulls (n = 469) and Removed bulls (n = 459).Body measurement parametersClassificationMeanStd. DeviationCV%Age at first measurement (days)Breeding-approved227.94a21.0879.3Removed223.28a60.98327.3Body weight (kg)Breeding-approved307.04b41.43413.5Removed269.81a47.24017.5Withers height (WH) (cm)Breeding-approved112.25b4.6464.1Removed109.10a5.8245.3Rump height (RH) (cm)Breeding-approved121.57b4.7063.9Removed118.41a5.8604.9Back length (BL) (cm)Breeding-approved61.80b3.6966.0Removed58.64a4.2107.2Shoulder width (SW) (cm)Breeding-approved32.73b2.6058.0Removed30.72a3.53411.5Hip width (HW) (cm)Breeding-approved37.97b2.4316.4Removed36.80a3.3409.1Pin width (PW) (cm)Breeding-approved12.52b1.0228.2Removed12.37a1.0228.3Body mass index (kg*100/WH)Breeding-approved272.81b29.70210.9Removed245.60a36.00514.7Overgrowth index (TH*100/WH)Breeding-approved108.32a1.2211.1Removed108.56b1.5671.4a,bDifferent letter within a row of each body measurement parameter between ‘Breeding-approved’ and ‘Removed’ indicates a significant difference (P < 0.05).⁎Std. deviation: standard deviation.CV% Coefficient of variation. Comparison of body measurement parameters between Breeding-approved bulls (n = 469) and Removed bulls (n = 459). Different letter within a row of each body measurement parameter between ‘Breeding-approved’ and ‘Removed’ indicates a significant difference (P < 0.05). Std. deviation: standard deviation. CV% Coefficient of variation. Table 9.aMean statistical differences (Levene's Test for Equality of Variances and t-test for Equality of Means) in body measurement parameters between the Breeding-approved and Removed bull groups.Body measurement parametersLevene's Test for Equality of Variancest-test for Equality of MeansFSig.tdfSignificance Two-Sided PMean DifferenceStd. Error DifferenceAge at first measurement (days)7.5800.0061.549563.9000.1224.6613.008Body weight (kg)6.9100.00912.746903.587<0.00137.2332.921Withers height (WH) (cm)20.0350.0019.106874.244<0.0013.1540.346Rump height (RH) (cm)17.1610.0019.051876.745<0.0013.1620.349Back length (BL) (cm)5.4690.02012.113905.381<0.0013.1530.260Shoulder width (SW) (cm)45.5770.0019.869841.725<0.0012.0150.204Hip width (HW) (cm)27.3400.0016.082836.193<0.0011.1680.192Pin width (PW) (cm)0.0030.9562.362925.5700.0180.1590.067Body mass index (kg*100/WH)9.3500.00212.545886.431<0.00127.2142.169Overgrowth index (TH*100/WH)4.0810.044−2.679865.1790.008−0.2470.092⁎df: degrees of freedom; Std. Error Difference: standard error of the difference. Mean statistical differences (Levene's Test for Equality of Variances and t-test for Equality of Means) in body measurement parameters between the Breeding-approved and Removed bull groups. df: degrees of freedom; Std. Error Difference: standard error of the difference. Detailed statistical differences between the groups classified as Breeding-approved and Removed bulls regarding Levene's Test for Equality of Variances and t-test for Equality of Means are presented in Table 9.a. The Levene’s test indicated that the assumption of homogeneity of variances was violated for most traits (P < 0.05), except for pin width, where equal variances could be assumed (P = 0.956). The t-test revealed statistically significant differences between the Breeding-approved and Removed bull groups for nearly all measured parameters (P < 0.05), except for age at first measurement (P = 0.122). The magnitude and direction of the mean differences indicate that the Breeding-approved group consistently exhibited higher values across most body dimensions, particularly for body weight and body mass index, while the negative mean difference observed for the overgrowth index suggests a relative proportional variation between height traits. Equality of variances was not rejected for PW (Levene’s test, P > 0.05); therefore, the equal-variances-assumed t-test result was reported for this trait. For all other traits, equality of variances was rejected (P < 0.05), and Welch’s t-test with adjusted degrees of freedom was used. Table 6a reports the corresponding t-values and degrees of freedom selected on the basis of Levene’s test. Table 9.b presents the Cohen’s d independent samples effect sizes and the 95% confidence interval of body measurement traits of Breeding-approved and Removed bulls.Table 9.bCohen’s d effect sizes and 95% confidence intervals (CI) for the comparison of Breeding-approved and Removed bull groups regarding body measurement parameters and related traits.TraitCohen's dInterpretation95% CIAge at first measurement0.103Small−0.026 to 0.231Body weight0.839Large0.704 to 0.973Withers height0.599Medium0.468 to 0.731Rump height0.596Medium0.464 to 0.727Back length0.796Medium0.663 to 0.930Shoulder width0.650Medium0.518 to 0.782Hip width0.401Medium0.271 to 0.531Pin width0.155Small0.026 to 0.284Body Mass Index0.825Large0.691 to 0.959Overgrowth index−0.176Small−0.305 to −0.047 Table 9.bCohen’s d effect sizes and 95% confidence intervals (CI) for the comparison of Breeding-approved and Removed bull groups regarding body measurement parameters and related traits.TraitCohen's dInterpretation95% CIAge at first measurement0.103Small−0.026 to 0.231Body weight0.839Large0.704 to 0.973Withers height0.599Medium0.468 to 0.731Rump height0.596Medium0.464 to 0.727Back length0.796Medium0.663 to 0.930Shoulder width0.650Medium0.518 to 0.782Hip width0.401Medium0.271 to 0.531Pin width0.155Small0.026 to 0.284Body Mass Index0.825Large0.691 to 0.959Overgrowth index−0.176Small−0.305 to −0.047 Cohen’s d effect sizes and 95% confidence intervals (CI) for the comparison of Breeding-approved and Removed bull groups regarding body measurement parameters and related traits. Data in Table 9.b show that differences between the two groups were small for age and the overgrowth index. In contrast, Breeding-approved bulls had significantly higher initial body weight and body mass index. For the remaining traits, differences between the groups were modest, with higher mean values observed in the Breeding-approved bulls. In the complete-case GLM (n = 927), farm and year were significantly associated with all examined traits, although their effect sizes varied among traits (Supplementary Table S1c). The largest farm-associated effect was observed for pin width (ηp² = 0.252), whereas the largest year-associated effect was found for hip width (ηp² = 0.217). After accounting for farm and year, final classification was significantly associated with body weight, all six body measurements, and body mass index (all P < 0.001; ηp² = 0.025–0.109). The largest classification-associated effect sizes were observed for body weight (ηp² = 0.109), body mass index (ηp² = 0.105), and shoulder width (ηp² = 0.077). No classification-associated differences were detected for age at first measurement (P = 0.558, ηp² < 0.001) or the overgrowth index (P = 0.349, ηp² = 0.001). Discussion Several studies have demonstrated that body weight and linear body measurements play a crucial role in selection decisions aimed at improving desirable genetic traits, growth potential, carcass yield, and overall animal health in cattle populations (Kamprasert et al., 2019; Naserkheil et al., 2020; Nurgiartiningsih et al., 2022). Consequently, the relationship between body parameter assessment and preselection in cattle breeding is an essential aspect of sustainable livestock production and genetic improvement strategies. As the demand for high-quality beef products increases, the focus on efficient selection methods becomes paramount. This study aimed to assess the relationship between body measurements recorded at weaning and at 13–14 months of age in Limousin cattle. Furthermore, the objective was to compare weaning-age body measurements between bull calves classified as Breeding-approved and those classified as Removed during the breeding selection process. Finally, we aimed to identify linear traits that exhibit sufficient phenotypic stability to serve as practical indicators for preselection. Correlations between body parts The results showed that there were strong correlations between corrected and raw data ranging between r = 0.847 and r = 0.999. CV% for most parameters were below 10% indicating low variability in data. There were moderate and strong correlations between body parameters and weight (SW1: r = 0.592, WH1: r = 0.773) at the time of weaning. However, we found a weak correlation between body weight and pin width (r = 0.293). Hip width (HW1) was moderately correlated with withers height (WH1), rump height (RH1), back length (BL1), and shoulder width (SW1) (r = 0.419–0.694). Cattle body measurements during growth are influenced by several factors, including genetics, nutrition, and management practices. For instance, Warman et al. (2023) performed a morphometric analysis of Limousin–Bali crossbred cattle, noting increases in withers height and hip height throughout their growth phases. The study highlighted that Limousin crosses generally exhibit larger body dimensions compared to their Bali counterparts. In the study undertaken by Yakubu (2010) in White Fulani cows aged between 1.5 and 2.4 years, the highest correlation was observed between withers height and rump height (r = 0.98) and the lowest was between rump height and shoulder width (r = 0.51). Similar to our findings, in Cumra, in Holstein cows between 26 and 36 months of age, Tasdemir et al. (2011) found medium correlations of 0.66 and 0.63 between body weight and withers height, and body weight and hip width, respectively. Rezende et al. (2022) found a correlation of 0.76 between body weight at 210 days of age and at 365 days of age in Limousin cattle, suggesting that early growth performance was a strong predictor of later body weight. In Indonesia, in Bali breed, Azis et al. (2023) reported strong positive correlations between body weight and hip height (r = 0.756), body weight and body length (r = 0.754), and body weight and chest girth (r = 0.877) highlighting the close relationship between live weight and body conformation traits. Previous studies suggest that body measurements may serve as reliable predictors of body weight through regression-based approaches, particularly in field conditions where direct weighing is not feasible. Similar results were established by Budianto et al. (2022), who observed positive correlations (r = 0.619–0.809) between body weight and several morphometric traits, including chest depth, width and circumference, abdominal circumference, withers height, pelvic height, hip height, and rump width in Limousin x Peranakan Ongole crossbred slaughter cattle. In Ongole cattle aged 2–5 years in South Lampung Regency, Dakhlan et al. (2024) reported very strong positive correlations between body weight and chest girth (r = 0.91), as well as body weight and body length (r = 0.91). These results further indicate that linear body measurements, such as withers height, hip height and body length, can be effectively used for predicting body weight and provide a practical basis for digital phenotyping systems (Ruchay et al., 2020). Correlation between the parameters of repeated measures Our study revealed a strong positive correlation between body measurements taken at 13–14 months of age and body weight (BL2: r = 0.726, SW2: r = 0.812) (n = 817, P ≤ 0.001), while PW2 showed a weak negative correlation with body weight (r = −0.009), it was not statistically substantiated (P > 0.05). A weak negative correlation (r = −0.192) was also observed between body weight and the overgrowth index (n = 817, P ≤ 0.001). The correlation between body measurements taken at weaning age (rAVG= 0.581) was weaker than that of 13–14 months of age (rAVG= 0.640, n = 817). Consequently, the present findings supported our hypothesis that body measurements would show stronger correlations and greater homogeneity at the later growth stage. The average correlation coefficient increased from weaning to 13–14 months of age, while the coefficient of variation remained low for most body measurements, indicating increased phenotypic consistency during growth. There were positive, moderate correlations for WH, RH, BL, and HW between the two measurement time points. In a study involving Hungarian Simmental, Hereford, Aberdeen Angus, Red Angus, Lincoln Red, Limousin, Charolais, Blonde d’Aquitaine and Shaver cows aged 4–12 years, withers height showed variable correlations with age depending on the breed. In Hungarian Simmental, a weak positive correlation was observed (r = 0.08), while Hereford also exhibited a weak positive correlation (r = 0.32). In contrast, a weak negative correlation was found in Angus (r = −0.06), and for the overall mean across all breeds, withers height showed a weak negative correlation (r = −0.01) with age (Bene et al., 2007). Consequently, the authors found somewhat contrasting results compared to our findings. Such discrepancies may be attributed to differences in breed, age, feeding, husbandry, and management conditions. In the study of Vargas et al. (2000), the phenotypic correlation of hip height (in our case rump height) at weaning and post-weaning (at 18 months of age) in Brahman cattle was r = −0.46 which was lower than in our case. The difference may be attributed to differences in breed, body weight, sex and management. However, a related study examined the correlation between body length and the rate of maturation, defined as the postnatal growth rate relative to the mature size, and reported a strong negative correlation of −0.90 (Hafiz et al., 2014). Furthermore, Bene et al. (2007) reported correlations between age and body length of r = 0.37 for Hungarian Simmental, r = 0.03 for Hereford, and r = 0.20 for Angus aged 4–12 years. These correlations are lower than those we observed; however, the correlation for Hungarian Simmental was relatively close to our result. Differences between studies may be related to alterations in breeds, sex, ages, and environmental conditions. Regarding hip width, age-related associations are available in the case of age at first calving in Hanwoo beef cows. Authors indicated very weak negative phenotypic correlations between HW and age at first calving (AFC) (r = −0.013 for < 24 months; r = −0.022 for > 24 months), suggesting minimal association between body size and reproductive timing (Shin et al., 2021). These findings suggest that HW may be a useful indicator of growth and body development, but it has limited predictive value for AFC, highlighting the need to consider additional traits when selecting for reproductive performance. Furthermore, moderate associations between withers and rump height and age indicate that these traits are less influenced by short-term environmental variation compared to others. These traits, therefore, can be used in a breeding strategy to plan corrective matings (Tőzsér et al., 2023). Associations between body weight, body measurement parameters and age aid in the early selection and optimization of breeding programmes for beef cattle (Feng et al., 2025). The measurement of body weight and correlated body parameters can be used to support selection for improved growth, enabling the simultaneous improvement of correlated traits in the breeding programme (Kamprasert et al., 2019). Comparison of breeding-approved and removed bulls at weaning age Based on the results, we found that, with the exception of the overgrowth index and pin width, Breeding-approved bulls showed significantly higher values than Removed bulls in all parameters (P < 0.05). The greatest difference was found in body weight, where Breeding-approved bulls had 37.233 kg higher body weight than Removed bulls. In terms of body size, there were slight differences between the groups, ranging from 1.168 to 3.162 cm with higher values for the Breeding-approved bulls in most body parameters. It should be noted that the comparison between the Breeding-approved and Removed bull groups does not represent an independent validation of the predictive value of body measurements. At the time of weaning, all bull calves were measured before the outcome of the breeding selection process was known, while final breeding approval was based on additional information, including pedigree, performance during the own performance test, and, where available, genomic information. Therefore, these results should be interpreted as evidence that body measurements are associated with the final breeding selection outcome rather than as proof that body measurements alone are sufficient for identifying future breeding bulls. Nevertheless, these findings highlight the potential value of morphometric traits as additional information sources in future data-driven selection approaches. Body measurements provide valuable information for both breeding and fattening, supporting selection, the assessment of body weight development, estimation of breeding value, slaughter value, and meat production, as well as the evaluation of animal health (Li et al., 2022). Bene et al. (2007) measured the overgrowth index of breeding-purpose cows from various cattle breeds including Hungarian Simmental, Hereford, Aberdeen Angus, Red Angus, Lincoln Red, Limousin, Charolais, Blonde d’Aquitaine and Shaver cows aged between 4 and 12 years, and found similar overgrowth indexes ranging between 101 and 104. For Limousin, the index was 103.84, showing slightly lower value compared to ours, which can be explained by the difference in sex. In Italy, Rezende et al. (2022) measured the body weight of Limousin cattle 247 kg at an age of 210 day sand 411 kg at 365 days. Animals had 60 kg lower weight and were 18 days younger compared to our study (307.04 kg with an age of 227.94 days). There was no information available on the grazing, feeding and climatic conditions for further conclusions. In Indonesia, in Sumba Ongole cattle, mean body weight of bulls at 673 days was 299 kg with 122 cm withers height indicating three times older animals with 8 kg lower body weight and 10 cm higher withers height. Cattle were fed Elephant grass (Pennisetum purpureum), rice straw and cassava meal ad libitum (Putra, 2020). In West Java, Limousin bulls aged between 1.5 and 13 years showed body weight between 384 and 1021 kg. However, the reported correlations were calculated for chest girth, body length, and shoulder height rather than for the traits investigated in the present study, limiting direct comparison (Adhianto et al., 2025). Hungarian Simmental, Hereford, Aberdeen Angus, Red Angus, Lincoln Red, Limousin, Charolais, Blonde d’Aquitaine and Shaver cows’ live weight and body measurements were recorded. Authors found that weights ranged between 500 (Hungarian Simmental) and 638 kg (Blonde d’Aquitaine), and Limousins showed 591 kg, which was approximately twice the mean value observed in the present study. However, animals in their experiment were fully developed at ages between 4 and 12 years. The feed was pasture, corn silage, hay and concentrate feed supplement (Bene et al., 2007). In Brahman, the body weight at an age of 200 days was 160.72 kg with 102.05 cm hip height. At an age of 400 days, body weight increased to 220.35 kg with 115.8 cm hip height. Cattle were fed with pasture supplemented with silage, hay or concentrate (Kamprasert et al., 2019). In North Central Nigeria, in White Fulani cows aged between 1.5 and 2.4 years, Yakubu (2010) established body weight of 116 kg, withers height of 83 cm, hip width of 12 cm, and shoulder width of 18 cm. The author measured much lower values compared to our measurements. No data were available on the feeding and husbandry conditions. Limousin female calves bred in Bulgaria showed 126.5 cm withers height and 13.2 cm pin width at 1 year of age. They measured 14 cm higher withers height and 0.7 cm greater pin width. The feed included pasture and concentrate fodder (Karamfilov et al., 2020). Differences between studies may be related to variation in genetics, age, sex, feed composition, climate, husbandry, and management conditions. Studies comparing breeding-approved and removed bulls following the breeding selection process are not available in the international literature. Our analysis, in this regard, addresses an important gap. In summary, our findings suggest that early body measurements are associated with subsequent growth and may support early phenotypic evaluation. The observed associations between morphometric traits and the final breeding selection outcome indicate that these measurements may provide valuable additional information for selection decisions. These traits may also be assessed using automated technologies, such as ultrasound- and image-based methods, which are reliable approaches for meat production assessment (Ruchay et al., 2025). RGB-D image capture systems offer a practical and time-saving approach for the objective assessment of morphological traits and may further support animal selection (Gritsenko et al., 2023). Furthermore, the development of farm-specific decision-support tools, and future integration of morphometric measurements with pedigree, genomic, and performance data, for example a breeder-oriented platform, could improve the efficiency of preselection. Limitations of the study and avenues for future research The aim of this study was to identify parameters describing phenotypic associations between body measurements recorded at weaning and at 13–14 months of age. By systematically collecting body size data, breeders may obtain objective information that could support the early phenotypic evaluation of breeding animals. Beyond the numerous strengths highlighted throughout the manuscript, it is also important to acknowledge several limitations inherent to the present study. These constraints should be considered when interpreting the findings and their broader applicability. One of the main challenges of the study is the lack of standardised experimental conditions. In Hungary, breeding populations are typically small, making it difficult to examine large numbers of animals (e.g., over 100 individuals) of the same age, measured at the same time, and raised under identical conditions. As the dataset was compiled from multiple farms, where animals were kept under broadly similar but not identical (i.e., differences in genetics, feeding, husbandry, and management) conditions and the timing of data collection varied across the year. In addition, a potential source of selection bias is that only farms available for participation were included in the study. Therefore, the results may not be completely representative of the wider Limousin population, limiting the generalizability of the findings. Although farm, year, and sex were included as fixed effects in the GLM analyses of the repeated-measurement dataset, and farm, year, and final classification were included as fixed effects in the bull-group analysis, potential residual clustering within farms and genetic relatedness among animals were not explicitly modelled as random effects. Furthermore, genetic parameters were not estimated. The lack of repeated measurements and assessment of intra-observer reliability represents another limitation of the present study, as the potential influence of measurement error on the results could not be quantified. Repeated measurements were not feasible because they would have required substantial additional labour and logistical resources from the participating breeders. Future research will aim to establish collaboration with one or more breeders to enable repeated measurements under controlled conditions and to evaluate the repeatability and intra-observer reliability of the measurement protocol. Finally, although the identified linear traits showed promising associations with later performance, their suitability as practical preselection indicators was not confirmed using an independent dataset or predictive validation approach. Future studies should address these limitations to confirm the robustness and generalizability of the findings. For future analyses, this research could be continued by expanding the dataset with data based on artificial intelligence and machine learning approaches to reveal deeper relationships. Such approaches may facilitate the development and validation of predictive models for early phenotypic evaluation. Conclusions In conclusion, withers height (WH) and rump height (RH) showed the strongest associations between corresponding measurements across the two time points and therefore appeared to be the most promising candidate traits for further evaluation as early-selection indicators. The higher correlations between body weight and linear body measurements at 13–14 months of age suggest that the relationship among these traits became stronger with advancing age, compared with the associations observed at weaning. Although differences between Breeding-approved and Removed bull groups were modest, Breeding-approved bulls showed higher values for most traits, indicating that current selection practices are reflected in measurable phenotypic characteristics. These findings suggest the potential value of linear body measurements to support early evaluation in Limousin beef cattle, providing additional phenotypic information to complement selection based on observed development. Ethics declaration This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. This study was approved by the Albert Kázmér Mosonmagyaróvár Faculty - István Széchenyi University .(Approval No. SZE-AKMK/MAB/012/2024).) Funding sources The research was supported by the Ministry of Culture and Innovation and the National Research, Development and Innovation Fund (application ID: 2024-2.1.2-EKÖP-KDP-2024-00016). CRediT authorship contribution statement Márton János Demény: Writing – original draft, Visualization, Resources. Dóra Lili Brassó: Writing – review & editing, Methodology, Conceptualization. Csaba Rácz: Data curation. Márton Szűcs: Validation, Methodology, Conceptualization. János Tőzsér: Validation, Supervision, Software, Methodology, Formal analysis, Conceptualization. Declaration of competing interest The authors declare no competing interest.