R elationships between age at

Size: px
Start display at page:

Download "R elationships between age at"

Transcription

1 The Professional Animal Scientist 29 (2013): American Registry of Professional Animal Scientists R elationships between age at first calving; herd management criteria; and lifetime milk, fat, and protein production in Holstein cattle R. D. Curran,* K. A. Weigel, P. C. Hoffman, 1 J. A. Marshall, C. K. Kuzdas, and W. K. Coblentz * Curran LLC, Madison, WI 53593; Department of Dairy Science, University of Wisconsin, Madison 53706; AgSource, Verona, WI 53593; and US Dairy Forage Research Center, Madison, WI ABSTRACT Data from 69,145 Holstein cows that calved in 2005 were evaluated to determine the influence of age at first calving (AFC) on first lactation and lifetime production in commercial dairy herds. A DHI database was divided into 4 herd management criteria (HMC). The 4 HMC were 1) 3X milking and rolling herd average (RHA) = 12,750 kg (3X-H), 2) 3X milking and RHA = 11,250 kg (3X-M), 3) 2X milking and RHA = 11,250 kg (2X-M), and 4) 2X milking and RHA = 9,250 kg (2X-L). For all HMC, a universal loss in firstlactation milk, fat, and protein yield was observed when AFC was <23 mo. Compared with 24 mo, first-lactation milk yield was decreased, 166, 369, and 654 kg for heifers calving at 22, 21, and 20 mo, respectively. In contrast, AFC HMC interactions were observed for all herd-life and lifetime production criteria. 1 Corresponding author: pchoffma@wisc. edu In 3X-H and 3X-M herds, herd-life and lifetime DIM increased with decreasing AFC. In 2X-M and 2X-L herds, herdlife and lifetime DIM increased when AFC was reduced from 30 to 24 mo, but herd-life and lifetime DIM were not increased when heifers calved <24 mo. In 2X-M and 2X-L herds, the combined effect of reduced first-lactation milk yield and with no corresponding benefit in lifetime DIM resulted in reduced lifetime milk, fat, and protein yields when AFC was <23 mo. In contrast, increased lifetime milk, fat, and protein yields were observed in 3X-H and 3X-M herds when AFC was <23 mo. Data suggest HMC alters the relationship between AFC and lifetime milk, fat, and protein production in commercial dairy herds. Key words: calving age, heifer, lifetime milk INTRODUCTION The goal of dairy producers and custom heifer-raising operators is to rear dairy heifers as efficiently as possible, without compromising future health, fertility, lactation performance, or longevity. Rearing costs combined with the influence of calving age on future milk production necessitates defining optimal ages at first calving (AFC) for dairy replacement heifers (Pirlo et al., 2000; Hare et al., 2006). Controlled research studies evaluating early AFC (Hoffman et al., 1996; Van Amburgh, et al., 1998) have commonly observed decreased first-lactation milk yields when heifers calve at ages <22 mo. In commercial dairy herds, Ettema and Santos (2004) concluded the optimum firstlactation economic return occurred when heifers calved between 23.0 and 24.5 mo of age. However, conjecture is often applied to these findings, hypothesizing that lifetime milk yield is improved by calving heifers at earlier ages because lifetime DIM are increased (Hoffman and Funk, 1992). Lin et al. (1988) studied 2 groups of early- and late-calving heifers and reported that early-calving heifers (22.6 mo) produced more milk (10,693

2 2 Curran et al. vs. 9,218 kg) from calving until 61 mo of age as compared with late-calving heifers (25.8 mo). The work of Lin et al. (1988), however, does not necessarily define effects of early calving on productive life or lifetime milk yield because measurements were truncated at 61 mo of life. Subsequently, Nilforooshan and Edriss (2004) reported that AFC had no profound effect on productive life when heifers calved between 22 and 32 mo of age. Problematic in the aforementioned studies is lack of exploration of potential herd management AFC interactions on productive life and lifetime milk production. Modern dairy herds are housed in free or tie-stall barns, bedded with or without sand, milked 2X or 3X, potentially administered bovine somatotropin, and fed TMR or individual feed components. These management factors are known to alter milk yield (Bewley et al., 2001) or the productive life of dairy cows (Weigel et al., 2003). Responses to AFC on productive life or lifetime production within categorized dairy herd management criteria (HMC) have not been investigated. As a result, the objective of the present study was to evaluate relationships between AFC, productive life, and lifetime milk production in commercial dairy herds stratified into differing HMC. Herd management criteria used were basic and defined from information routinely available from dairy herds that participate in DHIA milk-recording programs through Ag- Source Cooperative Services (Verona, WI). MATERIALS AND METHODS Dependent variables evaluated included measures of longevity, firstlactation performance, and lifetime performance for Holstein cows that calved in 2005 that were enrolled in DHI programs with AgSource Cooperative Services (Verona, WI). Data from 2005 were chosen to reflect the most current industry trends and management practices, while minimizing bias due to censoring of data from cows that were still alive. Data from cows that calved for the first time in 2005 were chosen because 2005 was the most recent year for which >90% of cows had complete (uncensored) lifetime production and longevity data. By comparison, the proportions of cows remaining in the herd that calved for the first time in 2003, 2004, 2006, and 2007 were 1.3, 3.3, 13.0, and 23.6%, respectively. Lifetime production and longevity for 2005 data included completed records of cows that died or were culled from the herd (92.4%), as well as totals to date for cows that remained in the herd in April 2011 (7.6%). Specific measures of longevity included number of lactations, days from first calving until exit, and age at time of exit (d). Measures of first-lactation performance included lactation length (d), milk yield (kg), fat content (%), fat yield (kg), protein content (%), and protein yield (kg). Measures of lifetime production included DIM from first calving until exit, milk yield (kg), fat yield (kg), and protein yield (kg). Records from cows with AFC <20 mo of age or >30 mo of age were excluded because of low population (20 mo) or lack of inference to modern dairy management (>30 mo). In addition, records from cows with estimated birth dates were excluded, as were records from cows whose birth dates failed to match the calving date of the dam, records from herds in which >20% of cows calved on the first day of the month, and records from herds in which >20% of cows had AFC of exactly 24 mo (in the latter 2 cases, it was presumed that birth dates were missing and plausible values were filled at a later date). After editing, data from 69,145 Holstein cows were available for analysis. Culling practices, veterinary care, cow comfort, and other management factors that might influence cow longevity and lifetime production are known to differ between herds. In this study, milking frequency and rolling herd average (RHA) for milk yield were used as proxies for herd management. Cows were grouped into HMC based on differences in milking frequency (2X or 3X) and RHA for milk yield. Because 3X herds were larger and had greater average RHA (11,818 kg) that 2X herds (10,455 kg), a balanced factorial, with similar distributions of AFC within milking frequency and RHA, was not possible; therefore, cows were grouped and evaluated within independent HMC. The 4 HMC were as follows: 1) 3X milking and RHA = 12,750 kg (3X- H), 2) 3X milking and RHA = 11,250 kg (3X-M), 3) 2X milking and RHA = 11,250 kg (2X-M), and 4) 2X milking and RHA = 9,250 kg (2X- L) with RHA of 12,750, 11,250, and 9,250 kg representing mean RHA of a pool of dairy herds. First-lactation production, lifetime production, and longevity data from 14,692, 12,912, 20,037, and 21,503 cows in 158, 213, 1,049, and 1,491 3X-H, 3X-M, 2X- M, and 2X-L herds were evaluated, respectively. Subsequently, data were divided into subgroups based on AFC: 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, and 30 mo. Least squares means were calculated by AFC for each of the aforementioned HMC, for first lactation, and lifetime performance variables. Age at calving was nested within milking frequency by RHA subclass (2X-L, 2X-M, 3X-M, or 3X-H), and statistical analysis was conducted using PROC MIXED (SAS Institute Inc., Cary, NC) with HMC, AFC, and the interaction in the model. Contrasts between least squares means and corresponding tests of significance were carried out between HMC. Contrasts between least squares means and corresponding tests of significance for AFC or the interaction of HMC and AFC were conducted using AFC of 24 mo as the reference population. Primary statistical analysis revealed that AFC (n = 11) and HMC (n = 4) interactions were present for all lifetime production traits. Because the primary statistical model yielded >300 possible mean combination differences, response surfaces (linear vs. quadratic) of 24 mo residuals (actual 24-mo value) of AFC HMC lifetime means were conducted using RSREG procedures of SAS Institute (2001) to aid interpreta-

3 Age at first calving 3 Table 1. The effect of age at first calving (AFC) and herd management criteria (HMC) on first-lactation milk, fat, and protein production 1 First lactation Item n AFC, mo SE Milk, kg Fat, % Fat, kg Protein, % Protein, kg Mean SE Mean SE Mean SE Mean SE Mean SE HMC 2 3X-H 14, a ,952 a a a a a 3.0 3X-M 12, b ,715 b b b a b 3.1 2X-M 20, c ,546 b c b b b 2.7 2X-L 21, d ,946 c c c b c 2.3 AFC, 3 mo ,928 a b a a a ,751 9,213 a b a a ,194 9,416 a a a ,569 9, ,466 9, ,168 9, ,481 9, b b ,132 9, b ,387 9, b ,150 9, b ,434 9, b Statistics P < HMC <0.001 <0.001 < < AFC <0.001 <0.001 < < HMC AFC a d Column values with unlike superscripts differ (P < 0.001). 1 All data are least squares means. 2 3X-M = 3X milking and 12,750 rolling herd average (RHA), 3X-M = 3X milking and 11,250 RHA, 2X-M = 2X milking and 11,250 RHA, and 2X-L = 2X milking and 9,250 RHA. 3 Column values with superscripts differ (P < 0.01) from 24-mo values; a = value <24-mo value, b = value >24-mo value. tion and discussion of interactions. For linear responses, heterogeneity of slope test (Little et al., 1991) was also conducted, and for quadratic responses, quadratic response surface test (Palmquist, 1993) was performed. Heterogeneity and quadratic response surface test did not reveal any profound differences in linear or quadratic surface responses to AFC and lifetime production traits between 3X-H versus 3X-M or 2X-M versus 2X-L herds; therefore, these data are not presented nor will be discussed. RESULTS AND DISCUSSION The effects of AFC and HMC on first-lactation performance of Holstein cows are defined in Table 1. There were no interactions (P > 0.06) observed between AFC and HMC for any first-lactation production trait measured; therefore, only main effect means for HMC and AFC are presented. First-lactation milk yield was greatest (P < 0.001) for cows in 3X-H herds and least (P < 0.001) in 2X-L herds, with first-lactation milk yields intermediate and not different (P > 0.01) in 3X-M and 2X-M herds. Likewise, first-lactation yield of milk fat and protein was greatest (P < 0.001) for cows in 3X-H herds, least (P < 0.001) in 2X-L herds, and intermediate and not different (P > 0.01) in 3X-M and 2X-M herds. These data simply reflect the intended stratification of HMC in the study. A derivation of study design was observed for first-lactation milk fat and protein concentrations. Milk fat and protein concentrations were lower (P < 0.001) in cows from 3X-H and 3X-M herds as compared with 2X-M and 2X-L herds. The main effects of AFC on firstlactation performance of Holstein cows are also defined in Table 1, with differences as compared with 24-mo values presented to simplify interpretation. First-lactation milk yields of heifers calving between 23 and 25 to 30 mo of age were not different (P > 0.01) than milk yield of heifers calving at 24 mo. As compared with 24 mo, first-lactation milk yield was decreased, 166, 369, and 654 kg for heifers calving at 22, 21, and 20 mo, respectively. First-lactation milk fat

4 4 Curran et al. yield was also decreased (P < 0.001) 6, 10, and 19 kg for cows calving between 22, 21, and 20 mo. No differences in milk fat yield were observed when AFC of cows was 23 or 25 to 30 mo as compared with AFC at 24 mo. Milk fat percentage of cows was greater (P < 0.001) than was milk fat percentage of cows calving at 24 mo when AFC was <22 mo or >25 mo. There were no differences (P > 0.05) in milk fat percentage of cows when calving between 22 and 25 mo of age. Similarly, first-lactation milk protein yield was decreased (P < 0.001) 4, 10, and 16 kg for cows calving between 22, 21, and 20 mo, with no differences in milk protein yield observed when AFC was 23 or 25 to 30 mo as compared with milk protein yield when heifers calved at 24 mo. Milk protein percentage of cows calving at 20 mo was greater (P < 0.001) than was milk protein percentage of cows calving at 24 mo. There were no differences (P > 0.05) in milk protein percentage of cows when calving between 21 and 30 mo of age as compared with cows calving at 24 mo. The effects of AFC and HMC on herd-life, DIM, and lifetime milk, fat, and protein production are defined in Table 2. Both AFC and HMC influenced (P < 0.001) all lifetime traits, and interactions (P < 0.007) between AFC and HMC were observed for all lifetime production traits. To aid interpretation of AFC HMC interactions, response surfaces of 24 mo residuals (actual 24-mo value) of all lifetime traits means were graphed and are presented in Figures 1 to 6. For comparative purposes, numeric inferences of response surfaces on the effects of AFC on 24-mo residuals of lifetime production traits for each HMC are presented in Table 3. As previously mentioned, heterogeneity of slope test for linear lifetime trait surface responses and AFC between 3X-H and 3X-M herds did not reveal any major differences in response to AFC. Likewise, quadratic response surface test did not reveal any differences in response to AFC and lifetime production traits between 2X-M and 2X-L herds. All AFC HMC interactions observed were primarily associated with differences in lifetime trait responses to AFC between 3X and 2X HMC. Postcalving herd-life increased linearly (P < 0.001) with decreasing AFC in 3X-H and 3X-M herds, with the maximum response occurring when AFC = 20 mo, yielding 52.5 and 76.6 more days of herdlife, respectively, as compared with heifers calving at 24 mo (Figure 1). Conversely, a quadratic response (P < 0.01) to AFC was observed in postcalving herd-life in 2X-M herds, with the maximum response occurring when heifers calved at 23 mo of age. The quadratic response surface predicted AFC was decreased 10.5 d at 23 mo as compared with heifers calving at 24 mo of age. The decrease of 10.5 d of heifers calving at 23 mo as compared with 24 mo is a statistical nuance. Data simply suggest that in 2X-M herds, there was no notable improvement in postcalving herd-life when AFC was <25 mo. In 2X-L herds, AFC had no effect (P > 0.12) on postcalving herd-life. Lifetime DIM increased linearly (P < 0.001) with decreasing AFC in 3X-H and 3X-M herds, with the maximum response occurring when AFC = 20 mo, yielding 44.5 and 65.9 more DIM, respectively, as compared with heifers calving at 24 mo (Figure 2). A quadratic response surface (P < 0.03) was observed for AFC and DIM in 2X-M herds. The maximum response was predicted to occur when heifers calved at 22.5 mo of age, but similar to postcalving herd-life, the quadratic response surface model predicted no true improvement ( 8.0 d) when heifers calved at 22.5 mo as compared with heifers calving at 24 mo of age. In general, there was no improvement in DIM when AFC was <26 mo in 2X-M herds. In 2X-L herds, AFC had no effect (P > 0.12) on lifetime DIM. For all HMC there were (P < 0.001) linear response surfaces between herd exit age and AFC (Figure 3). Common within all HMC, heifers that calved at ages <30 mo correspondingly left the herd at an earlier age. In 2X-M and 2X-L herds, reducing AFC 30 d resulted in cows leaving the herd 27 and 28 d earlier, respectively. In contrast, reducing AFC 30 d in 3X-H and 3X-M herds resulted in cows leaving the herd 13 and 16 d earlier, respectively. Lifetime milk yield increased linearly (P < 0.001) with decreasing AFC in 3X-H and 3X-M herds, with the maximum response occurring when AFC = 20 mo, yielding 1,888 and 1,811 kg more milk, respectively, as compared with heifers calving at 24 mo (Figure 4). A quadratic response (P < 0.001) surface was observed for AFC and lifetime milk yield in 2X-M and 2X-L herds. The maximum lifetime milk yield response was predicted to occur in 2X-M and 2X-L herds when heifers calved at 24.0 and 24.7 mo of age, respectively. Lifetime milk yield in 2X-M and 2X-L herds declined precipitously when AFC was <22 mo or >26 mo. The interaction of HMC and AFC on lifetime milk fat and protein yield was almost identical to milk yield (Figures 5 and 6). Lifetime milk fat and protein yield increased linearly (P < 0.001) with decreasing AFC in 3X-H and 3X-M herds, with the maximum response occurring when AFC = 20 mo. Cows produced 71.3 and 61.8 kg more fat and 57.6 and 60.9 kg more protein in 3X-H and 3X-M herds, respectively, when heifers calved at 20 mo as compared with 24 mo. A quadratic response (P < 0.004) surface was observed for AFC and lifetime milk fat and protein yield in 2X-M and 2X-L herds. The maximum lifetime milk fat and protein yield response was predicted to occur in 2X-M and 2X-L herds when heifers calved at approximately 24.3 ( ) mo of age. Similar to lifetime milk yield, lifetime milk fat and protein yield in 2X-M and 2X-L herds declined precipitously when AFC was <22 mo or >26 mo. Controlled research studies evaluating early AFC (Hoffman et al., 1996; Van Amburgh et al., 1998) have commonly observed decreased first-

5 Age at first calving 5 Table 2. The effect of age at first calving (AFC) and herd management criteria (HMC) on herd-life; DIM; and lifetime milk, fat, and protein production 1 Herd-life, d Lifetime, kg HMC 2 AFC, mo n Postcalving DIM Exit age Milk Fat Protein Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE 3X-H , , ,545 1, , , , , , , , ,355 1, , , , , ,385 1, , , , , ,013 1, , , , , ,072 1, , , , , ,285 1, , , , , , , , , , , , , , , ,889 1, , , ,110 1, , , X-M , , ,142 2, , , , , , , ,517 1, , , , , ,614 1, , , , ,644 1, , , , ,044 1, , , , ,460 1, , , , , , , , , , , , , , ,835 1, , X-M , , ,211 2, , , , ,190 1, , ,288 1, , , , ,144 1, , , , ,174 1, , , , ,553 1, , , , ,610 1, , , , ,997 1, , , , ,331 1, , , , , , , , , , X-L , , ,156 1, , , , ,034 1, , , ,426 1, , , ,635 1, , , , ,499 1, , , , ,126 1, , , , ,446 1, , , ,988 1, , , ,594 1, , , ,267 1, , , Statistics P < HMC <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 AFC <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 HMC AFC All data are least squares means. 2 3X-M = 3X milking and 12,750 rolling herd average (RHA), 3X-M = 3X milking and 11,250 RHA, 2X-M = 2X milking and 11,250 RHA, and 2X-L = 2X milking and 9,250 RHA.

6 6 Curran et al. Figure 1. Response surfaces of age at first calving (AFC) and residual 24 herdlife. Residual 24 herd-life represents the difference [herd-life, d (month x) herdlife at 24-mo AFC, d] between herd-life at month x versus herd-life at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ). Figure 2. Response surfaces of age at first calving (AFC) and residual 24 lifetime DIM. Residual 24 DIM represents the difference [lifetime DIM, d (month x) lifetime DIM at 24-mo AFC, d] between DIM at month x versus DIM at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ). Figure 3. Response surfaces of age at first calving (AFC) and residual 24 herd exit age. Residual 24 herd exit age represents the difference [herd exit age, d (month x) herd exit age at 24-mo AFC, d] between herd exit age at month x versus herd exit age at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ). Table 3. Response surface analysis of the effect of age of first calving (AFC) and herd management criteria (HMC) on herd-life; DIM; and lifetime milk, fat, and protein production 1 Herd-life Lifetime HMC 2 3X-H 3X-M 2X-M 2X-L Response Postcalving DIM Exit age Milk Fat Protein Linear, P < <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Quadratic, P < AFC at maximum response, mo Maximum response, value , Linear, P < <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Quadratic, P < AFC at maximum response, mo Maximum response, value , Linear, P < < Quadratic, P < AFC at maximum response, mo Maximum response, value Linear, P < < Quadratic, P < < AFC at maximum response, mo Maximum response, value Maximim response values for herd-life and lifetime are expressed in days and kilograms, respectively. 2 3X-M = 3X milking and 12,750 rolling herd average (RHA), 3X-M = 3X milking and 11,250 RHA, 2X-M = 2X milking and 11,250 RHA, and 2X-L = 2X milking and 9,250 RHA.

7 Age at first calving 7 Figure 4. Response surfaces of age at first calving (AFC) and residual 24 lifetime milk yield. Residual 24 lifetime milk yield represents the difference [lifetime milk yield, kg (month x) lifetime milk yield at 24-mo AFC, kg] between lifetime milk yield at month x versus lifetime milk yield at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ). lactation milk yields when heifers calve at ages <22 mo. Reductions in first-lactation milk yield when heifer calve at ages <22 mo have also been reported in field studies involving Figure 5. Response surfaces of age at first calving (AFC) and residual 24 lifetime fat yield. Residual 24 lifetime fat yield represents the difference [lifetime fat yield, kg (month x) lifetime fat yield at 24-mo AFC, kg] between lifetime fat yield at month x versus lifetime fat yield at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ). commercial dairy herds (Ettema and Santos, 2004). Data from this study support previous observations that first-lactation milk yield is decreased when heifers calve at ages <22 mo. As compared with 24 mo, first-lactation milk yield was decreased 166, 369, and 654 kg for heifers calving between 22, 21, and 20 mo, respectively, and the reduction of first-lactation milk yield was independent of HMC. However, data from this study do not universally support the hypothesis that lifetime DIM and milk yield is improved by calving heifers at earlier ages (Lin et al., 1988; Hoffman and Funk, 1992). In the present study, when commercial dairy herds where stratified into differing HMC, relationships between AFC and lifetime DIM were greatly divergent. In 3X-H and 3X-M herds, DIM increased with decreasing AFC, but this effect was not observed in 2X-M and 2X-L herds. In 2X-M herds, no appreciable change in lifetime DIM was observed with heifers calved at ages <26 mo, and in 2X-L herds, lifetime DIM was not related to AFC. Observations in 2X-M and 2X-L herds are similar to those of Vukasinovic et al. (2001), who reported that AFC did not have a large influence on length of productive life. Observations from 3X-H and 3X-M herds are similar to those of Nilforooshan and Edriss (2004), who reported a positive effect of reducing AFC on productive life. Data from this study suggest discrepancies in AFC and lifetime trait relationships may be due to nuances in the production-records database used to conduct the evaluation. Speculatively, if production-records databases are dominated by higher-producing herds milking 3X, positive effects of reducing AFC on productive life or DIM may be observed. In contrast, if the production-records database is dominated by lower-producing herds milking 2X, minimal or no effects of reducing AFC on productive life or DIM may be observed. A paradox was observed in the present study in regard to the effects of AFC on lifetime milk, fat, or protein yield. In 3X-H and 3X-M herds reducing AFC to <24 mo improved lifetime milk, fat, and protein production, which is similar to the findings of Nilforooshan and Edriss (2004). In 2X-M and 2X-L herds reducing AFC to <24 mo decreased lifetime milk, fat, and protein production, which is similar to the findings of Vukasinovic et al. (2001). Upon examination of the data in the present study, the paradox appears to be primarily rooted in differences in herd exit age between 3X and 2X HMC categories. In 2X-M and 2X-L herds, reducing AFC 30 d resulted in cows leaving the herd 27 and 28 d earlier, respectively. Although not statistically tested, this relationship is near unity (decreasing AFC 30 d = cows leaving the herd 30 d earlier), which results in little change in lifetime DIM. In contrast, reducing AFC 30 d in 3X-H and 3X-M herds resulted in cows leaving the herd 13 and 16 d earlier, yielding 17 and 14 additional DIM, respectively. Disparity in lifetime production trait responses as associated with AFC between herds milking 3X and 2X has not been previously published; Figure 6. Response surfaces of age at first calving (AFC) and residual 24 lifetime protein yield. Residual 24 lifetime protein yield represents the difference [lifetime protein yield, kg (month x) lifetime protein yield at 24-mo AFC, kg] between lifetime protein yield at month x versus lifetime protein yield at 24 mo in dairy herds with differing herd management criteria; 3X-H = 3X milking, 12,750 kg ( ); 3X-M = 3X milking, 11,250 kg (......); 2X-M = 2X milking, 11,250 kg ( ); and 2X-L = 2X milking, 9,250 kg ( ).

8 8 Curran et al. therefore, comparisons to previously published work cannot be made. On the surface, differing responses in the present study could be related to milking frequency, but within the study design, milking frequency is likely confounded with numerous other differences in dairy herd management. Therefore, this conclusion is not valid. In the present study, herds were simply classified as 3X or 2X from a DHI database, and 3X or 2X classifications may have represented numerous and divergent differences in dairy herd management beyond milking frequency. For example, 3X herds may have a greater propensity to use bovine somatotropin, TMR, freestall housing, or sand bedding as compared with 2X herds (Bewley et al., 2001). Detailing specific differences in dairy herd management practices between 3X and 2X herds was not possible in this study; therefore, the terms 3X and 2X are mere management classifications and not a refined evaluation of milking frequency. Differences observed in relationships between lifetime DIM and AFC in 3X and 2X herds may be explained by differences in housing systems represented within the 3X and 2X HMC classifications. In the present study, data were evaluated from 14,692, 12,912, 20,037, and 21,503 cows in 158, 213, 1,049, and 1,491 3X-H, 3X- M, 2X-M, and 2X-L herds, respectively. Cross-product cow and herd values suggest herd sizes in 3X herds were notably larger and more likely to be housed in freestall barns (Bewley et al., 2001). In freestall housing systems cow populations can be more dynamic because freestall housing systems can be overcrowed (Bewley et al., 2001). In contrast, in station or tie-stall barns the number of stalls is generally fixed (unless expanding), and an individual stall cannot house more than one animal. As a result, when a heifer has her first calf in stanchion or tie-stall housing system, a stall must be available and if occupied, vacated for the heifer to enter the milking herd. This simple static relationship in stanchion or tie-stall barns may explain why decreasing AFC in 2X herds resulted in strong linear relationship with herd exit age. The 2X HMC classifications may have represented more stanchion or tie-stall herds in which earlier AFC resulted in culling older cows to accommodate stall availability. In freestall housing, when a heifer has her first calf, a freestall containing another cow can, but does not have to, be vacated for her to enter the milking herd. Because freestall barns can be overcrowed (Bewley et al., 2001), the 3X HMC classifications in this study may have represented more freestall herds in which earlier AFC results in dynamic stall occupation or overcrowding. Although speculative, potential dynamics with housing and milking systems may better explain the results observed in this study and warrant more controlled investigations. IMPLICATIONS This study indicates that some dairy producers achieve lifetime production benefits by reducing AFC to approximately 21 to 22 mo but other dairy producers do not. Conversely, extending AFC beyond 24 mo offers no apparent gains in first-lactation milk production, or longevity and lifetime production tend to be compromised. Study results suggest that a 21- to 22-mo target for AFC may be appropriate for some dairy herds, but results also suggest that for many dairy herds, unknown management impediments exist that do not yield lifetime productive benefits to reducing calving age <23 mo. The present study was observational, and despite the large number of cows and herds that were used, the multitude of management practices used by dairy producers likely affected the results. Results from this study, however, suggest that universal AFC recommendations may not apply across all dairy herds. Data also suggest that controlled studies that divide heifers into AFC, milking frequency, and stalloccupancy scenarios may be beneficial in understanding the dynamic herd effects of AFC on lifetime productivity. ACKNOWLEDGMENTS The authors gratefully acknowledge AgSource Cooperative Services for allowing use of their database for this research. Without that cooperation, this study would not have been possible. Financial support provided to K. Weigel by the National Association of Animal Breeders (Columbia, MO) is gratefully acknowledged. LITERATURE CITED Bewley, J., R. W. Palmer, and D. B. Jackson- Smith An overview of experiences of Wisconsin dairy farmers who modernized their operations. J. Dairy Sci. 84: Ettema, J. F., and J. E. P. Santos Impact of age at calving on lactation, reproduction, health, and income in first-parity Holsteins on commercial farms. J. Dairy Sci. 87: Hare, E., H. D. Norman, and J. R. Wright Trends in calving ages and calving intervals for dairy cattle. J. Dairy Sci. 89: Hoffman, P. C., N. M. Brehm, S. G. Price, and A. Prill-Adams Effect of accelerated postpubertal growth and early calving on lactation performance of primiparous Holstein heifers. J. Dairy Sci. 79: Hoffman, P. C., and D. A. Funk Applied dynamics of dairy replacement growth and management. J. Dairy Sci. 75: Lin, C. Y., A. J. McAllister, T. R. Batra, A. J. Lee, G. L. Roy, J. A. Vesely, J. M. Wauthy, and K. A. Winter Effects of early and late breeding of heifers on multiple lactation performance of dairy cows. J. Dairy Sci. 71: Little, R. C., R. J. Freund, and P. C. Spector SAS System for Linear Models. 3rd ed. SAS Inst. Inc., Cary, NC. Nilforooshan, M. A., and M. A. Edriss Effect of age at first calving on some productive and longevity traits in Iranian Holsteins of the Isfahan Province. J. Dairy Sci. 78: Palmquist, D. E The use of two distance measures as a modified multiple comparison technique for obtaining response surface differences. MS Thesis. Univ. Nevada, Reno. Pirlo, G., F. Miglior, and M. Speroni Effect of age at first calving on production

9 Age at first calving 9 traits and difference between milk yield returns and rearing cost in Italian Holsteins. J. Dairy Sci. 83: SAS Institute SAS User s Guide: Statistics. SAS Inst. Inc., Cary, NC. Van Amburgh, M. E., D. M. Galton, D. E. Bauman, R. W. Everett, D. G. Fox, L. E. Chase, and H. N. Erb Effects of three prepuberty body growth rates on performance of Holstein heifers during first lactation. J. Dairy Sci. 81: Vukasinovic, N., J. Moll, and L. Casanova Implementation of a routine genetic evaluation for longevity based on survival analysis techniques in dairy cattle populations in Switzerland. J. Dairy Sci. 84: Weigel, K. A., R. W. Palmer, and D. Z. Caraviello Investigation of factors affecting voluntary and involuntary culling in expanding dairy herds in Wisconsin using survival analysis. J. Dairy Sci. 86: