Forage Availability x Heifer Phenotype Interactions for Brahman-Hereford F, Yearling Heifers Grazing Humid Pasture and Semiarid Rangeland' J. W. Holloway*, B. G. Warrington*, F. M. Rouquette, Jr.t, C. R. Long+, M. K. Owens*, and J. F. Bakertp2
ABSTRACT2 Yearling heifer growth data were obtained during 4 yr for 524 heifers allotted to either humid bermudagrass pastures (Overton) or semiarid rangeland (Uvalde). Each year, heifers were allotted on April 15 to four forage availability levels (400 to 2,800 kg of DM per 100 kg of BW at Uvalde and 80 to 260 kg of DM per 100 kg of BW a t Overton) maintained by varying the stocking rate monthly until mid-October of each year. Forage availability and yearling heifer characteristics (weight, condition score, and height at hooks, taken on April 15) were treated as continuous variables in regression analyses. Final heifer weight, height, and condition responses to increased forage allowance were related to yearling phenotypes differently for the two locations. Generally, a t Overton, forage availability influenced
final characteristics to a greater extent than did yearling variables, whereas the trend was the opposite a t Uvalde. At Uvalde, the yearling characteristic that had the largest effect on performance was height at hooks. Yearlings with large frames benefited from increased forage allowance by accumulating body fat a t a faster rate than those with small frames. In contrast, at Overton, the yearling characteristic that had the largest effect on performance was condition. Fat heifers responded to increased forage availability to gain even greater advantages in fatness a t the expense of potential growth in height and, thus, achieved early maturation. Yearling phenotypes were more broadly adapted to arrays of forage availability for humid, improved pastures than for semiarid rangeland.
Key Words: Phenotype Environment Interaction, Adaptation, Heifers, Growth, Forage Availability
J. Anim. Sci. 1992. 70:2658-2667
Introduction Many cow-calf enterprises are managed in potentially stressful conditions that animals must a t least partially ameliorate through physiological or behavioral adjustments to grow and reproduce. Fortunately, the cattle phenotypic array is almost as broad as the array of potentially stressful environments. Ecological niches can be broadened by strategic application of inputs to reduce stress, but often these inputs are either too costly or too
'Approved for publication as Journal Article no. 30169 of the Texas Agric. Exp. Sta. 2Present address: Dept. of Anim. Sci., Univ. of Georgia, P. 0. Box 748, Tifton 31793-0748. Received January 27, 1992. Accepted May 11, 1992.
risky to gain industry acceptance. Thus, stressful conditions are prevalent in production environments and form the basis for genotype and phenotype x environment interactions (Butts et al., 1971; Kress et al., 1971a,b; Holloway and Butts, 1983; Holloway et al., 1985a,b; Bolton et al., 1987; Holloway et al., 1988). Early identification of adapted cow phenotypes should minimize biological mismatches, thereby decreasing the proportion of lowly productive or nonproductive cows in the herd and increasing the efficiency of the system. Because forage availability and cattle type are variables that managers can manipulate most easily and frequently, our purpose was to determine yearling heifer phenotype x forage availability interactions for two environments. Inferences tested are within environments (locations), although we also make some general observations concerning the contrasting nature of these interac-
2658
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
*Texas A&M University Research and Extension Center, Uvalde 78801 and +Texas A&M University Research and Extension Center, Overton 75684
HEIFER PHENOTYPE x ENVIRONMENT
tions between environments. Because these environments involve a myriad of confounded, potentially stressful conditions, the purpose of this experiment was to describe the interactions and not to delineate their causes. We formulate some hypotheses that, though untested, are consistent with empirical observations.
Animal Management. Five hundred twentyfour Brahman-Hereford F1 weanling heifers (approximate age of 8 mol were purchased during October and November of each year during a 4-yr period (1984 to 1987). Heifers were randomly allotted within source (one to four sources per year) to either semiarid rangeland in Uvalde (345 heifers, 52 cm of annual precipitation) or humid, improved pasture in Overton (179 heifers, 113 cm of annual precipitation). Heifers were wintered in systems typical to those practiced by the local beef industry for the two locations. At Uvalde, heifers were allowed native rangeland and supplemented for approximately 50 d (January to March) each year with .4 kg.anima1-l .d-’ of cottonseed (1985 and 1986) or .45 kg . animal-’. d-l of cottonseed meal (41% CP 1987 and 1988). At Overton, heifers were allowed ad libitum access to Cynodon dactylon hay supplemented with 2.3 kg.anima1-l.d-l of ground corn and .68 kg.anima1-lad-’ of 41% CP commercial supplement from January 8 to March 4. From March 4 until the beginning of the experiment, heifers grazed Secale cereale var. ‘Elbon’ and Lolium multiflorum var. ‘Marshall’ mixture. Animals at both locations were allowed free access to trace mineralized salt3. Heifers were exposed to Braford bulls (single-sire groups a t Overton, multisire groups at Uvalde, depending on pasture size to provide 20 heifershull) from April 15 to July 1 of each year. Bulls were purchased from the same source and allotted at random to location and to forage availability. AZZotment to Forage Availability. During April of each year, heifers at each location were stratified by weight and randomly allotted to four levels of forage availability designed to include the range in forage allowances in normal beef industry practice for the respective areas (80to 260 and 400 to 2,800kg of DM per 100 kg of BW, respectively,
3Trace mineralized salt composition: 33% NaC1, 11% Ca, 8 % P, 1.9% Mg, 1.3% K, . 4 % Mn, . 5 % Zn, .8% Fe, .1% Cu, 154,000 IU/kg, vitamin A, and 1,100 IU/kg vitamin D3.
at Overton and Uvalde). Heifers at Uvalde were allotted to four extensively managed native rangeland pastures each year (1,950to 6,100ha) dominated by a n overstory of Acacia species and Prosopis glandulosa with a n understory of native grasses including Bouteloua curtipendula (sideoats grama) and B. trifida (red grama), Buchloe dactyloides cbuffalograss), Helaria belangeri (curley mesquite), Pappophorum bicolor (pink pappas), Aristida wrightii (Wright’s threeawn), and Setaria leucopila bristlegrass) and native forbs including Ambrosia psilostachya and Croton texensis. Heifers a t Overton were allotted to intensively managed Cynondon dactylon bermudagrass) pastures each year (7.5to 13.5ha) fertilized with 330:lOO:lOO;kg/ h a of N:P205:K20 in split applications. Under general industry practice, more forage is provided per animal at Uvalde than a t Overton because the forage is more sparsely distributed and, thus, although it is provided and theoretically is available, it may not be accessible (e.g., shrub pattern may restrict animal movement). Estimation of Forage AvailabiZity. Targeted forage availability was attained gradually from initiation of the trial (April 15) to approximately June 1 of each year as a precaution against the possibility of drouth and resultant forage availability being less than that targeted for the respective groups and, thus, forced removal of experimental animals from treatment. The process for estimating and attaining targeted forage availability was as follows. Forage availability was calculated monthly by using clipped quadrants. At Uvalde, grass and forb mass was defined as forage, even though cattle may a t times consume shrubs (Launchbaugh et al., 1990). Pastures were stratified by range site, and 20 to 90 .5-m2 quadrants were randomly located within each site, depending on pasture size and within-site variation in forage availability. Grass and forb mass was visually estimated for each quadrant and then, for a random 20% of quadrants, was hand-clipped to ground level, dried, and weighed. Weighed mass was regressed on visual estimates a t each sampling time (R2 from .85 to .981. The resulting equations were used to calculate DM/hectare for each pasture each month. At Overton, four randomly located .25-m2quadrants were hand-clipped to ground level for each pasture each month. Random samples from each of these quadrants were dried and weighed, and the DM/hectare was calculated. Stocking rate ranged from 17 to 146 kg of BW/ ha at Uvalde and 1,249 to 4,250 kg of BW/ha at Overton and was adjusted monthly to maintain targeted forage availability. Animals were weighed monthly and the number was adjusted as
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Materials and Methods
2659
HOLLOWAY ET AL.
2660
FIGURE lA. UVALDE
FORAGE
80
Figure 1. Relationship of yearling weight and forage availability to final weight (n = 345, R2 = .86, residual SD = 18.38 for Uvalde [Figure la]; n 179, Rz .76, residual SD = 23.21 for Overton [Figure lb]).
-
-
scores were taken according to the method of Long et al. (1979; emaciated = 1, extremely fat = 9) by the same scorer each month within location. All heifers were scored by both scorers each October to assure similarity in scoring method between locations. Yearling and final measurements were those taken in April and October, respectively, for each grazing season (1985 to 19881, and the change was calculated by subtracting the yearling value from the final value and dividing by elapsed days. Statistical Analysis. Experimental design was dictated by the following conditions: 1) variables of interest (forage availability and yearling heifer phenotype1 were those most often manipulated by resource managers; 2) these variables are continuous in nature; and 31 large numbers of animals and pastures (as well as large-sized pastures in semiarid conditions) were required to attain the necessary arrays in the variables of interest. Therefore, we executed a design similar in principle to those described by Riewe (lQSll, Bransby et al., (19881, and Drane (19891. This design employed years (four pastures per location per year) to accumulate the necessary arrays and, in the terminology of Drane (19891, “forage availability was taken as a surrogate for paddock effect. . , mapping the complex interaction of paddock x year x forage available.” Regression procedures (SAS, 19851 were used with the response variables final weight, height, and condition and changes in these variables during the grazing period. Linear models in the analyses of these response variables included fixed effects of year and location, linear, quadratic, and cubic effects of forage availability and initial yearling heifer characteristics (weight, height, or condition), and two- and three-factor interactions among the continuous effects and location. Initial analyses indicated that the location x forage availability x yearling heifer characteristics were generally important (P < . l o ) for all dependent variables. Subsequent analyses were performed within location using linear models including classification effects of year, polynomial effects of initial yearling characteristic (up to cubic), and all possible interactions among classification and covariates. Models were reduced as was appropriate, to include only those effects contributing (P e .lo1 to the regression sums of squares. Equations developed for final characteristics were similar to those for changes in these characteristics. Therefore, only equations predicting final characteristics are presented. Also, separate equations were developed for independent variables describing yearling phenotype (weight, height, or condition) because of autocorrelation among these variables.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
needed to maintain targeted kilograms of DM/ kilogram of BW. Calculated monthly forage availabilities were averaged to compute forage availability for the season. This procedure was employed to attain a n array of forage availability during the 4 yr so that forage availability could be considered a continuous variable. Therefore, deviations from average seasonal targeted forage availability were not considered to be serious errors. Grazing season ended on approximately October 15 each year. Animal Measurements. Heifer 12-h fasted weight, height a t hooks (tuber coxae), and condition score were recorded monthly during the grazing season (April 15 to October 15). Weights were used to compute monthly forage availability. Condition
HEIFER PHENOTYPE x ENVIRONMENT
Results and Discussion
noted for final weight and, thus, only the figure concerning final weight is presented (Figure I). In each case, a synergistic relationship was detected between yearling weight and forage availability in that an increase in each resulted in increased final weight keight and condition), and as both increased, final weight (height and condition) increased at an accelerating rate, resulting in significant interactions (Table 2, Figure 1). At Overton, forage availability affected (P e .061 final height but to a lesser extent than a t Uvalde, and heavier yearling heifers were taller the following fall than lighter heifers a s yearlings (Table 21. Apparently, under extensive conditions a t Uvalde, heavier yearling heifers responded to a n increase in forage availability in terms of final weight, height, and condition to a greater extent than did lighter yearling heifers. Thus, to ensure maximum growth response to an increase in forage availability under semiarid conditions, heavier yearling heifers would be required. Because the forage a t Uvalde was more sparse than that at Overton, heavier yearling heifers possibly had a n advantage in traversing the area and procuring forage, giving them nutritional advantages that resulted in increased growth responses. A contributing factor to the differences between locations in the shape of response surfaces could be differences in forage quality. Forage quality was not measured because we did not have adequate means to measure selectivity. Because forage on offer often deviates from that consumed, and because individual animals vary in selectivity, available analytical methods were either inappropriate or unfeasible. The contrast between Overton and Uvalde is largely a result of the magnitude of responses of the two dimensions a t the two locations. For
Table 1. Characteristics of experimental sample of heifers on humid bermudagrass pastures (Overton) and semiarid rangeland (UvaldeIa
Item
1985-1988
1985
1986
1987
1988
Overton Uvalde
Overton Uvalde
Overton Uvalde
Overton Uvalde
Overton Uvalde
Precipitation (April-October), cm 63.55 Yearling Condition score 6.1' Weight, kg 304.4' Height, cm 123.8' Final Condition score 6.7' Weight, kg 399.8' Height, cm 130.3' No. 179
47.38 5.1'
289.7d 120.6d 4.9d 358.2' 126.5d 345
68.73
46.25
68.01
52.83
54.97
70.99
42.44
19.46
RSDb
-
5.3' 5.2' 259.2Cd 304.2e 121.3Cde 118.1
6.0d 313.3e 123.Qfg
4.6e 270.4' 120.0'
7.1 341.8 123.Qf
5.9' 333.2 122.Bdg
6.1d 303.4e 127.1
4.5e 251.5' 121.3e
.86 29.16 4.00
4.7 377.2d 132.9' 88
7.0' 405.4Ce 128.4' 45
6.0 375.7' 126.5 72
8.6' 394.1ef 128.4e 47
6.4d 382.8df 124.5f 92
6.5' 387.3df 131.0df 58
2.4 297.3 123.2 93
.9 1 36.36 4.00
6.7"
411.7' 132.9'' 31
-
*Least squares means from the model: Y = year, location, year x location. bResidual standard deviation. cid,e,f-gMeanson same line across years or between locations that do not have a common superscript differ (P < .05, t-test).
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Characteristics of Experimental Sample. Although the heifers were randomly allotted within source to location at weaning time, differences between locations in winter management resulted in heifers a t Overton that were 14.7 kg heavier (P e .01),3.2 cm taller (P < .01) and had 1.0 unit more (P < . O l l condition than those at Uvalde (Table I). During the trial, they also gained more (P e .011 weight, condition, and height. Also, variation among years for all responses was much less a t Overton than a t Uvalde. These data, when considered with the large areas required to attain targeted forage levels and the complications provided by the presence of shrubs in inhibiting animal access to forage at Uvalde (Owens et al., 19911, indicate that nutritional and foraging stresses were much greater a t Uvalde than at Overton. Yearling Weight x Forage Availability Interactions. Yearling weight interaction with forage availability was important (P e ,101 for all dependent variables (final weight, height, and condition) a t Uvalde but only important (P e ,101 for final weight a t Overton (Table 2 and Figure 11. Interaction of yearling weight and forage availability on final weight was more dramatic a t Uvalde than at Overton (Figure 1). At Uvalde, heifers that were heavier a s yearlings responded in terms of final weight, height, and condition to increased forage availability more dramatically than did lighter heifers. At Overton, although the trend was the same for final weight as that at Uvalde, heavier yearlings responded to increased forage availability only to a slightly greater extent that did lighter yearlings (Figure 1). At Uvalde, the interactions of yearling weight and forage availability on final height and condition were similar to interactions
2661
HOLLOWAY ET AL. 2662
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
-
r
c
e
E
$ 8
a, h
9 cj
4
,-I
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
9 9 9 2 9 1 1 1 1 1 1 1
88gm8
2663 HEIFER PHENOTYPE x ENVIRONMENT a o0 c u m0 O O l - c u ~ ~ 999991999919
HOLLOWAY ET AL.
2664
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
HEIFER PHENOTYPE x ENVIRONMENT
Yearling Condition x Forage Availability Interactions. Interactions of forage availability and yearling heifer condition were generally more important at Uvalde than at Overton (Table 3 and Figure 2). Interactions were detected (P < .02)at Uvalde for both final weight and condition, but, because the relationships were similar, we believe the effect on weight to be a result of the effect on condition and, thus, only the latter is shown (Figure 21. At Uvalde, although yearling condition had a greater relative effect than forage availability on final condition, the two dimensions acted synergistically in influencing final condition. That is, increases in forage availability resulted in more dramatic increases in final condition for heifers of greater yearling condition than for those of lesser yearling condition. Also, cattle of greater yearling condition always had greater final condition, but the relationship was more pronounced for heifers
flGURE 2A WALDE
Figure 2. Relationship of yearling condition and forage availability to final condition (n = 345, R2 = .81, residual SD = .81 for Uvalde [Figure Za]; n = 179, R2 = .54, RSD = .58 for Overton [Figure Zb]).
Figure 3. Relationship of yearling height and forage availability to final condition (n = 345, R2 = .77, residual SD = .90 for Uvalde [Figure 3a]; n = 179, R2 = .48, RSD = .62 for Overton [Figure 3b]).
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
example, the rate of response for final weight was much greater in the forage availability dimension at Overton than in that a t Uvalde. That is, at Overton, a n increase of approximately 60 to 90 kg in final weight per 380-kg increase in forage availability was noted compared with a n increase of approximately 30 to 80 kg in final weight per 2,400-kg increase in forage availability a t Uvalde (Figure 1). For the animal phenotype dimension, the response was much greater for yearling heifer weight a t Uvalde than for that at Overton (Figure 1). The forage available a t Overton was apparently easily procured and utilized by all grazers, whereas at Uvalde the sparseness of the forage and the physical impediments provided by the interdispersed shrubs (Owens et al., 1991) possibly resulted in problems of forage search and procurement, especially for the lighter (smaller) yearling heifers.
2665
2666
HOLLOWAY ET AL.
later maturing heifers that had been taller as yearlings.
Summary For semiarid rangeland, yearling height, weight, and condition affected final characteristics more than did forage availability, whereas the opposite was the case for humid, improved pastures. A greater number of interactions were detected for semiarid rangeland than for humid improved pastures, and interactions were more dramatic for semiarid rangeland. In general, the nature of the interactions was that yearling characteristics and forage availability acted synergistically to each other in affecting response variables. The exception was yearling height for humid improved pastures (taller heifers as yearlings responded to a lesser degree to increases in forage availability than those that were shorter). Because of the generally synergistic and curvilinear nature of the relationships and because these relationships were more dramatic under semiarid rangeland, we conclude that yearling F1 phenotypes were broadly adapted to humid, improved pastures varying in forage availability; for semi-arid rangeland, however, phenotypes were more narrowly adapted in that the taller, heavier, fatter yearlings exhibited greater levels of adaptation than their counterparts especially at higher levels of forage availability.
Implications Yearling heifers that vary in frame size and degree of fatness do not respond similarly to increases in forage availability for humid, improved pastures and semiarid rangeland. Animals on semiarid rangeland require access to much more forage to perform similarly to those grazed on humid, improved pastures. Yearling weight and yearling height are important components of interactions with forage availability for yearling heifer growth on rangeland. Yearling phenotype was a less critical determinant of response for humid, intensively managed pastures, although shorter yearling females tend to take advantage of increased forage availability by increasing levels of final condition.
Literature Cited Allden, W. G., and 1.A.McD. Whittaker. 1970. The determinants of herbage intake by grazing sheep: The interrelationship of factors influencing herbage intake and availability. Aust. J. Agric. Res. 21:755.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
allowed greater amounts of forage than for those allowed lesser amounts. At Overton, forage availability did not interact with yearling condition in influencing final condition (or final weight or height, Table 31, and the effect of forage availability was much greater than the effect of yearling condition (Figure 2). Heifers at Overton had consistently greater P c .051final condition scores than those a t Uvalde (P c .05, Table 11, and their final condition was affected to a greater extent by forage availability than was the condition of those a t Uvalde (Table 3, Figure 2). Yearling Height x Forage Availability Interactions. Yearling height x forage availability interactions were important (P c .lo) for final height and condition a t Uvalde and for final weight and condition a t Overton (Table 4). At Overton, the primary variable influencing final weight and condition was forage availability, although the effect was greater for heifers that were shorter as yearlings than for those that were taller [Table 4 and Figure 3). At Uvalde, the primary variable that influenced final condition was yearling height (Table 4 and Figure 3). For final condition a t 400 kg of DM/ha, yearling height was negatively related to final condition, whereas at 2,800 kg of DM/ha, yearling height was positively related to final condition. Heifers that had the greatest final condition were the ones that had been tallest as yearlings and allowed the most forage (Figure 3). Taller heifers at Uvalde possibly had a greater ability to move in search for forage. When forage was available in large enough quantities to meet their nutrient demands, the search was successful and resulted in increased body condition. When allowed low levels of forage, however, these heifers continued to have greater capacity for moving to the forage than did shorter heifers, but the futile search possibly resulted in declines in condition. According to this hypothesis, height is indicative of frame size and ability to range. An alternative hypothesis is that taller heifers also had wider incisor breadth and, thus, were less able to subsist on shorter swards but were more able to subsist when forage was abundant. Allden and Whittaker (1970)reported that young sheep were better able to maintain intake on shorter swards than were mature sheep with larger mouths. Also, Illius and Gordon (1987) concluded that “due to allometric relations of bite size and metabolic requirements to body size, small animals are more able to subsist on shorter swards than tare1 large animals.” At Overton, relationships of yearling height to final weight and condition seem to be more related to normal course of maturation in that shorter yearling heifers allowed abundant forage tended to have more final condition than the relatively
HEIFER PHENOTYPE x ENVIRONMENT
835.
Holloway, J. W., and W. T. Butts, Jr. 1983. Phenotype x nutritional environment interactions in forage intake and efficiency of Angus cows grazing fescue-legume or fescue pastures. J. h i m . Sci. 56:960. Holloway,J. W., W. T. Butts, Jr., J. R. McCurley, E. E. Beaver, H. L. Peeler, and W. L. Backus. 1985a. Breed x nutritional environment interactions for beef female weight and fat. ness, milk production and calf growth. J. Anim. Sci. 61:1354. Holloway, J. W., W. T. Butts, Jr., J. R. McCurley, H. L. Peeler, E. E. Beaver, and W. L. Backus. 1985b. Breed x nutritional
environment interactions for intake and digestibility of forage grazed by lactating beef females. J. Anim. Sci. 81:1345. Illius, A. W., and J. J. Gordon. 1987. The allometry of food intake in grazing ruminants. J. h i m . Ecol. 563989. Kress, D. D., B. G. England, E. R. Hauser, and A. B. Chapman. 1971a. Genetic-environmental interactions in identical and fraternal twin beef cattle. 11. Feed efficiency, reproductive performance, conformation score and fat thickness. J. h i m . Sci. 33:1186. Kress, D. D., E. R. Hauser, and A. B. Chapman. 1971b. Geneticenvironmental interactions in identical and fraternal twin beef cattle. I. Growth from 7 to 24 months of age. J. Anim. sci. 33:1177. Launchbaugh, K. L., J. W. Stuth, and J. W. Holloway. 1990. Influence of range site on diet selection and nutrient intake of cattle. J. Range Manage. 43:lOQ. Long, C. R., T. S. Stewart, T. C. Cartwright, and T. G . Jenkins. 1979. Characterization of cattle of a five breed diallel: I. Measures of size, condition and growth in bulls. J. Anim. Sci. 49:418. Owens, M. K., K. L. Launchbaugh, and J. W. Holloway. 1991. Pasture characteristics affecting spatial distribution of utilization by cattle in mixed brush communities. J. Range Manage. 44:118. Riewe, M. E. 1961. Use of the relationship of stocking rate to gain of cattle in a n experimental design for grazing trails. Agron. J. 53309. SAS. 1985. SAS User’s Guide: Statistics. SAS Inst. Inc., Cary, NC.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Bolton, R. C., R. R. Frahm, J. W. Castree, and S.W. Coleman. 1987. Genotype x environment interactions involving proportion of Brahman breeding and season of birth. I. Calf growth to weaning. J. Anim. Sci. 65:42. Bransby, D. I., B. E. Conrad, H. M. Dicks, and J. W. Drane. 1988. Justification for grazing intensity experiments. Analyzing and interpreting grazing data. J. Range Manage. 41:274. Butts, W. T., M. Koger, 0. F. Pahnish, W. C. Burns, and E. J. Warwick. 1971. Performance of two lines of Hereford cattle in two environments. J. h i m . Sci. 33:923. Drane, J. W. 1989. Compromises and statistical designs for grazing experiments in grazing research: Design, methodology and analysis. pp 69-83. Crop Sci. SOC.Am. Spec. Agric. No. 16. Holloway, J. W., J. F. Baker, A. D. Chamrad, P. 0. Reardon, and L. W. Varner. 1988. Breed type x management treatment interactions of cows grazing rangelands. J. Anim. Sci. 66:
2667
ABSTRACT2 Yearling heifer growth data were obtained during 4 yr for 524 heifers allotted to either humid bermudagrass pastures (Overton) or semiarid rangeland (Uvalde). Each year, heifers were allotted on April 15 to four forage availability levels (400 to 2,800 kg of DM per 100 kg of BW at Uvalde and 80 to 260 kg of DM per 100 kg of BW a t Overton) maintained by varying the stocking rate monthly until mid-October of each year. Forage availability and yearling heifer characteristics (weight, condition score, and height at hooks, taken on April 15) were treated as continuous variables in regression analyses. Final heifer weight, height, and condition responses to increased forage allowance were related to yearling phenotypes differently for the two locations. Generally, a t Overton, forage availability influenced
final characteristics to a greater extent than did yearling variables, whereas the trend was the opposite a t Uvalde. At Uvalde, the yearling characteristic that had the largest effect on performance was height at hooks. Yearlings with large frames benefited from increased forage allowance by accumulating body fat a t a faster rate than those with small frames. In contrast, at Overton, the yearling characteristic that had the largest effect on performance was condition. Fat heifers responded to increased forage availability to gain even greater advantages in fatness a t the expense of potential growth in height and, thus, achieved early maturation. Yearling phenotypes were more broadly adapted to arrays of forage availability for humid, improved pastures than for semiarid rangeland.
Key Words: Phenotype Environment Interaction, Adaptation, Heifers, Growth, Forage Availability
J. Anim. Sci. 1992. 70:2658-2667
Introduction Many cow-calf enterprises are managed in potentially stressful conditions that animals must a t least partially ameliorate through physiological or behavioral adjustments to grow and reproduce. Fortunately, the cattle phenotypic array is almost as broad as the array of potentially stressful environments. Ecological niches can be broadened by strategic application of inputs to reduce stress, but often these inputs are either too costly or too
'Approved for publication as Journal Article no. 30169 of the Texas Agric. Exp. Sta. 2Present address: Dept. of Anim. Sci., Univ. of Georgia, P. 0. Box 748, Tifton 31793-0748. Received January 27, 1992. Accepted May 11, 1992.
risky to gain industry acceptance. Thus, stressful conditions are prevalent in production environments and form the basis for genotype and phenotype x environment interactions (Butts et al., 1971; Kress et al., 1971a,b; Holloway and Butts, 1983; Holloway et al., 1985a,b; Bolton et al., 1987; Holloway et al., 1988). Early identification of adapted cow phenotypes should minimize biological mismatches, thereby decreasing the proportion of lowly productive or nonproductive cows in the herd and increasing the efficiency of the system. Because forage availability and cattle type are variables that managers can manipulate most easily and frequently, our purpose was to determine yearling heifer phenotype x forage availability interactions for two environments. Inferences tested are within environments (locations), although we also make some general observations concerning the contrasting nature of these interac-
2658
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
*Texas A&M University Research and Extension Center, Uvalde 78801 and +Texas A&M University Research and Extension Center, Overton 75684
HEIFER PHENOTYPE x ENVIRONMENT
tions between environments. Because these environments involve a myriad of confounded, potentially stressful conditions, the purpose of this experiment was to describe the interactions and not to delineate their causes. We formulate some hypotheses that, though untested, are consistent with empirical observations.
Animal Management. Five hundred twentyfour Brahman-Hereford F1 weanling heifers (approximate age of 8 mol were purchased during October and November of each year during a 4-yr period (1984 to 1987). Heifers were randomly allotted within source (one to four sources per year) to either semiarid rangeland in Uvalde (345 heifers, 52 cm of annual precipitation) or humid, improved pasture in Overton (179 heifers, 113 cm of annual precipitation). Heifers were wintered in systems typical to those practiced by the local beef industry for the two locations. At Uvalde, heifers were allowed native rangeland and supplemented for approximately 50 d (January to March) each year with .4 kg.anima1-l .d-’ of cottonseed (1985 and 1986) or .45 kg . animal-’. d-l of cottonseed meal (41% CP 1987 and 1988). At Overton, heifers were allowed ad libitum access to Cynodon dactylon hay supplemented with 2.3 kg.anima1-l.d-l of ground corn and .68 kg.anima1-lad-’ of 41% CP commercial supplement from January 8 to March 4. From March 4 until the beginning of the experiment, heifers grazed Secale cereale var. ‘Elbon’ and Lolium multiflorum var. ‘Marshall’ mixture. Animals at both locations were allowed free access to trace mineralized salt3. Heifers were exposed to Braford bulls (single-sire groups a t Overton, multisire groups at Uvalde, depending on pasture size to provide 20 heifershull) from April 15 to July 1 of each year. Bulls were purchased from the same source and allotted at random to location and to forage availability. AZZotment to Forage Availability. During April of each year, heifers at each location were stratified by weight and randomly allotted to four levels of forage availability designed to include the range in forage allowances in normal beef industry practice for the respective areas (80to 260 and 400 to 2,800kg of DM per 100 kg of BW, respectively,
3Trace mineralized salt composition: 33% NaC1, 11% Ca, 8 % P, 1.9% Mg, 1.3% K, . 4 % Mn, . 5 % Zn, .8% Fe, .1% Cu, 154,000 IU/kg, vitamin A, and 1,100 IU/kg vitamin D3.
at Overton and Uvalde). Heifers at Uvalde were allotted to four extensively managed native rangeland pastures each year (1,950to 6,100ha) dominated by a n overstory of Acacia species and Prosopis glandulosa with a n understory of native grasses including Bouteloua curtipendula (sideoats grama) and B. trifida (red grama), Buchloe dactyloides cbuffalograss), Helaria belangeri (curley mesquite), Pappophorum bicolor (pink pappas), Aristida wrightii (Wright’s threeawn), and Setaria leucopila bristlegrass) and native forbs including Ambrosia psilostachya and Croton texensis. Heifers a t Overton were allotted to intensively managed Cynondon dactylon bermudagrass) pastures each year (7.5to 13.5ha) fertilized with 330:lOO:lOO;kg/ h a of N:P205:K20 in split applications. Under general industry practice, more forage is provided per animal at Uvalde than a t Overton because the forage is more sparsely distributed and, thus, although it is provided and theoretically is available, it may not be accessible (e.g., shrub pattern may restrict animal movement). Estimation of Forage AvailabiZity. Targeted forage availability was attained gradually from initiation of the trial (April 15) to approximately June 1 of each year as a precaution against the possibility of drouth and resultant forage availability being less than that targeted for the respective groups and, thus, forced removal of experimental animals from treatment. The process for estimating and attaining targeted forage availability was as follows. Forage availability was calculated monthly by using clipped quadrants. At Uvalde, grass and forb mass was defined as forage, even though cattle may a t times consume shrubs (Launchbaugh et al., 1990). Pastures were stratified by range site, and 20 to 90 .5-m2 quadrants were randomly located within each site, depending on pasture size and within-site variation in forage availability. Grass and forb mass was visually estimated for each quadrant and then, for a random 20% of quadrants, was hand-clipped to ground level, dried, and weighed. Weighed mass was regressed on visual estimates a t each sampling time (R2 from .85 to .981. The resulting equations were used to calculate DM/hectare for each pasture each month. At Overton, four randomly located .25-m2quadrants were hand-clipped to ground level for each pasture each month. Random samples from each of these quadrants were dried and weighed, and the DM/hectare was calculated. Stocking rate ranged from 17 to 146 kg of BW/ ha at Uvalde and 1,249 to 4,250 kg of BW/ha at Overton and was adjusted monthly to maintain targeted forage availability. Animals were weighed monthly and the number was adjusted as
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Materials and Methods
2659
HOLLOWAY ET AL.
2660
FIGURE lA. UVALDE
FORAGE
80
Figure 1. Relationship of yearling weight and forage availability to final weight (n = 345, R2 = .86, residual SD = 18.38 for Uvalde [Figure la]; n 179, Rz .76, residual SD = 23.21 for Overton [Figure lb]).
-
-
scores were taken according to the method of Long et al. (1979; emaciated = 1, extremely fat = 9) by the same scorer each month within location. All heifers were scored by both scorers each October to assure similarity in scoring method between locations. Yearling and final measurements were those taken in April and October, respectively, for each grazing season (1985 to 19881, and the change was calculated by subtracting the yearling value from the final value and dividing by elapsed days. Statistical Analysis. Experimental design was dictated by the following conditions: 1) variables of interest (forage availability and yearling heifer phenotype1 were those most often manipulated by resource managers; 2) these variables are continuous in nature; and 31 large numbers of animals and pastures (as well as large-sized pastures in semiarid conditions) were required to attain the necessary arrays in the variables of interest. Therefore, we executed a design similar in principle to those described by Riewe (lQSll, Bransby et al., (19881, and Drane (19891. This design employed years (four pastures per location per year) to accumulate the necessary arrays and, in the terminology of Drane (19891, “forage availability was taken as a surrogate for paddock effect. . , mapping the complex interaction of paddock x year x forage available.” Regression procedures (SAS, 19851 were used with the response variables final weight, height, and condition and changes in these variables during the grazing period. Linear models in the analyses of these response variables included fixed effects of year and location, linear, quadratic, and cubic effects of forage availability and initial yearling heifer characteristics (weight, height, or condition), and two- and three-factor interactions among the continuous effects and location. Initial analyses indicated that the location x forage availability x yearling heifer characteristics were generally important (P < . l o ) for all dependent variables. Subsequent analyses were performed within location using linear models including classification effects of year, polynomial effects of initial yearling characteristic (up to cubic), and all possible interactions among classification and covariates. Models were reduced as was appropriate, to include only those effects contributing (P e .lo1 to the regression sums of squares. Equations developed for final characteristics were similar to those for changes in these characteristics. Therefore, only equations predicting final characteristics are presented. Also, separate equations were developed for independent variables describing yearling phenotype (weight, height, or condition) because of autocorrelation among these variables.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
needed to maintain targeted kilograms of DM/ kilogram of BW. Calculated monthly forage availabilities were averaged to compute forage availability for the season. This procedure was employed to attain a n array of forage availability during the 4 yr so that forage availability could be considered a continuous variable. Therefore, deviations from average seasonal targeted forage availability were not considered to be serious errors. Grazing season ended on approximately October 15 each year. Animal Measurements. Heifer 12-h fasted weight, height a t hooks (tuber coxae), and condition score were recorded monthly during the grazing season (April 15 to October 15). Weights were used to compute monthly forage availability. Condition
HEIFER PHENOTYPE x ENVIRONMENT
Results and Discussion
noted for final weight and, thus, only the figure concerning final weight is presented (Figure I). In each case, a synergistic relationship was detected between yearling weight and forage availability in that an increase in each resulted in increased final weight keight and condition), and as both increased, final weight (height and condition) increased at an accelerating rate, resulting in significant interactions (Table 2, Figure 1). At Overton, forage availability affected (P e .061 final height but to a lesser extent than a t Uvalde, and heavier yearling heifers were taller the following fall than lighter heifers a s yearlings (Table 21. Apparently, under extensive conditions a t Uvalde, heavier yearling heifers responded to a n increase in forage availability in terms of final weight, height, and condition to a greater extent than did lighter yearling heifers. Thus, to ensure maximum growth response to an increase in forage availability under semiarid conditions, heavier yearling heifers would be required. Because the forage a t Uvalde was more sparse than that at Overton, heavier yearling heifers possibly had a n advantage in traversing the area and procuring forage, giving them nutritional advantages that resulted in increased growth responses. A contributing factor to the differences between locations in the shape of response surfaces could be differences in forage quality. Forage quality was not measured because we did not have adequate means to measure selectivity. Because forage on offer often deviates from that consumed, and because individual animals vary in selectivity, available analytical methods were either inappropriate or unfeasible. The contrast between Overton and Uvalde is largely a result of the magnitude of responses of the two dimensions a t the two locations. For
Table 1. Characteristics of experimental sample of heifers on humid bermudagrass pastures (Overton) and semiarid rangeland (UvaldeIa
Item
1985-1988
1985
1986
1987
1988
Overton Uvalde
Overton Uvalde
Overton Uvalde
Overton Uvalde
Overton Uvalde
Precipitation (April-October), cm 63.55 Yearling Condition score 6.1' Weight, kg 304.4' Height, cm 123.8' Final Condition score 6.7' Weight, kg 399.8' Height, cm 130.3' No. 179
47.38 5.1'
289.7d 120.6d 4.9d 358.2' 126.5d 345
68.73
46.25
68.01
52.83
54.97
70.99
42.44
19.46
RSDb
-
5.3' 5.2' 259.2Cd 304.2e 121.3Cde 118.1
6.0d 313.3e 123.Qfg
4.6e 270.4' 120.0'
7.1 341.8 123.Qf
5.9' 333.2 122.Bdg
6.1d 303.4e 127.1
4.5e 251.5' 121.3e
.86 29.16 4.00
4.7 377.2d 132.9' 88
7.0' 405.4Ce 128.4' 45
6.0 375.7' 126.5 72
8.6' 394.1ef 128.4e 47
6.4d 382.8df 124.5f 92
6.5' 387.3df 131.0df 58
2.4 297.3 123.2 93
.9 1 36.36 4.00
6.7"
411.7' 132.9'' 31
-
*Least squares means from the model: Y = year, location, year x location. bResidual standard deviation. cid,e,f-gMeanson same line across years or between locations that do not have a common superscript differ (P < .05, t-test).
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Characteristics of Experimental Sample. Although the heifers were randomly allotted within source to location at weaning time, differences between locations in winter management resulted in heifers a t Overton that were 14.7 kg heavier (P e .01),3.2 cm taller (P < .01) and had 1.0 unit more (P < . O l l condition than those at Uvalde (Table I). During the trial, they also gained more (P e .011 weight, condition, and height. Also, variation among years for all responses was much less a t Overton than a t Uvalde. These data, when considered with the large areas required to attain targeted forage levels and the complications provided by the presence of shrubs in inhibiting animal access to forage at Uvalde (Owens et al., 19911, indicate that nutritional and foraging stresses were much greater a t Uvalde than at Overton. Yearling Weight x Forage Availability Interactions. Yearling weight interaction with forage availability was important (P e ,101 for all dependent variables (final weight, height, and condition) a t Uvalde but only important (P e ,101 for final weight a t Overton (Table 2 and Figure 11. Interaction of yearling weight and forage availability on final weight was more dramatic a t Uvalde than at Overton (Figure 1). At Uvalde, heifers that were heavier a s yearlings responded in terms of final weight, height, and condition to increased forage availability more dramatically than did lighter heifers. At Overton, although the trend was the same for final weight as that at Uvalde, heavier yearlings responded to increased forage availability only to a slightly greater extent that did lighter yearlings (Figure 1). At Uvalde, the interactions of yearling weight and forage availability on final height and condition were similar to interactions
2661
HOLLOWAY ET AL. 2662
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
-
r
c
e
E
$ 8
a, h
9 cj
4
,-I
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
9 9 9 2 9 1 1 1 1 1 1 1
88gm8
2663 HEIFER PHENOTYPE x ENVIRONMENT a o0 c u m0 O O l - c u ~ ~ 999991999919
HOLLOWAY ET AL.
2664
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
HEIFER PHENOTYPE x ENVIRONMENT
Yearling Condition x Forage Availability Interactions. Interactions of forage availability and yearling heifer condition were generally more important at Uvalde than at Overton (Table 3 and Figure 2). Interactions were detected (P < .02)at Uvalde for both final weight and condition, but, because the relationships were similar, we believe the effect on weight to be a result of the effect on condition and, thus, only the latter is shown (Figure 21. At Uvalde, although yearling condition had a greater relative effect than forage availability on final condition, the two dimensions acted synergistically in influencing final condition. That is, increases in forage availability resulted in more dramatic increases in final condition for heifers of greater yearling condition than for those of lesser yearling condition. Also, cattle of greater yearling condition always had greater final condition, but the relationship was more pronounced for heifers
flGURE 2A WALDE
Figure 2. Relationship of yearling condition and forage availability to final condition (n = 345, R2 = .81, residual SD = .81 for Uvalde [Figure Za]; n = 179, R2 = .54, RSD = .58 for Overton [Figure Zb]).
Figure 3. Relationship of yearling height and forage availability to final condition (n = 345, R2 = .77, residual SD = .90 for Uvalde [Figure 3a]; n = 179, R2 = .48, RSD = .62 for Overton [Figure 3b]).
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
example, the rate of response for final weight was much greater in the forage availability dimension at Overton than in that a t Uvalde. That is, at Overton, a n increase of approximately 60 to 90 kg in final weight per 380-kg increase in forage availability was noted compared with a n increase of approximately 30 to 80 kg in final weight per 2,400-kg increase in forage availability a t Uvalde (Figure 1). For the animal phenotype dimension, the response was much greater for yearling heifer weight a t Uvalde than for that at Overton (Figure 1). The forage available a t Overton was apparently easily procured and utilized by all grazers, whereas at Uvalde the sparseness of the forage and the physical impediments provided by the interdispersed shrubs (Owens et al., 1991) possibly resulted in problems of forage search and procurement, especially for the lighter (smaller) yearling heifers.
2665
2666
HOLLOWAY ET AL.
later maturing heifers that had been taller as yearlings.
Summary For semiarid rangeland, yearling height, weight, and condition affected final characteristics more than did forage availability, whereas the opposite was the case for humid, improved pastures. A greater number of interactions were detected for semiarid rangeland than for humid improved pastures, and interactions were more dramatic for semiarid rangeland. In general, the nature of the interactions was that yearling characteristics and forage availability acted synergistically to each other in affecting response variables. The exception was yearling height for humid improved pastures (taller heifers as yearlings responded to a lesser degree to increases in forage availability than those that were shorter). Because of the generally synergistic and curvilinear nature of the relationships and because these relationships were more dramatic under semiarid rangeland, we conclude that yearling F1 phenotypes were broadly adapted to humid, improved pastures varying in forage availability; for semi-arid rangeland, however, phenotypes were more narrowly adapted in that the taller, heavier, fatter yearlings exhibited greater levels of adaptation than their counterparts especially at higher levels of forage availability.
Implications Yearling heifers that vary in frame size and degree of fatness do not respond similarly to increases in forage availability for humid, improved pastures and semiarid rangeland. Animals on semiarid rangeland require access to much more forage to perform similarly to those grazed on humid, improved pastures. Yearling weight and yearling height are important components of interactions with forage availability for yearling heifer growth on rangeland. Yearling phenotype was a less critical determinant of response for humid, intensively managed pastures, although shorter yearling females tend to take advantage of increased forage availability by increasing levels of final condition.
Literature Cited Allden, W. G., and 1.A.McD. Whittaker. 1970. The determinants of herbage intake by grazing sheep: The interrelationship of factors influencing herbage intake and availability. Aust. J. Agric. Res. 21:755.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
allowed greater amounts of forage than for those allowed lesser amounts. At Overton, forage availability did not interact with yearling condition in influencing final condition (or final weight or height, Table 31, and the effect of forage availability was much greater than the effect of yearling condition (Figure 2). Heifers at Overton had consistently greater P c .051final condition scores than those a t Uvalde (P c .05, Table 11, and their final condition was affected to a greater extent by forage availability than was the condition of those a t Uvalde (Table 3, Figure 2). Yearling Height x Forage Availability Interactions. Yearling height x forage availability interactions were important (P c .lo) for final height and condition a t Uvalde and for final weight and condition a t Overton (Table 4). At Overton, the primary variable influencing final weight and condition was forage availability, although the effect was greater for heifers that were shorter as yearlings than for those that were taller [Table 4 and Figure 3). At Uvalde, the primary variable that influenced final condition was yearling height (Table 4 and Figure 3). For final condition a t 400 kg of DM/ha, yearling height was negatively related to final condition, whereas at 2,800 kg of DM/ha, yearling height was positively related to final condition. Heifers that had the greatest final condition were the ones that had been tallest as yearlings and allowed the most forage (Figure 3). Taller heifers at Uvalde possibly had a greater ability to move in search for forage. When forage was available in large enough quantities to meet their nutrient demands, the search was successful and resulted in increased body condition. When allowed low levels of forage, however, these heifers continued to have greater capacity for moving to the forage than did shorter heifers, but the futile search possibly resulted in declines in condition. According to this hypothesis, height is indicative of frame size and ability to range. An alternative hypothesis is that taller heifers also had wider incisor breadth and, thus, were less able to subsist on shorter swards but were more able to subsist when forage was abundant. Allden and Whittaker (1970)reported that young sheep were better able to maintain intake on shorter swards than were mature sheep with larger mouths. Also, Illius and Gordon (1987) concluded that “due to allometric relations of bite size and metabolic requirements to body size, small animals are more able to subsist on shorter swards than tare1 large animals.” At Overton, relationships of yearling height to final weight and condition seem to be more related to normal course of maturation in that shorter yearling heifers allowed abundant forage tended to have more final condition than the relatively
HEIFER PHENOTYPE x ENVIRONMENT
835.
Holloway, J. W., and W. T. Butts, Jr. 1983. Phenotype x nutritional environment interactions in forage intake and efficiency of Angus cows grazing fescue-legume or fescue pastures. J. h i m . Sci. 56:960. Holloway,J. W., W. T. Butts, Jr., J. R. McCurley, E. E. Beaver, H. L. Peeler, and W. L. Backus. 1985a. Breed x nutritional environment interactions for beef female weight and fat. ness, milk production and calf growth. J. Anim. Sci. 61:1354. Holloway, J. W., W. T. Butts, Jr., J. R. McCurley, H. L. Peeler, E. E. Beaver, and W. L. Backus. 1985b. Breed x nutritional
environment interactions for intake and digestibility of forage grazed by lactating beef females. J. Anim. Sci. 81:1345. Illius, A. W., and J. J. Gordon. 1987. The allometry of food intake in grazing ruminants. J. h i m . Ecol. 563989. Kress, D. D., B. G. England, E. R. Hauser, and A. B. Chapman. 1971a. Genetic-environmental interactions in identical and fraternal twin beef cattle. 11. Feed efficiency, reproductive performance, conformation score and fat thickness. J. h i m . Sci. 33:1186. Kress, D. D., E. R. Hauser, and A. B. Chapman. 1971b. Geneticenvironmental interactions in identical and fraternal twin beef cattle. I. Growth from 7 to 24 months of age. J. Anim. sci. 33:1177. Launchbaugh, K. L., J. W. Stuth, and J. W. Holloway. 1990. Influence of range site on diet selection and nutrient intake of cattle. J. Range Manage. 43:lOQ. Long, C. R., T. S. Stewart, T. C. Cartwright, and T. G . Jenkins. 1979. Characterization of cattle of a five breed diallel: I. Measures of size, condition and growth in bulls. J. Anim. Sci. 49:418. Owens, M. K., K. L. Launchbaugh, and J. W. Holloway. 1991. Pasture characteristics affecting spatial distribution of utilization by cattle in mixed brush communities. J. Range Manage. 44:118. Riewe, M. E. 1961. Use of the relationship of stocking rate to gain of cattle in a n experimental design for grazing trails. Agron. J. 53309. SAS. 1985. SAS User’s Guide: Statistics. SAS Inst. Inc., Cary, NC.
Downloaded from https://academic.oup.com/jas/article-abstract/70/9/2658/4705111 by Iowa State University user on 17 January 2019
Bolton, R. C., R. R. Frahm, J. W. Castree, and S.W. Coleman. 1987. Genotype x environment interactions involving proportion of Brahman breeding and season of birth. I. Calf growth to weaning. J. Anim. Sci. 65:42. Bransby, D. I., B. E. Conrad, H. M. Dicks, and J. W. Drane. 1988. Justification for grazing intensity experiments. Analyzing and interpreting grazing data. J. Range Manage. 41:274. Butts, W. T., M. Koger, 0. F. Pahnish, W. C. Burns, and E. J. Warwick. 1971. Performance of two lines of Hereford cattle in two environments. J. h i m . Sci. 33:923. Drane, J. W. 1989. Compromises and statistical designs for grazing experiments in grazing research: Design, methodology and analysis. pp 69-83. Crop Sci. SOC.Am. Spec. Agric. No. 16. Holloway, J. W., J. F. Baker, A. D. Chamrad, P. 0. Reardon, and L. W. Varner. 1988. Breed type x management treatment interactions of cows grazing rangelands. J. Anim. Sci. 66:
2667
