The selection index--then, now, and for the future.

When improvement is desired for several traits that may differ in variability, heritability, economic importance, and in the correlation among their phenotypes and genotypes, simultaneous multiple-trait index selection was more effective than independent culling levels or sequential selection. Such comparisons required definition of aggregate breeding value determined jointly by breeding values and economic importance of the component traits. The economic weight should approximate the partial regression of cost per unit of enterprise output value on breeding value for each trait. These can vary with production and marketing system, with performance of traits, and with breed role (i.e., paternal, maternal, or general) in crossbreeding systems. Genetic gains desired to maintain competitive ranking also may define the relative importance of traits. Because information available to estimate breeding values varies among the ages and categories of individuals under selection and because means are unknown, regressed (BLUP) predictions of trait breeding values are useful. They allow appropriate economic weights to be applied as the last step for predicting aggregate breeding values for individuals of different age classes, and they simplify choosing the proportions of selected breeders from each age class that maximize rate of change in aggregate breeding values. Inappropriate economic weights or errors in the parameters used to predict trait breeding values overestimate realized response in true aggregate breeding value.

Economic evaluation of heterosis and culling policies for lifetime productivity in Hereford, Angus, shorthorn, and crossbred cows.

Experimental lifetime performance data obtained from 156 straightbred Hereford, Angus, and Shorthorn and 172 first-cross heifers were used to estimate heterosis for economic efficiency in a 100-cow herd at age equilibrium under three culling policies and at terminal ages from 6 to 12 yr. All nonpregnant heifers and cows greater than 9 yr of age were culled. The culling policy for removal of nonpregnant cows from second parity through 9 yr of age were 1) no culling, 2) after two consecutive years (actual), and 3) all (imposed). Efficiency was calculated as input cost per unit of output value. A 10-yr average was used for costs of replacement heifers, cow units, and the ratio of calf:cull cow prices (PR), plus higher and lower PR. Input included costs for both cow units and purchased replacements. Output value included both weaned calves and cull cows. Optimum terminal age was mainly a function of PR: 9 yr for average and high PR, but 6 through 9 yr when PR was low, regardless of culling policy or breed groups. Efficiency differences among culling policies were small for high or average PR, but more culling for infertility was beneficial when PR was low. Estimated reductions in unit costs of output value under any culling policy or terminal age were approximately 6% from crossbred cows plus another 6% from crossbred calves, or a total of 12% from specific three-breed crossing of these British breeds. Cost reductions would be somewhat less for rotation crossbreeding but greater for mating smaller crossbred cows with sires of superior growth-carcass breeds.

Heterosis for lifetime production in Hereford, Angus, shorthorn, and crossbred cows.

The effects of maternal heterosis and maternal and grandmaternal breed effects on cumulative lifetime number and weight of calves weaned per cow entering the breeding herd were evaluated for 172 reciprocal crossbred and 156 straightbred cows of the Hereford, Angus, and Shorthorn breeds. Cows born in 1960 and 1961 were developed and mated to calve first at 3 yr of age and those born in 1962 and 1963 at 2 yr of age. Performance under actual culling of cows nonpregnant in two consecutive years and imposed culling of any nonpregnant cows were analyzed. Reproductive rates and weaning weight per calf and per cow exposed increased (P less than .05) as cows advanced from 2 through 5 yr of age, peaked at ages 5 through 9 yr, and decreased from 9 through 12 yr. Effects of heterosis did not interact (P greater than .05) with age at first calving management. During the 12-yr span in the 2-yr-old first-calving management system, crossbred cows produced nearly one more calf than straightbred cows under the actual culling policy (.97 calves, P less than .10) and .82 more calves (P less than .10) if all nonpregnant cows were culled. Corresponding cumulative calf weight weaned was 272 kg (P less than .01), or 25% more, and 232 kg (P less than .01), or 30% more, for crossbred cows than for straightbred cows. The 12-yr cumulative calf weight weaned by straightbred Angus cows exceeded that of Shorthorn cows (P less than .05) and tended to be greater than that of Hereford cows.

Reproductive rate and genetic variation in composite and parental populations: experimental results in sheep.

Mean and genetic variability for ewe performance of two composite populations were compared with those of their contemporary parent lines to assess relative potential for continued response from selection. Data were from the U.S. Meat Animal Research Center, 1973 to 1983. The 4,415 ewes by 508 sires of Finnsheep (F), Rambouillet (R), Dorset (D), and Composite 1 (C1 = F/2 + R/4 + D/4) lines were mated in April, August, and December of each year. The 1,763 ewes by 320 sires of Finnsheep (F), Suffolk (S), Targhee (T), and Composite 2 (C2 = F/2 + S/4 + T/4) were mated in October only. Data were adjusted for effects of year, season, and age of ewe at breeding, plus interval between lambings, for the accelerated lambing lines and for year and age at breeding for the annual lambing lines. Superiority of composite over parent lines in lambs weaned per ewe exposed was 36% for accelerated and 44% for annual lambing lines; that superiority arose largely from dominant heterotic effects for fertility and viability. Weighted mean paternal half-sib estimates of heritability for mean lifetime ewe performance of composite vs parental lines were 32 vs 23% for litter birth weight, 25 vs 28% for lamb birth weight, 26 vs 11% for number born, 11 vs 23% for conception rate, 5 vs 10% for number weaned, -1 vs 19% for lamb weight at 42-d weaning, but 6 vs 23% for number and 8 vs 28% for litter weight weaned per ewe exposed. Heritability was higher in composites than in parental lines for numbers born, but it was lower for the more heterotic traits (fertility, lamb viability) and, hence, for total weaned lamb output, without change in phenotypic variation. Thus, even an equal response in composites from continuing selection for weaned lamb output may depend mainly on more intense selection permitted by a higher reproductive rate.

A deterministic computer simulation model of life-cycle lamb and wool production.

A deterministic mathematical computer model was developed to simulate effects on life-cycle efficiency of lamb and wool production from genetic improvement of performance traits under alternative management systems. Genetic input parameters can be varied for age at puberty, length of anestrus, fertility, precocity of fertility, number born, milk yield, mortality, growth rate, body fat, and wool growth. Management options include mating systems, lambing intervals, feeding levels, creep feeding, weaning age, marketing age or weight, and culling policy. Simulated growth of animals is linear from birth to inflection point, then slows asymptotically to specified mature empty BW and fat content when nutrition is not limiting. The ME intake requirement to maintain normal condition is calculated daily or weekly for maintenance, protein and fat deposition, wool growth, gestation, and lactation. Simulated feed intake is the minimum of availability, DM physical limit, or ME physiological limit. Tissue catabolism occurs when intake is below the requirement for essential functions. Mortality increases when BW is depressed. Equations developed for calculations of biological functions were validated with published and unpublished experimental data. Lifetime totals are accumulated for TDN, DM, and protein intake and for market lamb equivalent output values of empty body or carcass lean and wool from both lambs and ewes. These measures of efficiency for combinations of genetic, management, and marketing variables can provide the relative economic weighting of traits needed to derive optimal criteria for genetic selection among and within breeds under defined industry production systems.

Simulation of life-cycle efficiency of lamb and wool production for genetic levels of component traits and alternative management options.

A deterministic computer model was used to predict effects of genetic improvements in performance and of management options on life-cycle flock TDN input per unit of empty body weight (EBW) or carcass lean (CLN) equivalent value of market lamb, cull ewe, and wool output from a pure breeding system of sheep production for sheep fed to maintain normal weight. Relative values per kilogram for market lambs, cull ewes, and clean wool were 1 to .33 to 2.04 in lamb empty body equivalent, but 1 to .33 to 1 in lamb carcass lean equivalent. A 10% increase in lamb viability improved TDN/EBW or CLN by -15 to -20% for high to low lambing rates. Corresponding smaller gains were -7 to -11% for fertility, -3 to -13% for lambing rate, -1 to -3% for wool growth rate, -1 to -5% for milk production without creep feeding, -2 to -1% for mature size to about 70 kg, and -.6 to -.9% for precocity of fertility. Increasing leanness 10% improved TDN/CLN by -3 to -1% but increased TDN/EBW 3% because of higher maintenance requirements of leaner sheep. Higher protein requirements for increased lambing rate, milk production or leanness, or greater increases in non-feed than in feed costs, would mean only slightly less reduction of TDN/output than shown. Creep feeding was beneficial only for prolific, low-milking stock. Flushing reduced adverse effects of restricted feeding. Different values for wool vs meat or for costs of feed vs non-feed inputs would change results. These estimates for relative economic importance of traits apply to derivation of optimum criteria for selection among breeds or crosses, or within-breeds used in rotation crossbreeding, but would differ for specialized terminal-sire or maternal breed roles.

Simulated effects of reproductive performance on life-cycle efficiency of lamb and wool production at three lambing intervals.

A deterministic computer model of a purebreeding sheep production system was used to simulate effects on life-cycle efficiency from genetic changes in lambing rate (LB), fertility (FERT), precocity of fertility (PREC), and estrus season length (ES) under annual (Y), 4-mo (A), and 73-d (S) breeding intervals. Ewes and lambs of a 1,000-ewe flock were fed to requirement through five yr, beginning with ewe lambs at the optimal breeding season and accumulating feed input (TDN) and market lamb equivalent weight of output (EBW), including cull ewes and wool. For annual lambing, improvement in cost (TDN/EBW equivalence) per 10% increase in genetic potential was -9.8, -8.2, -1.3, and -.8% for LB, FERT, PREC, and ES, respectively. For the A and S lambing intervals, corresponding gains in TDN efficiency were -19 for FERT and -9% for LB when ES was 72 d, but only -6 and -5% when ES was 192 d. Increasing PREC by 10% improved cost -3% under accelerated systems when ES was 72 d, but less than -1% when ES was 192 d. Cost reduction from 10% longer ES under A and S breeding intervals was greater when other reproductive performance was low (-11 and -17%) rather than high (-2 and -3%). Reduction in costs for A and S vs Y lambing intervals was greatest with long ES and was greater with low (-43 and -53%) than with high reproductive performance (-20 and -24%). When ES was 72 d, A and S intervals would reduce costs for high (-9 and -11%) but increase costs for low reproductive traits. Corresponding increases in CP/TDN required in diet would reduce gains in TDN/EBW only slightly. These results should aid in development of optimum selection criteria for stocks used for pure or rotation crossbreeding or as maternal parents of terminal crosses.

Body composition and fasting heat production from birth to 14 months of age for three biological types of beef heifers.

Thirty heifers were randomly sampled at five ages from each of the Hereford, Charolais, and Simmental herds at the U.S. Meat Animal Research Center to estimate association of fasting heat production (FHP) with body composition. Replicated measures of respiratory exchange were obtained for six heifers per breed at ages 2 d, 3 mo, 7 mo, 10 mo, and 14 mo using open-circuit calorimetry. Regression adjustment of FHP/live weight.75 to zero activity (AFHP) reduced the mean by 12% and variance among periods for the same animal by 42%. Daily AFHP (kcal/kg.75) was highest at 2 d (122), lowest at 10 mo (92) (P less than .01), and intermediate (103 to 106) at other ages and averaged 109, 106, and 102 for Charolais, Simmental and Hereford over all ages (P less than .05). Pooled within-age correlations of AFHP were .77 with weight of carcass (CAR) nonfat or water and .75 with live and empty body weight (EBW) but were only .13 with fat weight. Prediction of AFHP within age groups was most accurate from multiple regression on the nonfat weight in visceral organs and blood (VOB), gastrointestinal tract (GIF), head, hide, and shanks (HHS), and CAR fractions (R2 = 61%, error SD = 21.5), from regression on nonfat in CAR alone (60%, 21.6), or from regression on chemical components in each of the four fractions (59%, 22.0), relative to EBW (55%, 22.9) or its four chemical components (58%, 22.3). Partial regressions were largest for water or nonfat (P less than .01) and were negligible for fat. Importance in predicting AFHP was two to eight times greater for nonfat in CAR than in other fractions because CAR was 60 to 65% of EBW. Lean mass is clearly a major predictor of nutrient requirement that is useful to evaluate effects of body composition on the efficiency of beef production.

Genetic variation in composite and parental populations: expectations for levels of dominance and gene frequency.

Usefulness of multibreed composites depends on: 1) adaptation of the average breed effects combined, 2) amount of initial heterosis retained, and 3) the rate of further improvement from selection. Potential improvement could be increased, relative to parental stocks, by retained heterosis in reproductive rate and by any increase in additive genetic variation in the composite. A single-locus, two-allele model with additive plus dominant gene effects was used to clarify expected changes from varying: 1) degree of dominance (d) from none to over-dominance, 2) variance among parental lines in frequency of a dominant allele (Vq), and 3) mean gene frequency in the composite (qc). Simulation for a three-breed composite (1/2, 1/4, 1/4) showed that expected heterosis retained in the composite is proportional to d and Vq and is highest for any d and Vq when qc is intermediate. Percentage of change of the composite from parental mean in additive genetic variance (delta Vac,%) increases most with parental diversity (Vq) when d = 0 but changes with higher d from positive when q is less than .5 to negative when q is greater than .5. Therefore, the expected association of delta Vac with level of heterosis retained (delta H,%) is always smaller for higher d and changes from positive when qc less than .5 to negative for the more likely higher equilibrium values of qc. Thus, greater selection intensity (from a higher reproductive rate) in composites than in parent stocks may not be accompanied by increased genetic variability and change in response per unit of selection applied may be limited.

Lifetime production of beef heifers calving first at two vs three years of age.

Records of Angus, Hereford, Shorthorn, and first-cross cows born from 1960 to 1963 were used to study the effect of age at first calving and culling policy on cumulative survival (SU), number of breeding seasons (BS), pregnancies (PG), calves born (CB) and alive at 72 h (C72) and at weaning (CW), calf weaning weights (WW), and input/output efficiency, up to 12 yr of age. The 155 cows born in 1960 and 1961 calved first at 3 yr of age (M3) and the 173 born in 1962 and 1963 calved at 2 yr of age (M2). Analyses included both actual culling of cows open in two consecutive years (AC) and imposed culling of any open cow (IC). Also, economic efficiency for no culling (NC) of cows for infertility after first calving and up to 10 yr of age was calculated. The model included cow birth year-sire breed of cow (Y-S), sires within Y-S, breed of cow's dam (D), and D x Y-S. Linear functions of Y-S were used to estimate means for M2 and M3 cows. Under AC, by 12 yr of age M2 exceeded M3 cows by 6.8% SU, 1.1 BS, 1.2 PG, 1.1 CB, 1.0 C72, .9 CW, and 138 kg WW (P less than .10). Under IC, M2 exceeded M3 cows by only .7% SU, .3 BS, .3 PG, .3 CB, .2 C72, and .1 CW; and M3 produced 24 kg more WW than M2 cows (P greater than .10). Economic efficiency was higher for M2 than for M3 cows (greater than 10% at 6 yr to greater than 5% at 12 yr culling age), regardless of culling policy. Economic efficiency was maximum when terminal age of cows was 6 to 9 yr for M2 and 8 to 9 yr for M3 cows. Repeatability of pregnancy was higher (P less than .05) for M3 (19%) than for M2 (2%) cows. Economic efficiency for M2 or M3 cows under NC was slightly poorer (2 to 3%) than under either AC or IC policies. Economic efficiency can be improved by managing heifers to calve first as 2-yr-olds under either mild or intense culling of open cows.

Life-cycle biological efficiency of Bos indicus x Bos taurus and Bos taurus crossbred cow-calf production to weaning.

A deterministic model was developed that accounted for all biological inputs and outputs for a theoretical herd of F1 females at age equilibrium mated to produce three-way terminal-cross calves and for the required proportion of straightbred cows needed to produce replacements. Two Bos indicus x Bos taurus vs two Bos taurus x Bos taurus types of crossbred cows were compared in the production environment of south-central Nebraska. The four types of F1 females were from Hereford (H) or Angus (A) dams and by H or A (HA), Pinzgauer (Pz), Brahman (Bm), or Sahiwal (Sw) sires. The crossbred females were assumed mated to Red Poll (R) sires for their first calving and to Simmental (S) sires thereafter. Two evaluations of efficiency for each of the four breeding systems were total cow and calf feed energy input 1) per unit of only weaned calf weight output (CALFEFF, Mcal/kg) and 2) per unit of weaned calf plus .55 x cull cow weight output (TVALEFF, Mcal/kg). Results for a terminal age of 7 yr in systems using HA, Pz, Bm, and Sw crossbred cows, respectively, were 64.9, 64.5, 60.9, and 59.3 Mcal/kg for CALFEFF and 45.7, 46.4, 44.1, and 43.7 Mcal/kg for TVALEFF. Changing terminal age to 11 yr reduced CALFEFF about 6% but increased TVALEFF about 7%, because total inputs increased more than output value (10 vs 3%) from 7 to 11 yr terminal ages. These results suggest differences in efficiency among these breed crosses favoring the Bos indicus crossbred cows by over 4% in this particular environment.

Output/input differences among nonpregnant, lactating Bos indicus-Bos taurus and Bos taurus-Bos taurus F1 cross cows.

Nonpregnant F1 crossbred cows, progeny of either Hereford (H) or Angus (A) dams and sired by Brahman (Bm), Sahiwal (Sw), Pinzgauer (Pz), H, or A sires, were fed to maintain initial weight while rearing Charolais (C)-sired progeny for a period of 126 d in drylot commencing at about 48 d postpartum. Cow-calf pairs were assigned to equalize cow age, calf sex, and breed of cow's dam among three replicate pens of approximately 12 pairs each. Cows and calves were weighed every 2 wk and feed intake was adjusted to minimize change in cow weight. Metabolizable energy (ME) consumption for zero cow weight change was estimated by regression. Milk production was estimated by weight-suckle-weigh at 58, 85, 125, and 170 d of lactation. Calf gain (GAIN, kg) relative to cow weight (CWT1, kg) was higher (P less than .01) for calves from Bm-X (139.5/585) and Sw-X (132.2/534) than for calves from Pz-X (127.2/552) and HA-X (116.9/547) cows. Estimated mean daily production of milk was 7.40, 7.15, 7.28, and 6.37 kg for the Bm-X, Sw-X, Pz-X, and HA-X, respectively. Total cow ME intake (TMEcow) for breed groups ranked (P less than .05) with cow size and milk production, and calf creep-feed intake (FMEcalf) was inversely related to estimated milk intake. Proportion of total feed ME (TMEcow+calf) consumed by calves was higher (P less than .05) for HA-X cows (18%) than for the others (14%). Total efficiency of calf gain in weight (GAIN/TMEcow+calf) was 11% greater (P less than .05) for crossbred cows of Bos indicus X Bos taurus (Bm-X, Sw-X) than for Bos taurus X Bos taurus (Pz-X, HA-X) cows (35 vs 32 g/Mcal) in the 126-d lactation period.

Effects of breed, heterosis, age of dam, litter size, and birth weight on lamb mortality.

Sources of individual plus maternal effects on lamb mortality were studied in data collected at the U.S. Meat Animal Research Center from 1980 to 1985 for paternal and maternal breeds lambing yearly and for maternal breeds lambing at 8-mo intervals. Records included 16,881 lambs born. Breeds included were Finnsheep (F), Dorset (D), Rambouillet (R), Suffolk (S), Targhee (T), Composite 1 (C1 = F/2, D/4, R/4), Composite 2 (C2 = F/2, S/4, T/4), and Composite 3 (C3 = Columbia/2, S/4, Hampshire/4), Traits analyzed were perinatal, postnatal, and total mortality to 60 d of age and postnatal respiratory, digestive, starvation, injury, and other or unknown causes of mortality. The least squares analyses included breed, year, sire within breed-year, sex, linear and quadratic season and age of dam covariates (Model 1), plus litter size (Model 2), plus birth weight (Model 3), and significant two-way interactions. Age of dam, litter size, and birth weight all had important, often quadratic, effects that differed among breeds. Models 1, 2, and 3, respectively, reduced variation by 8, 10, and 16% for perinatal, 7, 8, and 12% for postnatal, and 9, 11, and 20% for total mortality. In Model 1, breed means ranged from 3.5 to 16.2% for perinatal, 7.2 to 21.1% for postnatal, and 16.7 to 32.8% for total mortality. Respiratory and starvation problems were major causes of postnatal mortality. Heterosis for lamb survival retained in composites was 9% for C1 and 18% for C2. Mortality was 1 to 5% higher for males than for females.(ABSTRACT TRUNCATED AT 250 WORDS)

Genetic and phenotypic variation in sources of preweaning lamb mortality.

Information on causes of mortality to 8 wk weaning for 16,881 lambs of 10 line-breed groups by 594 sires born over a 5-yr period was analyzed by least squares procedures to estimate sire variance and covariance components within line and year-season. Heritability estimates were approximately 5% for total mortality in the binomial scale and near 10% when converted to a normal scale. Average heritability estimates in the binomial and normal scale were, respectively, 4 and 12% for perinatal and 3 and 7% for postnatal mortality. The heritability estimates for respiratory problems were inconsistent; binomial-scale estimates ranged from 0 to 7% in different populations. Regression adjustment for effects of variation in litter size and birth weight caused only minor changes in the heritability estimates. Comparison of covariances among half-sibs vs full-sibs indicated very large maternal effects on perinatal mortality but smaller and less consistent effects on sources of postnatal mortality. Phenotypic correlations among sources of mortality indicated a 26% greater influence of postnatal than perinatal mortality on variation in total mortality, but both the genetic and phenotypic correlations between those two components were slightly negative. These results suggest that the use of family and progeny test selection for transmitted effects on both perinatal and postnatal lamb viability should be effective.

Comparison of Booroola Merino and Finnsheep: effects on productivity of mates and performance of crossbred lambs.

Booroola Merino (BM) and Finnsheep (FS) rams were mated to FS and crossbred (CO) ewes to compare their effects on the productivity of their mates and the performance of their progeny. Effects of breed of ram and the interaction of breed of ram X breed of ewe were no significant for fertility, total number born, number born alive, litter birth weight, or number of lambs weaned by the ewe or in the nursery. Birth weights of FS- and BM-sired lambs were equal, but FS-sired lambs were heavier at weaning and 147 d of age. Compared with BM-sired lambs, survival from birth to weaning of dam-reared lambs was greater for FS-sired lambs produced by CO dams than for FS-sired lambs produced by FS dams. At a constant live weight, FS-sired lambs had a heavier carcass with less external fat thickness than BM-sired lambs. Percentage of ewe lambs reaching puberty was lowest and age at puberty was highest for the BM X CO cross; relatively small differences existed between the other three genotypes for these traits. Ewe lambs sired by BM had a greater ovulation rate than FS-sired ewe lambs. Embryo mortality was greater in BM X CO ewe lambs than in ewe lambs of the other three genotypes; the difference was larger for ewes with three ovulations than for ewes with two ovulations. The net result was that BM X CO ewe lambs averaged slightly fewer lambs born than FS X CO ewes. The BM X FS ewe lambs had larger litters than the FS ewes due to a substantially higher ovulation rate and only a slightly lower overall embryo survival rate. With the exception of ovulation rate and litter size, FS rams produced lambs with a more desirable level of performance than did BM rams.

Heterosis for survival and dentition in Hereford, Angus, Shorthorn, and crossbred cows.

Records of 328 Angus, Hereford, Shorthorn, and first-cross cows produced from 1960 to 1963 were used to estimate heterosis (h) and individual (gI) and maternal (gM) average breed effects on cumulative survival (CS) by ages to 12 yr, longevity (L), and size (SI) and condition of incisors at 10 to 15 yr of age. Reasons for cow disposal were also studied. The 155 cows born in 1960 and 1961 first calved at 3 yr of age, and the 173 born in 1962 and 1963 first calved at 2 yr of age. Analyses for CS and L were done for both actual culling of cows open in two consecutive years (AC) and imposed culling of any open cow (IC). The model for CS and L included cow birth year-breed of cow's sire (Y-S), sires within Y-S, breed of dam (D), and D x Y-S. For SI, the model included breed of cow, year of recording, and quadratic effect of age. Linear contrasts were used to estimate h, gI, and gM for quantitative variables, and chi-square was used for discrete variables. Under both AC (P less than .05) and IC, crossbreds exceeded straightbreds in CS to 12 yr (16.7 and 8.5%) and L (1.36 and .99 yr). Estimates of gI and gM for CS and L were generally small, except for the low (P less than .05) gM of Shorthorn cows under IC. Mortality and culling for emaciation, cancer eye, or prolapse was less for crossbreds than for straightbreds and increased with age. Crossbreds had longer (P less than .01) and better (P less than .001) teeth than straightbreds. The longer productive life of crossbred cows reduces rearing costs for replacements and increases sales of calves and cull cows.

Embryo-transfer twinning and performance efficiency in beef production.

Effects of twinning on efficiency of beef production were estimated from results of bilateral transfer of two Angus x Hereford (AxH) embryos into each of 241 heifers and 84 cows (H, A, HxA or Holstein x H) over 4 yr. Calves were weaned at 180 d and fed either 220 d in a feedlot (1977) or 170 d on forage and 140 d in a feedlot (1978 to 1980). Effects of parity, twinning and sex of calf were estimated as covariates within year-breed of dam. Pregnancy at 45 to 60 d of gestation was 68% in heifers (H) and 74% in cows (C), with 40% single (S) and 60% twin (T) births. Dystocia was 28% in H vs 10% in C (P less than .05), and tended to be less (P greater than .05) for T than S in H. More placentas were retained (P less than .05) for T than for S in both H (35 vs 12%) and C (24 vs 4%). Twin gestations averaged 3 d shorter and subsequent calving intervals 13 d longer (P less than .05), but total calf mortality was slightly higher (P greater than .05). Abortions were 4% in H only. Twinning females lost maternal weight during late gestation (P less than .05) when crowding limited voluntary feed intake, while fetal requirements were 60% higher (P less than .01). Twins increased milk output 25% (P less than .05), but 11% higher feed intake maintained cow weight during lactation. Twinning reduced birth weight 13% and weaning weight 17% (P less than .05), but 400-d feedlot weight only 9% because of compensating feedlot gain. Twins gained 18% faster than S during postweaning 170-d forage feeding, but 5% slower in feedlot to 8% lighter 490-d weight (P less than .05). Assuming 40% higher veterinary and labor costs for twins, estimated integrated herd costs per unit of age-constant output value would be lower for T than for S production by about 24% for marketing either at weaning or at 400 d.