The herds on each of the forage programs included approximately 60 cows. Each herd included five breed groups as follows: ten Brahman (B), ten 3/4 Brahman-1/4 Shorthorn (B₃S ₁), twenty Brahman-Shorthorn first crosses (F₁), ten 3/4 Shorthorn-1/4 Brahman (S₃ B₁) and ten Shorthorn (S). The trial was conducted in two 5-year phases. During the first phase the B, the B₃ S₁, and one-half of the F₁ cows were mated to Brahman bulls while the remaining groups were mated to Shorthorn bulls. This procedure resulted in a balanced design considering breed of calf but resulted in confounding of breed of bull with breed of cow since only one-half of the possible breed combinations were included during one phase. The procedure was dictated, however, by the availability of replacement females for the study and the decision to maintain 10 animals per breed group x year sub-cell rather than five, which would have resulted had breed of cow and breed of sire been used annually in all possible combinations. During the second phase, breed of sire for the different cow groups was reversed from that of the first phase. Thus, including both phases each breed of cow was mated to bulls of both breeds. By including two mating groups (30 F₁ cows mated each to B and S bulls) across all years it was possible to remove year effects and obtain a satisfactory comparison of all breed combinations.

Pasture-Management Programs -

Program 1 - This program involved native grasses, mainly Aristida and Andropogon spp. (wiregrass and broomsedge) During the first 5-year period, the 60 cows and their calves grazed 812 acres of native range divided into five fields. Eighty acres of each of these five fields were burned on alternate years. A reduction to 772 acres in the program was made during the second phase.

Supplemental feed was provided on native range only to the extent necessary to prevent extreme weight loss in the cows. Cows were supplemented during the winter months with 41% cottonseed pellets and a limited amount of hay in the first phase. In phase 2 an average of 555 pounds of hay, 52 pounds cottonseed meal (41%), and 52 pounds of citrus meal were fed over a period averaging 109 days annually beginning in December.

Program 2 - This program included a combination of 73 acres of improved pasture plus 315 acres of native grasses for the 60 cow herd. The improved pasture included 40 acres of Pensacola bahiagrass (Paspalum notatum Flugge.) 15 acres of Coastal bermudagrass (Cynodon daclylon (L.) Pers.) and 18 acres of Pangola digitgrass (Digitaria decumbens Stent. ) during the first phase. For the second phase the bermudagrass was replaced by digitgrass. Approximately 20 acres of the bahiagrass was grown in conjunction with Hairy Indigo (Indigof era hirsuta Linn.) with the remaining 20 acres being in combination with non-irrigated white clover (Trifolium repens L.) and Hubam sweetclover (Melilotus alba Desv.). The cattle had access to the native grasses continuously and generally to one field of improved pasture.

The annual fertilizer program per acre for improved grass pastures was approximately 400 pounds of 8-8-8 (N-P₂O₅-K₂O) plus 150 pounds of ammonium nitrate (33.3% N) for pangolagrass, or a like amount of complete fertilizer plus 100 pounds of ammonium nitrate for bahiagrass pastures. The legume pastures received 250 pounds per acre of 0-8-24 fertilizer annually with an extra application of muriate of potash averaging 32 pounds K₂O per acre on the white clover.

Cattle were given suplemental feed only in the winter of 1957-58 during phase 1. During the last 5-year period the cattle on Program 2 received supplemental feed (43 pounds hay, 65 pounds cottonseed meal and 31 pounds citrus meal per cow) for an average of 72 days annually from January to April.

Program 3 - This program included only improved pastures. During the first phase it included 76 acres of pangolagrass, 20 acres of which were irrigated and over-planted with white clover. Due to pasture becoming less productive the area was increased to 107 acres for the second phase. This change resulted in 70 acres of all pangola digitgrass, 27 acres of digitgrass overplanted with white clover, and 10 acres of bahiagrass. Pasture fertilization in Program 3 followed the same pattern as described for Program 2.

Cattle in Program 3 were maintained in a high nutritional status at all times. For the most part, the cattle maintained good weights from pasture, as only an average of 280 pounds of hay per cow was fed for each of the five winters.

General Management -

The cattle on all programs were bred in a restricted season of 105 days extending from March 15 to July 1. All the calves were weaned at one time in early September.

Replacement heifers were grown together from weaning until being placed in their respective herds just prior to the breeding season at two years of age.

Cows were culled on the basis of repeated reproductive failure or failure to raise a calf.

Data Analyses -

Individual performance data were recorded annually for pregnancy status, calf survival, weaning rate, age of calf at weaning, weaning weight, 205-day weight, and market grade of calf. The research was conducted in two phases, five years each. The separate analyses for both reproduction and production traits are discussed.

The observed data for pregnancy status, calf survival, and weaning rate were recorded as 1 or 0 (zero). These individual data were processed initially in separate analyses for each phase. Least squares techniques as outlined by Harvey (4) were followed, employing the same model for all analyses. All effects were assumed to be fixed. Since subclass numbers were approximately proportional the above analyses yielded efficient estimates of effects.

A third analysis was performed by pooling data for the two phases. The least squares means for the breed x pasture program subclasses from the above analyses were adjusted for differences between phase and combined in a singled analysis. The adjustments required were small for pregnancy and weaning rates and nil for calf survival. The adjustment for combining the two phases into a single analysis was made on the basis of differences between phases for the two breed groups of F₁ cows which were included during both periods of the study. These breed groups represented a total of approximately 300 matings per phase. There were no indications of interactions between the two phases. Thus, it was concluded that the combined analyses yielded unbiased estimates of effects. The error mean square used in testing significance was the pooled error term for the two phases coded by the harmonic mean of subclass numbers. The model for the combined analyses for reproduction traits included breed of dam (D), breed of sire (S), pasture program (P) and their interactions (Table 2).