3. If the basis for making early selections is relatively inaccurate, the progeny test is more likely to be effective. Therefore, the progeny test would be more apt to improve the annual progress in traits where heritability is low, than in traits where it is high. Thus, the relative value of the progeny test is determined by a combination of circumstances that are largely beyond the breeder's control, and a regular plan of progeny testing is unlikely to increase ( and may reduce) genetic progress unless the progeny-test information becomes available early in the animal's lifetime, the reproductive rate is low, and the basis for making early selections is relatively inaccurate. Dickerson and Hazel point out that improvement from selection is nearly maximum for most traits when culling is based on individual performance, family average, and pedigree, and when the interval between generations is kept short.

Dickerson and Hazel also studied the effectiveness of different methods of selecting for two specific characters in swine, growth rate of pigs and productivity of sows, and they have made some recommendations concerning the procedures that should be most effective. In selecting for growth rate, they recommend that 8 to 10 times as many boars and about 3 times as many gilts as are needed for breeding should be retained long enough after weaning (such as 180 days of age) to obtain a more reliable measure of growth rate than weaning weight. The rest may be culled without reducing appreciably the effectiveness of selection.

Several plans for culling were compared. Yearly progress from selection is greatest when sows are culled after the first litter, the best one-third to one-half being kept for a second litter 6 months later. Another plan, which is almost as effective, is to delay culling until after the second litter, and keep the best one-fifth to one-fourth of the sows for a third litter at 2 years of age. Progress is retarded by retaining more than the optimum proportion of older sows, because the less intense culling of sows and the longer interval between generations is only partly offset by the more severe culling of gilts and the greater accuracy of sow culling.

Having sows farrow two litters a year results in more rapid genetic improvement in productivity, since it permits the accuracy of selection of boars and gilts to be improved by basing the dam's productivity on two litters instead of one. It also permits the more productive sows to be kept for additional litters, with a minimum increase in the average interval between generations.

It is important that the breeder have effective yardsticks of merit, regardless of the selection procedures and breeding system he uses.

Evaluation of the fitted animal in the show ring has long been considered an important part of livestock improvement in the United States. In recent years it has become increasingly apparent that this procedure has many shortcomings as a tool in selection of improved breeding stock. For obvious reasons, only a small portion of the animals raised each generation can be prepared for evaluation. The condition of the animals at the time of the show is usually highly artificial, and quite often is very different from the condition that is desired in practice. Undue attention is often given to so-called fine points of little or no economic importance. Some traits, such as milk yield and efficiency of feed utilization, cannot be accurately evaluated by visual inspection. The practice of excessive fitting has been carried over to the conditioning of breeding stock for sale, and is found to a marked degree even in bulls and rams that are to be sold for use on western ranges. Thus the breeder spends an undue amount for feed to put excessive fat on the animals, for which the buyer must pay, but for which he has no use. Also, the excessive fat may obscure defects in conformation, a point that is aptly stated in the common phrase, "Fat is a pretty color.”

Recognizing the need for improved yardsticks, many workers have turned their attention to the development of measures of the economically important characters. Some characters, like litter size in swine, may be observed directly. Others, like body size, rate of growth, milk yield, yield of grease wool, and length of staple, can be weighed or measured directly. Others, for example face covering, skin folds, and body conformation in sheep, require indirect methods of evaluation and the assignment of a score to represent the degree of development in each animal. Devices have been developed for measuring such characters as length of wool fibers, tenderness of meat (muscle), and diameter of wool fibers, density of wool fibers, and hardness of fat.

Much attention also has been given to the measuring of functional traits, such as efficiency of feed utilization in beef cattle and swine, physiological response of horses and mules to exercise, and performance of work by draft horses and mules and by light horses in carriage and under saddle. Many of the developments are still in the experimental stage, but active research is continuing at many institutions to test existing procedures, to develop new ones, and to simplify experimental procedures so they may be applied by breeders in evaluating and selecting their stock.