Managing Animal Variation
Excerp submitted by Eric van Heugten, Swine Nutrition Specialist, North Carolina State University - Before the widespread adoption of all-in-all-out (AIAO) production systems, variation in growth was largely a “hidden” cost.
![]() Swine Nutrition Specialist North Carolina State University |
Introduction
Pigs were selected from pens when they reached market weight, and the fact that some pigs required 4 to 6 weeks longer than others [to reach that weight] went largely unnoticed, or at least ignored. Furthermore, in continuous-flow systems, downtime due to variable growth rates affects pen usage, while in AIAO systems, it affects room or barn usage. The economic impact is therefore much greater in AIAO systems. There is also the annoyance factor, because from a labor and management perspective, dealing with tail-end pigs in a room or barn is also greater than in pens. Thus, the cost of variation has always been with us, but it is now exaggerated and much more obvious.
Variation also costs producers money due to its impact on sort losses. The extent of these losses will depend on the nature of the settlement agreement with the packer. Because increased attention is being paid to variability in growth, the industry’s previous singular focus on “average” growth rate is now increasingly being replaced by “average” growth rate and the “range” in that growth.
The industry has developed an array of management techniques to deal with problems associated with variability. Perhaps the most obvious is cross-fostering, in which sows with smaller litters receive piglets from sows with larger litters to “even out” lactation performance and reduce stress on the sow. However, because variability in performance in all phases of pork production is so costly, there is increasing motivation to develop and implement additional strategies to minimize its economic impact.
Measures of variation
Statistically, variation can be defined in a variety of ways; the most common terms are standard deviation (SD) and coefficient of variation (CV), although range may also be useful. The standard deviation is a measure of dispersion. The greater the variation in weight of a group of pigs, the larger will be the standard deviation. In a normal distribution, one standard deviation about the mean will include 68.3 percent of the pigs in the total group. Two standard deviations will take in 95.45 percent of the pigs, and three standard deviations will include 99.73 percent.
The coefficient of variation is calculated by dividing the standard deviation by the mean, and then multiplying by 100. The standard deviation becomes larger as the mean increases. Therefore, in order to determine if relative variation is increasing or decreasing as pigs grow, the coefficient of variation is often used. The standard deviation increases as the pigs become heavier; the coefficient of variation decreases, indicating that “relative” variability among pigs under commercial conditions typically declines as they grow.
However, the use of the CV can be misleading. For example, when evaluating the impact of a particular management practice on variability at a given stage in the pig’s life, the standard deviation may be preferable. If the management practice improves overall average performance, then the variation expressed as CV may decrease, even if the standard deviation is unchanged, because the mean changes. We have seen a number of instances where conclusions regarding the reduction in variability were based on a reduced CV when, in fact, variation did not change because the average, not the standard deviation, was affected.
Addressing Variability
There are a variety of management techniques that producers can adopt to address the issue of variability. We believe they fall into two categories: “reducing” variability and “managing” variability. Suffice to say that if variability within a barn is already quite low, producers are more likely to achieve success by seeking ways to “manage” variability. On the other hand, if variability is high, then clearly there are deficiencies within the barn that are impairing performance and should be addressed. In this instance, it should be possible to “reduce” variability.
Reasonable targets for variability in the barn can be based on CV as described above. On the basis of our admittedly limited information on “normal” variation, we suggest the following thresholds for CV: 20 percent of weaning weights, 12 to 15 percent for nursery exit weights, and 8 to 12 percent for weight at first pull from the finishing barn. In other words, if the CV for body weight at weaning is around 20 percent or at nursery exit is 12 to 15 percent or at first pull is 8 to 12 percent, then it makes sense for barn managers to seek ways to “manage“ variability rather than “reduce” it, as we suspect it is already close to the lower practical limit. These targets may change in the future as we develop more information on CVs under commercial conditions.
Reducing variability
It must be very clear that reducing variability refers to increasing the uniformity of ALL pigs within a group. Removing small or large pigs from the group does not reduce variability; it merely separates it into subgroups of increased uniformity. Overall variability remains the same. However, if the smaller pigs can be made to grow faster, or the heavier pigs made to grow slower, uniformity will be enhanced. Certainly, there is absolutely no logic to slowing the growth of the faster-growing pigs, so the objective of the barn manager, in terms of reducing variability, is to enhance the performance of the slower-growing pigs.
Improving herd health is clearly an essential objective, if it can be achieved in a sustainable and economical manner. De Grau et al. (2001) and Dewey et al. (2001) conducted a survey on 9 commercial farms of varying size, health status, and management types. They reported that the presence of disease increased weight variation and that multisite production and AIAO management were associated with a lower CV. Removing or reducing the impact of pathogens on individual pigs will not only improve performance of the overall group, it will also create a “tighter” group of pigs because the tail-enders will be minimized. Improving herd health can be achieved, for example, by not mixing pigs from multiple sources, by isolating barns from other pig barns to the greatest extent possible, or by sourcing pigs from high-health sow barns. If CV within a barn is too high, herd health should be the first target for attention.
A related practice would be to minimize the impact of existing disease in a barn through careful attention to nutrition, ventilation, herd health management, and the removal of social and environmental stressors to the greatest extent possible. Access to feed and water is also important. Restricted access to the feeder or to the drinker will adversely affect some pigs in the group, resulting in increased variability. It would appear that restricted access must be carefully defined. For example, Smith et al. (2004) reported that adjusting nursery feeders too tightly reduced growth rate and increased body weight CV (Table 1).

Managing variability
There are many ways to manage variability; however, the key issues become capital and labor costs to implement them. It is possible, for example, to improve barn throughput by culling the bottom 5 or 10 percent of the pigs at weaning or at nursery closeout, but the cost of such sorting and the cost of obtaining alternative housing for the culled animals must be balanced against potential savings. With this balance in mind, the following list identifies a number of strategies that can be used to manage variability so that its impact on profits is minimized.
Resorting pigs at marketing is often employed as a means of improving barn space utilization. However, after regrouping, pig performance temporarily declines—in the order of 10 to 15 percent (Stookey and Gonyou, 1994)—so the practice may actually end up reducing barn utilization (Table 2).

Preplanned segregation—Preplanned segregation refers to the practice of splitting a group of pigs into subgroups on the basis of expected differences in future performance. For example, split-sexed housing is a common example of preplanned segregation because everyone knows that barrows will go to market, on average, 5 to 7 days before gilts (Cooper et al., 2001). Thus, pens, rooms, or barns housing barrows will turn over about one week faster than those housing gilts, such that facility usage is increased. This works only if, for example, rooms are reduced in size to accommodate housing sexes separately; if not, this practice only works well with very large sow herds, as smaller operations cannot fill a room or barn with one sex within one fill period. This practice also flies in the face of the current trend of housing animals in much larger groups than has previously been the case.
Another example of preplanned segregation is to remove smaller pigs at weaning or at nursery closeout. In any AIAO system, it is the slower-growing pigs that dictate the turnover rate of a room or barn. Removing the lighter-weight pigs will help to remove those pigs most likely to be the tail-enders and thus the pigs that slow barn throughput. The logic of this practice was tested to determine if the removal of the lighter pigs would, in fact, allow the remaining pigs to grow at their normal rate and thus achieve closeout sooner, or would the smaller pigs in the remaining group themselves slow down (Patience et al., 2004). We found no evidence that removing faster-growing (or slower-growing) pigs from a group altered the performance of the remaining animals.
Lighter pigs certainly contribute to slower barn throughput. Table 3 illustrates the performance of various subgroups within a total population of 2,110 pigs, sorted according to their weaning weights. The lightest 12 percent of the pigs weighed only 4.1 kilograms (kg) at weaning, compared to an overall population average of 5.9 kg, and weighed 26.7 kg at nursery exit, compared to an overall population average of 31.5 kg (Beaulieu, unpublished). Although these data end at nursery closeout, it is clear that lighter entry weights result in slower growth in the nursery and, subsequently, in the growout barn (Mahan, 1993).

Parity segregation—Parity segregation refers to the separate housing of gilts and their offspring, as compared to those of older parity animals. Depending on the system, the gilts will be mixed with older parity animals at the time of breeding or farrowing of their second litter. The report of Moore (2001) indicated that when housed separately, gilt offspring will grow faster as compared to conventional blended housing systems. Moore suggests that gilt offspring are less competent immunologically than offspring of older sows, and that mixing of parities places them at a disadvantage.
Increasing weaning age—It has been suggested that weaning at an older age will also reduce variability (Main et al., 2004). Without question, pigs weaned at 21 days of age had a lower CV at marketing than those weaned at 12 days of age; however, they were also heavier pigs, so the smaller CV is partly an artifact of the mathematics of calculating CV (Table 4). The standard deviation at growout offtest was unchanged, indicating that weaning older pigs does not reduce variability. It may well be worth weaning at any older age for other reasons, but reducing variation should not be one of them.

Increasing overall weight gain—As mentioned previously, it is the tail-end pigs that dictate the rate at which a barn or room can turn over. It is therefore quite logical that if all of the pigs in a group grow faster, the impact of the tail-enders will be lessened. Indeed, for every 50 g/d increase in average daily gain in the growout period, there is a 50 percent reduction in the number of tail-end pigs in a room. In this instance, a tail-end pig is defined as one that will not meet the minimum core marketing weight within the defined growout period in the barn. Of course, if growth rate increases, there is a great temptation to build fewer growout weeks into future construction to reduce costs, with the net result that the problem of tail-enders returns.
Listing the many strategies that can be undertaken to increase growth rates is a talk unto itself (Patience et al., 2001a). Suffice to say that, in our experience, the opportunity exists in many barns to increase growth rate in an economical fashion.
Weigh pigs at marketing—Weighing market pigs is probably as popular as power washing; yet, the economic benefits can be substantial if there is a substantial packer penalty for sort losses. Unfortunately, many growout facilities have not been designed for easy weighing of pigs, which compounds the problem. Autosort technologies are becoming increasingly popular, and although there are many questions regarding the details of their use, they may be an alternative to allow accurate selection of animals for market.
Practices that do not reduce or increase variability
Because the industry’s focus on variability is relatively recent, some incorrect conclusions have been drawn regarding practices that increase or decease variability. This is not surprising, as it is part of the normal learning process. Nonetheless, it is important to separate what we currently believe to be correct (we are under no illusion that as we learn more about the subject, some of our current “truisms” will be refined or even proven incorrect) from what we know for certain is incorrect.
The first misconception is that sorting pigs into pens of uniform weights at weaning or at growout entry will reduce variability. This has been proven wrong in numerous studies, so there is no need to waste time and effort in sorting pigs for this purpose (Gonyou et al., 1986). The only reason to sort pigs into pens is to achieve split-sex feeding or to house pigs according to a predetermined management need; for example, at weaning, smaller pigs are often sorted from larger pigs so they can be given greater quantities of the Phase 1 and Phase 2 starter diets.
Another misconception is that crowding increases variability. In truth, as long as access to feed and water is not limited, and if the degree of crowding is not extreme, crowding has not been shown to increase variability. It certainly reduces performance, but not variability.

Source: North Carolina State University Extension Swine Husbandry - December 2006