If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The article reviews ruminant ecology and evolution and shows insights they offer into livestock research. The first ruminants evolved about 50 million years ago and were small (<5 kg) forest-dwelling omnivores. Today there are almost 200 living ruminant species in 6 families. Wild ruminants number about 75 million, range from about 2 to more than 800 kg, and generally prefer at least some browse in their diets. Nine species have been domesticated within the last 10,000 yr. Their combined population currently numbers 3.6 billion. In contrast to wild ruminants, domestic species naturally prefer at least some grass in their diets, are of large body weight (BW; roughly from 35 to 800 kg), and, excepting reindeer, belong to one family (Bovidae). Wild ruminants thus have a comparatively rich ecological diversity and long evolutionary history. Studying them gives a broad perspective that can augment and challenge the status quo of ruminant research and production. Allometric equations, often used in ecology, relate BW to physiological measurements from several species (typically both wild and domestic). They are chiefly used to predict or explain values of physiological parameters from BW alone. Results of one such equation suggest that artificial selection has increased peak milk energy yield by 250% over its natural level. Voluntary feed intake is proportional to BW0.9 across wild and domestic ruminant species. This proportionality suggests that physical and metabolic factors regulate intake simultaneously, not mutually exclusively as often presumed. Studying the omasum in wild species suggests it functions primarily in particle separation and retention and only secondarily in absorption and other roles. Studies on the African Serengeti show that multiple species, when grazed together, feed such that they use grasslands more completely. They support the use of mixed-species grazing systems in production agriculture. When under metabolic stress, wild species will not rebreed, but rather will extend lactation (to nourish their current offspring). This bolsters the suggestion that lactation length be extended in dairy operations. Cooperation between animal scientists and ecologists could generate more valuable insight.
As an applied field, animal science borrows variously from classical physiology, endocrinology, biochemistry, genetics, and nutrition, among other disciplines. It draws less often from ecological and evolutionary research. For ruminant livestock research, several excellent publications exist on these 2 overlooked subjects, including classic works such as those by
presents a brief overview of ruminant ecology and evolution. As a brief overview, it necessarily lacks detail and synthesizes few explicit connections between ruminant ecology and evolution and applied livestock research. Some individual manuscripts (e.g.,
) have drawn more detailed and direct connections. Although valuable in their own right, these manuscripts do not include a thorough introduction to ruminant ecology and evolution.
The aim of this review is to 1) summarize key points of ruminant ecology and evolution, and then 2) show where these points offer insight into livestock research and production, drawing on our original ideas and some isolated ones presented previously. The focus on nutrition and physiology in this review reflects our own expertise, not a lack of importance of other fields (e.g., reproduction, behavior, genetics).
Ecology and Evolutionary History of Wild Ruminants
For the purpose of this review, a ruminant includes any artiodactyl (member of the mammalian order Artiodactyla) possessing a rumen, reticulum, omasum, or isthmus homologous to the omasum, and abomasum. Ruminants also possess certain skeletal features—such as loss of upper incisors, the presence of incisiform lower canines, and fusion of the cubiod and navicular bones in the tarsus—that are useful in fossil identification (e.g.,
The 6 extant (i.e., nonextinct) ruminant families include the Tragulidae, Moschidae, Bovidae, Giraffidae, Antilocapridae, and Cervidae. Table 1 provides a description of these families, including the number of species and genera (from
The Tragulidae (chevrotains; 4 species) are small, reclusive, forest-dwelling, deer-like ruminants (Figure 1). They are the most primitive of all living families and have changed little morphologically over evolutionary history; this has led them to being called “living fossils” (
). Although still considered ruminants, the Tragulidae are not included in the same infraorder (Pecora) as other ruminant families (Moschidae, Bovidae, Giraffidae, Antilocapridae, Cervidae) because of these ancestral features.
Figure 1A greater Malay chevrotain (Tragulus napu), a member of the family Tragulidae and one of the most primitive ruminants. Note the small size (approximately 3 kg), short limbs, and absence of horns, all of which are characteristic of early ruminants. Enlarged upper canines are absent because this specimen is a female. Photo courtesy of Ellen S. Dierenfeld (Novus International Inc., St. Charles, MO).
The Moschidae (musk deer; 5 species) are small, Asiatic forest deer whose males possess a musk gland anterior to the genitals. Like tragulids, the moschids are hornless (other families possess horns or antlers), and the males have large upper canines instead. The remaining families, the Bovidae (e.g., cattle, sheep, goats, antelope; 140 species), Giraffidae (giraffe and okapi), Cervidae (true deer; e.g., white-tailed deer, red deer elk, caribou, moose; 41 species), and Antilocapridae (pronghorn), include species familiar to most readers. Standard mammalogy textbooks (e.g.,
): the Hypertragulidae, Leptomerycidae, Gelocidae, Palaeomerycidae, and Dromomerycidae. The Hypertragulidae, Leptomerycidae, and Gelocidae were small, hornless ruminants that probably most closely resembled moschids or tragulids (
Figure 2A phylogeny of ruminant families. Families included are the Hypertragulidae, Tragulidae, Leptomerycidae, Gelocidae, Moschidae, Dromomerycidae, Palaeomerycidae, Antilocapridae, Giraffidae, Cervidae, and Bovidae—that is, those recognized by
). Their skull and dental morphology (low-crowned teeth, small incisors, long and narrow skulls) were optimal for consuming fruits, shoots, and insects (
). This evidence, in addition to the observed habitat and diet of living tragulid and moschid species (which are taken as rough analogs for these first groups), suggests that the first ruminants were small, reclusive, forest-dwelling omnivores (
). The first ruminants did not ruminate or have a functional rumen flora (viz., for fiber fermentation) until about 40 Ma, as indicated by dental morphology (
), most of which are Bovidae and Cervidae (Table 1). A conservative estimate places the world population of wild ruminants at 75.3 million, with 0.28 million tragulids, 0.28 million moschids, 44.6 million cervids, 29.1 million bovids, 0.15 million giraffids, and 0.88 million antilocaprids (Supplemental Table 1, http://www.journalofdairyscience.org/; Appendix). The majority of wild ruminants, in terms of species and population numbers, are thus bovids and cervids.
Following their distribution in the fossil record, living ruminants are natively found on all continents except Antarctica and Australia, although most species are found in Africa and Eurasia (Table 2, constructed from data in
). The Bovidae and Cervidae both enjoy an almost worldwide distribution, whereas the range of the remaining families is much more restricted (Table 2). Only the Cervidae are found in South America (Table 2).
Table 2Native distribution of species (% of total within family) across continents and habitat and climate types
Ruminants not only have a wide geographic distribution, but are also found across many climates and habitats. Although the classification system of habitats and climates used in this review (adopted from
) is admittedly crude, it still gives a general sense of this distribution. As a whole, ruminant species are evenly spread across open, ecotone, and forested habitats, but they prefer warm to other types of climates (Table 2). The distribution of the Bovidae and Cervidae is generally representative of this overall pattern, whereas other families individually inhabit a more restricted range of habitats and climates (Table 2).
), the median BW of modern ruminants is 45 kg. Body mass ranges greatly, from approximately 2 kg [Salt's dik-dik (Madoqua saltiana), royal antelope (Neotragus pygmaeus), lesser Malay mouse deer (Tragulus javanicus)] to 800 kg [American bison (Bison bison), wisent (Bison bonasus), gaur (Bos gaurus), Asian water buffalo (Bubalus bubalis), kouprey (Bos sauveli);
]. Although not shown in Table 3, some individuals from the largest species achieve BW ≥1,000 kg, with the maximum size of a male reticulated giraffe (Giraffa camelopardalis) reaching 1,400 kg (
). By family, the Giraffidae are the largest; Antilocapridae, Bovidae, Cervidae are intermediate; and Moschidae and Tragulidae are the smallest (Table 3). The Bovidae and Cervidae have species at or near these BW extremes, whereas the other families display a much more restricted range in BW (Table 3).
Table 3Body mass of wild ruminant species by family
Ruminant species display innate dietary preferences, and these differ greatly across species. A concise way of classifying these preferences is with the feeding class system (first proposed by
), which categorizes species as 1) browsers, which innately prefer browse such as fruits, shoots, and leaves (typically from shrubs, forbs, and trees); 2) grazers, which innately prefer grasses; or 3) intermediate feeders, which switch between browse and grass, usually depending on their seasonal availability. For most of their evolutionary history, ruminant species were predominately or exclusively browsers. Today, a plurality of ruminant species are still classified as browsers (Table 4), and only about a quarter are grazers. The Bovidae and Cervidae have species represented in all 3 feeding classes; the other families are exclusively browsers.
Table 4Dietary preferences of species (% of total species within family), according to their assignment as browsers (BR), intermediate feeders (IM), or grazers (GR)
The first livestock species to be domesticated (ruminant or nonruminant) was the goat, at approximately 8,000 BC in the Fertile Crescent of the Near East (
). The goat was initially domesticated to supply meat to burgeoning, congested human populations whose hunting had depleted large prey populations in the wild (
). Most of the other 8 domesticated ruminant species (sheep, European and Zebu cattle, water buffaloes, mithans, reindeer, yaks, Bali cattle) were brought under human control by 2,500 BC in either the Near East or southern Asia (
). Some of these species, such as the goat, were initially domesticated for meat, but reasons for domestication varied greatly, including for milk, draft, transportation, sacrifice, and barter (
) have determined that each domestic species is probably derived from several wild species; at least 12 species can claim ancestry to the 9 domesticated species. Of the multitude of available wild species, these 12 were chosen for domestication because they were gregarious, submissive to human captors, unexcitable, and easy to breed (
Points in the discussion below are summarized in Table 5. The total population size of domestic species is 3.57 billion, nearly 50-fold larger than that of wild ruminants. As might be anticipated, cattle, sheep, and goats constitute most (about 95%) of the domestic ruminant population. All but reindeer belong to the family Bovidae.
Table 5Characteristics of domestic species, including population size, mean BW, and feeding class
Data for goats, sheep, cattle, and water buffaloes from the Food and Agriculture Organization of the United Nations (2008a); for reindeer from Ulvevadet and Klokov (2004); for yaks from Wiener et al. (2003); and for Bali cattle from the Food and Agriculture Organization of the United Nations (2008b).
Data for sheep, cattle, and goats are from typical literature studies (those summarized in Tables 5 and 6 of Hummel et al., 2005); for water buffaloes from Popenoe (1981); for yaks from Wiener et al. (2003); and for all other species from van Wieren (1996), taking BW of wild ancestors.
Data for goats, cattle, and water buffaloes from Hofmann (1989); for sheep from Hofmann (1989) and Pfister and Malechek (1986); and for all other species from van Wieren (1996), taking feeding class of wild ancestors.
Goat (Capra hircus)
850
35
IM
Sheep (Ovis aries)
1,113
50
IM/GR
European and Zebu cattle (Bos taurus, Bos indicus)
Food and Agriculture Organization of the United Nations. 2008a. Domestic Animal Diversity Information System. Food Agric. Org. United Nations, Rome, Italy. http://lprdad.fao.org/ Accessed Sep. 9, 2008.
Ulvevadet B. Klokov K. Family-Based Reindeer Herding and Hunting Economies and the Status and Management of Wild Reindeer/Caribou Populations. Centre for Saami Studies,
Univ of Tromsø, Tromsø, Norway2004
Food and Agriculture Organization of the United Nations. 2008b. FAO Statistical Databases. Food Agric. Org. United Nations, Rome, Italy. http://faostat.fao.org/ Accessed Sep. 9, 2008.
) have argued that they are instead intermediate feeders.
Although BW varies greatly by sex and across breeds, the rough averages in Table 5 demonstrate that the BW of domestic ruminants are large in comparison with many wild ruminants. The smallest species (sheep, goats) are near the median BW of wild ruminants (45 kg) and many species (cattle, mithans, Bali cattle) approach the maximum observed in the wild (800 kg; Table 3).
Perspectives Relevant to Modern Production Systems
Wild ruminants, past and present, prove much more diverse (in terms of phylogeny, behavior, diet, and otherwise) than domestic ruminants. Further, the 50-million-year evolutionary history of the ruminant extends far before domestication. Studying ruminant ecology and evolution gives a broad perspective of what ruminants are and how they came to be—much broader than achieved through studying domestic species alone. This broad perspective can augment or even challenge the status quo of livestock research and management, which has been established by using a much narrower perspective. The following discussion presents some examples of how principles in ruminant ecology and evolution can offer insight into livestock research and production.
Predicting Values of Physiological Parameters from BW
Body mass has a clear influence on the value of many physiological parameters. For example, the greater feed intake of a cow relative to a goat, intuitively, is largely attributable to the greater BW of the cow. The exact, quantitative relationship between a physiological parameter and BW is often less obvious, however.
) quantitatively express these relationships with the formula
where y is the value of a physiological parameter, a is the allometric intercept (value of y at BW = 1), and b is the scaling parameter. Values of a and b are found empirically by regressing BW against y for several species (Figure 3). By using observations from multiple ruminant species (including wild ones), one is provided a widely applicable equation that gives a benchmark prediction for a physiological parameter from BW alone. Some uses of these predictions for livestock research include 1) serving as a first approximation for a physiological value for a species when one has not been measured directly and 2) explaining to what extent observed differences between livestock species are attributable to BW (i.e., act as a control for BW in comparisons).
Figure 3Graph of the allometric relationship between BW (kg) and wet mass of reticulorumen tissue (kg), illustrating the general principles of allometric equations. Tissue mass observations of individual ruminant species are shown with symbols (♦), with observations of sheep and cattle labeled. The solid line indicates the best-fit allometric equation, with the allometric intercept a and scaling parameter b defined graphically. Note that the plot is semilogarithmic. Data are for 18 wild and domestic species from 10 studies (listed in Table 6.1 of
) is widely known and applied, but many others are much less so. Examples of allometric equations are given in Table 6, including those for predicting anatomical, ingestive and digestive, energetic, reproductive, and other physiological parameters. For illustration, expected values of these parameters for 2 different BW (50 and 500 kg) are shown. Note that many equations have a very high R2 (>0.95) and thus can be expected to be precise; others have much lower R2 values (as low as 0.18) and should be applied more cautiously.
Table 6Some allometric equations useful for predicting physiological parameter values from BW
The phrase “data from” preceding a source, where present, indicates that we performed the allometric regression using ruminant data (averaged by species) originally reported by that source.
a and predicted values reported in this table are for the 58% NDF, 19% CP mature alfalfa hay of van Wieren (1996); in the original equation, a is adjusted by a fixed-effect term for diet, which allows intake to be predicted for other diet types, but this term was omitted here for simplicity.
2 Unit, logarithmic, and other conversions have been applied to equations in original sources to standardize their units and form.
3 The phrase “data from” preceding a source, where present, indicates that we performed the allometric regression using ruminant data (averaged by species) originally reported by that source.
4 a and predicted values reported in this table are for the 58% NDF, 19% CP mature alfalfa hay of
; in the original equation, a is adjusted by a fixed-effect term for diet, which allows intake to be predicted for other diet types, but this term was omitted here for simplicity.
5 Equation applies to artiodactyls.
6 NA = not available.
7 Equation applies to ungulates with single offspring.
The equations listed in Table 6 represent a broad survey of the many available in the literature. For more allometric equations, the reader should refer to the sources referenced in the table, as well as
The equations given in Table 6 have been derived using both wild and domestic ruminant species. Note that one can parameterize allometric equations using other approaches. Domestic species alone can be used because only 2 species observations are technically required to estimate the 2 parameters (3 if error is to be estimated). However, with so few species observations, an outlier from any one species can unduly affect parameter estimates. Wild species should thus be included to increase the number of observations and the robustness of parameter estimates. At the opposite end, nonruminant species (wild or domestic) can be included in addition to ruminants. In this case, phylogenetic differences across the wide range of species, if not controlled for statistically, may make the resultant equation very general, but also very imprecise and potentially biased. For example, adding marsupials to the allometric equation for metabolic rate would make the equation more widely applicable (viz., for marsupials) but would skew the regression because marsupials have characteristically low metabolic rates (
). A good balance between precision, generality, and robustness is thus found by including domestic and wild ruminant species and these alone.
One must be careful in using allometric equations when a physiological parameter is affected by artificial selection. In such cases, allometric equations often predict poorly—namely, for domestic breeds for which the parameter has been targeted by artificial selection. For example, the equation for peak milk energy yield in Table 6—formulated with wild species and meat-producing domestic breeds fed predominantly forage—predicts a yield of 7.76 Mcal/d for a 550-kg cow. By contrast, observed peak yields for Holstein cows, for which milk production has been intensively selected, are approximately 20 Mcal/d on a similar (predominantly forage) diet (
). Although allometric equations have little direct predictive ability in such cases, another valuable purpose emerges: to estimate the impact of artificial selection on a parameter. For peak milk energy yield, assuming that the higher-than-expected yield of the cow is primarily caused by artificial selection, we can infer that artificial selection has increased peak yield in the dairy cow by about 250% (20 Mcal/d observed vs. 7.76 Mcal/d expected).
Role of Physical and Metabolic Factors in Regulating Feed Intake
In addition to the 2 general uses of allometric equations explained above, some allometric equations can be applied to draw deeper, more complex inferences. For example, the allometric equation for voluntary feed intake (VFI) demonstrates that forage intake is regulated by physical and metabolic factors simultaneously.
Physiological regulation of feed intake is important to livestock production systems because feed intake affects animal performance and operation costs. Of the many proposed regulation mechanisms (
) are often suggested to predominate. Generally, these 2 regulation mechanisms (physical, metabolic) are considered to operate on a mutually exclusive basis, with intake of low-energy, bulky diets (usually forages) regulated only physically and high-energy, highly degradable diets regulated only metabolically (
). An allometric examination of VFI of wild and domestic ruminants considerably strengthens the case for simultaneous regulation. Because reticulorumen digesta contents and volume are nearly proportional to BW1 (i.e., proportionally do not change with BW; Table 6), one would also expect VFI to be proportional to BW1, if physical regulation only were operating. On other hand, because metabolic rate is proportional to BW0.75 (Table 6), one would also expect VFI to be proportional to BW0.75, if metabolic regulation acted alone. To test which proportionality is observed (if either), we derived an allometric equation for forage intake by using data on 19 wild and domestic species from 5 studies (listed in Table 6.1 of
). We performed a regression similar to that shown in Figure 3, except we included an additional, fixed-effect term that adjusts the allometric intercept (a) by diet. The equation we derived (Table 6) indicates that forage intake is proportional to BW0.875±0.032, with lower and upper 95% confidence limits of BW0.810 and BW0.941. This agrees with the finding that, for livestock species (cattle, sheep, goats), forage intake scales with BW0.9 (
The mean value of the scaling parameter (0.9) and its 95% confidence limits fall in between values expected if intake were regulated only physically (1) and metabolically (0.75). Intuitively, this suggests that physical and metabolic regulation operate simultaneously for forage diets, contrary to the classical suggestion that physical regulation alone should occur; the mechanistic modeling described in
confirmed this suggestion. Although a lack of controlled data disallows a similar examination with diets that are not all forage, the wide range in quality of the forage diets (predicted NEM ranged from 0.71 to 1.77 Mcal/kg of DM;
) suggests the results are not simply constrained to a narrow range of forages. Because this simultaneous regulation is apparent across wild and domestic species alike, it can be inferred that it is highly conserved evolutionarily and is deeply seated.
, who originally proposed that metabolic and physical regulation are mutually exclusive, also used allometry to support their arguments. They concluded that for high-producing dairy cows, VFI scaled with BW1 for low digestibility (<66.7 digestible DM) and with BW0.73 for high-digestibility (>66.7% digestible DM) diets, consistent with mutually exclusive regulation. However, for high-digestibility diets, scaling parameter values were approximately 0.73 for only 2 of their 5 regressions; all others were lower (≤0.62). The upper 95% confidence limits of these 2 favorable regressions (0.962 and 1.03) do not rule out VFI scaling with BW0.9 or possibly even BW1. This fact, and considering that VFI was only poorly related to BW (R2 = 0.074 or lower), indicates that this data set is poor for discriminating intake scaling patterns. Detailed results are not presented for low-digestibility diets, but the above discussion suggests one should remain skeptical of their conclusion that VFI scales with BW1 for these diets. Although intriguing for their time, the allometric analysis of
and conclusions based thereon must be rejected in favor of our more discriminating analysis.
Primary Function of the Omasum
Whereas functions of the rumen, reticulum, and abomasum are well delineated, the chief function of the omasum remains somewhat a mystery. It may help retain and separate particles because 1) large particles tend to become trapped between the omasal laminae, whereas small particles and liquids pass through quickly (
Ehrlein, H. J. 1980. Forestomach motility in ruminants. Publikationen zu Wissenschagtlichen Filmen, Sektion Medizin, Serie 5, Film No 9/C 1328. IWF, Goettingen, Germany.
). It may also be an absorptive organ; the omasum absorbs approximately 12.5, 50, 35, 25, 10, and 50% of water, VFA, ammonia, sodium, potassium, and carbon dioxide that enter (
) also occurs in the omasum. Finally, some claim the omasum reduces particle size (via purported grinding of digesta between laminae), although evidence for this function is at best circumstantial (
To establish the primary function of the omasum, we review the omasal form and function of several wild species and suggest how and why the omasum evolved. In tragulids, a true omasum is not present at all; where it should be found, only an isthmus exists instead (
). Considering this evidence and that the Tragulidae are otherwise primitive (Figure 2), this isthmus probably resembles a very early form of omasum, as concluded by other authors (see
). Because the isthmus lacks structures to retain digesta within it, its contribution to absorption, fiber fermentation, absorption, and particle size reduction must be minimal. Its poor structural development also precludes it from acting as a suction pump to regulate digesta flow. However, the isthmus likely helps retain particles because its small aperture should allow only fine particles to pass into the abomasum and subsequently through the rest of the tract (
In small, browsing Pecoran species, which are more advanced than the tragulids, the omasum forms a distinct compartment, but it still tends to be small and has few laminae (
concluded that its simple structure permits it to serve as little more than a “sieving screen” that prevents large particles from entering the abomasum. In large, grazing ruminants, the omasum is large and with many laminae (
) and presumably other more advanced functions (e.g., fiber fermentation).
We thus find a progression in omasal form and function from the tragulids to browsing and then grazing Pecoran ruminants. This progression suggests the omasum originally evolved as a simple isthmus that acted as a “floodgate” (
) to retain particles. Compartmentalization, well-developed laminae, and complex motor activity subsequently evolved to support absorption, fiber fermentation, and suction-based digesta flow control. These more derived functions indeed have some adaptive benefit, particularly for grazers, for which they may help process large amounts of refractory fiber (
). Nevertheless, considering that the particle retention function is pervasive across all species and is the impetus for the evolution of the omasum, this function would seem to be of primary importance.
Dietary Niche Separation and Mixed-Species Grazing
Certain elements of this argument have also been presented by
. Studies on the African Serengeti support the use of mixed-species grazing systems—that is, systems in which pastures are stocked with more than one livestock species simultaneously. In their seminal studies,
showed that when an occupied area of the Serengeti plains becomes overgrazed, African buffalo and zebra are the first to migrate into ungrazed regions of long, poor-quality grass. As they remove the top herbage layer (stems and leaves of mature grasses), they expose lower, high-quality layers (stems, leaves, and fruits of immature grass and browse), which wildebeest, topi, and Thompson's gazelle then graze as they move in succession. By occupying different dietary niches, these ruminants and zebras together not only successfully occupy the same habitat, but also use grasslands more efficiently and completely.
Two primary reasons explain why these species occupy different dietary niches and thus can exist sympatrically (i.e., in the same geographic area). First, these species exhibit innately different dietary selectivities (i.e., feeding classes) that would immediately suggest the observed dietary niches; the buffalo and zebra are strict grazers, the wildebeest and topi are more selective grazers, and the Thompson's gazelle is an intermediate feeder (
). Because VFI scales with BW0.9 whereas metabolic requirements scale only with BW0.75 (Table 6), nutrient intake increases relative to metabolic requirements with increasing BW. Consequently, large species can adapt to poor-quality material because they can consume relatively large amounts to meet their metabolic requirements, whereas smaller species are constrained to higher quality material because they can eat relatively little (
Geist, V. 1974. On the relationship of ecology and behavior in the evolution of ungulates: Theoretical considerations. Pages 235–246 in The Behavior of Ungulates and Its Relation to Management. V. Geist and F. R. Walther, ed. New Series No. 24. IUCN Publications, Morges, Switzerland.
) further reinforces innate selectivity differences to establish different dietary niches.
Significantly, major livestock species (goat, sheep, cattle) differ greatly in feeding class, BW, or both (Table 5). Probably as a combination of these factors, species choose diets that overlap incompletely: as goats choose more browse, cattle choose more grass, and sheep are intermediate (Figure 4). Diets should be even less similar than Figure 4 might suggest because 1) Figure 4 shows ranges that apply to single-species grazing, and cattle shift to poorer quality diets in their range when grazing with sheep (
); and 3) the rough expression of botanical composition shown in Figure 4 ignores other ways diets can differ (e.g., by plant species or part). Livestock species thus have ample opportunity to separate their dietary niches when grazed concurrently.
Figure 4Botanical composition (% of grass and browse) of diets chosen by goats (n = 13), sheep (n = 105), and cattle (n = 121) on pasture. Bars delineate mean ± SD. Data are from
Given this probable dietary niche separation, one might expect mixed-grazing systems to lead to more complete utilization of pasture and a higher combined level of animal productivity (because more pasture is transformed into animal tissue). In support,
). However, the enhanced production and efficiency of the systems compared with conventional ones—as expected from the archetypal mixed-grazing system on the Serengeti—challenge these barriers.
. The aim of many dairy cow operations is a 305-d lactation with a 12- to 14-mo calving interval. This requires cattle to be rebred by 100 d after parturition, soon after peak lactation, when they experience a negative energy balance (
and others, is to purposely extend lactation as long as possible and thus avoid early rebreeding. This approach, compared with current practice, indeed appears to be better supported by ecological observations. When in severe metabolic stress (poor body condition or nutritional plane) during lactation, many wild species [muskoxen (
, extending lactation is presumably a maternal strategy to maximize fitness (the most important biological drive of organisms); investing in current offspring by extending lactation is safer and, in the long term, more profitable than producing new offspring when necessary nutritional resources are inadequate.
These observations of wild species suggest that when under metabolic stress, the ruminant animal is evolutionarily entrained to continue lactation rather than rebreed. Attempting to rebreed the high-producing dairy cow soon (60 d) after parturition is a direct fight against this entrained response. If producing replacement heifers is not a major production goal, it might make more sense to rebreed less frequently and exploit the physiological capacity and drive for extended lactation shown in wild ruminants.
Although this argument for extended lactation is largely conceptual, growing experimental evidence suggests practicing extended lactation is viable and profitable. Lactation has been maintained naturally in goats for 4 yr (
). Milk production far into an extended lactation is less than around peak, but it is more than with production of a rebred animal during late pregnancy and its dry period (Figure 5). All predesigned experiments have demonstrated that, over the long term, extending lactation up to 16 mo (and sometimes longer) does not decrease daily production (
) were observational or theoretical, not experimental.] Although more research is clearly needed, these preliminary results suggest that the biological principles of extended lactation, as elucidated by wild ruminants, may be of great service to livestock production systems.
Figure 5Milk production of extended versus conventional lactations of multiparous Muriciano-Granadina goats milked once daily. Open (○) and closed (●) symbols represent production of goats managed for kidding intervals of 12 mo (K12) and 24 mo (K24), respectively. Arrows labeled a and b highlight where milk production of an extended lactation is greater (during pregnancy and the dry period of K12 goats) and less (during peak lactation of K12 goats) than conventional lactation. The first and second asterisks near the x-axis indicate times that K12 and all goats were rebred, respectively. Figure modified from
By offering a comparative vantage point, ruminant ecological and evolutionary research can offer valuable insight into livestock research. This research can reinforce and augment some conventional livestock practices, and if we allow, can challenge and help revise others. With further dialogue and cooperation between animal scientists and ecologists, the insights that ruminant ecology and evolution have to offer should grow in number and usefulness.
Acknowledgments
We acknowledge Benjamin Nielson (University of Missouri–Columbia) for laboratory support. Financial support is from the Missouri Agricultural Experiment Station (Columbia). Additional financial support for T. J. H. was provided by the C. W. Turner Graduate Scholarship (University of Missouri–Columbia).
Appendix.
Global population estimates of wild ruminants were compiled from several sources (
East, R. 1999. African Antelope Database 1998. International Union for Conservation of Nature/Species Survival Commission Antelope Specialist Group, Gland, Switzerland.
Ulvevadet B. Klokov K. Family-Based Reindeer Herding and Hunting Economies and the Status and Management of Wild Reindeer/Caribou Populations. Centre for Saami Studies,
Univ of Tromsø, Tromsø, Norway2004
), many of which are compilations themselves. We excluded from our estimates domesticated, feral, captive, and nonnatively introduced populations or species. We also excluded species for which estimates were judged very fragmentary (that included only a few isolated or subspecies populations) and were unlikely to approach anything of a global estimate.
In all, we obtained estimates for 150 species (78% of total). This includes 1 species from Antilocapridae (100% of family total), 116 from Bovidae (83% of total), 26 from Cervidae (63% of total), 2 from Giraffidae (100% of total), 4 from Moschidae (80% of total), and 1 from Tragulidae (25% of total). Poor or nonexistent census data account for missing species. In addition, note that estimates for many Asian species include numbers only in China, again because of poor census data.
The population sizes reported here are clear underestimates. In total, they are still more comprehensive and up-to-date than the last apparent global census (
), which estimated population numbers for only 11 species (excluding feral and currently unrecognized species) in 2 families (Bovidae, Cervidae), for a total of 27 million ruminants overall.
East, R. 1999. African Antelope Database 1998. International Union for Conservation of Nature/Species Survival Commission Antelope Specialist Group, Gland, Switzerland.
Ehrlein, H. J. 1980. Forestomach motility in ruminants. Publikationen zu Wissenschagtlichen Filmen, Sektion Medizin, Serie 5, Film No 9/C 1328. IWF, Goettingen, Germany.
Food and Agriculture Organization of the United Nations. 2008a. Domestic Animal Diversity Information System. Food Agric. Org. United Nations, Rome, Italy. http://lprdad.fao.org/ Accessed Sep. 9, 2008.
Food and Agriculture Organization of the United Nations. 2008b. FAO Statistical Databases. Food Agric. Org. United Nations, Rome, Italy. http://faostat.fao.org/ Accessed Sep. 9, 2008.
Geist, V. 1974. On the relationship of ecology and behavior in the evolution of ungulates: Theoretical considerations. Pages 235–246 in The Behavior of Ungulates and Its Relation to Management. V. Geist and F. R. Walther, ed. New Series No. 24. IUCN Publications, Morges, Switzerland.
Ulvevadet B. Klokov K. Family-Based Reindeer Herding and Hunting Economies and the Status and Management of Wild Reindeer/Caribou Populations. Centre for Saami Studies,
Univ of Tromsø, Tromsø, Norway2004
00105-0&id=gr1.jpg){kind=link}
00105-0&id=gr2.jpg){kind=link}
00105-0&id=gr3.jpg){kind=link}
00105-0&id=gr4.jpg){kind=link}
00105-0&id=gr5.jpg){kind=link}