Life history and reproduction

Suckling as a defining feature

Patterns of reproduction are truly fundamental to mammal biology. This is at once apparent from the word mammal itself. In all species of the class Mammalia (monotremes, marsupials, and placentals), females suckle their offspring, and almost all of them have teats (mammae) to deliver the products gathered from the milk-generating glands. As defining features of the class Mammalia, mammary glands and milk production (lactation) are clearly central to mammalian evolution. Indeed, these features undoubtedly appeared at an earlier stage than the birth of live offspring (vivipary). Whereas marsupials and placentals give birth to live offspring after a period of development within the mother's body, monotremes (platypuses and echidnas) still lay large, yolk-rich eggs. Although the milk-generating glands of monotremes release their products through milk patches in the pouch rather than through a small number of teats, suckling of the offspring is clearly evident. Hence, suckling occurs in all extant mammals, and most species show a characteristic duration of this behavior (lactation period) as one of their reproductive hallmarks. Unfortunately for biologists concerned with the reconstruction of mammalian evolution, reproductive features are very rarely preserved in fossils. For this reason, the origin of mammals is defined for practical purposes by the emergence of a new jaw hinge between the dentary and the squamosal, replacing the original reptilian jaw hinge between the articular and the quadrate. Even in the dentition, however, there are features that reflect the mammalian pattern of reproduction. One defining dental feature of mammals is the presence of only two sets of teeth throughout life (diphyodonty), contrasting with the typical reptilian pattern of continuous, wave-like replacement of teeth (polyphyodonty). Young mammals usually have an initial set of deciduous teeth containing only incisors, canines, and premolars, which is replaced by a permanent set of teeth in which molars are also added. It is in itself revealing that the deciduous teeth of young mammals are referred to as "milk teeth", although the replacement of teeth may continue well after the end of the lactation period.

An intriguing question that arises is why lactation and suckling of offspring are consistently limited to female mammals. In principle, it should be possible for male mammals to produce milk as well and thus contribute directly to the survival of their offspring. This question is all the more appropriate because most male mammals (including the human male) have teats that serve no apparent function. As a rule, anatomical structures that have no function tend to disappear in the course of evolution, one striking example being the reduction and eventual loss of eyes in cave-living animals that live in the dark. Yet teats are so widespread among male mammals that it seems quite likely that male teats were present in the common ancestor of the marsupials and placentals. So why are they still present in most male mammals today, after some 150 million years of evolution? We are still awaiting a satisfactory answer to this enigma. All that can be said is that the initial structure that eventually gives rise to teats and (in females) to functional mammary glands appears very early in fetal development in both sexes, in the form of so-called milk lines, one on each side of the ventral surface of the body. Such early appearance in development is, in fact, a further indication of the basic importance of lactation and suckling in the evolution of mammals. But we still have no explanation for the fact that milk lines develop not only in the female fetus but also in the male.

The development of mammary glands is linked to another universal feature of mammals, namely the possession of hair and associated sweat glands, both developed in connection with the evolution of a relatively constant body temperature (homeothermy). It seems highly likely that mammary glands were derived from sweat glands through a secondary conversion that enabled them to produce a nutrient fluid instead of sweat.

Reproductive tract

Two key features in the initial development of the mammalian reproductive tract are crucial for understanding its evolution in both males and females. First, the system begins development as essentially separate left and right halves that are virtually mirror images of one another (bilateral symmetry). Second, there is a very close connection between the development of the reproductive tract and that of the kidneys and allied structures (renal system). According to species, these initial conditions are progressively reduced to varying extents in the course of development, but they provide important

clues for distinguishing between primitive and advanced features.

The basic features of the female reproductive tract are common to all mammals. On each side of the body, there is an ovary that discharges the egg(s) into an oviduct, which leads to a uterus that is in turn connected with the vagina. Like other land-living vertebrates, all mammals have internal fertilization, which requires insertion of the male's erectile penis through the external opening (vulva) into the vagina. One variable feature of the vaginal region of the female reproductive tract in mammals concerns the relationship with the outlet of the urinary system (urethra) on either side of the body. In most mammals, the ureter enters the female tract at some distance from the vulva and there is a combined urinary and reproductive passage (urogenital sinus). This is the case for most small mammals, such as numerous marsupials, many insectivores, tree shrews, many rodents and carnivores. Perhaps the most spectacular case is the female elephant, which has a urogenital sinus that is close to 2 ft (61 cm) in length. In contrast, there is no urogenital sinus and the ureter has a separate opening adjacent to the vulval orifice in primates (including humans) and certain other mammals. Because a close connection between the urinary and reproductive systems is known to be primitive, the loss of the urogenital sinus is undoubtedly a secondary development. It should be noted, incidentally, that in many mammals (e.g. monotremes, marsupials, and some placentals) the reproductive, urinary, and digestive tracts all open into a common structure known as the cloaca. Indeed, the word monotreme means "one-holed", referring to the fact that there is a single cloacal opening. In several groups of placental mammals, such as hoofed mammals and primates, the cloaca has been completely lost and there is a wide separation between the anus and the reproductive/urinary outlets.

The form of the uterus also shows substantial variation among mammals. In the widespread primitive condition, there are two separate uterine chambers (bicornuate uterus), reflecting the initial development of two completely separate reproductive tracts in the female. A bicornuate uterus is found in marsupials and most placentals, although there is variation in the degree of separation between the two uterine chambers. In contrast, some placental mammals have the two original uterine chambers completely fused to form a single midline structure (simplex uterus). This relatively unusual condition is found in simian primates (monkeys, apes, and humans), but not in prosimian primates (lemurs, lorises, and tarsiers), which have retained the primitive bicornuate condition. A single-chambered, simplex uterus is also found in some edentates (armadillos and sloths) and in a few bat species. Interestingly, several bats show conditions intermediate between

the typical bicornuate uterus and the advanced, single-chambered form, thus providing clues to the evolution of the simplex uterus. All placental mammals show some degree of midline fusion of the right and left female reproductive tracts in that there is a single vaginal passage (with or without a urogenital sinus). In marsupials, there are separate left and right vaginal passages and the penis is correspondingly bifid in males. This difference between marsupials and placental mammals may have arisen through a chance difference in development. In marsupials, the ureters pass between the paired vaginal tracts, thus preventing full midline fusion, whereas in placentals, the ureters are located lateral to the vagina and do not stand in the way of midline fusion.

In the male reproductive system, sperm are produced in the testis, stored in the epididymis and transported into the penis by the vas deferens. As a further consequence of the early close connection between the urinary and reproductive systems, the bladder and the vas deferens from each testis open into a common channel in the penis, which conveys both urine and seminal fluid to the outside world. An unusual phenomenon found in most marsupials and placental mammals (but not in monotremes) is descent of the testes into special scrotal sacs outside of the main abdominal cavity. Once again reflecting the close developmental connection between the urinary and reproductive systems, the testes initially develop close to the kidneys. In most mammals, they subsequently migrate into external scrotal sacs. In addition to monotremes, mammals that show no descent of the testes, or only partial migration within the abdominal cavity, notably include burrowing marsupials, insectivores and rodents, various aquatic mammals (certain marsupials, insectivores, and rodents, along with seals, sea-cows, hippopotamus, whales, and dolphins) and heavily built pachyderms (elephant and rhinoceros). It has long been suspected that descent of the testes is in some way connected with avoidance of the relatively constant, elevated core body temperature that characterizes mammals. However, recent evidence suggests that it is not the actual production of sperm (spermatogenesis) that requires a lower temperature but rather sperm storage. For instance, in some mammal species that lack descent of the testis as such, the tail of the epididymis migrates to a position close to the ventral abdominal wall. This adaptation, which ensures a lower temperature for sperm storage but not for sperm production, is found, for example, in hyraxes and elephant shrews.

A further special feature of both male and female reproductive organs in many mammals is the presence of a baculum. This is commonly present both in the penis of the male (os penis) and in the clitoris of the female (os clitoridis), although the baculum is typically significantly larger in the male. Various mammals have secondarily lost the baculum. This is, for example, true of certain higher primates, including humans. The function of the baculum is still unclear, although there is some evidence that the os penis may play a role in stimulating the female during copulation. However, rather like teats in males, no function has been proposed for the os clitoridis in females.

Reproductive processes

The primary reproductive process in female mammals is the production of eggs (ova) from follicles in the ovary. In a non-pregnant female mammal, production of eggs is typically a cyclical process, although there are varying degrees of seasonal restriction such that some female mammals do not show repeated cycles. Seasonality of reproduction in mammals is mainly governed by annual variation in rainfall and vegetation, and hence becomes increasingly common at high latitudes. In many mammals, seasonality of reproduction is indirectly triggered by annual variation in day length, but in some mammals, it is a direct response to rainfall or food availability.

The typical ovarian cycle of mammals begins when one or more follicles ripen to the point where the egg can be released (ovulation). Following ovulation, the residue of the follicle is converted into a corpus luteum (yellow body), which produces progesterone that maintains at least the early part of pregnancy. The basic stages of the ovarian cycle are common to all mammals, with a follicular phase preceding ovulation and a luteal phase afterwards. However, there is a fundamental difference between different mammal groups with respect to the occurrence of ovulation and changes in the ovary. In many mammals, ovulation occurs only if mating takes place (induced ovulation), and therefore formation of a corpus luteum also requires mating. Some species show a slightly different condition in which ovulation takes place without mating, but mating is necessary for formation of a corpus luteum (induced luteinization). In both cases, mating is required for a corpus luteum to form, such that without mating ovarian cycles are confined to follicular phases and are correspondingly short. By contrast, in other mammals both ovulation and formation of a corpus luteum occur regardless of whether mating takes place (spontaneous ovulation). In these species, cycles always include combined follicular and luteal phases and are correspondingly long. Induced ovulation and induced luteinization are found in mammals that breed relatively rapidly, as is the case with most insectivores, tree shrews, many rodents, and numerous carnivores. On the other hand, spontaneous ovulation is typical of mammals characterized by slow breeding, such as hoofed mammals, cetaceans (whales and dolphins), hystricomorph rodents, and primates.

Following ovulation in marsupials and placentals, the egg travels down the oviduct, where fertilization will take place if the female has been inseminated. The fertilized egg (zygote) begins to divide as it completes its journey down the oviduct. By the time the zygote reaches the uterus, it has transformed into a hollow ball of cells (blastocyst). In placentals, the blastocyst is ready to implant in the wall of the uterus as the first stage in the development of placentation that will nourish the developing embryo/fetus. (By definition, a developing embryo becomes a fetus when recognizable organs are formed.) In both marsupials and placentals, development of the embryo/fetus within the uterus involves four embryonic membranes that play different roles. The chorion is the outermost membrane and remains intact throughout development right up to birth. Hence, any nutrients supplied by the mother to the developing offspring must first of all pass through the chorion. In all placental mammals, the chorion is in intimate contact with the wall of the uterus in the placenta. A second embryonic membrane, the amnion, surrounds the developing embryo/fetus throughout pregnancy and its fluid content (amniotic fluid) provides a protective hydrostatic cushion. The

remaining two embryonic membranes, the yolk sac (vitelline sac) and the allantois, play a crucial role in transfer of nutrients from the mother and in transfer of waste products in the opposite direction across the placenta, to be disposed of by the mother. In the enclosed egg of reptiles, which has been retained by monotremes, the vitelline sac contains a nutrient-rich yolk that is absorbed by blood vessels running over the surface of the sac, while the allantois stores waste products deposited by similar superficial waste products. When a reptile or montreme emerges from the egg, the waste-filled allantois is shed. In the development of the embryo and fetus from the yolk-poor egg in marsupials and placentals, nutrients must be provided directly by the mother and waste products must be removed in some way. In a fascinating reversal of function, the superficial blood vessels of the yolk sac in marsupials and placentals absorb maternal nutrients arriving from outside rather than absorbing yolk from inside. In many cases, the superficial blood vessels of the allantois also absorb nutrients coming from the mother as a substitute for the original function of depositing waste products inside the sac. In accordance with the original functional adaptations, in marsupials and placentals the blood vessels of the yolk sac typically develop their exchange role first (chorio-vitelline placenta), while blood vessels of the allantois do so secondarily (chorio-allantoic placenta).

Because monotremes still lay eggs, they are commonly labeled Prototheria to distinguish them from the Theria (marsupials and placentals), which all have live births. Although the fertilized egg is retained within the mother's body for the initial phase of development in marsupials, the impression is often given that there is no placentation in marsupials. It is, indeed, true that in all marsupials a shell membrane is present over the chorion at least for the major part of pregnancy. Widespread use of the name "placental mammal" has unfortunately tended to reinforce the false impression that placentation is lacking in all marsupials. In fact, some form of placentation is developed in certain marsupials and a few of them, such as the bandicoot (Perameles), even develop a relatively advanced chorio-allantoic form of placentation. For this reason, many mammalogists prefer the terms metatherian for marsupials and eutherian for placentals, derived from the formal names Metatheria and Eutheria. The fact remains, however, that proper formation of a placenta is characteristic of all eutherians, whereas it has secondarily been developed only in some marsupials, so continued use of the easily under-standable term "placental mammal" is surely acceptable.

In placental mammals, there is considerable variation in the form of the definitive chorio-allantoic placenta, although as a general rule each order, or at least suborder, of mammals tends to have a particular kind of placentation. Following Grosser (1909), a basic classification of types of placentation into three major categories, reflecting different degrees of invasiveness, can be made with respect to the relationship between the chorion and the inner wall of the uterus. In the least invasive type of placentation, the placenta is diffuse and the chorion is simply apposed to the inner epithelial lining of the uterus. It is labeled epitheliochorial placentation. In the other two kinds of placentation, invasion of the uterine wall occurs to some degree and the placenta is accordingly relatively localized (discoid). When moderate invasion occurs, the uterine wall is broken down in the region of the placenta and the chorion comes into contact with the walls of maternal blood vessels (endothelium). This type of placentation is called endotheliochorial. In the most invasive form of placentation, the walls of

the maternal blood vessels are themselves broken down in the region of the placenta, such that the chorion is directly bathed by maternal blood (haemochorial placentation). Epitheliochorial placentation is found in a few insectivores and it is uniformly characteristic of hoofed mammals, whales and dolphins, hyraxes, and strepsirrhine primates (lemurs and lorises). Endotheliochorial placentation is found in some insectivores, tree shrews, carnivores, sloths, anteaters, armadillos, elephants, and sea cows. Haemochorial placentation is found in many insectivores, rodents, bats, and haplorhine primates (tarsiers, monkeys, apes, and humans).

The evolutionary history of the three basic types of placentation is still subject to debate. It is often stated that the least invasive, epitheliochorial kind of placentation is the most primitive. This seems to be only logical, as the initial development of placentation must surely involve simple superficial contact between the chorion and the inner lining of the uterus. Because it is regarded as primitive, the epitheliochorial placenta is also often believed to be inefficient, notably with respect to development of the brain. By contrast, the highly invasive, haemochorial type of placentation is commonly thought to be very advanced and efficient. Human beings have the largest brain size (relative to body size) found among mammals and they also have highly invasive haemochorial placentation, so this is often seen as proof of the advanced nature of that very invasive type of placentation. However, many mammals with endotheliochorial or haemochorial placentation have relatively small brains, while dolphins, which have noninvasive epitheliochorial placentation, come a close second to humans with respect to relative brain size. In fact, there is much to be said for the alternative interpretation that ancestral placental mammals already had a moderately invasive type of placenta, following a long previous history of development. According to this view, noninvasive epitheliochorial and highly invasive haemochorial types of placentation represent divergent specializations away from a moderately invasive ancestral condition. It is noteworthy that during pregnancy, mammals with epitheliochorial placentation show great proliferation of uterine glands in the wall of the uterus. These uterine glands produce a nutrient secretion (so-called uterine milk) that is absorbed by special structures (chrionic vesicles) on the surface of the chorion. The selective advantages of the different basic types of placentation have yet to be identified. However, it is clear that the degree of invasiveness of the placentation has little to do with development of large-bodied offspring or of offspring with relatively large brains. Instead, it seems likely that the degree of invasiveness of the placenta reflects a trade-off between the advantages of an intimate placental connection between the mother and her developing offspring and the disadvantages of potential immunological conflict between the mother and her embryo/fetus.

The process of spermatogenesis typically takes place throughout the life span of male mammals, although it may be subject to periodic interruption in those species with a seasonal pattern of breeding. Spermatogenesis occurs as a wave-like process along the seminiferous tubules and completion of

sperm development in any one region takes between several days and a number of weeks. After transfer to the epididymis, the sperm are then stored until ejaculation takes place.

Gestation and neonate type

With only a few exceptions, each mammal species has a characteristic gestation period that shows remarkably little variation. In comparisons between species, gestation periods tend to increase as body size increases. However, effective comparisons of gestation periods among mammals must take into account a fundamental distinction in the state of the neonate at birth. As a general rule, it is possible to distinguish fairly clearly between mammals that give birth to several poorly developed (altricial) offspring and those that give birth to a few (usually just one) well-developed (precocial) offspring. Altricial offspring are largely hairless at birth and their eyes and ears are sealed with membranes. They are relatively helpless at birth and are typically deposited in a nest. By contrast, precocial offspring, which are usually born with a well-developed coat of hair and with their eyes and ears open, are typically able to move around quite actively at birth and are rarely kept in a nest. Other things being equal, it is obvious that the gestation period should be relatively longer for precocial offspring than for altricial offspring. In principle, it might be expected there would be a smooth continuum between altricial and precocial offspring. In practice, however, there is a fairly sharp division between them. When the relationship between gestation period and body size is examined for altricial and precocial mammals separately, it is found that there is a wide gap between them. At any given maternal body size, the gestation period for precocial offspring is about three times as long as that for altricial offspring. Furthermore, each main mammal group (order or suborder) is distinguished by the typical condition of the neonate and the relative length of the gestation period. Most insectivores, tree shrews, carnivores, and many rodents (myomorphs and sciuromorphs) give birth to altricial offspring after a relatively short gestation period, whereas hoofed mammals, hyraxes, elephants, cetaceans, pinnipeds, primates, and hystricomorph rodents give birth to precocial offspring after a relatively long gestation period. This is one of the few reproductive characters for which there is supporting evidence. Pregnant fossil horses from the Eocene and Miocene have consistently been found to have only one fetus, while an Eocene fossil bat has been found with twin fetuses. This shows that the small litter size of horses and bats, at least, have characterized those groups for at least 45 million years.

An inverse relationship between the average number of offspring produced at birth (litter size) and the gestation period is only to be expected. For a given uterus volume, there is clearly a trade-off between the number of developing offspring and the extent to which they can develop prior to birth. One corollary of this is that, for any given adult body size, altricial offspring must grow more after birth than precocial offspring.

Postnatal development

Across mammal species generally, there is a fairly consistent relationship between the average litter size and the typical number of teats possessed by the mother. As a rule, it can be said that there is one pair of teats for each offspring in the typical litter. However, suckling of the offspring is just one aspect of parental care in mammals. Maternal care, which can include nest building, grooming of the offspring, and infant carriage, is found in all mammals. Paternal care is relatively rare and is usually restricted to grooming and/or carriage. Predictably, paternal care in mammals is usually restricted to monogamous species in which there is a relatively high level of certainty of paternity.

Once the effects of body size have been taken into account, it emerges that the pattern of maternal care for any mammal species is quite closely reflected in milk composition. Three principal components of mammalian milk are carbohydrates, fats, and proteins. As a crude approximation, it can be said that the carbohydrate content of milk reflects immediate energy needs of the offspring, while the fat content indicates energy needs over a longer term. The protein content of milk provides a fairly good indication of requirements for growth. Milk composition also provides an indication of maternal behavior. Here, a major distinction can be drawn between mothers that feed on schedule and those that feed on demand. For offspring that are fed on schedule, it is the mother that determines the suckling frequency. Commonly, this applies to offspring that are left in a nest. These tend to be fast growing but relatively inactive altricial offspring that must maintain their body temperature unaided in the absence of the mother. As a result, the milk of such species tends to be high in protein and fat but relatively low in sugar. The most extreme example of suckling on schedule known for mammals is found in certain tree shrews species that keep their offspring in a separate nest and suckle them only once every 48 hours. The milk of these tree shrews is extremely concentrated, containing more than 20% fat and 10% protein. By contrast, in mammals that feed on demand, it is the offspring (usually a singleton) that determines suckling frequency. Usually, this requires close proximity between the offspring and its mother, so that it can signal its intention to suckle. Suckling on demand is, for example, generally typical of primates, most of which show parental carriage of the infant. Because infants that suckle on demand are usually slow growing but quite active precocial singeletons, the milk tends to be low in protein and fat but relatively high in sugar. This is the case with human milk, which is evidently naturally adapted for suckling on demand.

Eventually, provision of milk by the mother comes to an end and the offspring are weaned. In altricial mammal species, there is usually a fairly constant lactation period, and weaning tends to occur within a few weeks after birth. In precocial mammals, the lactation period can be quite variable and it may last months or even years. In fact, there is some indication that for certain hoofed mammals and primates there is a feedback relationship between the frequency of suckling and the mother's resumption of fertility, driven by the level of maternal nutrition. If food availability is low, the mother produces more dilute milk, which results in an increased suckling frequency. A higher frequency of suckling can suppress maternal fertility and also lead to an extension of the lactation period.

After weaning, the developing offspring must forage for food independently in order to meet its nutrient requirements for growth and maintenance. Eventually, it will attain sexual maturity and enter the breeding population. Here, too, altricial mammal species tend to have fairly standard ages for the attainment of sexual maturity, whereas precocial mammals can show more flexibility, according to prevailing environmental conditions. Despite such variability, the age of sexual maturity is an important milestone for comparisons among mammal species. It should be noted, incidentally, that sexual maturity may or may not coincide with the attainment of the adult condition in other respects, for example in body size and/or skeletal and dental maturity. In some cases, individuals may continue to grow for some time after achieving sexual maturity. It is noteworthy, however, that mammals (like birds) differ from reptiles, amphibians, and fish in showing a target body size. In each species, individuals tend to cease growing at a fairly standard size.

Life histories

All of the basic parameters that can be identified in the life cycle of any given mammal species, such as gestation period, litter size, lactation period, time taken to reach sexual maturity, and life span (longevity), contribute to its overall life history. Numerous lines of evidence suggest that these individual components of the life history of a species together constitute an adaptive complex that has been shaped by natural selection. For instance, one fundamental finding is that species that are subject to relatively heavy mortality under natural conditions tend to breed earlier. Because early breeding typically translates into a higher reproductive turnover, it can be concluded that heavy mortality promotes rapid breeding. In such comparisons between species, it has become commonplace to refer to "reproductive strategies." However, this is an anthropomorphic term and it is perhaps better to use a more neutral term like "life-history pattern."

In examining life-history patterns across species, it is essential to take account of the scaling influence of body size. It is only to be expected that large-bodied species will breed more slowly than small-bodied species, as it is generally likely that the time taken to grow to maturity will increase with increasing body size. However, it is also possible for mammals to show divergent life-history patterns even when body size is the same. David Western has aptly referred to these two kinds of difference as "first-order strategies" and "second-order strategies." For any given animal population, increase in population size is typically geometric until the population reaches a particular level (carrying capacity) determined by limiting factors (e.g. food supply, predation) in its environment. During the geometric phase of population growth, population increase depends on the intrinsic rate of increase (r) that is permitted by the reproductive parameters of the species. Because it is very difficult and time-consuming to obtain field data on the intrinsic rate of increase for natural populations, especially for large-bodied species, comparisons are often based on the maximum possible intrinsic rate of increase (r_max) that is permitted by standard reproductive values. A number of basic parameters such as age at sexual maturity, gestation period, litter size, interbirth interval, and longevity are used to calculate the r_max value for each species. As expected, for mammals r_max generally declines with increasing body size. In addition, there are distinctions between groups in the value found at any given body size. Primates, for example, generally have markedly lower values of r_max than other mammals. The two parameters that have the greatest influence in the calculation of r_max values are litter size and age at sexual maturity. Hence, the distinction between altricial and precocial mammals, in which litter size is a crucial feature, is clearly connected to a divergence in life-history patterns.

Various attempts have been made to develop general theories to explain the evolution of life-history patterns in mammals and other organisms. One basic problem that is encountered is that differences found between species do not fit well with the patterns that are observed within species, For instance, within a mammal species a particularly long gestation period is typically associated with a decrease in the duration of postnatal growth. In comparisons between species, however, it is found that species with long gestation periods tend to have extended periods of postnatal growth as well. It is therefore unclear how natural selection can shape life-history patterns. One model that has been suggested is "r-and K-selection." In this, it is proposed that species living in unpredictable environments with occasional catastrophic mortality will be subject to selection to increase r_max (r-selection). By contrast, species that live in predictable environments with moderate mortality will be subject to selection to increase competitiveness at carrying capacity (K-selection). An alternative model is that of "bet-hedging," in which it is suggested that high mortality among juveniles will favor slow breeding whereas high mortality among adults will favor rapid breeding. In fact, neither of these models fits all of the facts, so a convincing overall model remains elusive.

One point that deserves special mention is a potential link between life span (longevity) and relative brain size. Various authors have reported that mammals with relatively large

brains tend to have particularly long life spans. Although this proposed link is controversial, especially because of claims that it may be based on a secondary correlation, there is certainly enough evidence to indicate that some kind of connection exists. Hence, slow-breeding, long-lived mammals may also have relatively large brains as part of their overall life-history patterns.

Mating systems

Mammals exhibit a wide variety of mating systems, which can be basically divided into promiscuous, monogamous, polygynous, and multi-male. In mammals that live in gregarious social groups, the mating system is commonly (but not always) reflected by the composition of those groups, whereas in dispersed mammals patterns of mating must be determined from observations of interactions between separately ranging, "solitary" individuals. In all cases, however, it must be remembered that social systems and mating systems do not necessarily coincide. Even with mammals that are seemingly monogamous, genetic tests of paternity are quite likely to produce surprises just as they have already done for several bird species.

Monogamy, which is the predominant pattern of social organization and widely assumed to be the dominant mating system among birds, is relatively rare among mammals. It is somewhat more common in carnivores and primates than in other mammals, but even in those groups it is found in only a minority of species. It seems likely that promiscuous mating was present in ancestral mammals in association with their likely nocturnal, dispersed habits, as it is in various relatively primitive nocturnal mammals today that lack any obvious social networks (e.g. many marsupials, insectivores, carnivores, and rodents). Another common mating pattern among mammals is polygyny, in which a single male has exclusive or almost exclusive mating access to a number of females. Polygyny is commonly found, for example, in hoofed mammals, pinnipeds, and elephants. By contrast, it is relatively rare to find multi-male systems in which several adult males are present in a social network or group competing for mating access to females, although it is widespread among higher primates. A key issue here is the potential occurrence of competition between sperm from different males. In mating systems in which a single male has clear priority of access to one or more females (monogamous and polygynous systems), the probability of sperm competition is presumably low, whereas in promiscuous and multi-male systems there is likely to be a high incidence of sperm competition. This expectation has been confirmed by studies of the relative size of the testes in mammals. Species with promiscuous or multi-male mating systems generally have significantly larger testes, relative to body size, than species with monogamous or polygynous systems. This indicates that males show increased levels of sperm production in cases where sperm competition is relatively intense.

Sexual dimorphism

Male and female mammals obviously differ in various features that are directly linked to reproduction, as is the case with the sex organs of both sexes and the mammary glands of females (primary sexual characteristics). However, males and females can also differ in a variety of features that are not directly associated with reproduction (secondary sexual characteristics). Such secondary differences between males and females, like the human facial beard, are collectively labeled sexual dimorphism. The simplest form of sexual dimorphism distinguishing male and female mammals involves overall adult body size. As a general rule in mammals, males tend to be bigger than females, but there are some cases in which females are bigger than males (reverse sexual dimorphism). Differences in body size between the sexes are often relatively mild, as is the case in humans, where adult males are about 20% heavier than females; but there are also some striking contrasts. In the most extreme case of dimorphism in body size found among mammals, namely in the elephant seal (Mirounga), adult males are about four times heavier than females (8,000 lb [3,629 kg] compared to 2,000 lb [907 kg]). Sexual dimorphism in mammals can also affect other features, notably involving differences in external appearance (e.g. coat coloration), the size of the canine teeth and special appendages such as the antlers of deer (Cervidae). Overall, there seems to be a general tendency for the degree of sexual dimorphism in size of the body or its appendages to increase with increasing body size (Rensch's Rule), although the validity of this generalization has been questioned. An additional generalization that can be made is that sexual dimorphism in body size, canine size, and the size of such appendages as horns is generally lacking from species with a monogamous pattern of social organization. However, this does not apply to sexual dimorphism in coat coloration, as is shown by striking differences in external appearance between males and females in certain species of monogamous gibbons and lemurs.

The baseline expectation for mammals is that males and females will be similar in size and other features unless some special selective factor intervenes. However, there is no real reason why males and females should be similar in the size of the body and its external appearance or appendages. Given the major inequality in contribution to reproduction that characterizes all mammals, because gestation and lactation are exclusive to females, the baseline expectation should surely be that male and female strategies are quite likely to diverge. We should be more surprised by the numerous cases in which males and females are very similar in size and appearance than we are by sexual dimorphism. The standard explanation for sexual dimorphism in mammals is that selection acts on the male to increase the size of the body or its appendages because of competition for mating access to females (sexual selection). In elephant seals, for example, the big bull males fight one another to establish mating territories and maintain harems that may contain three dozen females. The large body size of males and their large canine teeth are therefore reasonably interpreted as features that increase male success in competition for females. A similar explanation is provided for the development of large antlers in male deer. Among primates, this interpretation is also applied to various species that show conspicuous sexual dimorphism. A prime example is the mandrill (Mandrillus sphinx), in which males are more than twice as heavy as females, vividly colored and equipped with very prominent canine teeth.

There has been a general tendency to overlook the potential part played by selection on females in the evolution of sexual dimorphism in mammals. In the first place, it is obvious that a single unitary explanation for sexual dimorphism, such as improved competitive ability of males, is inadequate because the different kinds of sexual dimorphism (e.g. body size, canine size, and coloration) can vary independently to a large degree. In primates, for example, it is possible to find marked dimorphism in coat coloration and mild variation in canine size without any matching difference in body size. Furthermore, most lemur species lack any kind of sexual dimorphism despite sometimes fierce competition among males for mating access to females. When males and females of a species differ in body size, it is commonly assumed that selection has operated to increase male body size, but it should be remembered that selection might also act to reduce (or increase) female body size. Because many reproductive features are scaled to body size (e.g. neonate size and age at sexual reproduction), one effect of reduction of female body size will be to decrease nutritional requirements for reproduction and increase the rate of reproductive turnover. In principle, sexual dimorphism between males and females in adult body size could be achieved by an increased rate of growth in males, by an extension of the growth period in males, or by some combination of these two possibilities. In practice, sexual dimorphism in mammalian body size is always associated with at least some delay in the attainment of sexual maturity in males relative to females, so there are consequences for reproductive dynamics in every case. Hence, it seems likely that sexual dimorphism reflects the effects of diverging selection pressures operating on both sexes.

Resources

Books

Austin, Colin R., and Roger V. Short. Reproduction in Mammals. Book VI: The Evolution of Reproduction. Cambridge: Cambridge University Press, 1976.

Austin, Colin R., and Roger V. Short. Reproduction in Mammals. Book I: Germ Cells and Fertilization. Cambridge: Cambridge University Press, 1982.

Bronson, Frank H. Mammalian Reproductive Biology. Chicago: University of Chicago Press, 1989.

Hayssen, V., and A. Van Tienhoven. Asdell's Patterns of Mammalian Reproduction: A Compendium of Species-Specific Data. Ithaca, New York: Comstock Publishing Associates, Cornell University Press, 1993.

Johnson, Martin H., and Barry J. Everitt. Essential Reproduction. Oxford: Blackwell Scientific Publications, 1980.

Martin, Robert D., Lesley A. Willner, and Andrea Dettling. "The evolution of sexual size dimorphism in primates." In The Differences between the Sexes, edited by Roger V. Short and Evan Balaban. Cambridge: Cambridge University Press, 1994, pp. 159–200.

Mossman, Harland W. Vertebrate Fetal Membranes. New Brunswick, New Jersey: Macmillan/Rutgers University Press, 1987.

Perry, John S. The Ovarian Cycle of Mammals. Edinburgh: Oliver & Boyd, 1971.

Stearns, Steven C. The Evolution of Life Histories. Oxford: Oxford University Press, 1992.

Steven, D. H. Comparative Placentation: Essays in Structure and Function. London: Academic Press, 1975.

Periodicals

Ben Shaul, D. M. "The composition of the milk of wild animals." International Zoology Yearbook 4 (1962): 333–342.

Charnov, E. L. "Evolution of life history variation among female mammals." Proceedings of the National Academy of Sciences of the United States of America–Biological Sciences 88 (1991): 1134–1137.

Cowles, R. B. "The evolutionary significance of the scrotum." Evolution 12 (1958): 417–418.

Daly, M. "Why don't male mammals lactate?" Journal of Theoretical Biology 78 (1979): 325–345.

Derrickson, E. M. "Comparative reproductive strategies of altricial and precocial eutherian mammals." Functional Ecology 6 (1992): 57–65.

Fenchel, T. "Intrinsic rate of natural increase: The relationship with body size." Oecologia 14 (1974): 317–326.

Harvey, Paul H., and Zammuto, R. M. "Patterns of mortality and age at first reproduction in natural populations of mammals." Nature (London) 315 (1985): 319–320.

Kenagy, G. J., and S. C. Trombulak. "Size and function of mammalian testes in relation to body size." Journal of Mammalogy 67 (1986): 1–22.

Kleiman, Devra G. "Monogamy in mammals." Quarterly Review of Biology 52 (1977): 39–69.

Long, C. A. "Two hypotheses on the origin of lactation." American Naturalist 106 (1972): 141–144.

Martin, Robert D. "Scaling effects and adaptive strategies in mammalian lactation." Symposium of of the Zoological Society of London 51 (1984): 87–117.

Martin, Robert D., and Ann M. MacLarnon. "Gestation period, neonatal size and maternal investment in placental mammals." Nature (London) 313 (1985): 220–223.

Møller, Anders P. "Ejaculate quality, testes size and sperm production in mammals." Functional Ecology 3 (1989): 91–96.

Oftedal, O. T. "Milk composition, milk yield, and energy output at peak lactation: a comparative review." Symposium of the Zoological Society of London 51 (1989): 33–85.

Pianka, E. R. "On r-and K-selection." American Naturalist 104 (1970): 592–597.

Pond, Caroline M. "The signficance of lactation in the evolution of mammals." Evolution 31 (1977): 177–199.

Ralls, K. "Mammals in which females are larger than males." Quarterly Review of Biology 51 (1976): 245–276.

Read, A. F., and Paul H. Harvey. "Life history differences among the eutherian radiations." Journal of Zoology, London 219 (1989): 329–353.

Rose, R. W., Claire M. Nevison, and Alan F. Dixson. "Testes weight, body weight and mating systems in marsupials and monotremes." Journal of Zoology, London 243 (1997): 523–531.

Sacher, George A., and E. Staffeldt. "Relation of gestation time to brain weight for placental mammals." American Naturalist 108 (1974): 593–615.

Schaffer, William M. "Optimal reproductive effort in fluctuating environments." American Naturalist 108 (1974): 783–798.

Western, David. "Size, life history and ecology in mammals." African Journal of Ecology 17 (1979): 185–204.

Robert D. Martin, PhD