Most pregnancies that survive the first few months make it to birth. But it is at the time of birth that we see evidence of some of the most significant trade-offs resulting from the human evolutionary process. The most obvious one is fatal head size versus mother’s pelvis. Put very simply, the relatively narrow pelvis that is good for efficient bipedalism is not very good for babies to pass through because of their large heads. In many instances, they can only be delivered with cultural or technological assistance.
Bipedalism and Birth
Perhaps no other aspect of human evolutionary history has had a greater impact on birth than bipedalism. Bipedalism has been linked with many other human adaptations, including expanding brain size and intelligence, tool using/making, hunting, highly dependent infants, dietary changes toward increased animal consumption, and increased energy efficiency. Whatever the ‘‘prime mover,’’ the anatomical changes that accompanied the evolution of this unusual form of locomotion were significant and almost every one had an impact on birth because they all influenced the shape of the passageway through which almost every baby is born. A full description of the anatomical changes that were associated with the transition to bipedalism is probably not necessary, but there are a number of important adaptations that are revealed when the pelvises of a two-legged (biped) and four-legged (quadruped) animal are compared. These include, for the biped, an overall narrowing of the birth canal and pronounced spines that protrude into the birth canal. The hominin sacrum is also broader, which serves to provide stronger and better support for the upright trunk. The top of the sacrum (sacral promontory) also protrudes forward into the birth canal, reducing the opening size further and constricting the back portion. The actual birth canal or passageway has modified itself from a quadrupedal ‘‘shallow bony ring’’ to a bipedal ‘‘deep curved tube.’’ These changes are related to muscle alterations for successful striding and maintaining balance above the centre of gravity associated with bipedalism. Human males and females are physically different in a number of ways, a phenomenon known as sexual dimorphism. This is most clearly seen in overall body size, secondary sexual characteristics (breasts, buttocks, facial hair), distribution of body fat, and degree of muscularity. For osteologists, however, the most reliable part of the skeleton to use for assessment of sex is the pelvis. This is because the pelvis, in addition to its functional role in locomotion, also reflects modifications for childbirth, which are, of course, restricted to females. In general, efficient bipedalism is best with a narrow pelvis, but successful birth is best with a broad pelvis. The female pelvis that has evolved in response to these competing demands is, in some sense, compromised on both accounts: women are not quite as accomplished at bipedal running as men are (the fastest runner in the world will forever be a man), nor is delivery of an infant as easy for her as it is for our close relatives the chimpanzees. In an evolutionary sense, she has traded high-speed bipedal running for the ability to deliver large-brained infants.
At first glance, it seems that the slight sacrifice in efficiency that a wider pelvis would afford would not be as costly as the risks that the narrow pelvis places on childbirth. Why not just grow a very wide pelvis that makes childbirth easy? One can invoke the need for speed in escaping predators, but even that does not seem like a powerful enough selective force. While it may be true that a healthy woman in reasonably good physical shape is a pretty good biped, when we consider that adult females in the evolutionary past spent most of their adult years either pregnant or carrying infants, the selective advantage of efficient bipedalism becomes more apparent. In the latter stages of pregnancy and early months of nursing and carrying infants, energetic demands on females increase significantly, placing greater selective emphasis on efficient bipedalism. The resulting balance is a pelvis that allows for a high degree of energetic efficiency in locomotion but also allows a large-brained infant to pass at birth, albeit with a tight squeeze. Here is what an osteologist looks for in determining if a pelvis is female greater angle and concave curvature in the area under the pubic arch; wider birth canal in both side-to-side and front-to-back dimensions; greater sideways flare to the hipbones; a more rounded pelvic opening when seen from above (the typical male pelvic opening is heart-shaped); ischial spines on the side rather than toward the front; a top of the sacrum that does not project into the birth canal; and a slightly broader sacrum with less curvature. Certainly these characteristics are highly variable from woman to woman and depend on her overall body size, activity levels, and health and diet during growth, but if several are present, they indicate femaleness, and, of course, they all serve to enlarge the birth canal and reflect adaptation to childbirth. As obstetrician Maurice Abitbol has pointed out, the characteristics of the pelvis that distinguish humans from chimps are those that are attributable to bipedalism. Those that distinguish males from females are attributable to increased brain size (encephalization) and the requirements of childbirth. From the perspective of human evolutionary history, the earliest hominin pelvis (australopithecine) was adapted to bipedalism and the more recent (Homo species) is adapted to birth of a large-brained infant. Our discussion so far is to distinguish quadrupedal from bipedal pelvises and male from female pelvises, and it neglects the great variability seen in pelvises among women. Most medical research has been done on healthy women who grew up in reasonably good environments with adequate nutrition and health care. Their pelvises tend to be well developed for childbirth. For women who grow up in poverty with poor nutrition and overall poor health, however, pelvic dimensions described in medical textbooks and ‘‘birthing’’ books may not describe them. Also, taller women tend to have pelvises that better fit the babies they bear, whereas cephalopelvic disproportion or CPD (when the head is too large for the pelvis) is more common in shorter women. There may be a genetic component to short stature, but as we have seen, it is more often due to under nutrition and poor health during childhood. (It should be noted that there is equivalent variation in male pelvises, whose shapes are also affected by nutrition and health factors.)
Birth in Our Ancestors
Fossilized skeletal remains of early hominins have been extensively studied to ascertain how and when bipedalism evolved. Although there are very few pelvic remains from the earliest group, known as australopithecines, there are two almost complete pelvises and, as luck would have it for our interest in birth, they both appear to be from females. The major changes in the pelvis occurred when the lines leading to modern chimpanzees and humans diverged, approximately 5–7 million years ago. Birth is not ‘‘easy’’ for most primates, given the close correspondence between the size of the neonatal head and the maternal bony pelvis, but bipedalism placed even more constraints on the birth process through alteration of the dimensions of the pelvic entrance and exit. For quadrupedal primates, the birth canal has its greatest breadth in the front-to-back dimension, but with bipedalism, the birth canal is ‘‘twisted’’ in the middle so that the greatest dimension of the entrance is side-to-side and the greatest dimension of the exit is front-to-back. Because the passage is such a tight fit, the neonatal head must line up with each of the dimensions, meaning that the baby enters the birth canal with its head facing to the side and rotates approximately 90 degrees in the middle so that the head lines up with the broader front-to-back dimension of the exit. One way to envision this is to make a fist with your hand. The side-to-side (thumb to little finger) dimension is most likely greater than the front-to-back dimension. Now imagine a narrow oval passageway that is twisted in the middle so that the entrance is perpendicular to the exit. If you envision pushing your hand through the passageway, you will see that it has to rotate to pass through. If This is pretty much what a baby’s head has to do to get through the birth canal. But it is not just the baby’s head that has to pass through the birth canal. Following the head in a vertex (head first) presentation are the shoulders. Unlike monkeys whose shoulders are somewhat narrow (a good adaptation to running on all fours) and can collapse when passing through the birth canal, humans have broad, somewhat rigid shoulders that are not as flexible. They are also perpendicular to the long dimension of the head; so not only does the head rotate in the middle to line up with the long axis of the pelvic outlet, it must rotate again as it exits the birth canal to enable the shoulders to pass through the transverse dimension of the inlet. Once the head has emerged, it undergoes further rotation to bring the shoulders through the pelvic outlet. With large babies, a phenomenon known as shoulder dystocia in which the shoulders get stuck at the pubic bones may occur following emergence of the head. This can be life threatening, although it may not have been a significant cause of mortality in the past because infants were unlikely to have been large at birth. I will discuss this in more detail in the section on birth complications.
Additionally, adaptation to bipedalism has resulted in the top of the sacrum (sacral promontory) protruding into the birth canal so that the front of the maternal pelvis is more spacious than the back. Because the back of the baby’s head is the broadest, those two dimensions line up so that the baby usually emerges from the birth canal facing toward the mother’s back. If your image of childbirth is of a woman lying on her back on a bed or delivery table, this means that the baby is born facing the mattress or pad rather than facing the ceiling (midwives I have worked with call this ‘‘sunny side up’’). In quadrupedal primates, the back of the baby’s head lines up with the back of the mother’s birth canal and passes straight through without rotating to emerge facing the front of the mother’s body. When the baby’s head emerges from the birth canal, the mother monkey or ape reaches down and guides it out along the normal contours of its body. In humans, however, the mother often finds it difficult to reach behind her or between her legs to complete the delivery and risks pulling the infant against the angle of flexion in a way that damages nerves and muscles. I have argued that the tendency for the human neonate to be born facing the mother’s back accounts for the almost universal practice of seeking assistance at birth. For most primates and most mammals, birth is a solitary event and females typically seek isolation rather than companionship when they begin labour. Human females, in contrast, tend to seek companionship at birth in almost all cultures. Certainly it is possible for women to give birth unattended, but my anthropology colleague Karen Rosenberg and I have argued that throughout human evolutionary history, those who sought assistance at birth had more surviving offspring than those who delivered their infants alone. Even a small reduction in mortality and morbidity rates over hundreds of generations could account for the near-universal practice of accompanied birth that we see today. Our ancestors probably sought out others not because of a conscious awareness that it would reduce mortality; instead, it is more likely that they felt anxiety and uncertainty about labour and delivery and sought companionship for emotional support. For most of human history, those emotional needs at birth have been met by friends and family members, most of whom were probably women. These people may not have had midwifery skills, but they cared about the woman and her new-born and were able, simply by being present, to reduce emotional stresses that could interfere with labour and delivery. But today, many women give birth in unfamiliar surroundings (in hospitals or birth centres), often with people they do not know, and their emotional needs are often not met. Accompanying that experience are fear and anxiety in labour that are considered to be problems (or ‘‘defects’’ in the vocabulary of evolutionary medicine) that need to be dealt with medically (with pain-relieving drugs, for example), rather than the ‘‘defences’’ that they once were when they led women to seek the companionship and assistance at delivery that reduced morbidity and mortality. Of course, when anxiety and fear become extreme, medical intervention may be necessary for a healthy birth outcome, but companionship may prevent the anxiety from becoming extreme in the first place. An evolutionary perspective argues that the anxiety many women feel at birth is a legacy from our ancestral history of a time when it helped to reduce mortality. In Randy Nesse’s words, ‘‘what good is feeling bad’’ about birth is increased reproductive success ; in other words, women who felt ‘‘bad’’ (anxiety) about giving birth alone may have had more surviving offspring.
Brain Size at Birth
Another consequence of altered pelvic shape in the evolution of bipedalism is that an upper limit was placed on the size of the neonatal head that could pass through at birth. This meant that the birth process was in direct opposition to selection for increased brain size in human evolution and some sort of evolutionary compromise was required. (Obstetrician Philip Steer calls this a ‘‘conflict between walking and thinking.’’ ) Expansion of the birth canal to allow delivery of a large-brained infant would have meant that women would have sacrificed efficiency in bipedalism. As noted earlier, while this may not be a problem today with low activity levels and motorized transit systems, females who could not walk great distances carrying food and infants because of inefficient bipedalism would have been at a selective disadvantage. The only way for the conflicting trends of a narrowing of the pelvic opening and increasing size of the brain to be resolved was for more and more brain development to be postponed until after birth. For most mammalian species, approximately half of brain growth has occurred by the time of birth. In most primate species, however, more growth takes place after birth than before and infants are born with less than half of their expected adult brain size. Although there are virtually no fossilized neonatal hominins, anthropologists Jeremy De Silva and Julie Lesnik have developed a method for estimating neonatal brain size based on cranial capacity of fossilized adult skulls and statistical analyses of neonatal and adult brain sizes of contemporary primates. Their conclusions indicate that, among other things, modern humans have ‘‘precisely the brain size at birth expected’’ for a primate of our body size. What is notable about humans is not how large the brain is at birth but how much more the brain grows after birth. The percentage of brain growth that is completed at the time of birth gradually decreases from the earliest australopithecines to modern humans, related most likely to the increasing size of adult brains over that same period.
The great amount of brain growth that occurs after birth means different things for the infant and the parents. For the mother, energetic investment is greater postnatally than prenatally (pregnancy ‘‘costs’’ an extra 300 calories per day, while lactation ‘‘costs’’ an extra 500 per day beyond her normal needs). And the only way that more dependent infants could survive would be for mothers and fathers to invest more heavily in them in the first several months after birth while the neurological systems necessary for independent function develop. Delayed development of the brain enables birth to occur through the narrow bipedal pelvis. Another characteristic related to undeveloped fatal brains that makes birth easier is that the cranial plates are not fused and can slide over each other at the time of birth in vertex presentations. Some of the ligaments that hold the mother’s pelvis together can also relax a bit at the time of delivery under the influence of the aptly named hormone relaxin. Quadrupedal monkeys experience a lot of relaxation of the ligaments and joints of the pelvis at the time of delivery, easing the passage of the large-headed infant through the somewhat small birth canal. For bipedal humans, however, too much pelvic relaxation provides risks for the muscles and structures that maintain bipedal locomotion, so the degree of flexibility at delivery is limited. Another unusual characteristic of modern humans that may contribute to successful childbirth is that the place where the two pubic bones come together does not fully fuse at the end of adolescence as it does in other primates. This means that the pelvic opening can continue to expand even after growth has ceased in the rest of the skeleton. In fact, the pubic bones do not fuse until after the most fecund years (until the early 30s) are over. This characteristic has not been observed in australopithecines, suggesting that its evolution is related to birth and increased brain size rather than bipedalism. Most of the discussion so far has revolved around selection for less brain maturity at birth so that the baby’s head can get through the tight squeeze of the bipedal birth canal. Certainly this crude mechanical constraint played an important role, but it appears that there are numerous advantages for the infant that derive from being born before growth is completed. One of the most significant has to do with language learning at a time when the brain is still developing. In a general sense, being developmentally immature at birth provides greater flexibility in cognitive development overall. Infant motor immaturity may seem like a disadvantage at first because the energetic costs to the mother of carrying the young are significant, but it sets the stage for the development of complex cognitive and social skills. Arguably, the world outside the womb is much more stimulating for infant learning than inside the womb, although that is not to say that prenatal learning is not important or extensive.
Medical Consequences of Bipedalism
The professions of obstetrics and midwifery probably would not exist were it not for the inherent conflict between the bipedal pelvis and the large brained infant.
Pelvic Organ Prolapse (POP)
Pelvic organ prolapse is a much more common problem for women than for men, due in part to the expansion of the pelvic opening for birth; it is especially common in older women who have given birth several times. It occurs when the pelvic floor weakens or is damaged and can no longer support the internal organs. The most likely culprit is damage due to overstretching of the levator ani muscle during birth; this is the muscle that wagged the tail in pre-hominin ancestors and now forms part of the pelvic floor. Heavy lifting and chronic coughing or straining can also contribute to weakening. Pelvic floor prolapse is not usually life threatening but can cause pain and discomfort and may require surgery for relief. It also interferes with sexual functioning. A recent survey determined that in the United States, almost a quarter of women suffered some form of pelvic floor disorder, most prominent of which are incontinence and organ prolapse. As another example of trade-offs in evolution, women with wider pelvic inlets and shorter birth canals, which make for easier birth, are at greater risk. Women whose pelvic inlets are very wide (called a platypelloid pelvis in the medical literature) are at greater risk for pelvic floor disorders than those with narrow inlets (‘‘anthropoid pelvises’’.) Bipedalism requires the pelvic floor to support the pelvic organs, including the bladder, rectum, and so on, but it also requires changes in processes related to elimination. A narrow bony pelvic ring helps with the support function. Rather than being formed primarily by muscles, the bulk of the pelvic floor is made up of strong connective tissue, which provides better support in that it does not require muscles to constantly constrict the sphincters of the anus and bladder to keep them closed. Bipedalism requires a ‘‘more refined control of the visceral apertures.’’ In summary, the pelvic floor has evolved from a tail-moving function in quadrupeds to a weight-supporting and sphincter-controlling function in bipedal humans.30 It is probably not a stretch to say that constipation, incontinence, haemorrhoids, and herniation are all by-products of bipedalism. Furthermore, the relatively small pelvic basin also holds the uterus and fetus during pregnancy. The abdominal wall is fairly rigid and does not allow much extension. There is a limit to how large the fetus can grow, a fact that may partially explain why multiple gestations tend to be shortened—the combined sizes of the twins or triplets exceed the capacity for the abdomen to hold them toward the end of pregnancy. Expanding fatal size in pregnancy also puts pressure on the blood vessels, which compromises both maternal and fatal health. Significantly, the forward protrusion of the fetus into the abdomen also changes the centre of gravity for the woman’s body and provides challenges to maintaining upright posture while standing and walking, as noted in the previous. For pregnant quadrupeds, where the fetus is balanced among the four legs, this is not a problem. So it is not just birth that bipedalism challenges, but the latter part of pregnancy also.
Another medical consequence of bipedalism for some women is the need to deliver an infant via caesarean section. As noted, birth is another point when the interests of the mother and the fetus may not be the same. Larger babies (up to a point) are more likely to survive the birth process and the neonatal period than small babies, but their delivery is much more difficult for both the mother and the infant. It is in her best interest to give birth to a sufficiently large infant that can survive, but not so large an infant that the mother’s own health is compromised. The optimal birth weight for the infant (the weight at which perinatal mortality is lowest) is between 3,800 and 4,200 grams, according to studies in health-rich nations. There are no data reporting ease of delivery for birth,weight categories, but it is probably safe to say that the smaller the baby, the easier it is to deliver. Of course, the mother’s reproductive success depends on giving birth to a healthy baby, so it is clearly not in her interest to give birth to an infant too small to survive, no matter how easy it is for her to deliver. In fact, the average birth weight is less than the birth weight that is optimal for infant survival, suggesting that the mother’s needs override those of the infant. This makes sense in that in the past, if the mother died, there was very little chance that the infant would survive, but if the infant was a bit smaller than the optimum, his or her survival chances were still pretty good. In other words, the costs to the infants are less than the costs to the mother in this case. Babies that are too large are often delivered by caesarean section, which may increase the infant’s survival chances, but the mother’s risk is greater with surgical delivery than with vaginal delivery. Of course, in the past, both mother and infant would have probably died in the absence of surgical delivery options. Infant birth weight is an excellent example of stabilizing selection—higher mortality is associated with weights too large or too small, so natural selection keeps birth weight balanced between about 2,500 and 5,000 grams. To some extent, however, the constraints against birth weight have been removed by technology that keeps alive very low birth weight babies and allows caesarean section for very large babies. It would not be surprising to see a gradual increase in the mean birth weight in some populations now that overall health and nutrition have improved and surgical deliveries are possible. In fact, in the United States over an 18-year period, average birth weight increased by 40 grams, a 1.2% increase. National U.S. data from 1960 to 1997 show a 2% increase in babies weighing 3,500–3,999 grams and a 1% increase in those weighing 4,000–4,999 grams. This indicates that the average birth weight is drawing closer to the optimum because the risks to the mother of having a large baby have been reduced by C-section. There is even the suggestion that if this continues, virtually all babies will have to be born by C-section in the future. For example, it appears that the English bulldog, bred for squat bodies and large heads, has reached the point that virtually all are delivered by C-section.
Caesarean section rates are on the increase worldwide (29% in the United States; 23% in the UK; 22% in Canada; 32%–70% in Brazil; 23% in Australia; 35% in Italy and Taiwan; 40% in Chile). Most investigators suggest an ‘‘expected’’ rate of 10%– 15% for C-section and this is the rate recommended by the World Health Organization. Risk factors for surgical delivery include short stature, obesity, contracted pelvis (usually from poor nutrition during growth), and infection. It is possible that the ‘‘natural rate’’ may be increasing because of under nutrition during growth that results in short stature and, surprisingly, obesity in adulthood. For example, in Western Australia, women shorter than 160 cm had a C-section rate four times that of women taller than 164 cm. Women with a body-mass index (BMI) above 30 have a three times greater risk of C-section than women with BMI below 20. Obese women tend to have obese babies, which may explain the higher C-section rates. A final ‘‘natural’’ reason for increased C-section rates may lie in maternal age, which tends to be later in most of the countries with high rates. Of course, risks of caesarean section are still present and include death (seven times the risk of vaginal delivery in a Netherlands study), haemorrhage, pulmonary embolism, sepsis, and anaesthesia complications. Recovery time is longer and breastfeeding and mother-infant bonding are often compromised. Caesarean section may also compromise a woman’s future fertility. For example, there is evidence that stillbirths increase in pregnancies following a C-section, as do the rates of placental problems. It is also common for women who have one caesarean section delivery to have all future children this way, whether indicated or not. Furthermore, if a woman has a C-section for cephalo-pelvic disproportion (CPD) or dysfunctional labour, the chances are higher that her daughters will also require C-sections. Maternal-fatal conflict theory might argue that a big baby also reduces competition for resources from future siblings when the mother’s chances of having more children are reduced. Caesarean section also presents problems for infants, including higher incidence of respiratory distress, disrupted sleep rhythm in the early postnatal days, and greater challenges to self-regulation, the process by which the infant begins to maintain behavioural and physiological balance following the disruptions caused by birth. Because the mother has experienced major surgery, she may not be in a position to help her infant achieve self-regulation through the usual interactive processes of holding, stroking, skin-to-skin contact, and vocal and visual engagement. In one study of women two to four weeks after delivery, the brains of mothers who delivered vaginally were significantly more responsive to their babies’ cries than those who delivered by C-section as shown in fMRI (functional magnetic resonance imaging).
Some of these complications may be due to drugs that the mother receives during surgery, but there also may be disadvantages for the infant who ‘‘skips’’ labour. Although we are just beginning to ask these questions, there appear to be long-term effects of caesarean delivery on the infant that may cast doubt on the advisability of elective C-sections. A particularly alarming review concludes that the risk of developing type 1 diabetes is 20% greater in children delivered by C-section and that no other known factors can account for this increase.On the other hand, caesarean section has saved lives of countless mothers and infants in cases of cephalo-pelvic disproportion (CPD), mal presentation, placenta Previa, detached placenta, cord prolapse, uterine rupture, fatal heart rate abnormalities, and severe infections. Caesarean sections are increasingly recommended for women who are HIV positive or who have active outbreaks of herpes simplex virus (HSV); benefits of C-section for other infections like hepatitis B virus, human papilloma virus, and hepatitis C virus are more controversial. Urinary incontinence is less problematic for women who had C-sections. For women who have extreme fear of childbirth, the option for surgical delivery is comforting. Finally, it is often more convenient to have a scheduled, elective caesarean delivery for women who find their lives are too complicated for ‘‘natural’’ childbirth. Hospital births with electronic fatal monitoring, induction and augmentation of labour, epidurals, forceps, episiotomies, and unfamiliar attendants hardly seem more natural or desirable to women who have the option of a scheduled surgical delivery. Furthermore, the type of delivery with the greatest risk is an emergency caesarean section, which a planned C-section can avoid, especially if the baby presents breech or is past term or if the mother had a previous C-section.
Is Birth ‘‘Good for’’ Babies?
As noted, an increasingly common technological way of dealing with the tight squeeze of the baby through the birth canal and the anxiety and fear that some women feel is to deliver the baby via caesarean section. But with the rapidly rising rates of caesarean section and concerns about maternal and infant health following what is undeniably major surgery, one might wonder if there are advantages to being born vaginally. In other words, is there a benefit to being pressed by uterine contractions, navigating the tight quarters of the birth canal, occasionally being deprived of oxygen, and having your head compressed? Anthropologist Ashley Montagu was one of the first to suggest that relatively long labours were ‘‘good for’’ human infants. In fact, he argued that the stroking and massaging provided by the uterine contractions replace the functional significance of licking the young seen in virtually all other mammalian species. Among mammals, licking new-born infants stimulates development of the respiratory and digestive systems and infants that are not licked often die because of complications in these systems. The stimulation that the human baby’s skin receives during labour also influences organ development and proper functioning of the nervous system. Montagu notes that respiratory and digestive problems and poor control of the bladder and sphincter are common ailments in infants born prematurely or by caesarean section, those who experience brief or no labour contractions. There is evidence that infants born by caesarean after a trial of labour have better functioning systems than those born without the benefit of contractions. Another reason that labour might be good for babies is that the stress hormones that both the mother and fetus produce during labour trigger the production of catecholamine’s (adrenalin/epinephrine and dopamine) in the infant that help it cope with life outside the womb. Specifically, these ‘‘stress hormones’’ increase production of agents that speed up maturation of the lungs and enable the lungs to expand to keep amniotic and other fluids from filling them up. They also serve to increase fatal blood flow, especially to the brain; increase availability of calories to the baby; and increase white blood cells for immune protection. Further, catecholamine’s help the fetus withstand hypoxia during delivery. They appear to be important for promoting breathing immediately after birth, and infants born by elective caesarean section without labour (and thus not producing the high levels of catecholamine) often have breathing problems. Blood sugar levels are low in babies delivered by elective caesarean section, reinforcing the significance of catecholamine for mobilizing energy. This may be particularly important for infant survival in the first few days after birth in cultures that believe mothers should withhold the first secretions from the breasts (known as colostrum) and delay breastfeeding until the milk comes in.