Reasons for diversity of placental structure.

Many early embryonic stages are nearly indistinguishable in different Eutheria. However, implantation stages and placental morphological types vary tremendously. A number of factors favor this conflicting diversity. 1. Whereas embryo development takes place in the isolation of the amniotic cavity, the extraembryonic membranes of the conceptus develop in close association with the uterus of a genetically different organism. 2. Early conditions for the developing conceptus are anaerobic whereas later in development efficient aerobic conditions are essential for continued growth of the fetus. 3. Developing extraembryonic membranes have the potential to form two partially sequential placentas. The yolk sac can participate in forming a choriovitelline placenta, including an interhemal region, and can be adapted to various non-respiratory functions as gestation proceeds. Development of the chorioallantoic placenta begins later than the choriovitelline placenta but can overlap with this before supplanting it. The original development of the extraembryonic membranes occurs when the conceptus is sufficiently small that neither its nutritional requirements nor its respiratory needs are the burden to the maternal organism that they are later in gestation. Despite these developmental factors promoting different methods of forming the definitive placenta, the placental type is consistent within most families indicating that the divergence in placental structure accompanied evolution of differences between groups.

Cytological events involved in protein synthesis in cellular and syncytial trophoblast of human placenta. An electron microscope autoradiographic study of [3H]leucine incorporation.

Electron microscope autoradiography has been used to study protein synthesis in syncytial and cellular trophoblast of term human placental villi incubated in vitro with tritiated leucine ([3H]leu). Autoradiographs were analyzed using the hypothetical grain analysis of Blackett and Parry (1973. J. Cell Biol. 57:9-15). The results of this study demonstrated that both cellular and syncytial trophoblast have marked capacities for protein synthesis. Cellular trophoblast synthesized protein in both its rough endoplasmic reticulum (RER) and its ground plasm which contained abundant free ribosomes. The vast majority of 3H-proteins remained within the cell, with some of the proteins synthesized ultimately appearing in the nucleus. A small percentage of grains was ultimately associated with the trophoblast basement membrane. In syncytial trophoblast, the RER was the dominant site for protein synthesis. The autoradiographic data suggested that, as in the cellular trophoblast, the vast majority of 3H-proteins synthesized by the syncytial trophoblast remained within the syncytial trophoblast throughout the incubation period. The major portion of [3H]leu-labeling present in the syncytial trophoblast of villi incubated the longest times (4 h+) remained in association with the RER. Labeled proteins did not become concentrated in syncytial trophoblast Golgi apparatus, vesicles, or granules. In contrast to cellular trophoblast, the nuclei in the syncytium did not contain 3H-proteins at any time-point studied.

Cytological events involved in glycoprotein synthesis in cellular and syncytial trophoblast of human placenta. An electron microscope autoradiographic study of [3H]galactose incorporation.

Electron microscope autoradiography was used to study glycoprotein synthesis in cellular trophoblast (cytotrophoblast) and syncytial trophoblast of term human placental villi incubated in vitro with D-[1-3H]galactose ([3H]gal). Autoradiographs were analyzed using the hypothetical grain analysis of Blackett and Parry (1973. J. Cell Biol. 57:9-15). The results of this study indicated that [3H]gal incorporation into term placental villi was predominantly localized to cytotrophoblast. Utilization of [3H]gal by term syncytial trophoblast was extremely low and yielded too few grains for a quantitative grain analysis. This result is in striking contrast to that found in the preceding study of [3H]leucine incorporation (Nelson, D. M., A. C. Enders, and B. F. King. 1978). Within cytotrophoblast, the rough endoplasmic reticulum incorporated the most [3H]gal into glycoprotein. The Golgi apparatus was another site of [3H]gal incorporation. The vast majority of the [3H]gal incorporated into cytotrophoblast during the pulse incubation remained intracellular through the duration of the experiment. There was little autoradiographic evidence for secretion of tritiated macromolecules. Cytotrophoblast incubated for the longest time period studied (4 h+) showed a substantial concentration of tritiated macromolecules in the Golgi complex and in the ground plasm but not in the rough endoplasmic reticulum.

Female reproductive tract and placentation in sucker-footed bats (Chiroptera: Myzopodidae) endemic to Madagascar.

The reproductive tract was examined in four non-pregnant and two gravid specimens of Myzopoda. The ovaries had little interstitial tissue. The uterus was bicornuate and the lenticular placental disk was situated mesometrially in one horn. The interhaemal barrier of the placental labyrinth was of the endotheliomonochorial type. There was a single layer of trophoblast cells. The cells of the maternal endothelium were large and basophilic, contained abundant rough endoplasmic reticulum, and rested on an irregular basement membrane. Blunt projections of endometrium extended into the placental disk and clusters of large cells occurred between the endometrial stroma and labyrinth. At the margins of the disk folds of trophoblast occurred and at the cranial end they formed an haemophagous region. The folds lateral to the disk included some peculiar tubular-appearing structures. There was a persistent yolk sac containing large endodermal cells around a largely collapsed lumen. Several features of placentation, such as the interhaemal barrier and the haemophagous region, are consistent with an association of Myzopodidae with Emballonuridae. No support was found for alternative hypotheses that include Myzopodidae in the noctilionoid or vespertilionoid lineages.

Implantation in the macaque: expansion of the implantation site during the first week of implantation.

Data accumulated over several years of investigating implantation in macaque monkeys have been used to estimate the rate of expansion of the initial implantation site, the increase in volume of the site, and the rate of arterial invasion by cytotrophoblast columns. In addition the expansion of the secondary implantation site has also been estimated. The primary implantation site expands from an average diameter of 0.268 mm on day 10 to 4.93 mm on day 16-17. It expands in thickness from 0.064 mm on day 10 to 0.96 mm on day 16-17, and in volume from 0.0036 mm(3) on day 10 to 18.34 mm(3) on day 16-17. During this period of rapid expansion in extent and volume of the implantation site, trophoblast invades the endometrium, forms a lacunar stage, and initiates villus formation; consequently these very considerable changes in structure occur when the implantation is still very small yet growing rapidly. The secondary site expands from 0.23 mm in diameter on day 12 to 2.8mm on day 17. The rate of penetration of cytotrophoblast into endometrial arteries diminishes from 0.602 mm per day on day 12 to 0.171 mm per day on day 16, using the straight-line method of estimation. This diminution in rate is consistent with the hypothesis that cytotrophoblast cells generated at the anchoring villi migrate over cytotrophoblast cells that have attached to the endothelium of the endometrial arteries in advancing the intraluminal columns of trophoblast. It is hoped that the summaries provided will be useful to investigators using macaque monkeys to analyze aspects of implantation in primates.

Placental diversity in malagasy tenrecs: placentation in shrew tenrecs (Microgale spp.), the mole-like rice tenrec (Oryzorictes hova) and the web-footed tenrec (Limnogale mergulus).

Placentation in tenrecs of the subfamily Oryzorictinae, family Tenrecidae, has not been described previously. The structure of the placenta of this group and especially of the genus Microgale was investigated to determine its similarity or dissimilarity to previously described placentas of the tenrec subfamilies Potamogalinae and Tenrecinae. Fifteen specimens of the genus Microgale ranging from an early yolk sac stage to near term were available for study. Placentation in Microgale was found to be different from other tenrecids in that there is an early simple lateral rather than central haemophagous region. In addition, a more villous portion of the placental disk forms before the formation of a more compact labyrinth. Although the definitive placenta is cellular haemomonochorial, it lacks the spongy zone found in the Tenrecinae. Neither does it resemble the endotheliochorial condition found in the Potamogalinae. Of the two genera of the subfamily Oryzorictinae represented by single specimens, the placenta of Limnogale resembled that of the Microgale but Oryzorictes had several differences including a lobulated placental disk. It is concluded that there is more variation in placentation both within the subfamily Oryzorictinae and within the family Tenrecidae than would ordinarily be expected.

Comparative placentation: some interesting modifications for histotrophic nutrition -- a review.

In considering the diversity of Eutherian mammalian placental structure, it is helpful to keep in mind that both phylogenetically and ontogenetically a functional yolk sac placenta precedes development of the chorioallantoic placenta. Usually the chorioallantoic placenta progressively displaces the area of contact of the yolk sac placenta with the endometrium. It is also closely applied to the endometrium, increasing respiratory efficiency but tending to decrease histotrophic nutrition. Catarrhine primates have minimal histotrophic uptake, using haemotrophic mechanisms throughout most of gestation. Rodents, by using partial or complete inversion of the yolk sac, have extensive regions available to histotroph in later pregnancy. Most mammals, however, have developed specialized regions of the chorioallantoic placenta for ingestion of uterine secretions, cell debris and erythrocytes. These regions vary from simple saccular areolae to more complex areolae such as the chorionic vesicles of prosimian primates to specialized haemophagous areas. All such structures are consistently lined by columnar cellular trophoblast. Examples of these structures, many of which can ingest both glandular secretions and erythrocytes (i.e. they are heterophagous), can be found in species whose definitive placenta is epitheliochorial. They are common but not universal in species whose definitive placenta is endotheliochorial and are even found in a few species with definitive haemochorial placentas. Restriction of phagocytosis to regions of columnar cells provides polarized cells appropriately oriented for ingestion, breakdown and transport, and limits the type of ingestion of both beneficial and potentially toxic materials to expendable individual cells.

The fetal membranes of the otter shrews and a synapomorphy for afrotheria.

The otter shrews of mainland Africa are the closest relatives of the Madagascar tenrecs. We sought for similarities in placentation between the two groups and, in a wider context, with other mammals of the Afrotheria clade. Specimens of the Nimba otter shrew (Micropotamogale lamottei) were obtained from the Ivory Coast and examples of the giant otter shrew (Potamogale velox) from the Hill Collection. The Nimba otter shrew has a central haemophagous organ similar to that in tenrecs. The labyrinth of the Nimba otter shrew, however, is endotheliochorial with syncytial trophoblast enclosing the maternal vessels. On the other hand tenrecs have cellular haemomonochorial placentae and an associated spongy zone, which is not present in the Nimba otter shrew. The placenta of the giant otter shrew is also endotheliochorial. The central region of its placenta is particularly interesting, since the juxtafetal portion is clearly a haemophagous region whereas the labyrinth feeding this region is endotheliochorial. Thus there is considerable variation in placental morphology within Tenrecidae. Importantly, however, both otter shrews have a large allantoic sac divided into four intercommunicating lobes by two pairs of septal folds. A similar arrangement has been described for representatives of each of the remaining five orders within Afrotheria. This is significant because previous anatomical studies have failed to establish a single synapomorphy in support of Afrotheria.

Suppression of extravillous trophoblast invasion of uterine spiral arteries by estrogen during early baboon pregnancy.

The present study determined whether estrogen plays a role in regulating invasion and remodeling of the uterine spiral arteries by extravillous trophoblasts during early baboon pregnancy. The level of trophoblast invasion of spiral arteries was assessed on day 60 of gestation (term is 184 days) in baboons untreated or treated on days 25-59 with estradiol or aromatizable androstenedione. The administration of estradiol or androstenedione increased (P<0.01) maternal serum estradiol levels approximately 3-fold above normal. The mean+/-SE percentage of spiral arteries/arterioles invaded by extravillous cytotrophoblasts in estradiol-treated baboons for vessels with diameters of 26-50 microm (0.0+/-0.0), 51-100 microm (1.2+/-0.7) and >100 microm (13.2+/-5.5) was 100%, 90%, and 75% lower (P<0.001), respectively, than in untreated baboons (2.4+/-1.2%; 11.0+/-5.5%, and 54.5+/-8.5%, respectively). Similar results were obtained with androstenedione treatment. However, the distribution of uterine spiral arteries grouped by diameter or number of arteries per basal plate area, i.e. microvessel density, were similar in untreated and estrogen-treated baboons. We suggest, therefore, that the low levels of estrogen exhibited during early primate pregnancy are required to permit normal progression of trophoblast vascular invasion and that the surge in estrogen which occurs during the second-third of normal pregnancy has a physiological role in suppressing further arterial trophoblast invasion. Consequently, we propose that the estrogen-dependent restraint of spiral artery invasion/remodeling ensures optimal blood flow dynamics across the uteroplacental vascular bed to promote normal fetal growth and development.

Early studies of placental ultrastructure by electron microscopy.

BACKGROUND: Transmission electron microscopy (TEM) was first applied to study placental ultrastructure in the 1950's. We review those early studies and mention the scientists that employed or encouraged the use of TEM. FINDINGS: Among the pioneers Edward W. Dempsey was a key figure who attracted many other scientists to Washington University in St. Louis. Work on human placental ultrastructure was initiated at Cambridge and Kyoto whilst domestic animals were initially studied by Björkman in Stockholm and electron micrographs of bat placenta were published by Wimsatt of Cornell University. CONCLUSIONS: Prior to the introduction of better fixation techniques, TEM images were of modest technical quality. Nevertheless they gave important insights into placental ultrastructure, particularly the nature of the maternal-fetal interface.

Placentation in mammals: Definitive placenta, yolk sac, and paraplacenta.

An overview is given of variations in placentation with particular focus on yolk sac, paraplacenta, and other structures important to histotrophic nutrition. The placenta proper varies in general shape, internal structure, and the number of tissues in the interhemal barrier. Yolk sac membranes persist to term in insectivores, colugos, rodents, and lagomorphs. In the latter two orders, they are of known importance for maternal-fetal transfer of antibodies, vitamins, lipids, and proteins. The detached yolk sac of bats is also active throughout gestation. A vascular paraplacenta, or smooth chorioallantois, has known functions in ruminants and carnivores and is found in several other orders of mammal where its function has yet to be explored. In human gestation, the chorion (avascular chorioallantois) is important for hormone synthesis. The true chorion of squirrels and hedgehogs is avascular but may nevertheless allow transfer from mother to fetus through the exocelom. Hemophagous areas with columnar trophoblast are paraplacental structures in carnivores and elephants but occur also within the placenta as in hyenas and moles. In shrews, it is the yolk sac that ingests and processes red cells. Areolas and chorionic vesicles are other structures important for absorption of uterine secretions and ingestion of cellular debris. In conclusion, we find that paraplacental structures, while showing less variation than the placenta proper, contribute not just to the integrity of overall placentation, but in various ways to maternal-fetal interrelationships.

Fine structural abnormalities of the placenta in diabetic rats.

In the streptozocin-induced diabetic rat, the placenta is larger and the fetus is smaller than normal. To study cellular differences that might contribute to the size and functional disparity between diabetic and control placentas, a light- and electron-microscopic analysis was performed on 14-, 18-, and 22-day (term) control and diabetic placentas. Diabetic placentas, especially later in gestation, were marked by the presence of large numbers of glycogen-distended cells in the basal zone. Within the placental labyrinth, the trophoblastic layers of the interhemal membrane were significantly thicker in the diabetic placentas on days 18 and 22, and large accumulations of liid droplets were present, especially in the inner two trophoblastic layers. In normal placentas there is a marked thinning of the placental barrier in the labyrinth by day 22 of gestation, making the thickness of the labyrinthine layers in age-matched diabetic placentas even more impressive. Finally, the labyrinth of 22-day diabetic placentas contained more glycogen and rough endoplasmic reticulum in the inner trophoblastic layer, a feature that is found in less-mature (18-day) control placentas. Thus, the diabetic placentas have a number of features that are consistent with functional immaturity/dysmaturity. Cytologic evidence confirms the presence of increased glycogen and lipid reserves in both the junctional zone and the cellular area between maternal and fetal blood.(ABSTRACT TRUNCATED AT 250 WORDS)

Development of the haemophagous region and labyrinth of the placenta of the tenrec, Echinops telfairi.

Our purpose was to determine how the central haemophagous region and cellular haemomonochorial labyrinth of the tenrec placenta are formed. The haemophagous region is preceded by a region of invasion of the endometrium by trophoblast comprising a cytotrophoblast layer covered by syncytial trophoblast and contiguous with numerous masses of multinucleate trophoblast. The trophoblast intrudes into the endometrium, eliminating the stroma, although small vessels and clumps of glandular epithelium persist. This extensive central region is connected to the forming disk by a ring of chorioallantois covered by a single layer of columnar trophoblast. Later the multinucleate masses and syncytial trophoblast degenerate. The unilaminar cytotrophoblast remains, is elaborated into folds, and phagocytoses glandular secretion, cell debris and erythrocytes. As the central area is transforming, fetal capillaries move into the cytotrophoblast pads surrounding the central zone. Prior to this, the cytotrophoblast has formed a multilayered structure and interrupted maternal vessels to create an anastomotic network of blood spaces lined by cytotrophoblast. The invasion of fetal capillaries transforms this preplacental pad into a cellular haemomonochorial labyrinth with the uninvaded portion forming an underlying spongy zone. Thus interaction of the trophoblast with the endometrium is substantially different in the central zone compared to the area of the preplacental pad.

Luteotropic regulation of dispersed rat luteal cells in early pregnancy.

Viable and functional luteal cells were prepared, using a combination of hyaluronidase, collagenase, and a low concentration of trypsin in a Dulbecco's modified Eagle medium containing 0.5% bovine serum albumin and 3.3 mM Ca++, from corpora lutea taken from 2-day pregnant rats. The viability and functional capacity of the dispersed cells were evaluated by electronmicroscopy and by measuring steroidogenic capicity during perifusion. Dispersed luteal cells previously exposed in vivo to biphasic prolactin (PRL) surges were found to respond during perifusion to as little as 0.5 ng/ml LH by increased steroid secretion. The net progesterone synthesis and secretion remained elevated over a time course of 2 1/2 hours perifusion, and the magnitude of the luteotropic stimulation was dose dependent on LH. However, luteotropic stimulation of LH could not be maintained beyond 2 1/2 h without renewed (in vitro) PRL exposure. PRL by itself maintained the low initial secretion rate of progesterone but demonstrated no stimulatory effect. Different steroidogenic responses were noted during the in vitro administration of LH alone and the administration of LH plus PRL. In the former case, the decreasing rate of progesterone secretion was accompanied by an increasing 20 alpha-dihydroprogesterone secretion, suggesting that luteal 20 alpha-hydroxysteroid dehydrogenase activity was not suppressed. In the latter case, progesterone secretion was maintained and 20 alpha-dihydroprogesterone secretion fell suggesting an inhibitory action by PRL against 20 alpha-hydroxysteroid dehydrogenase activity. Dispersed luteal cells, preincubated at 36 C in medium containing only PRL, retained viability and functional capacity in response to LH-PRL stimulation for periods of time up to 48 h. Preincubation with LH alone did not prolong cell viability.

Current topic: structural responses of the primate endometrium to implantation.

Morphological responses to implantation and pregnancy in the human and nonhuman primates include decidualization of the endometrial fibroblasts and accumulation of large numbers of large granular lymphocytes, formerly called endometrial granular cells, and may also include an epithelial plaque response and endothelial cell hypertrophy. Although stromal decidualization occurs in all species, it develops slower in macaques than in the human, and slower in the baboon than in the macaques. Cytologically, however, in all of these species there is extensive modification of the decidual cells by midgestation. Hypertrophy of luminal and gland neck epithelial cells is common in most monkeys and is also seen in the baboon, but has not been reported in humans. Large granular lymphocytes undergo changes in morphology during the first week after implantation in the rhesus monkey, the only species in which they have been studied in the immediate postimplantation period. Later in pregnancy many of the large granular lymphocytes are surrounded by decidual cells in this species. All of the responses can be elicited by trauma and appropriate hormonal conditions, but the epithelial plaque response forms first in the rhesus monkey and baboon. It is suggested that more complete fine structural and immunocytochemical studies of the different decidual regions at different gestational ages, combined with studies of the synthetic and antigenic nature of the different cell types, would allow determination of whether or not there are subpopulations of decidual cells in primates as well as suggesting possible functions of the cells in pregnancy.

Implantation in the nine-banded armadillo: how does a single blastocyst form four embryos?

In the course of a study on reproduction in the nine-banded armadillo, conceptuses between the beginning of implantation and primitive streak formation were examined to determine the manner of trophoblast differentiation during invasion of the endometrium and the sequence involved in formation of four identical quadruplets. The armadillo blastocyst implants in the fundic recess of the uterus. A single amnion and cup-shaped epiblastic plate are formed, and an exocelom develops between the amnion and trophoblast of the implantation site. Loss of the abembryonic trophoblast exposes both visceral and parietal endoderm to the uterine lumen, inverting the yolk sac. Continued expansion of the exocelom facilitates the intrusion of the forming conceptus into the uterine lumen and is accompanied by enlargement of the epiblastic plate. Separate areas of condensations of epiblast cells are the first indication of formation of the four identical quadruplets. The single layer of microvillous trophoblast with basal infoldings (designated absorptive trophoblast) is most likely to contribute extensively to movement of fluid into the exocelom. The resulting expansion of the exocelom not only enlarges the implantation site but also displaces the collapsing common amnion, limiting the amnion to the areas of the forming embryos.

Formation of monozygotic twins: when does it occur?

An individual embryo from day 15 of gestation in the rhesus monkey showed partial separation of the cranial end of the embryonic shield into two separate cell masses with separate amnions but with a common yolk sac. Although what the fate of this conceptus might have been is not known, it demonstrates that the physical features of implantation in this primate are permissive of relatively late division of the epiblast, which suggests that monozygotic twinning may occur as late as the embryonic shield stage just prior to primitive streak formation.

What can comparative studies of placental structure tell us?--A review.

The diversity of placental structures in Eutherian mammals is such that drawing generalizations from the definitive forms is problematic. There are always areas of reduced interhaemal distance whether the placenta is epitheliochorial, synepitheliochorial, endotheliochorial or haemochorial. However, the thinning may be achieved by different means. The presence of a haemophagous area as an iron transport facilitator is generally associated with endotheliochorial placentae but is also found in sheep and goats (synepitheliochorial) and in tenrecs and hyaenas (haemochorial). Although similar chorioallantoic placentae are found within families, structure begins to diverge at the ordinal level and there is little correlation at the supraordinal level of phylogeny. Differences in formation and function of the yolk sac provide additional variation. There would appear to be considerable adaptive pressure for development or retention of the haemochorial type of chorioallantoic placenta. This type of placenta has several possible drawbacks including more ready passage of fetal cells to the maternal organism and, should the haemochorial condition be achieved early, oxidative stress. At any rate no animal larger than the human and gorilla has this type of placenta. The endotheliochorial condition is found in animals as large as the bears, manatee and elephants. In addition to the ungulates, the epitheliochorial condition is present in the largest animals with the longest gestation periods, the whales. Considering the length of time since the early stages of mammalian evolution, it is probable that few unmodified structural features are present in any currently surviving mammal. Nevertheless, more complete studies of divergent types of mammalian placenta should help our understanding of mammalian interrelationships as well as placental function.

A unique exocelom-like space during early pregnancy in the horse.

The free allantois and allantochorion of conceptuses from 17 mares between 20 and 90 days of gestation were examined to determine the manner in which the associated mesodermal derivatives differentiated. It was found that a robust basement membrane developed under the allantoic endoderm, and that this basement membrane was partially isolated from the vascular layer of the allantois by a mesothelial layer and an exocelom-like space. The exocelom-like space persisted until approximately the stage of villous formation, and remnants of the space persisted over larger allantoic vessels even later. It is suggested that originally the presence of the mesothelial layer facilitates type I collagen formation in the thick basement membrane under the basal lamina of the allantoic endoderm. The basement membrane maintains the integrity of the allantois while the underlying space allows some slippage between the allantois and its vascular layer which is more closely associated with the endometrial and exocelomic surfaces, respectively.

Structure of the definitive placenta of the tenrec, Echinops telfairi.

Until recently, tenrecs were classified with insectivores in the order Lipotyphla, but nucleotide sequence data suggest they have closer affinities with a group of African mammals called Afrotheria. The placenta of Echinops has not been described and no studies involving electron microscopy of the placenta of any species of tenrec have been published. We used light and transmission electron microscopy to examine fixed placentae of embryos ranging from 25-66 mm in length. The placental disk is situated in the antimesometrial portion of the bicornuate uterus. The greater part of the disk consists of a labyrinth underlain by a spongy zone. The interhaemal barrier is unusual in that the trophoblastic component is a single layer of cytotrophoblast. These trophoblast cells have thick areas especially near the nuclei and extensive thin flanges but only occasionally have membrane-closed pore regions. The luminal surface has isolated patches of microvilli, and pinocytotic vesicles are numerous both apically and basally. In the centre of the placental disk is an elaborately folded haemophagous region. The primary folds have allantoic endoderm at one surface and columnar cytotrophoblast at the other. These trophoblast cells have numerous lipid droplets and vesicles, and often contain large yellow pigment crystalloids. The labyrinthine zone ends abruptly at the margins of the placental disk. However, the endoderm and connective tissue of the allantois and a layer of cytotrophoblast extend beyond the placental disk as a paraplacental region. Some of these distinctive features of Echinops placenta are shared with individual afrotherians, but no significant characteristic of definitive placentation is shared by all the Afrotheria.