Comparative patterns of craniofacial development in Eutherian and Metatherian mammals.
Despite the extensive literature on the evolutionary significance of the different reproductive patterns, less attention has been paid to the developmental differences characterizing the two groups. Although several studies on the development of marsupials exist (e.g., Esdaile 1916; Hartman 1919; McCrady 1938; Hill and Hill 1955; Nesslinger 1956; Bridge and Allbrook 1970; Hughes and Hall 1984; 1988; Hall and Hughes 1987; Klima 1987; Maier 1987a,b, 1993; Nelson 1987; Gemmell et al. 1988; Gemmell and Nelson 1988, 1992; Clark 1990; Filan 1991; Clark and Smith 1993; Gemmell and Selwood 1994; Smith 1994; Frigo and Woolley 1996) few compare development in detail with eutherians (although see Muller 1972a-c, 1973).
Of particular interest is the development of the craniofacial region. The craniofacial region is exceedingly complex functionally, morphologically and developmentally and many systems, such as the central nervous system, the cranial sense organs, and the skeletal and muscular systems develop in a highly coordinated manner. In marsupials a number of elements of the craniofacial region must be functional at a time when much of this region is only just beginning morphogenesis. It has long been recognized that these functional requirements have led to specific heterochronies and that some components of the face and oral region develop early and rapidly in marsupials relative to eutherians. However, a comprehensive comparison of development has been hampered by the lack of methods to analyze many developmental events in many taxa with different overall rates of development.
In this paper, I present a method that allows such comparison and I apply the method to questions about the evolution of developmental sequences and associations in the craniofacial region in marsupial and placental mammals. Several specific issues are considered. The first is methodological. Heterochrony, or changes in the timing of developmental events relative to an ancestral condition or sister taxon, has received a great deal of attention as a means to produce evolutionary change. The majority of studies of heterochrony examine changes in the relative rates of differentiation or growth (e.g., Alberch et al. 1979; McKinney 1988). Few focus on changes in developmental sequences and no well-established method for such studies exists (e.g., Wake and Hanken 1982; Alberch 1985; Strauss 1990; Clark and Smith 1993; Mabee 1993; Velhagen 1995; Cubbage and Mabee 1996; Mabee and Trendler 1996). Here I present a new approach to the comparative analysis of developmental sequence. This approach provides a means to distinguish the elements of developmental sequences that are conserved across the Theria, those that are unique to marsupials and placentals each, and those that are variable with no clear phylogenetic pattern.
TABLE 1. Taxa examined; the stages, sizes, or ages (as available) of the series; and the location and specific specimens examined. Within each collection, the specimens are arranged in chronological order, from the youngest specimen examined to the oldest. The listing is the original catalogue number of the specific collection. Abbreviations: IU for an age in metatherians represents an intrauterine specimen of undetermined age (all eutherian specimens are intrauterine); P after an age represents days postnatal; E represents days embryonic; mm represents crown-rump length of the original specimen in millimeters; in Macropus the length represents the head length (hl). The staging system for Manis may be found in Huisman (1933), for Mus in Theiler (1989), for Sus in Butler and Juurlink (1987), for Tupaia in Lange and Nierstrasz (1932), and for Dasyurus in Hill and Hill (1955). The staging systems, even within eutherians, are not comparable. KK Smith/DUCEC represents author's personal collection housed with the Duke University Comparative Embryology Collection. For both M. domestica and M. musculus a subset of the specimens examined, which covers the entire range of events, is provided. Complete listings and information about access to the collection, may be obtained from the author. Taxon and series ordering Location and specimen numbers Eutheria Felis domestica Cornell University: 8.8 mm-90 mm cat43, cat55, cat58, cat63, cat71, <20E-[similar to]46E cat72, cat78, cat82, cat87, cat90 Hubrecht Laboratory: car64, car66, car73, car74, car75, car76; Felis 20A Manis javanica Hubrecht Laboratory: 6 mm-90 mm Manis 25, Manis 30, Manis 110, Manis stages 13-[greater than]22 35, Manis 21, Manis 26, Manis 34, Manis 162, Manis 111, Manis 13, Manis 19, Manis 18 Mus musculus KK Smith/DUCEC: 10E-birth KS 150, KS 243, KS 174, KS 152, KS stages 18-26 153, KS 155, KS 110, KS 154, KS 176, KS 115, KS 156, KS 157, KS 138, KS 139, KS 137, KS 140, KS 239, KS 241, KS 42, KS 118 (This list is a subset of the more than 50 specimens examined.) Sus scrofa KK Smith/DUCEC: 13 mm-43 mm KS 186, KS 187, KS 180, KS 189, KS stages 19-25 190, KS 191, KS 192, KS 199, KS 200, KS 198 Tupaia javanica Hubrecht Laboratory: stages 14-[greater than]24 Tupaia 286a, Tupaia 270, Tupaia 285b, Tupaia 48b, Tupaia 45c, Tupaia 277b, Tupaia 278a, Tupaia 275a, Tupaia 280a, Tupaia 279, Tupaia 310a, Tupaia 310b, Tupaia 102a, Tupaia 291a, Tu- paia 576, Tupaia 568a, Tupaia 308a, Tupaia 303a, Tupaia 567a, Tupaia 309, Tupaia 58a, Tupaia 565a, Tupaia 564, Tupaia 260 Metatheria Dasyurus viverrinus Hubrecht Laboratory: 4.7 mm-29 mm ms128; ms129, ms131, ms132, ms133, stages IU-K ms134, ms138, ms142, ms148, ms150, IU-41 P ms154, ms158, ms160, ms165, ms172, ms176, ms177, ms179 Macropus eugenii KK Smith/DUCEC: hl 6 mm-27 mm KS 213, KS 212, CC 42, KS 214, KS IU-55 P 247, KS 246, CC 41, KS 215, CC 45, KS 249, CC 59, KS 248, CC 56, CC 47, CC 58, CC 55, CC 46, CC 52, CC 54, CC 53, KS 255 Monodelphis domestica KK Smith/DUCEC: 10 mm-50 mm KS 108, KS 109, KS 144, KS 201, KS 14E-35 P 168, KS 67, KS 180, KS 45, KS 166, KS 165, KS 104, KS 129, KS 206, KS 181, KS 182, KS 280, KS 130, KS 83, KS 185, KS 147, KS 208, KS 233, KS 184, KS 216, KS 217, KS 218, KS 219, KS 256 (This list represents a subset of the more than 80 specimens examined.)
Second, few studies of heterochrony examine shifts in timing of individual characters or sets of characters to understand the interactions of particular developmental events, and the changes in these interactions across evolution. Here the results of the study of heterochrony are applied to questions about the reasons for the differences in development between marsupial and placental mammals. A hypothesis on the developmental consequences of the different reproductive patterns is presented and hypotheses on the primitive reproductive condition are discussed. Finally, patterns of apparent plasticity and conservatism in craniofacial development derived from the comparative study are used to discuss hypotheses on the interaction of various cranial elements during morphogenesis in therian mammals.
MATERIALS AND METHODS
Relatively complete developmental series of five eutherians, Mus musculus (Rodentia), Felis domestica (Carnivora), Sus scrofa (Artiodactyla), Manis javanica (Pholidota), and Tupaia javanica (Scandentia), and four metatherians, Monodelphis domestica (Didelphidae), Macropus eugenii (Macropodidae), Dasyurus viverrinus (Dasyuridae), and Perameles nasuta (Peramelidae) were examined. For each of the nine species listed above at least 10 stages were available for the period between the first and last developmental event. The taxa, sources, stages and specific specimens examined are listed in Table 1. It should be noted that the terms marsupialplacental and metherian-eutherian are used informally and interchangeably throughout the text. Neither pair is completely appropriate. Because this paper discusses consequences of the different reproductive strategies, it is only applicable to extant taxa, and no inferences can be made on the condition of animals classified as Theria, Metatheria, or Eutheria and known only in the fossil record. The terms marsupial and placental are also unsatisfactory because not all extant marsupials have a marsupium (pouch), most have placentae, and some (permelids) possess a true chorioallantoic placenta (Tyndale-Biscoe and Renfree 1987). The Mus and Sus embryos were prepared in my laboratory (for preparation details, see Smith, 1994) and are housed with the Duke University Comparative Embryology Collection (DUCEC). Embryos (or in the case of mice, pregnant females) were obtained from commercial sources. Manis and Tupaia were examined at the comparative embryology collections at the Hubrecht Laboratory of the International Embryological Institute (Utrecht, The Netherlands). Felis embryos were examined at the Cornell College of Veterinary Medicine (Ithaca, NY) comparative embryological collection and the Hubrecht collection. Two marsupial taxa, Monodelphis and Macropus were also prepared in my laboratory and are housed with the DUCEC (for details on specimens and preparation techniques, see Clark and Smith 1993; Smith 1994). The Monodelphis specimens were collected from a breeding colony maintained at Duke University, and the Macropus specimens were obtained from M. Renfree (Melbourne University, Australia). The Dasvurus and Perameles collections were originally collected by J. P. Hill and are deposited at the Hubrecht laboratory. All animals examined were serially sectioned embryos, originally embedded in paraffin and stained with common histological stains.
Several well-known problems arise when developmental events are compared across taxa. The first problem is that age is not a useful marker because of different absolute rates of development. This is generally overcome by establishing staging systems. However, numerous staging systems exist and despite the efforts of a few workers, the correlation of different staging systems is problematic (e.g., Gribnau and Geijberts 1981; Butler and Juurlink 1987; Theiler 1989; Nelson 1992). Further, even under the best conditions, it is quite difficult to correlate systems when animals differ significantly, such as occurs in comparisons between eutherians and metatherians. To obtain sufficient taxonomic diversity, very often collections made by prior investigators must be used and in such cases exact stage or age data often are not available. Finally, in many cases collections found in museums or comparative embryology laboratories contain uneven sampling of development, that is, the available embryos may not sample the developmental period of interest at equally spaced intervals. All of these factors make analyses of comparative rates or absolute timing difficult, if not impossible. In this study neither absolute age nor stage was examined, but instead the relative sequence of events within each taxon was documented. This approach is therefore independent of rate and knowledge of absolute or relative timing of each event is not necessary. As long as specimens within a taxon may be serially ordered, by any system, different taxa can be compared. Finally, even though a focus on sequence does not measure rate of development, some inferences on relative rates may be obtained by comparing the relative proportion of any specific series of events to another series of events or to the total period under consideration.
In each collection, specimens were ordered by age, size or stage (Table 1). Because in mammals these are generally well correlated (e.g., Noden and de Lahunta 1985; Butler and Juurlink 1987), it was assumed that all series reflected a true time series. These assumptions were examined in the taxa for which the best data were available, M. musculus and M. domestica. Because both series were large, with multiple specimens for each age/stage, and because absolute age was relatively well known for each, it was possible to examine the relation of age, stage, and size. In both cases in the time period under consideration there was minimal variability in the sequence of developmental events. Most of the variability is due to the fact that in each collection age was known with no more than approximately 24-hour accuracy. Such uniformity, and in particular the relatively constant relation between age, size, and level of maturation, is present in mammals but generally is not true for organisms that develop under ectothermal conditions (e.g., see Mabee and Trendler 1996).
TABLE 2. List of developmental events examined. First ossification of Dentary Premaxilla Maxilla Frontal Jugal Parietal Squamosal Alisphenoid Basioccipital Basisphenoid Exoccipital Periotic Other skeletal elements Membrane bones approach midline Cartilage in basioccipital region Closure of secondary palate First appearance of tooth buds Differentiation of malleus and incus Differentiation of condylar cartilage Joint capsule at dentary-squamosal joint Muscle First alignment of myoblasts (tongue) First appearance of striations Craniofacial muscles distinguishable Central nervous system Evagination of telencephalon Differentiation of pigment in retina Connections between olfactory nerve and epithelium Layering in cortex Swelling of thalamus and hypothalamus Primary lens cells fill lens vesicle
Developmental Events Examined
Each specimen was examined to determine the state of 28 specific elements of the cranial skeletal, muscular, and central nervous systems (Table 2). The object of the study was to document the initial differentiation and morphogenesis of craniofacial structures, and structures were chosen to provide data on a range of events that could capture the relative maturation of the major systems. Particular attention was paid to elements and events thought to be of functional significance in neonatal marsupials and also to elements that were undergoing rapid change during the period of interest. In marsupials the period examined represents the time immediately prior to birth (generally one to three days prenatal) and the early postnatal period (approximately the first quarter of "pouch" life) and in eutherians, the period is entirely intrauterine, spanning the onset of histogenesis through organogenesis. This period in mice, for example, is embryonic day 10 to just before birth (= embryonic day 17-18). In all eutherians and one marsupial (Dasyurus) the series included specimens that preceded the appearance of the first morphological structure of interest. Thus the beginning of the sequence was completely resolved. Several series ended with the timing of one to three elements unresolved. However, these are generally the same structures (e.g., in almost all specimens the ossification of the periotic is the last event) and none of these late elements is significant in sorting the taxa.
The time in which ossification first appeared was noted for a number of bones of the membranous and endochondral skeletons (Table 2). First ossification was mapped by noting the specimen in which ossification appeared anywhere in the bone. The specimen in which the membrane bones of the cranial roof - primarily the frontal and parietal - approached the midline and each other was noted. This date did not indicate fusion, but approximate contact. Several other events in skeletal morphogenesis were examined, including differentiation of cartilage in the basioccipital-basisphenoid region, closure of the secondary palate, first appearance of tooth buds, the differentiation of the malleus and incus at the posterior region of Meckel's cartilage, the differentiation of secondary cartilage on the mandibular condyle, and the development of a joint capsule at the dentary-squamosal joint.
The emergence of all craniofacial muscles was examined, and the sequence of events was found to be fairly uniform across taxa and highly overlapping among muscles. Therefore the data were condensed into the following three stages. First, the specimen in which the myoblasts first began elongation, alignment, and fusion was noted. In all taxa this alignment occurred first in the intrinsic and extrinsic musculature of the posterior tongue. Second, the specimen in which the first striations appeared in any muscle was recorded; again, striations appeared first in the tongue, although in most taxa this was simultaneous with first arch and pharyngeal musculature. Finally, the specimen in which all craniofacial muscles were distinguishable was registered. The criteria for this condition included the presence of striated myotubes with peripheral nuclei and the differentiation of major internal divisions of all major cranial muscles (for more detail on the assessment of muscle development see Smith 1994).
Finally, the appearance of several structures indicating the maturation of the central nervous system (CNS) and cranial sense organs was noted. These elements included the evagination of the telencephalic vesicles, the differentiation of the retina into neural and pigment layers (i.e., the appearance of pigment in the retina), the establishment of connections between the olfactory nerve and the olfactory epithelium (determined by physical contact and not by any physiological or cellular measure), the appearance of at least four distinct layers in the cortex, the appearance of swellings beginning the differentiation of the thalamus and hypothalamus in the diencephalon, and the first stage in which primary lens cells filled the lumen of the lens vesicle.
Of all the events, the timing of the first ossification was most subject to measurement error. In the collections of embryos prepared by other workers, staining was often either not designed to distinguish bone, faded, or both. This was a particular problem with the Hill and Hubrecht collections, which were prepared in the early parts of the 20th century. Further, some bones were almost impossible to distinguish at the earliest stages of ossification in some section planes. Therefore, in particular for the membrane bones of the cranial roof, it is possible that actual first ossification occurred earlier than is documented here. A few conditions, such as contact of membrane bones or appearance of thalamic swellings, are somewhat subjective. However, there are no particular biases that would separate marsupials and placentals in these errors. Other elements are fairly easy to recognize, and the major source of errors in determining sequence is sampling interval.
Comparative Analysis of Developmental Sequences
To analyze the patterns of distribution of developmental events in nine different taxa, the timing of each of these events was assessed by comparing the relative timing of pairs of elements. To do this a matrix was constructed for each species in which the timing of the above 28 events relative to every other event could be expressed [ILLUSTRATION FOR FIGURE 1 OMITTED]. This resulted in 378 pairs of developmental events for each taxon. These pairs of events were defined as characters; the character state reflected the relative timing of the two events. (To avoid terminological confusion, "developmental events" or "morphological condition" refers to the 28 specific morphological elements examined. The use of the term "character" is restricted to a pair of developmental events.) Three character states were defined: state 0 indicates that a given event occurred before a second event; state 1 indicates that the given event occurred in the same stage as the second; state 2 indicates that the given event occurred after the second event (similar approaches have been independently, e.g., Velhagen 1995; Mabee and Trendler 1996).
It is unlikely that many events occur at exactly the same time in development, and if development were sampled continuously, few state 1 characters would be observed. Therefore, state 1 represents a combination of unresolved pairs and true simultaneous occurrences. A predominance of state 1 characters may indicate that the sampling interval was long relative to the actual pace of development so that many events appear to occur simultaneously. Table 3 lists the relative percentages of states 0, 1, and 2 for each taxon. It can be seen that state 1 is generally a small percentage of the total (never more than 17%, and generally around 13%). Further, the percentage is similar in all taxa, suggesting that sampling intervals, relative to the events and period of interest, were comparable across taxa.
Values of the character states for each taxon were then entered into the data matrix in the MacClade computer program (Vers. 3, Maddison and Maddison 1992), and mapped onto independently determined phylogenies for both metatherians and eutherians (Novacek and Wyss 1986; Marshall et al. 1990; Novacek 1990; Luckett 1994; Szalay 1994). Care was taken to choose taxa that sampled the Eutheria or Metatheria as broadly as possible. The eutherians include representatives of five orders that are presumed to have separated near the basal radiation of eutherians [ILLUSTRATION FOR FIGURE 2 OMITTED], with four of the orders representing an unresolved tetrachotomy: the Artiodactyla, Carnivora, Scandentia, and Rodentia (Novacek and Wyss 1986 and references therein; Novacek 1990). Thus no taxa within the Eutheria share particularly recent ancestry or close branching points. Likewise, the four metatherian taxa represent major basal clades of marsupials. Although there is some disagreement about the specific branching points of the families studied here, most workers agree that these taxa represent major branching points near the base of the marsupial radiation and last shared a common ancestor in the Paleocene (e.g., Kirsch 1977c; Novacek 1990; Nowak 1991; Luckett 1994; Szalay 1994).
TABLE 3. Distribution of character states, expressed as a percentage of the 378 characters or event pairs. Manis does not add to 100%, as data on the first appearance of tooth buds are missing; Manis is edentate. Taxon State 0 State 1 State 2 Felis 40 10 50 Manis 30 13 50 Mus 39 15 46 Sus 36 7 57 Tupaia 30 15 55 Dasyurus 48 11 41 Macropus 48 17 35 Monodelphis 50 12 38 Perameles 45 13 42
The maps were examined character by character to determine the distribution of character states within and between metatherians and eutherians, with the aim of identifying four different character state distributions: (1) those characters that were conserved (uniform) across the Theria; (2) those characters that had a unique pattern in metatherians relative to eutherians; (3) those characters that had a unique pattern in eutherians relative to marsupials; and (4) those characters that showed variability either across all taxa, or within either the marsupials or placentals. Characters were considered to be conserved across the Theria if they were uniform in their character state across all nine taxa, or were uniform with one or two exceptions. Marsupials and placentals were each used to evaluate the condition in the other group. For example, if marsupials were uniformly state 0, then eutherians were considered unique if the character distribution consisted of character states 1 and 2. But, if marsupials exhibited a character state distribution of 0 and 1, then placentals would be considered unique only if their character state distribution consisted entirely of character state 2. Because character state 1 represents, at least in part, unresolved conditions, further data may modify these assessments. Because data on outgroups to the Theria are not yet available, no assessment of evolutionary polarity can be made. Any comment on advancement or delay of characters refers only to comparative timing and is intended to imply nothing about the primitive condition (see discussion).
For the purpose of the comparison between marsupials and placentals, only characters where both marsupials and placentals exhibited unique and different patterns (or were distinguished by only a single inversion) were considered significant [ILLUSTRATION FOR FIGURE 3 OMITTED]. For example, if all marsupials were either character state 0 or 1, and all placentals were state 2, it was considered complete discrimination. A single inversion might be a case in which the above held, with one placental taxon exhibiting character state 1. Any case in which there was more than one inversion was considered not to discriminate the taxa. This is an arbitrary basis for considering significant differences and, in particular when inversions included state 1, is probably conservative. It is important to note that none of these assessments is a statistical measure of significance, as this is not a statistical analysis (such an analysis is discussed in Nunn and Smith, unpubl. data).
However, to assess the results presented here, it is important to have some evaluation of the probabilities of these distribution patterns occurring by chance alone (the null model). These probabilities can be assessed by assuming that the probability of being in one of the three states in any particular taxon is proportional to the overall probability of being in that state, and that the probability of being in a particular state in one taxon is independent of the states in the other taxa. For example, the probability of a specific group of four taxa all being either state 0 or 2 and another group of five taxa all being in the opposite state is given by the equation
([[p.sub.0].sup.4][[p.sub.2].sup.5]) + ([[p.sub.0].sup.5][[p.sub.2].sup.4]) (1)
where [p.sub.0] equals the probability of being in state 0 and [p.sub.2] equals the probability of being in state 2. For the data presented here, [p.sub.0] = 0.409; [p.sub.1] = 0.127, and [p.sub.2] = 0.464 (the averages of the frequencies of these states in Table 3). The likelihood of achieving complete character state segregation into state 0 and 2 by chance alone for a single character is 0.00113. The chance of segregating a number (n) of character states is calculated by raising that value to the nth power, and it becomes astronomically small for many characters. Thus it is improbable that the distributions found here would have occurred by chance alone.
Of the 378 event pairs (characters) examined in the phylogenetic analysis, 163 have character states that are uniformly distributed across all taxa; 82 are uniform with one or two exceptions; 28 absolutely distinguish metatherians and eutherians, and another 28 distinguish the two groups with one inversion. The remaining 78 have two or more inversions between the two groups and are considered to exhibit no pattern. In no instance is state 1 (two events appearing at the same time) uniform across all nine taxa. In 77% (37 of 48) of the cases that are uniform with one exception and 68% (23 of 34) of those that are uniform with two exceptions, the exceptions are state 1, which suggests that in these cases the exceptions are due to an unresolved condition, which additional data may modify. The cases in which there were more than two exceptions from uniformity or from complete sorting also include a large number of state 1 characters.
Fifty-six characters distinguish marsupials and placentals either absolutely or with a single inversion. These characters are clustered so that only a few structures and structure combinations are involved in the majority of changes in the relative timing or sequence of events (Table 4). These clusters of events tend to share shifts in relative timing of differentiation. The structures of the CNS sort into two groups: (1) the evagination of the telencephalon, which is always the first event in CNS differentiation; and (2) later events, which include the filling of the lens by the primary lens cells, the connection of the olfactory nerve with the olfactory epithelium, the differentiation of the thalamus and hypothalamus, and the development of layering in the cortex. The appearance of pigment in the retina is associated more closely with the evagination of the telencephalon than with other events. Within the skeletal system there are three major sets of bones involved in the majority of shifts in relative timing: (1) the membrane bones of the face, specifically, the premaxilla, maxilla, and dentary; (2) other, primarily membranous, bones of the calvaria; and (3) the exoccipital. Other events in the facial region involved in the shifts in timing that differentiate taxa include the appearance of tooth buds, the differentiation of ear ossicles, the closure of the secondary palate, the differentiation of basicranial cartilage and several aspects of muscle development. These clusters of characters demonstrate the following patterns of change in relative timing (heterochrony).
TABLE 4. Element pairs (characters) that sort eutherians and metatherians. In all pairs the first event is early in marsupials relative to the condition in placentals. Those pairs followed by an asterisk differentiate the two taxa with no exceptions; those that have no asterisk differentiate the taxa with a single exception (meaning that either one marsupial exhibits a pattern seen in placentals, or one placental exhibits a pattern seen in marsupials). In many cases these exceptions involve character state 1, which means they are unresolved. The event pairs in bold face are characters in which a somatic structure is accelerated relative to an event of central nervous system differentiation in marsupials. The four sets of groupings reflect clustering of events of the somatic system. The top group represents pairs in which the premaxilla, maxilla, and dentary bones are advanced; the second group represents characters in which other bones are advanced; the third group represents characters in which other skeletal structures are advanced; and the final group represents those characters in which muscles are advanced. dentary-telencephalon(*) dentary-thalamus(*) dentary-pigment in eye dentary-primary lens cells(*) dentary-olfactory nerve maxilla-primary lens cells premaxilla-olfactory maxilla-pigment in eye(*) nerve(*) premaxilla-thalamus(*) maxilla-telencephaion(*) premaxilla-primary lens maxilla-thalamus(*) cells premaxilla-pigment in maxilla-olfactory nerve(*) eye(*) premaxilla-telencepha- maxilla-tooth buds lon(*) premaxilla-tooth buds maxilla-first muscle maxilla-ear ossicles frontal-primary lens frontal-thalamus(*) cells(*) parietal-thalamus squamosal-primary lens cells jugal-primary lens cells(*) jugal-cortex jugal-thalamus alisphenoid-cortex exoccipital-cortex alisphenoid-thalamus exoccipital-primary lens basioccipital-thalamus cells(*) exoccipital-thalamus squamosal-thalamus exoccipital-alisphenoid jugal-parietal exoccipital-basioccipital basisphenoid-membrane bones meet exoccipital-squamosal jaw cartilage-membrane bones meet exoccipital-parietal(*) tooth buds-thalamus(*) secondary palate-olfactory nerve(*) ear ossicles-thalamus(*) secondary palate-thalamus(*) ear ossicles-olfactory secondary palate-primary lens nerve cells(*) ear ossicles-primary lens tooth buds-primary lens cells cells(*) basicranial cartilage-tel- encephalon(*) striations-thalamus(*) first muscle-telencephalon(*) striations-primary lens craniofacial muscles-cortex(*) cells striations-olfactory craniofacial muscles-thalamus(*) nerve(*) striations-cortex
When marsupials are compared to placentals, the membrane bones of the face begin ossification early relative to all events in the morphogenesis of the central nervous system. They begin ossification early relative to both the evagination of the telencephalon, as well as to the other events in CNS differentiation. Two other events in marsupials occur early relative to the evagination of the telencephalon: condensation of the basicranial cartilage and first appearance of muscle. Many events of the somatic system are advanced in marsupials relative to the later CNS events. These include the first ossification of a variety of cranial bones, the appearance of tooth buds, the differentiation of the ear ossicles, the closure of the secondary palate, and steps in the maturation of cranial muscles (Table 4). Within the ossification of cranial bones, only one element exhibits a major difference in its relative place in the developmental sequence: in marsupials, the exoccipital begins ossification relatively early. As can be seen in Table 4, the vast majority of element pairs that differentiate marsupials from placentals involve advancement of the development of cranial somatic structures relative to the development of CNS structures.
While the above patterns are of interest, it is important to note that the character state of most (65%) element pairs was relatively uniform (with either complete uniformity or with one or two exceptions, usually state 1) across these nine taxa. Therefore the majority of event pairs can be considered to be conserved in their relative timing across the Theria. For example, with the exception of the relative timing of ossification of the exoccipital bone, there is only minor variation in the sequence of onset of ossification of cranial bones among taxa. While the time of ossification of the bones of the front of the face (dentary, maxilla, and premaxilla) is advanced in development relative to the evagination of the telencephalon in marsupials, these three bones are the first to ossify in all mammals. In all taxa, membrane bones typically begin ossification before endochondral bones, although in both marsupials and placentals there is some overlap in the period in which these bones begin ossification. Mus, in particular, exhibits significant overlap in the period of ossification of dermal and endochondral bones (see below). Further the sequence of events of nervous system development is virtually constant across these taxa, although, rates of CNS maturation differ. The difference in rate is particularly notable in the forebrain, which in eutherians develops early and rapidly relative to metatherians.
In all taxa the sequence of muscle morphogenesis is the same: the tongue musculature is the first to begin differentiation and all craniofacial muscles follow quickly. Unlike the skeletal or central nervous systems, where regional and perhaps functional differences in relative rates appear, the events of muscular development appear to occupy a small and similar proportion of the entire period under consideration in all taxa. Further, for the most part the sequence in which muscles differentiate is fairly consistent across taxa. The posterior tongue musculature is always the first to begin alignment and the first in which striations appear. Differentiation and maturation of first arch and pharyngeal muscles follow immediately (inevitably in the same or the next available stage), and facial and ocular muscles differentiate last.
Patterns of Development within the Metatheria and Eutheria
Among marsupials 258 characters (of the total 378) were uniformly distributed and an additional 49 were uniform with a single exception. Therefore over 80% of the characters were essentially uniform across the four marsupial taxa examined here. In over half of the cases in which a single taxon differed from the rest, that taxon exhibited state 1, which may indicate that the departure from uniformity reflects insufficient resolution rather than actual differences in development. Dasyurus viverrinus was the single exception in 55% of the 49 cases. The remaining departures from uniformity were more or less evenly distributed among taxa or pairs of taxa with no obvious clumping. Of the 258 characters that were uniformly distributed across all marsupials, only four characters were uniformly character state 1. These four characters were relative ossification times of the following pairs of bones: dentary-maxilla, dentary-premaxilla, maxilla-premaxilla, and frontal-exoccipital.
Among eutherians 215 characters had character states that were uniform across the group. In an additional 77 characters, only one taxon differed from the other eutherians. Of these, Mus was the taxon that differed approximately 60% of the time. Many of the departures from uniformity were character state 1 as in marsupials, but in no cases were all eutherians scored state 1.
The Hierarchical Distribution of Developmental Patterns
The primary aim of the analysis was to identify patterns of developmental sequence that were consistent within each of the various taxonomic levels, and then to compare the distribution of these patterns across the taxonomic levels. Therefore the method relies on a two-step mapping of characters in a phylogenetic context. The first step distinguishes between the characters with states that are conserved across Theria, those that are conserved within the Metatheria (but differ from the Eutheria), those that are conserved within the Eutheria (and are different from the Metatheria), and those that vary, either across all Theria or within one of the superorders. Although it is virtually always assumed that development is conserved to some extent at various taxonomic levels, it is impossible to know the taxon-specific patterns a priori. The data presented here demonstrate that such taxonspecific characteristics of development exist and are distributed in a hierarchical manner. Of the 378 event pairs discussed here, 57% are uniform (or uniform with a single exception) in the nine therian taxa examined; 81% are uniform (or have a single exception) among the four metatherians, and 77% are uniform (or with a single exception) in the five eutherians.
Because the Eutheria and Metatheria are both represented by an array of taxa that do not reflect higher order branching (i.e., none of the species in the sample shares close relation to another species in the sample) and that cover the diversity within the superorder fairly broadly (i.e., the families and orders represented are among the basal radiations of the groups), it is reasonably likely that the patterns that emerge are representative of marsupials and placentals. This type of analysis is therefore a considerable improvement on the model taxa approach, in which it is simply assumed that a taxon is representative of its higher group.
The second stage in the analysis was to look for characters whose relative position in the developmental sequence consistently distinguishes marsupials and placentals. These characters were those element pairs that exhibited consistent patterns within but distinct patterns between the Eutheria and Metatheria. A total of 56 characters, or 15% of the event pairs examined, distinguished the two taxa; they were summarized above and will be discussed below.
As this is a novel approach to the comparative analysis of development, it is useful to discuss some of the strengths and weaknesses of the technique proposed here. No techniques for the comparison of developmental sequence have been established, but if development is ever to be studied in broad comparative or phylogenetic contexts, then such techniques must be developed. There are two major aspects to the current approach. The first is the definition of characters that are comparable across taxa; the second is the method of assessment of pattern. These two aspects will be discussed in turn.
To compare the relative timing of events in a series of animals with widely differing rates of development, (and for which information on absolute timing is generally not available), the sequence of events was studied. The developmental sequence was converted into a set of characters that allowed comparison across taxa by expressing the relative timing of all potential pairs of events as one of three character states: before, at the same time, and after. By looking at all possible character pairs in this manner, the specific events that shift their relative timing and the directions of these shifts are revealed. While many previous analyses of marsupial development have mentioned the early ossification of the bones of the face, the approach presented here can, for the first time, answer questions such as "early relative to what?" The analysis of event pairs thus, in addition to allowing an unambiguous coding of relative time of development, provides information that might not be available by just looking at the raw numbers in a sequence. It does not necessarily identify series of characters that may change sequence together, although such data may be derived by careful study of the results. For example, Table 4 demonstrates that the maxilla, premaxilla, and dentary tend to shift their relative timing as a unit.
The most significant disadvantage of the analysis of event pairs as characters is that the characters are not independent. For example, if a single event is shifted significantly in its relative timing, it will be reflected by change in many characters. The examination of the interactions of events is the goal of the current study. But, if these data were to be used in studies in which it is assumed that each of the 378 characters were independent (e.g., phylogeny reconstruction or statistical treatments of results) the nonindependence of the characters would introduce considerable problems in the analysis. Therefore such analyses should be avoided with this kind of dataset.
The second part of the analysis concerns the methods to assess the distribution of characters (event pairs) in a phylogenetic context. On one level this is a fairly simple matter of mapping various character states relative to phylogenetic grouping (either visually on a program such as MacClade, or numerically on a spreadsheet). Difficulties arise in determining what constitutes a significant difference. In this analysis it was decided that if the two groups sorted (i.e., each exhibited unique and different sets of character states) either absolutely or with one inversion, the two groups were different. This is a fairly arbitrary decision and the results therefore must be interpreted with the arbitrary nature of the boundary kept in mind. Finally, the major problem with this method is that it is exceedingly cumbersome. Twenty-eight events in nine taxa produced over 3400 characters. The addition of either taxa or characters to the analysis increases this number rapidly.
Developmental Consequences of Eutherian and Metatherian Reproduction
The characters distinguishing eutherian and metatherian mammals indicate two major kinds of heterochronies, or shifts in the sequence of developmental events. The first occurs within the somatic or mesenchymally derived elements of the head. Marsupials and placentals exhibit slightly different patterns in which the onset and rate of development of some elements of the head skeleton are advanced (or delayed) relative to others. The second and more pervasive change in developmental sequence is a whole-scale shift in the relative timing of the differentiation of the CNS and associated special sense organs of the head relative to the somatic elements. Although both phenomena have been noted previously, the current study includes a broad phylogenetic sample and also a method for the specific definition of the elements that display changes in relative timing of differentiation.
Shifts in the timing of development of somatic elements (especially the tongue, the facial skeleton and the secondary palate) have received most previous attention (e.g., Esdaile 1916; Hill and Hill 1955; Sharman 1973; Hall and Hughes 1987; Maier 1987a; Hughes and Hall 1988; Filan 1991; Nelson 1992; Clark and Smith 1993; Renfree 1993; Smith 1994). Previous studies discuss the fact that these elements appear at a relatively early stage in marsupial development, and for the most part interpret the early differentiation as an adaptation to the functional requirements of the neonate. The marsupial neonate must at a minimum have a closed secondary palate (providing support to and separation of the oral and respiratory pathways), sufficient ossification and chondrification (to provide cranial rigidity and support muscle attachment), functional oral musculature (primarily tongue and oral-pharyngeal muscles), and a means to support the lower jaw (because at birth there is no dentary-squamosal contact).
Previous studies have not, however, been able to identify the precise elements involved in shifts in relative timing. Although structures of the face develop more rapidly than many other cranial somatic structures such as the bones surrounding the brain, the current study demonstrates that there are very few changes in sequence involving pairs of elements of the somatic system. For example, the secondary palate shows little evidence of advancement relative to other somatic elements. Likewise, although the tongue appears to be relatively large at birth in marsupials, tongue muscle development is not notably advanced relative to either other muscles or other events in the craniofacial region in marsupials compared to placentals. The tongue is the first muscle to differentiate and mature in all mammals (see Smith 1994). Finally, the ear ossicles, which perform a specialized function in support of the jaw in the neonatal marsupial, show little evidence of advancement in development (Smith 1996). Therefore few of the shifts in relative timing or sequence involve events within the skeletal-muscular system.
Instead, the most significant difference in the sequence of craniofacial development involves the timing of the differentiation of the somatic structures of the head relative to the differentiation of the CNS. Over 80% of the characters identified as differentiating marsupials and placentals involve the relative timing of cranial musculo-skeletal elements relative to the timing of the CNS and associated sense organs (even though the CNS and sense organs make up only 21% of the elements examined). The somatic elements that are particularly advanced in marsupials include the onset of ossification of the maxilla, premaxilla, and dentary; the onset of organization of craniofacial musculature; and the differentiation of basicranial cartilages. These elements are advanced relative to even the earliest events in CNS differentiation (e,g., the evagination of the telencephalon and the differentiation of pigment in the retina). Many more somatic elements in marsupials (e.g., ossification of the frontal and jugal bones, closure of the secondary palate, appearance of striations in muscles) are advanced relative to later events in nervous system development (e.g., differentiation of the thalamus and hypothalamus and layering in the cortex), indicating different relative rates of differentiation of the two sets of structures in the two taxa.
The changes in relative timing of developmental events documented in the current study are shown schematically in Figure 4. In eutherians the onset of morphogenesis of the CNS structures examined here begins long before the onset of morphogenesis of the somatic system, and development of these structures is completed before most somatic structures have begun differentiation. In marsupials the converse is true: the events of CNS morphogenesis begin later and extend throughout the period of the morphogenesis of the cranial skeleton and musculature. The fact that CNS development is largely postnatal in marsupials is well known (e.g., Holt et al. 1981; Renfree et al. 1982; Nelson 1988, 1992; Reynolds and Saunders 1988; Saunders et al. 1989; Krause and Saunders 1994). However, the current analysis makes it clear that the shift in the relative timing of the differentiation of CNS and somatic structures is the most significant heterochrony distinguishing craniofacial development in eutherian and metatherian mammals.
One hypothesis for the whole-scale shift in timing is that it is a result of the intersection of the requirements of two fundamental developmental processes. The first requirement arises out of the sensitivity of nervous tissue to energetic fluctuations during differentiation. Neural tissue is highly expensive to construct and to maintain, requiring not only significant energy, but also a large number of specific nutritional elements (e.g., Dobbing 1972; Winick et al. 1972; Cheek 1975; Winick 1976, 1979; Dhopeschwarkar 1983; Herschkowitz 1989). Numerous studies have shown that the absence of these elements during organogenesis of the brain can lead to long-lasting neural deficiencies (Dodge et al. 1975; Shoemaker and Bloom 1977; Hetzel and Smith 1981). In eutherian mammals, the onset of neurogenesis and the initial period of growth of the CNS begins early, when there is little competition from other tissues, and growth and differentiation extend throughout the fetal and embryonic periods, when nutrition is relatively constant.
The second set of requirements arises out of cellular processes at the onset of skeletogenesis. It long has been known that many of the processes most important in determining the form of both the cartilaginous and osseous skeleton occur at the membranous or precondensation stage. Gruneberg (1963) documented numerous cases of skeletal malformation that can be traced to defects in the allocation or patterning of cells at this stage. Of greatest significance in the present context is the fact that many skeletogenic processes require a minimum condensation size before skeletogenesis will even proceed (Atchley and Hall 1991; Hall 1991; Hall and Miyake 1992; Dunlop and Hall 1995; Miyake et al. 1996). Further, Atchley and Hall (1991) argue that the initial condensation size is the most critical factor in determining the later form of skeletal elements. Thus the very processes of skeletogenesis require that the developing embryo allocate a minimum number of cells, and presumably energy, to the somatic system at the very earliest stage of differentiation.
In marsupials these two processes intersect because the period between the onset of organogenesis and birth is exceedingly compressed. In the short period of intrauterine organogenesis, which occurs roughly between the appearance of the primitive streak and birth (which averages six days and ranges from nine days in some macropodids and less than three days in dasyurids; Tyndale-Biscoe and Renfree 1987) apparently all available resources are directed toward systems most critical for independent survival: the mouth and feeding system, the forelimb, and the organ systems that must be at least minimally functional at birth (circulatory, respiratory, digestive, and excretory). The skeletal elements at least require specific and minimal cellular investment before histogenesis may proceed. In marsupials the bulk of neurogenesis is delayed in favor of the structures most critical for immediate survival and function of the neonate. Virtually all neurogenesis occurs after birth, during lactation.
Sacher and Staffeldt (1974) have previously hypothesized that neural tissue is the rate-limiting tissue during development in eutherians. They demonstrate that gestation times in eutherian mammals are best predicted by brain weight at birth, and show that while it is possible to extend gestation beyond that predicted by brain weight, few animals have gestations shorter than would be predicted by the limiting factor of brain growth. Sacher and Staffeldt (1974) present three general hypotheses about eutherian gestation: (1) the brain is the slowest-growing organ in eutherians; (2) the brain is the pacemaker for growth of other somatic tissues; and (3) brain growth proceeds at the maximum rate allowed by its nutrition and intrinsic growth parameters. They show that at least in eutherians data on development support these hypotheses. If these hypotheses are general for therians, it follows that in metatherians the extremely short gestation period, the need to achieve minimal functionality of somatic systems, the necessity for sufficient allocation to these systems for morphogenesis, and the rate-limiting nature of neurogenesis are incompatible. If so, marsupials avoid the constraints arising from these competing demands by shifting the bulk of neural differentiation to the postnatal period, where somatic growth may be slowed once more.
Finally, it is possible that the exceedingly long period of lactation in marsupials is a response to the shifting of neurogenesis to this period. Nelson (1988) has suggested that the nutrient composition of milk may be more subject to fluctuations in maternal diet than are nutrients delivered through intrauterine exchange. He further suggests that neurogenesis during lactation is vulnerable to these nutritional fluctuations, including caloric intake, trace element requirements, and the ingestion of possible toxins. However, it is also possible that the slow rate of differentiation and growth of neural tissue that characterizes marsupials represents a response to such fluctuations. Cheek (1975) has shown that in taxa with very slow rates of neurogenesis, such as higher primates (including humans), there appears to be less long-term neural deficiency arising from nutritional fluctuations (Cheek 1975). Therefore an extended period of neurogenesis may protect the brain from nutritional fluctuations because any specific short period is less critical. Thus the slow and extended neurogenesis that characterizes marsupials may serve to buffer the brain during development even if nutrient fluctuations appear during the lactational period; this buffering effect may be one of the reasons for the relatively long postnatal period of marsupials.
When the major shifts in the relative timing of development identified in the current study are viewed together with the basic requirements of morphogenesis of the nervous and somatic systems, the differences between marsupials and placentals can be fully appreciated. Differences between the marsupial and the placental development are not sufficiently explained by the particular adaptations of the highly altricial neonate, because this does not explain the major shifts in the relative timing of somatic and neural tissues. Short gestation could merely lead to altricial young, and not a developmental pattern in which major systems were differentially advanced and delayed. The shifts in relative developmental timing are best viewed as the intersection of an underlying and general constraint on the rate of neurogenesis, coming into direct conflict with the requirements for rapid development of some somatic systems - because both processes require specific and minimal resource allocations from the embryo - in the context of the exceedingly short period of organogenesis in marsupials.
The Primitive Condition
Because data are not as yet available on taxa beyond the Eutheria and the Metatheria, it is impossible to infer the primitive condition. Events in marsupials that are advanced or accelerated relative to eutherians have been identified, but no evolutionary polarity can be determined and the same events could be described for eutherians as delayed or decelerated relative to marsupials. It is commonly assumed that the metatherian condition resembles the primitive condition (e.g., Lillegraven et. al. 1987; Maier 1993), although others believe it is derived and that the primitive condition is a more intermediate one (e.g., Kirsch 1977a,b; Hayssen et al. 1985; Tyndale-Biscoe and Renfree 1987; Cockburn 1989; Renfree 1993). The primitive reproductive pattern of therians is of interest because without a view of the primitive condition, it is difficult to cast the heterochronies, or shifts in the relative sequence of events, in the context of any evolutionary scenario. At least two possible scenarios exist. First it is possible that marsupials have delayed the onset of neurogenesis relative to a primitive condition in response to a changing reproductive strategy that emphasizes very short gestation. Alternatively, it is possible that eutherians, released from some kind of gestational constraint, may have advanced the onset of neurogenesis relative to the primitive condition. Once the primitive condition is identified, it may be possible to distinguish between these hypotheses and then identify the kinds of constraints or adaptations that have emerged during therian evolution.
As it is unlikely that characters providing definitive data on the reproductive condition of primitive therians will ever appear in the fossil record, inferences on the primitive condition might best be understood in a study of extant outgroups. The current study, in which new data on the correlates of the reproductive pattern (i.e., developmental sequence) are provided and analyzed in an explicitly phylogenetic context, offers new characters for such an outgroup analysis. Monotremes are the obvious choice for future research as all other extant outgroups to the Theria are distantly related. Further, it is usually assumed that the oviparous monotreme reproductive pattern is primitive relative to the viviparity seen in extant therians. The data presented here provide for specific predictions about the condition of monotremes. If the marsupial condition is most like the primitive condition, then monotremes should resemble metatherians in the accelerated development of somatic tissues relative to the CNS. The opposite would be predicted if the marsupial condition were derived and placentals preserved a more primitive condition: considerable CNS development would be expected to precede development of facial structures in monotremes. Examination of the patterns of neural and somatic morphogenesis in additional outgroups (e.g., Aves, Chelonia, Squamata, or Crocodylia) would aid in determining major patterns of cranial development and evolution in amniotes. However, these groups have each undergone considerable independent evolution since their last common ancestor with mammals (which lived over 300 M.Y.B.P.; Carroll 1988) and thus cannot be assumed to be representative of a primitive condition.
Variation within the Metatheria and Eutheria
Although marsupials are relatively uniform in their degree of development at birth and the group does not exhibit the wide range of altricial and precocial young seen in eutherians, there are differences in the relative level of neonatal maturation (Hughes and Hall 1988). In D. viverrinus and other dasyurids differentiation is minimal at birth: the eyes are barely visible, retinal pigmentation is absent, and the oral apparatus is poorly defined. In P. nasuta and M. domestica the general level of maturation is intermediate and the above structures are more mature at birth. Macropus eugenii is most advanced at birth. These three levels, which Hughes and Hall (1988) label G1, G2, and G3, generally correlate with the time from the formation of the primitive streak to birth (the period of organogenesis) and with neonatal size. The levels of maturation of craniofacial structures at birth observed in these animals in the present is consistent with these grades; there do not, however, appear to be any shifts in developmental sequence that distinguish these levels.
Examination of craniofacial features in the newborn Dasyurus seems to confirm that at birth the level of differentiation is probably the bare minimum for survival. For example, the posterior part of Meckel's cartilage and the auditory ossicles (forming the jaw articulation) have not yet differentiated, pigment is sparsely present in the retina, if at all, and the telencephalon has only just begun evagination. The cranial muscles are also rudimentary, and although tongue and first arch muscles are differentiated at birth, these muscles are composed of only a few fibers surrounded by undifferentiated mesenchyme. However even this level of development is remarkable: dasyurids pass from the primitive streak stage to birth in approximately 3-4 days. Dasyurids are exceptionally small at birth (as a percentage of maternal weight) and may require special adaptations such as the specialized cervical swelling to attach to the teat (e.g., Hill and Hill 1955; Cockburn 1989; Nelson 1992).
Over one-half of the cases in which one marsupial species varied from the rest in the current study involved D. viverrinus. One interpretation of this observation is that these differences are due to the extreme altricial condition. But examination of the specific characters differentiating Dasyurus shows that they largely consist of characters that appear very early in the sequence and that are stage 1 (same time or unresolved) in the other marsupials, but are resolved in Dasyurus. Because Dasyurus is the sole marsupial taxon in which specimens were available that preceded the first developmental event, it is the only taxon in which it is certain that the beginning of the sequence is completely resolved. I believe that the differences between Dasyurus and other taxa are largely an artifact of sampling, although resolution of this issue awaits further data on early events in other taxa.
Eutherians are fairly uniform in the sequence of morphogenesis with the exception of Mus. Of 77 cases in which one taxon differed from all other eutherians, 46 involved Mus. Twenty-nine of these 46 characters reflect an advanced onset of ossification of endochondral bones relative to other structures of the head - skeletal and nonskeletal. The characters separating mice from other eutherians also reflect exceptional overlap in the period of ossification of cranial bones (i.e., they all began ossification within a very short period and there is significant overlap of the events of the somatic systems considered middle and late in [ILLUSTRATION FOR FIGURE 4 OMITTED]). Fifteen characters reflect relatively advanced maturation of the cranial skeleton in Mus relative to the maturation of craniofacial muscles. It therefore appears that craniofacial development, in particular cranial skeletal development, is exceptionally rapid and compressed in Mus. This compression also appears to be true of several other murid rodents such as Cricetus and Rattus (Smith, unpubl. data). It is not likely that Mus differs from other eutherians because of sampling differences, as Mus was the best sampled of all eutherians: both ends of the sequence were fully resolved and the sampling interval was even and fine.
The fact that Mus and other murid rodents differ from other eutherians may relate to the rapid development in these taxa. Myomorph (and sciuromorph) rodents have the fastest rates of development in any eutherian (data summarized in Eisenberg 1981). This applies to total gestation period, time from primitive streak to birth, and also the period from conception to eye opening (which corrects for the relative altriciality of the neonatal mouse or hamster), and holds true both absolutely and relative to body size. This rapid development appears to have produced exceptional compression and overlap of developmental events.
The observation that craniofacial development in Mus is not typical for eutherians is of general significance because the vast majority of knowledge about development in mammals is derived from observational, experimental, molecular, and genetic studies of Mus. However, if the course or timing of development in Mus is different from other eutherians, any generalization about craniofacial development derived from Mus must be cautiously interpreted. It is particularly important to be aware of the fact that events are compressed in Mus and occur with an exceptional amount of temporal overlap. Some aspects, such as the ossification of certain bones and the relations between skeletal and muscular development that may appear to be simultaneous or spatially or temporally linked in Mus may not be spatially or temporally related in other eutherians. Such spatial and temporal relations are particularly important in hypotheses of epigenetic mechanisms, and if the timing of cranial development in Mus differs from other mammals, then hypotheses about such relations derived from Mus may not be easily extended to other taxa.
Relations between Craniofacial Elements during Development
The discussion thus far has focused on the implications of these results for understanding development and evolution in marsupials and placentals. However, these data can also be used to investigate the mechanisms and patterning of craniofacial development in therians. Alberch (1985) pointed out that only causal sequences - sequences in which the occurrence of one event is mechanistically related to another - will be expected to be conserved in evolution. The analysis of developmental sequence within a phylogenetic context allows the identification of conserved elements and may provide new data on potential mechanistic relations during development.
In this context, the shifting of CNS development relative to differentiation of cranial skeletal elements is of greatest interest. Many studies have demonstrated the central role of CNS tissues in cranial skeletal morphogenesis. These include possible influences of the CNS on the prepatterning of cartilage (Thorogood et al. 1986; Thorogood 1988; Wood et al. 1991), CNS involvement in the induction of membranous bone of the calvaria (Schowing 1961, 1968; Tyler 1978, 1983, 1988; Hall 1984, 1987), and also mechanical relations between the growing brain and the growing cranium (e.g., Moss 1968; Moss and Salentijin 1969; Hanken 1983; Hall and Herring 1990; Herring 1993a,b). If it is true that mechanically linked events retain conservative sequence relations, then the skeletal elements that retain a temporal association with the CNS during development across therians, despite the overall shifts in timing of the CNS, might be causally linked to CNS development. Conversely, those skeletal elements that have shifted their development relative to the CNS are not likely to be linked mechanistically to the CNS. The relative patterns of linkage and independence can point to particular relations for possible experimental studies. For example, the relatively late development of CNS tissues in marsupials is accompanied by a relatively slow ossification of the membrane bones of the cranial roof. These results corroborate data from a wide variety of sources suggesting that the bones of the cranium are integrated with the CNS through the mechanisms listed above (e.g., Hall 1987; Hanken and Thorogood 1993; Herring 1993a,b and references therein).
The exoccipital bone, however, has achieved developmental independence from the rest of the bones surrounding the cranium. It begins ossification early in marsupials (at the same time as many of the bones of the face), well in advance of CNS differentiation or ossification of other elements of the braincase. In eutherians it begins ossification with the other endochondral bones of the braincase, after much CNS differentiation has proceeded. It is the only bone to exhibit major shifts in the ossification sequence in marsupials and placentals (see also Clark and Smith 1993). From this we can conclude that in marsupials its ossification is not determined by the CNS, but instead is under the influence of local mechanical forces. Specifically, it appears that ossification is due to a mechanical influence imposed by cervical muscles, which are important in supporting the head during the migration to the teat, and also while the neonate is attached to the teat. Thus these local forces have overridden the dominating effect of the CNS in the ossification of this single bone of the braincase.
A relation of particular interest in this context is the CNS and the patterning of the cartilages of the braincase. One of the important shifts in relative sequence was the advancement of the condensation of the cartilages of the cranial base relative to all measured events in the CNS. Previous studies have suggested that the CNS is responsible for laying down a matrix that forms the prepattern of the cartilage of the cranial base (e.g., Thorogood 1988; Wood et al. 1991). The data presented here would lead to the prediction that differentiation of the specific neural tissues responsible for this prepatterning would be accelerated in metatherians relative to other neural tissues. In this manner, the results of phylogenetic comparisons of sequence may be useful in designing and testing hypotheses on fundamental relations between tissues, organs, and structures during development.
I thank W. M. Kier, P. Mabee, A. van Nievelt, and C. Wall for comments on earlier drafts of this paper. I am particularly grateful to D. Noden for providing access to the comparative embryology collection at the Cornell College of Veterinary Medicine, Ithaca, New York; J. Bluemink, for access to the collections at the Hubrecht Laboratory of the International Embryological Institute in Utrecht, The Netherlands; M. Renfree of Melbourne University, Australia, for providing specimens of Macropus eugenii; and to C. Schnurr and A. van Nievelt for valuable assistance in all stages of this work. This project was supported by National Science Foundation grants DEB 9208514 and IBN 9407616 and funds from the Duke University Arts and Sciences Research Council.
ALBERCH, P. 1985. Problems with the interpretation of developmental sequences. Syst. Zool. 34:46-58.
ALBERCH, P., S. J. GOULD, G. F. OSTER, AND D. B. WAKE. 1979. Size and shape in ontogeny and phylogeny. Paleobiology 5:296-317.
ATCHLEY, W. R., AND B. K. HALL. 1991. A model for development and evolution of complex morphological structures. Biol. Rev. 66:101-157.
BRIDGE, D. T., AND D. ALLBROOK. 1970. Growth of striated muscle in an Australian marsupial (Setonix brachyurus). J. Anat. 106: 285-295.
BUTLER, H., AND B. J. J. JUURLINK. 1987. An atlas for staging mammalian and chick embryos. CRC Press, Boca Raton, FL.
CARROLL, R. L. 1988. Vertebrate paleontology and evolution. Freeman, New York.
CHEEK, D. B. 1975. Fetal and postnatal cellular growth. Wiley, New York.
CLARK, C. T. 1990. A comparative study of cranial skeletal ontogeny in two marsupials, Monodelphis domestica (Didelphidae) and Macropus eugenii (Macropodidae). Ph.D. diss. Duke Univ., Durham, NC.
CLARK, C. T., AND K. K. SMITH. 1993. Cranial osteogenesis in Monodelphis domestica (Didelphidae) and Macropus eugenii (Macropodidae). J. Morph. 215:119-149.
COCKBURN, A. 1989. Adaptive patterns in marsupial reproduction. Trends Ecol. Evol. 4:126-130.
CUBBAGE, C. C., AND P. M. MABEE. 1996. Development of the cranium and paired fins in the zebrafish Danio rerio (Ostariophysi, Cyprinidae). J. Morph. 229:121-160.
DHOPESHWARKAR, G. A. 1983. Nutrition and brain development. Plenum Press, New York.
DOBBING, J. 1972. Vulnerable periods in brain development. Pp. 9-20 in K. Elliot and J. Knight, eds. Lipids, malnutrition and the developing brain. Ciba Foundation Symposium. Vol. 3. Elsevier, Amsterdam, The Netherlands.
DODGE, P. R., A. L. PRENSKY, AND R. D. FEIGIN. 1975. Nutrition and the developing nervous system. Mosby Co., Saint Louis, MO.
DUNLOP, L.-L. T., AND B. K. HALL. 1995. Relationships between cellular condensation, preosteoblast formation and epithelialmesenchymal interactions in initiation of osteogenesis. Int. J. Devel. Bio. 39:357-371.
EISENBERG, J. R. 1981. The mammalian radiations. Univ. of Chicago Press, Chicago.
ESDAILE, P. C. 1916. On the structure and development of the skull and laryngeal cartilages of Perameles with notes on the cranial nerves. Phil. Trans. R. Soc., Ser. B Biol. Sci. 207:439-479.
FILAN, S. L. 1991. Development of the middle ear region in Monodelphis domestica (Marsupialia, Didelphidae): marsupial solutions to early birth. J. Zool., Lond. 225:577-588.
FRIGO, L., AND P. A. WOOLLEY. 1996. Development of the skeleton of the stripe-faced Dunnart, Sminthopsis macroura (Marsupialia: Dasyuridae). Aust. J. Zool. 44:155-164.
GEMMELL, R. T., AND J. NELSON. 1988. Ultrastructure of the olfactory system of three newborn marsupial species. Anat. Rec. 221:655-662.
-----. 1992. Development of the vestibular and auditory system of the northern native cat, Dasyurus hallucatus. Anat. Rec. 234: 136-143.
GEMMELL, R. T., AND L. SELWOOD. 1994. Structural development in the newborn marsupial, the stripe-faced dunnart, Sminthopsis macroura. Acta Anat. 149:1-12.
GEMMELL, R. T., G. JOHNSTON, AND M. M. BRYDEN. 1988. Osteogenesis in two marsupial species, the bandicoot Isoodon macrourus and the possum Trichosurus vulpecula. J. Anat. 159:155-164.
GRIBNAU, A. A. M., AND L. G. M. GEIJBERTS. 1981. Developmental stages in the Rhesus monkey (Macaca mulatta). Adv. Anat., Emb. Cell Bio. 68:1-84.
GRUNEBERG, H. 1963. The pathology of development. Wiley, New York.
HALL, B. K. 1984. Genetic and epigenetic control of connective tissues in the craniofacial structures. Birth Defects: Original Article Series 20:1-17.
-----. 1987. Tissue interactions in the development and evolution of the vertebrate head. Pp. 215-259 in P. F. A. Maderson, ed. Developmental and evolutionary aspects of the neural crest. Wiley-Interscience, New York.
-----. 1991. Cellular interactions during cartilage and bone development. J. Craniofac. Genet. Devel. Bio. 11:238-250.
HALL, B. K., AND S. W. HERRING. 1990. Paralysis and growth of the musculoskeletal system in the embryonic chick. J. Morph. 206:45-56.
HALL, B. K., AND T. MIYAKE. 1992. The membranous skeleton: the role of cell condensations in vertebrate skeletogenesis. Anat. Embry. 186:107-124.
HALL, L. S., AND R. L. HUGHES. 1987. An evolutionary perspective of structural adaptations for environmental perception and utilization by the neonatal marsupials Trichosurus vulpecula (Phalangeridae) and Didelphis virginiana (Didelphidae). Pp. 257-271 in M. Archer, ed. Possums and opossums: studies in evolution. Surrey Beatty and Sons, Sydney.
HANKEN, J. 1983. Miniaturization and its effects on cranial morphology in Plethodontid salamanders, genus Thorius (Amphibia: Plethodontidae). II. The fate of the brain and sense organs and their role in skull morphogenesis and evolution. J. Morph. 177: 255-268.
HANKEN, J., AND P. THOROGOOD. 1993. Evolution and development of the vertebrate skull: the role of pattern formation. Trends Ecol. Evol. 8:9-15.
HARTMAN, C G. 1919. Studies on the development of the opossum (Didelphis virginiana L.). J. Morph. 32:1-144.
HAYSSEN, V., R. C. LACY, AND P. J. PARKER. 1985. Metatherian reproduction: transitional or transcending? Am. Nat. 126:617-632.
HERRING, S. W. 1993a. Epigenetic and functional influences on skull growth. Pp. 153-206 in J. Hanken and B. K. Hall, eds. The skull. Vol. 1. Univ. of Chicago Press, Chicago.
-----. 1993b. Formation of the vertebrate face: epigenetic and functional influences Am. Zool. 33:472-483.
HERSCHKOWITZ, N. 1989. Brain development and nutrition Pp. 297-304 in P Evrard and A. Minkowski, eds. Developmental neurobiology. Vevey/Raven Press, New York.
HETZEL, B. S., AND R. M. SMITH. 1981. Fetal brain disorders. Elsevier/North Holland, Amsterdam, The Netherlands.
HILL, J. P., AND W. C. O. HILL. 1955. The growth stages of the pouch young of the native cat (Dasyurus viverrinus) together with observations on the anatomy of the newborn young. Trans Zool. Soc., Lond. 28:349-453.
HOLT, A. B., M. B. RENFREE, AND D. B. CHEEK. 1981. Comparative aspects of brain growth: a critical evaluation of mammalian species used in brain growth research with emphasis on the Tammar wallaby. Pp. 17-43 in B. S. Hetzel, and R. M. Smith, eds. Fetal brain disorders - recent approaches to the problem of mental deficiency. Elsevier/North-Holland, Amsterdam, The Netherlands.
HUGHES, R. L., AND L. S. HALL. 1984. Embryonic development in the common brushtail possum Trichosurus vulpecula. Pp. 197212 in A. P Smith and I. D. Hume eds. Possums and gliders Australian Mammal Society, Sydney.
-----. 1988. Structural adaptations of the newborn marsupial. Pp. 8-27 in C. H. Tyndale-Biscoe and P. A. Janssens. eds. The developing marsupial. Models for biomedical research. Springer, Berlin.
HUISMAN, F. J. 1933. Tabellarische uebersicht der Entwicklung von Manis javanica Desm. A. Oosthoek, Utrecht.
KIRSCH, J. A. W. 1977a. Biological aspects of the marsupial-placental dichotomy: a reply to Lillegraven. Evolution 31:898-900.
-----. 1977b. The six-percent solution: second thoughts on the adaptedness of the Marsupialia. Am. Sci. 65:276-288.
-----. 1977c. The comparative serology of Marsupialia, and a classification of marsupials. Aust. J. Zool., Suppl. 52:1-152.
KLIMA, M. 1987. Early development of the shoulder girdle and sternum in marsupials (Mammalia: Metatheria). Springer, Berlin.
KRAUSE, W. J., AND N. R. SAUNDERS. 1994. Brain growth and neocortical development in the opossum. Ann. Anat. 176:395-407.
LANGE, D. DE., AND H. F. NIERSTRASZ. 1932. Tabellarische uebersicht der Entwicklung von Tupaia javanica Horsf. A. Oosthoek, Utrecht.
LEE, A. K., AND A. COCKBURN. 1985. Evolutionary ecology of marsupials Cambridge Univ. Press, Cambridge.
LILLEGRAVEN, J. A. 1975 Biological considerations of the marsupial-placental dichotomy. Evolution 29:707-722.
-----. 1979. Reproduction in Mesozoic mammals. Pp. 259-276 in J. A. Lillegraven, Z. Kielan-Jaworowska, and W. A. Clemens, eds. Mesozoic mammals. Univ. of California Press, Berkeley.
-----. 1984. Why was there a "marsupial-placental dichotomy?" Pp. 72-86 in P. D. Gingerich and C. E. Badgley, eds. Mammals: notes for a short course. University of Tennessee studies in geology 8. Univ. of Tennessee, Knoxville.
LILLEGRAVEN, J. A., S. D. THOMPSON, B. K. MCNAB, AND J. L. PATTON. 1987. The origin of eutherian mammals. Biol. J. Linn. Soc. 32:281-336.
LUCKETT, W. P. 1994. Suprafamilial relationships within Marsupialia: resolution and discordance from multidisciplinary data. J. Mamm. Evol. 2:255-283.
MABEE, P. M. 1993. Phylogenetic interpretation of ontogenetic change: sorting out the actual and artifactual in an empirical case study of centrarchid fishes. Zool. J. Linn. Soc. 107:175-291.
MABEE, P. M., AND T. A. TRENDLER. 1996. Development of the cranium and paired fins in Betta splendens (Teleostei: Percomorpha): intraspecific variation and interspecific comparisons. J. Morph. 227:249-287.
MADDISON, W. P, AND D. R. MADDISON. 1992. MacClade. Vers. 3. Sinauer, Sunderland, MA.
MAIER, W. 1987a. The ontogenetic development of the orbitotemporal region in the skull of Monodelphis domestica (Didelphidae, Marsupialia), and the problem of the mammalian alisphenoid. Pp. 71-90 in J.-J. Zeller and U. Kuhn, eds. Morphogenesis of the mammalian skull. Paul Pary, Hamburg.
-----. 1987b. Der Processes angularis bei Monodelphis domestica (Didelphidae; Marsupialia) und seine Beziehungen zum Mittelohr: eine ontogenetische und evolutionsmorphologische Untersuchung. Gegen. Morph. Jahr. 133:123-161.
-----. 1993. Cranial morphology of the therian common ancestor, as suggested by the adaptations of neonatal marsupials. Pp. 165181 in F. S. Szalay, M. J. Novacek, and M. C. McKenna, eds. Mammal phylogeny - Mesozoic differentiation, multituberculates, monotremes, early therians and marsupials. Springer, New York.
MARSHALL, L. G., J. A. CASE, AND M. O. WOODBURNE. 1990. Phylogenetic relationships of the families of marsupials. Pp. 433502 in. H. H. Genoways, ed. Current mammalogy. Vol. 2. Plenum, New York.
MCCRADY, E. 1938. The embryology of the opossum. Memoirs of the Wistar Institution, Philadelphia, PA.
MCKINNEY, M. L. 1988. Heterochrony in evolution. Plenum, New York.
MIYAKE, T., A. M. CAMERON, AND B. K. HALL. 1996. Stage-specific onset of condensation and matrix deposition for Meckel's and other first arch cartilages in inbred C57B1/6 mice. J. Craniof. Genet. Devel. Biol. 16:32-47.
MOSS, M. L. 1968. A theoretical analysis of the functional matrix. Acta Biotheo. 18:195-202.
MOSS, M. L., AND L. SALENTIJIN. 1969. The primary role of functional matrices in facial growth. Am. J. Orthodon.s 55:566-577.
MULLER, F. 1972a. Zur stammesgeschichtlichen Veranderung der Eutheria-Ontogenesen. Versuch einer ubersicht aufgrund vergleichend morphologischer Studien an Marsupialia und Eutheria. I. Rev. Suisse Zool. 79:1-97.
-----. 1972b. Zur stammesgeschichtlichen Veranderung der Eutheria-Ontogenesen. Versuch einer ubersicht aufgrund vergleichend morphologischer Studien an Marsupialia und Eutheria. II. Rev. Suisse Zool. 79:501-566.
-----. 1972c. Zur stammesgeschichtlichen Veranderung der Eutheria-Ontogenesen. Versuch einer Ubersicht aufgrund vergleichend morphologischer Studien an Marsupialia und Eutheria. III. Rev. Suisse Zool. 79:567-611.
-----. 1973. Zur Stammesgeschichtlichen Veranderung der Eutheria-Ontogenesen. Versuch einer Ubersicht aufgrund vergleichend morphologischer Studien an Marsupialia und Eutheria. IV. Rev. Suisse Zool. 79:1599-1685.
NELSON, J. E. 1987. The early development of the eye of the pouchyoung of the marsupial Dasyurus hallucatus. Anat. Embry. 175: 387-398.
-----. 1988. Growth of the Brain. Pp. 86-100 in C. H. Tyndale-Biscoe and P. A. Janssens, eds. The developing marsupial. Springer, Berlin.
-----. 1992. Developmental staging in a marsupial Dasyurus hallucatus. Anat. Embry. 185:335-354.
NESSLINGER, C. L. 1956. Ossification centers and skeletal development in the postnatal Virginia opossum. J. Mamm. 37:382-394.
NODEN, D. M., AND A. DE LAHUNTA. 1985. The embryology of domestic animals. Williams and Wilkins, Baltimore, MD.
NOVACEK, M. J. 1990. Morphology, paleontology and the higher clades of mammals. Pp. 507-543 in H. H. Genoways, ed. Current mammalogy. Vol. 2. Plenum, New York.
NOVACEK, M. J., AND A. R. WYSS. 1986. Higher-level relationships of the recent eutherian orders: morphological evidence. Cladistics 2:257-287.
NOWAK, R. M., 1991. Walker's mammals of the world. 5th ed. Vol. 1. Johns Hopkins Univ. Press, Baltimore, MD.
PARKER, P. 1977. An ecological comparison of marsupial and placental patterns of reproduction. Pp. 273-286 in B. Stonehouse and D. Gilmore, eds. The biology of marsupials. Macmillan, London.
RENFREE, M. B. 1983. Marsupial reproduction: The choice between placentation and lactation. Pp. 1-29 in C. A. Finn, ed. Oxford reviews of reproductive biology. Vol. 5. Oxford Univ. Press, Oxford.
-----. 1993. Ontogeny, genetic control, and phylogeny of female reproduction in monotreme and therian mammals. Pp. 4-20 in F. S. Szalay, M. J. Novacek, and M. C. McKenna, eds. Mammal phylogeny - Mesozoic, differentiation, multituberculates, monotremes, early therians and marsupials. Springer, New York.
RENFREE, M. B., A. B. HOLT, S. W. GREEN, J. P. CARR, AND D. B. CHEEK. 1982. Ontogeny of the brain in a marsupial (Macropus eugenii) throughout pouch life. Brain, Beha. Evol. 20:57-71.
REYNOLDS, M. L., AND N. R. SAUNDERS. 1988. Differentiation of the neocortex. Pp. 101-116 in C. H. Tyndale-Biscoe and P. A. Janssens, eds. The developing marsupial. Springer, Berlin.
RUSSELL, E. 1982a. Patterns of parental care and parental investment in marsupials. Biol. Rev. 57:423-486.
-----. 1982b. Parental investment and desertion of young in marsupials. Am. Nat. 119:744-748.
SACHER, G. A., AND E. F. STAFFELDT. 1974. Relation of gestation time to brain weight for placental mammals: implications for the theory of vertebrate growth. Am. Nat. 108:593-615.
SAUNDERS, N. R., E. ADAM, M. READER, AND K. MOLLGARD. 1989. Monodelphis domestica (gray short-tailed opossum): an accessible model for studies of early neocortical development. Anat. Embry. 180:227-236.
SCHOWING, J. 1961. Influence inductrice de l'encephale et de la chorde sur la morphogenese du squellette cranien chez l'embryon de Poulet. J. Embry. Exp. Morph. 9:326-334.
-----. 1968. Mise en evidence du role inducteur de l'encephale dans l'osteogenese du crane embryonnaire du poulet. J. Embry. Exp. Morph. 19:88-93.
SHARMAN, G. B. 1973. Adaptations of marsupial pouch young for extrauterine existence. Pp. 67-90 in C. R. Austin, ed. The mammalian fetus in vitro Chapman and Hall, London.
SHOEMAKER, W. J., AND F. E. BLOOM. 1977. Effect of undernutrition on brain morphology. Pp. 147-192 in. R. J. Wurtman and J. J. Wurtman eds. Nutrition and the brain. Vol. 2. Raven Press, New York.
SMITH, K. K. 1994. The development of craniofacial musculature in Monodelphis domestica (Didelphidae, Marsupialia). J. Morph. 222:149-173.
-----. 1996. Integration of craniofacial structures during development in mammals. Am. Zool. 36:70-79.
STRAUSS, R. E. 1990. Heterochronic variation in the developmental timing of cranial ossification in poeciliid fishes (Cryprinodontiformes). Evolution 44:1558-1567.
SZALAY, F. S. 1994. Evolutionary history of the marsupials and an analysis of osteological characters. Cambridge Univ. Press, Cambridge.
THEILER, K. 1989. The house mouse. Developmental and normal stages from fertilization to 4 weeks of age. Springer, Berlin.
THOMPSON, S. D. 1987. Body size, duration of parental care, and the intrinsic rate of natural increase in eutherian and metatherian mammals. Oecologia 71:201-209.
THOMPSON, S. D., AND M. E. NICOLL. 1986. Basal metabolic rate and energetics of reproduction in therian mammals. Nature, Lond. 321:690-693.
THOROGOOD, P. 1988. The developmental specification of the vertebrate skull. Development (supplement) 103:141-153.
THOROGOOD, P., J. BEE, AND K. VON DER MARK. 1986. Transient expression of collagen type II at epitheliomesenchymal interfaces during morphogenesis of the cartilaginous neurocranium. Devel. Biol. 116:497-509.
TYLER, M. S. 1978. Epithelial influences on membrane bone formation in the maxilla of the embryonic chick. Anat. Rec. 192: 225-234.
-----. 1983. Development of the frontal bone and cranial meninges in the embryonic chick: an experimental study of tissue interactions. Anat. Rec. 206:61-70
-----. 1988. Development of osteogenic and chondrogenic potentials along the mediolateral axis of the embryonic chick mandible. Arch. Oral Biol. 33:443-449.
TYNDALE-BISCOE, C. H. 1973. Life of marsupials. Edward Arnold, London.
TYNDALE-BISCOE, C. H., AND M. RENFREE. 1987. Reproductive physiology of marsupials. Cambridge Univ. Press, Cambridge.
VELHAGEN, W. A. 1995. A comparative study of cranial development in the thamnophiine snakes (Serpentes: Colubridae). Ph.D. diss. Duke Univ. Durham, NC.
WAKE, M. H., AND J. HANKEN. 1982. The development of the skull of Dermophis mexicanus (Amphibia: Gymnophiona), with comments on skull kinesis and amphibian relationships. J. Morph. 173:203-223.
WINICK, M. 1976. Malnutrition and brain development. Oxford Univ. Press, New York.
-----. 1979. Nutrition, pre- and postnatal development. Plenum, New York
WINICK, M., P. ROSSO, AND J. A. BRASEL. 1972. Malnutrition and cellular growth in the brain: existence of critical periods. Pp. 200-206 in Lipids, malnutrition and the developing brain, Ciba Foundation symposium. Elsevier/North Holland, Amsterdam.
WOOD, A., D. E. ASHHURST, A. CORBETT, AND P. THOROGOOD. 1991. The transient expression of type II collagen at tissue interfaces during mammalian craniofacial development. Development 111: 955-968.
|Printer friendly Cite/link Email Feedback|
|Author:||Smith, Kathleen K.|
|Date:||Oct 1, 1997|
|Previous Article:||Gene effects on a quantitative trait: two-locus epistatic effects measured at microsatellite markers and at estimated QTL.|
|Next Article:||Effects of pollen quantity on progeny vigor: evidence from the dessert mustard Lesquerella fendleri.|