On the Origins of New Forms of Life

9.12: Marsupials

EUGENE M. MCCARTHY, PHD GENETICS

Site Search

(Continued from the previous page)

If the bones of a fossil animal are the same as those of an extant form classified as a marsupial, then one can confidently classify the fossil type as a marsupial. But such is not the case when a fossil form is now extinct. R. A. Barbour (1977: 243), an expert on marsupial anatomy, says

Even among living mammals, traits typical of marsupials are not found in every kind of marsupial. For example, the angle of the lower jaw is not always inflected in marsupials (Kielan-Jaworowska et al. 1979: 222), nor do all marsupials have epipubic bones (ibid.).¹ Therefore, when we look at the fossil remains of an extinct animal, we cannot be certain the bones in question are those of a marsupial. Even on the basis of soft anatomy, Marsupialia is a rather poorly defined category. For example, many forms classified as marsupials lack a pouch (see comments in Note 2 below).

And yet, when remains of extinct mammals are discovered in regions where marsupials predominate today (e.g., Australia and New Guinea), or are assumed to have predominated in times past (South America), it is often assumed the fossils in question are those of marsupials. Mammalian specimens found outside those regions are typically categorized as placental.³

In Section 9, we have sought plausible ancestors for each of the various mammalian types. Here, however, since the marsupial type is itself poorly defined, there will be no effort to identify such forms. We will merely attempt to gain some idea of when and where marsupial traits first began proliferating.

Some authors⁴ have suggested that marsupials first arose in Laurasia, which recall is the ancient precursor landmass that gave rise to North America and Eurasia. But this position seems untenable because:

1. outside the southern continent complex (Australian Region, Antarctica, South America), the only extant marsupials not introduced by human agency are the seven didelphid genera that invaded North America over the Isthmus of Panama within the last five million years;
2. in strata predating the collision of North and South America, the fossils of extant marsupials are known only from the southern complex;
3. extinct marsupials cannot be reliably identified as marsupials from their bony remains even when those remains are intact;
4. fossils from outside the southern complex, which are supposed to represent the earliest known marsupials, known almost exclusively from the late Cretaceous (Romer 1966: 379),⁵ are so extremely fragmentary we can in no way feel assured of their marsupial status.

As Clemens (1977: 192) points out,

And yet debate has centered, not upon whether these rudimentary specimens actually were marsupials, but rather upon the question of how the alleged descendants of these supposed early marsupials reached the southern continents.

Thus, the only firm evidence — that is, the geographic distribution of known (extant) marsupials and their fossils — is strongly consistent with the idea that marsupials never existed outside the southern continent complex until South and North America came into contact a few million years ago. That is, the only evidence carrying any weight suggests extant marsupial forms represent the remnants of a similarity set that was until recently restricted to the southern complex of formerly connected continents, and that appeared much later than did the earliest placental mammals (probably after Africa separated from South America and the rest of the southern continent complex in the middle Cretaceous). It was within this southern similarity set that the traits characteristic of marsupials first proliferated. Our line of reasoning thus far has strongly suggested placental mammals date back at least to the early Jurassic. In other words, the placental type appears to be ancient, while the marsupial type seems to be a relatively recent innovation.

Conclusion. The example (Class Mammalia) discussed in Section 9 suggests that numerous widely accepted notions should be discarded (such as "radiation" and “primitive” “generalized” forms). In particular, if the conclusions reached in this section are, in the main, correct, there is no reason to think reptiles predated mammals or that they "dominated" them during the Mesozoic. Numerous synapsids, many of them large, existed during the early Mesozoic—and even as early as the Permian—and these we may reasonably consider to be early mammals. Other forms, many of them large and "dominant," existed far back into the "reptile" age, and yet can be interpreted as ancestors of the various modern mammalian orders (e.g., mosasaurs, plesiosaurs, multituberculates, pterosaurs). Evidence (offered in the discussions of ungulates and proboscideans) suggests that early elephants and rhinoceroses, as large or larger than most dinosaurs, already existed in the early Cretaceous. We have already seen there is reason to suppose certain animals actually classified as dinosaurs (e.g., ankylosaurs, stegosaurids) can be construed as direct ancestors of creatures classified as mammals (e.g., armadillos, pangolins). Perhaps the most remarkable conclusion reached in this section is that the famed "extinction of the dinosaurs" appears in large part to be a matter of mistaken classification.

Available evidence seems to exclude the idea that mammalian evolution has been monophyletic (i.e., that all modern mammals are descended from a single, “generalized” ancestor living in the Cretaceous Period). It is certainly more plausible to assume each recent type has arisen from a similar ancient type (e.g., seals from earlier plesiosaurs, bats from pterosaurs). Many of these general types can be traced back to synapsid times. This perspective is consistent, not with divergence and adaptation to vacated "niches" (“adaptive radiation”) and gradual change in isolation, but with new forms of life arising from similar preexisting forms via stabilization processes, a kind of evolution that can be discussed without reference to “branching trees” and that can instead be conceptualized in terms of sets of related forms that give rise to new sets of similar forms. Instead of gradually evolving, isolated lineages (neo-Darwinism's perspective), under stabilization theory we can picture an evolving community of forms. Nor would there be any radiation of placental mammals from a common ancestor living in the late Cretaceous. In fact, the evidence suggests many of the major types of placental mammals arose long before the Cretaceous, in the Triassic or perhaps even in the late Paleozoic and that marsupials arose within a previously diversified placental similarity set. The relative positions of the southern continents in times past seem to indicate that marsupials arose much more recently, subsequent to the separation of Africa and South America (approx. 105 mya). Thus, if we judge only by relative times of origin, marsupials would be a more “advanced” than placentals.

In general, stabilization theory suggests we should seek to explain the origin of any given set of similar forms in terms of their derivation from some preexisting set of forms with similar traits. Under this view no radiation is expected. No single ancestral form is presumed. Instead each similarity set goes on evolving as it generates new member forms via stabilization processes (and as old members become extinct). In this respect, stabilization theory is far more plausible than neo-Darwinian theory, which makes far-fetched assertions about "generalized" ancestors being rapidly and radically altered into a variety of totally dissimilar types (e.g., shrew to whale, shrew to bat, shrew to deer, shrew to monkey, shrew to pangolin, shrew to pig, shrew to seal, etc.). The facts and arguments that have been offered in this section, what might be called a scientific taming of the shrew, make patent the implausibility of such claims. This theoretical emperor simply isn't wearing any clothes. NEXT PAGE >>

Notes (Works Cited):

1. Moreover, many early mammals (and not just those that have been posited as early marsupials) had epipubic bones (von Koenigswald and Storch 1992: 155). No marsupial has a baculum, but many placental mammals also lack this feature (e.g., Homo sapiens). True, a post-orbital bar is absent from the skull of most (perhaps all) marsupials, but the same is also true for a wide variety of placental mammals (e.g., it is typically absent in rodents, as well as in many edentates, carnivores, seals, sirenians, cetaceans, bats, perissodactyls, and even in some artiodactyls (e.g., wart hogs).

2. Even on the basis of soft anatomy, Marsupialia is a rather poorly defined category. The usual dictionary definition says marsupials are "pouched mammals." But, in some forms classified as marsupials, the pouch consists merely of small folds of skin near the teats, for example, certain didelphids and dasyurids (Walker 1983: 10). Other marsupials possess a pouch for only part of the year (some dasyurids); and at least one-third of all forms treated as marsupial species lack a pouch entirely (Kirsch 1977b; Walker 1983). According to Walker (1983) and Kirsch (1977a) these include the caenolestids, certain dasyurids, marmosids, monodelphids, and such genera as Caluromysiops, Caluromys, Glironia, Lutreolina, Metachirus, Myrmecobius, and Philander. In particular, the vast majority of American marsupial species lack this feature. Some marsupials even have longer gestation periods than those of some placentals. For example, rabbit (Oryctolagus): 28–33 days; rat (Rattus): 21–26 days; mouse (Mus): 19–21 days (Nowak 1999). The kowari (Dasyuroides byrnei), a marsupial mouse, has a gestation period of 30–36 days (Lee and Carrick 1989; Nowak 1999). A North American chipmunk, Tamias striatus, is the same size as a kowari, and its gestation period of 31 days (Nowak 1999) is comparable a kowari's . Another trait generally considered characteristic of marsupials, tenacious attachment of the young to the mother's nipples, is known to occur in a wide variety of (placental) rodent genera (Gilbert 1995). In addition to the genera listed by Gilbert, Walker (1983) lists the following genera as having this trait: Conilurus (p. 707), Leporillus (p. 709), Zysomys (p. 709), and Mastacomys (p. 716). Some sources say marsupials develop no placenta during the course of reproduction. But, as marsupial experts Hugh Tyndale-Biscoe and Marilyn Renfree (1987: 310) point out, "There has been much debate over whether marsupials possess a true placenta." (Ramsey 1982; Walker 1983: 39). The current consensus of opinion on this point seems to be that 1) early in the development of the embryo, a yolk-sac placenta forms in both placental mammals and marsupials, and 2) a chorio-allantoic placenta develops thereafter, not only in the typical placental mammal, but also in certain marsupials, for example, Peramelidae, Thylacomyidae, and Phascolarctos (Ramsey 1982: 19, 75; Walker 1983: 45; Strahan 1995: 167; Tyndale-Biscoe and Renfree (1987: 311). Additional marsupials may also have a chorio-allantoic placenta since little information — or even none at all—is available on the reproductive anatomy of most marsupials (Tyndale-Biscoe and Renfree 1987: 14).

3. Traditional mammalian classification has treated marsupials and placental mammals as two separate major categories within Class Mammalia, but the difficulties of definition just mentioned have led some recent workers (e.g., Nowak 1999) to drop the category Marsupialia altogether, to break up the forms traditionally assigned to that category into various orders, and simply to list them along with the various placental orders.

4. For example, Martin (1977).

5. It is true, however, that a few alleged marsupials dating to the early Tertiary have been found in the Old World (Crochet 1986; Gabunia et al. 1985; von Koenigswald 1982). With the exception of those described by von Koenigswald (1982), these are tiny, isolated teeth, about the size of a large grain of sand. On the other hand, von Koenigswald’s “opossums” from the Messel oil shale are largely complete, but have apparently only been identified as marsupial because they have an incisor count of 5/4 (von Koenigswald and Storch 1992: 155). Incisor counts are a shaky basis of diagnosis for two reasons: 1) Many, perhaps even most, modern marsupials do not have such a count (for example, the Virginia Opossum, Didelphis virginiana, has an incisor count of 3/3); 2). As Ziegler (1971) has emphasized, 5/4 was a common incisor count not just in early marsupials, but in all early mammals (on this basis he has suggested 5/4 was the primitive state). See also Barbour (1977: 242).

8.1: Dubious Assumptions - © Macroevolution.net