Introduction

The title of this chapter has a double entendre embedded in it. It is a truism that biological history, in addition to ongoing adaptive demands, is decisive in shaping properties of lineages. But it is also uncontestable that precedent notions, influential contributors, or specific papers, right or wrong, channel and continue to profoundly influence thinking on many issues in science. There is a difference, however, in these two processes of canalization. In the evolutionary dynamic, there is no right or wrong, and the inherited attributes are the initial and boundary conditions that define the avenues open for subsequent phylogenetic/adaptive change. These paths do not only constrain but facilitate as well. At any rate, whales are not fish, so history fundamentally

Frederick S. Szalay • Anthropology, and Ecology and Evolutionary Biology, CUNY, and Department of Anthropology Hunter College, CUNY, 695 Park Avenue, New York 10021

matters in all biological science. Genotype encoded factors which sum up history, beyond the maternal contribution in the egg, guide the change, adaptive or not, which is phylogeny.

Such largely adaptive phylogeny is an ongoing probabilistic outcome of environmental demands, which determine the frequency of individuals that make it through the survival and reproductive bottlenecks of each generation. The necessity to consider this theoretical foundation should be, therefore, neither surprising nor burdensome for natural historians who study morphological attributes or behaviors. Sundry disciplines, in particular research from behavioral ecology, provide fundamentally important plausibility hypotheses for paleobiologists, who seek such questions as this conference set out to do. The task, however, to reconstruct adaptive phylogeny is within the realm of paleontologists and morphologists who must tie the fossil record, through a variety of procedures referred to as modeling (see Szalay and Sargis, 2001), to information and ideas from neontology. This is done by testing specific historical-narrative explanations (i.e., phylogeny or taxon hypotheses; see Figure 1 in Szalay, 2000) against various areas of information. Historical narratives of science are tested against evidence of all sorts (Bock, 1981)—an activity not indulged in by Kipling. So contrary to Popperian thinking, much of science consists of historical-narrative explanations offered within the confines of law-like explanations, in juxtaposition to Cartmill's (1990) opinion that only the law-like statements are scientific. Law-like statements must be part of the context within which the various topics of becoming are explained (Bock, 1981; Szalay and Bock, 1991). But nomological-deductive explanations (law-like statements) alone, obviously, do not suffice in any science where history played a role. It is all those specific and contextual historical "mistakes" in the law-like workings of chemistry that result in consequences for replication, transcription, and translation of nucleic acids where the science of evolutionary change begins and couples with the vicissitudes of the environment.

The aim of this chapter is relatively straightforward, but because of space constraints, it is more of a review of some literature debates and an outline of some issues related to the origin of both the Plesiadapiformes and Euprimates (perhaps best considered as sister orders at present) rather than detailed documentation. To achieve these goals I will (1) examine, in a historical framework, selected examples of hypotheses in which conceptual methods, as well as empirical emphasis or de-emphasis, have had a significant role in the construction of these hypotheses, as well as their consequences, for analyzing the phylogeny of adaptations for plesiadapiforms and euprimates; (2) present my views on modeling in paleobiology and the testing of homology hypotheses; (3) remark on the evidence related to locomotor strategies of Cretaceous and some recent therians that are relevant to the assessment of the ancestral pattern in the Eutheria, and of a clade within that group, the Placentalia; and (4) reassert the importance of the "morphotype locomotor mode" concept as a critical connection between phylogeny estimation and adaptational (functional, in a broad sense) assessment. As an example, I point to some evidence from hard anatomy for the ancestral euprimate locomotor mode. The latter was first referred to as "grasp-leaping" in Szalay and Delson (1979) and subsequently more fully developed in Szalay and Dagosto (1980, 1988). R. H. Crompton (1995) appears to strongly second this view.

The Placentalia, diagnosed elsewhere based on tarsal attributes and four premolars, is the taxon that stems from (back in the Cretaceous) the last common ancestor of the Cenozoic and surviving eutherians. This is not just the living crown group because it includes now extinct orders as well. The corresponding stem group of the Eutheria from which the Placentalia arose is the paraphyletic Eoeutheria that diverged from Metatheria at least 125 MYA.

GLIMPSES OF HISTORY OF RESEARCHES REGARDING ARCHONTAN, PLESIADAPIFORM, AND EUPRIMATE MORPHOTYPE LOCOMOTOR STRATEGIES, AND THEIR INFLUENCE

The customary empirical efforts to study extant forms and fossils often break down into two approaches, the functional (in a broad sense) and the phylo-genetic (see Szalay, 2000, for review). As a consequence, the conceptual methods that should guide the analysis of the various facets of a problem become simplified either into functional undertakings or synapomorphy sorting through parsimony analysis, or other phyletic approaches. In addition, it is not unusual at all for many scientists leaning in one or another direction regarding morphological analysis to completely barricade themselves into either of these two, often walled-off, compartments, stating they are not really interested in the "other" questions. This is not an exaggerated rendering of the state of affairs, particularly either for functional anatomy or parsimony cladistics-based studies, with the subsequent distortion of the questions and a loss of the evolutionary explanation that one is interested in.

There have been valiant undertakings to somehow combine function and taxonomic position in one fell swoop, although some of these quantitative efforts have, unfortunately, resulted in such empirical conflation of data that all that followed were strikingly visual "species stamps," rather than any illumination of the role of heritage in the evolution of functional complexes (e.g., Oxnard et al., 1990). While the aim of these studies was laudable, the setting aside of the complex but feasible and complementary interrelationship of functional-adaptive and phylogenetic methods for the analysis of evolutionary origins (problems of transformation from one stage to another in the history of lineages) suffered, or simply was not part of the analysis.

As attested to by this conference and many others before it, primates generate great interest among an inordinately large number of natural historians of all sorts, morphologists among them. This is understandable but it makes for an enormous literature, and extremes of conceptual approaches to problems of adaptation of ancestral conditions and disagreements about the specifics of an ancestral lineage. At this conference (and before), for example, I or Dagosto viewed the ancestral euprimate as the phyletic antecedent of the reasonably well-known Eocene strep-sirhines and haplorhines, whereas others considered a cheirogaleid such as Microcebus as a stand-in for this ancestor. In spite of such a difference in perception, which is almost never explicitly stated, what is less understandable is how several past contributions on the deep adaptive history of primates were based sometimes on a lack of expertise in evolutionary morphology, on highly selected literature contributions, or on a neglect of the specifics of extant species. Some of these publications were often by primatologists, who have written about bones and fossils with little experience either in the theoretical issues surrounding evolutionary analysis of morphology or the fossil record. This state of affairs, however, has considerably improved recently due to competitive pressures resulting from an upwelling of young talent specializing in these complex and intertwined fields of analysis. But the past has shown its powerful constraint on the collective minds of a whole subfield. Some textbooks and reviews have helped to perpetuate uncorroborated ideas about locomotor inferences regarding protoplacentalians, plesiadapiforms, and the stem euprimates. In the review given in later section, I will comment briefly on some such examples.

Arboreality as a Novel Strategy for the Stem of Archonta

Over and beyond the obvious specifics of arboreal heritage in the morphology of living primates, in the 1960s, the debate over this heritage has entered a new phase with an admittedly confusing framework for considering primates with or without the archaic primarily Paleocene radiation of the Plesiadapiformes. While the evidence now is overwhelming regarding the arboreality of plesi-adapiforms (and their close phyletic ties to euprimates without the interference of dermopterans; see Bloch and Boyer, 2001; Bloch et al., 2000, 2001a,b,c 2002; Boyer and Bloch, 2000; Boyer et al., 2001), the history of the literature regarding archontans, tupaiids, and plesiadapiforms is highly instructive.

The initial and widely read impetus (if one was to start somewhere in a quasi-historical assessment such as this) that euprimates owe their particular morphological (and functional in a broad sense) divergence from their ancestry due to a particular locomotor behavior that involved leaping, was the contribution of Napier and Walker (1967)—a study that advocated vertical clinging and leaping as the initial stage of euprimate locomotor evolution. This restatement of previous views on leaping but with greater force and examples were significant because they went beyond the customarily evoked arboreality as an explanation for euprimate attributes. Much of the development of the insight regarding leaping in euprimate ancestry was largely due to the seminal studies of Walker (1967) on the subfossil and extant osteology of the Malagasy strepsirhines. The extended debate about vertical clinging and leaping that ensued is interesting history, but not directly relevant here. The theoretical underpinnings of the ecomorphological assessment of the vertical clinging Malagasy lemurs, galagos, and tarsiers were obviously sound. Much of the following debate focused, correctly, on the applicability of those conclusions to the fossil postcranial morphology, the area of anatomy that should have been logically the most significant for locomotor assessment of the fossil record. But that is not what happened.

Visual Predation as the Strategy for the Stem Lineage of Euprimates

Cartmill (1972) has presented the ambitious "visual-predation hypothesis" based on cranial attributes and grasping hands that was to explain the whole diagnostic structural make up of the protoeuprimate, and, at the same time, came to de-emphasize not arboreal locomotion as such, but the importance of grasping related leaping that shaped this ancestor. The whole argument was an attack on the straw man of "arboreality," without any consideration given to the multitude of ways that adaptations may be required to fulfill various kinds of positional regimes, arboreal or otherwise. The fact that clawed hands work very well in all predatory mammals, from opossums to cats as a tool for prey capture, however detected, was coupled with the need to climb cautiously and grasp tightly on small branches. With an emphasis on grasping hands and the loss of the claws coordinated with stereoscopy and appropriate neurology, Cartmill has relegated the powerful and larger grasping hindfeet (compared to the hands) as a means to allow "...to move cautiously up to insect prey and hold securely onto narrow supports when using both hands to catch the prey" (p. 440). What was largely missing from this overarching hypothesis is the accounting for the skeletal evidence known by then for a number of early euprimate lineages. Cartmill (1975) further developed his views along similar lines. It should be emphasized here that Cartmill's (1972, 1974) views (or those of Hamrick, 1998) regarding the reduction of claws are not supported by the targeted selective loss of the falcula on the hallux in didelphids and descendants. The correlation in extant marsupials appears to be with the powerful grasp of the pes and a postulated selective disadvantage of the sharp falcula on the hallux on smaller branches.

What also complicates matters of historical reconstruction regarding the evolution of various published perspectives on primate morphotype locomotion, and the implicit assumptions that these views rest on, is the apparent inconsistency of some published views. Issues of phylogeny, latent in any adaptive hypothesis, but almost always implicit when they should be explicit, point to some critical inconsistencies in the presentation of the visual predation hypothesis of Cartmill. For example, Cartmill (1974: 74) has given confusing testimony about the historical context of his views on claw "loss" (part of a transformation series of the homologues called digital ungulae) in the protoeuprimate. In fact, Cartmill (1974) and later Hershkovitz (1977) have strongly supported the transformation of falculae (claws) into the tegulae of platyrrhines independently from other euprimates, with Cartmill, in particular, arguing for "greatest parsimony." This is particularly puzzling because Cartmill's "visual-predation" hypothesis launched in 1972, and expanded in 1975 was critically dependent on the assumption of a nailed condition in the euprimate stem. Regarding the loss of claws in euprimate ancestry Cartmill (1974: 74) says that: "The comparative anatomical evidence indicates that the hands and feet of the last common ancestor of the extant primates [i.e., Euprimates] must have resembled those of the opossum; claws have been lost independently in four or five parallel lineages of primates." On the same page further down Cartmill explicitly supports the notion that claw loss can be the result of a "...trend toward increased size in animals inhabiting the higher strata of tropical forest, or from the restriction to the lower strata of a relatively treeless heath or scrub floral community." It is also relevant here that Lewis (1989, based on a series of articles published in 1980) explicitly supported an arboreal ancestry for the last common ancestor of the fossil and living placentalian mammals—a view that Martin (1990) has continued to champion. This appears to be decidedly untrue for the Placentalia, and probably also for the stem of the Early Cretaceous Eutheria as well.

To put it bluntly, contrary to pronouncements, the comparative anatomical evidence never "indicates" anything; one explicitly tests and interprets homology hypotheses, which Cartmill did in an unacceptable way (see detailed discussion of this in Szalay, 1981b: 40-44). But the most striking feature at that time, given the "visual-predation" hypothesis (which one might have thought was based on a homology-based phylogenetic position, i.e., "claw-loss" and postorbital bars) was the concept of parallelisms in Cartmill's theoretical and historical-narrative explanations (i.e., the recurrence of parallel trends in the evolution of euprimates).

Added to this, I believe, was a connection to the "Plesitarsioidea" versus "Anthrolemuroidea" view of primate phylogeny, an interesting historical curiosity (Gingerich, 1974; 1975a,b; see also Krishtalka and Schwartz, 1978; and Schwartz et al., 1978) which is relevant here. This view of primate phylogeny, which posited an unacceptable wedding (then or now) of the plesiadapiforms and one of the early euprimate groups (the Tarsiiformes) as a clade, represented at that time a significant manifestation of primate evolutionary studies in contrast to the strepsirhine-haplorhine dichotomy advocated by others. The disregard for the very accessible postcranial evidence of fossils (Szalay et al., 1975) and the extant postcranial osteology by both the proponents of the "visual-predation" hypothesis and the taxonomic notion of the "Plesitarsiiformes" (this latter derived from, and synonymous with, the "Plesitarsioidea") points out that postcranial attributes (at the level perceived by these authors) were considered (if examined at all) as rife with "parallelisms," hence not very reliable.

But subsequent to Cartmill (1972), Szalay and Decker (1974), and Szalay et al. (1975) have assessed the then known skeletal collections of Plesiadapis

(the former study emphasizing the tarsus the latter the remainder of the skeleton) and concluded that the only reasonable explanation of the evidence was unquestionable arboreality for the archaic plesiadapiforms. Szalay and Decker (1974), Decker and Szalay (1974), and Szalay et al. (1975) emphasized in particular both the similarities to but also the differences in arboreal adaptations in the tarsus between plesiadapiforms and euprimates in contrast to latest Cretaceous eutherians. These conclusions were dismissed as doubtful by Cartmill (1975: 32), without any indication that he considered the evidence. But then, it appears, that Cartmill was wedded to the notion that arboreality was primitive for the ancestry of living placentalian mammals, and therefore, his attacks on the arboreal theory of primate origins, as he called it, were justified only on the grounds that attributes related to other than some specific arboreal locomotion were necessary to explain the origin of both the Plesiadapiformes and the Euprimates. Martin (1990), in his text also insisted that the plesiadapiforms simply retained arboreal modifications already present in a remote placental ancestor. This unfortunate disregard for the fossil evidence (dubbed as "special problems of the fossil record" by Martin, 1986: 4) was also evident earlier. [Martin's statement (1986: 23) about Plesiadapis that its hallux "might have been totally lacking," is particularly revealing in light of the fact that in the same volume Gingerich illustrates and makes a note about the preserved big toe, suggesting a lack of familiarity with the record. Yet, this unfamiliarity with the specifics of fossil evidence did not prevent that author to present high profile discourse about fossil primates elsewhere as well (see for example Martin, 1993)].

There can be little doubt that there was a nearly complete disconnect between phylogenetic thinking and adaptive assessment by Cartmill (1975: 32-33) when one reads that "[if] the characteristic primate traits are the result of progressive adaptation for arboreal visual predation in one line of descent from an early plesiadapoid... thrusting the plesiadapoids...back into the ancestral order Insectivora would make the order Primates more coherent, However, we must not forget...[that]...If, for instance, it turns out that anaptomorphids arose from very early paromomyids, while adapids evolved separately out of the earliest plesiadapids, it might still prove true that the lines leading to the Eocene families went through an adaptive shift to visual predation, in parallel in two different lineages...". It is difficult to see how the more complex areas of the skeleton, particularly the carpus and tarsus, failed to convince these authors both about the unequivocal arboreality in the plesiadapiform ancestry and the unquestioned monophyly of the Euprimates, except if one considers the overwhelming "scenario" bias by Cartmill and a then prevalent dental mindset by Gingerich and associates. After the widely available postcranial evidence had been repeatedly pointed out in the literature in the 1970s and 1980s (Szalay, 1972) the polyphyletic notion of the euprimates was finally abandoned.

It should be added here that parsimony (a useful notion if properly applied to not only relevant "facts" but to all the complex interpretations necessary in the construction of tested hypotheses regarding properties) was much used then as it is now. Such procedures, however, rapidly (and properly) turn into a series of Bayesian considerations. This is an approach not much appreciated by Popperian systematists in primatology who became advocates of a falsifica-tionist approach to cladogeny based on algorithm research, as opposed to an incremental research program leading to phylogeny estimation (e.g., Szalay, 2000). The unfortunate reality has been, however, that either erudite and literary rhetoric about scenarios or unexamined character lists require more than "parsimonious thinking" or scholastic Aristotelian logic (algorithmic or not) for nonmonotonic testing procedures in evolutionary morphology and the testing of historical-narrative explanations. The arguments about plesi-adapiform and euprimate relationships and adaptations, and the methods of assessment, continued in the literature.

Kay and Cartmill (1977: 19) in their restudy of a crushed skull of the Torrejonian Paleocene Palaechthon concluded that while euprimates were derived from plesiadapiforms, the cranial adaptations of the latter (exemplifying primitive plesiadapiforms) reflect a "...predominantly terrestrial insect-eater, guided largely by tactile, auditory, and olfactory sensation in its pursuit of prey." Even more interestingly (and in stark contradiction to Cartmill's views on ancestral placental arboreality), they noted that "Adaptations to living in trees and feeding on plants probably developed in parallel in more than one lineage descended from the ancestral plesiadapoids." It was pointed out subsequently in a critique by Szalay (1981a: 157) that Kay and Cartmill in their analysis of the cranial evidence based their conclusions regarding plesi-adapiform adaptations on: (a) nonphylogenetic and static assumptions, (b) misinterpretation of the form and mechanics of the attributes analyzed, and (c) employment of irrelevant characters for the establishment of substrate preference (e.g., infraorbital foramen size). Szalay criticized the general outlines for adaptational analysis espoused by Kay and Cartmill, and the positions taken by these contending parties on the type of character choices and functional interpretations are still the general positions that endure in many debates today. Namely, in dealing with fossils, how should one approach the difficult issue of adaptational assessment (see Szalay, 2000, contra the arguments offered by Anthony and Kay, 1993: 374)?

My arguments in 1981 were in juxtaposition to the practice of indiscriminate use of ancestral characters that could be correlated with some habitat in living animals (e.g., the relative size of the infraorbital foramen in archaic primates used by Kay and Cartmill, 1977, to argue for terrestriality in paro-momyid plesiadapiforms). While the persistence of functional correlates of even primitive traits can be useful in framing an adaptational analysis, primitive traits are often revealing of ancestrally acquired adaptations within a different context. The human thorax, shoulder complex, and elbow joint are good examples. These heritage traits, a group's synapomorphies, set the limits for various trajectories of the more derived features. For example, the contact of the fibula with the femur, and also via the parafibula (the fibular fabella), correlates only with some aspect of therian primitiveness in the knee complex, but no ecologically meaningful differentiating function can be associated with it in marsupials that show different habits today. Both the most arboreal and terrestrial marsupials have this as part of the knee complex, although instructively, with different conformation of the proximal fibula. The extreme narrowing of the proximal fibula (and attendant muscular and mechanical correlates) occurs only in highly terrestrial metatherians (see later section). Similarly, the repeated narrowing of the lateral femoral condyle in terrestrial didelphids, bandicoots, basal, and all other kangaroos, as well as in the ancestral placentalian, also closely predicts terrestriality (Szalay and Sargis, 2001). But the narrowing of the proximal fibula that also occurred in proto-placentalians does not rewiden again in Cenozoic and recent arboreal euthe-rians, nor does the medial femoral condyle changes its proportions. The extant eutherian lineages (and their fossil relatives which postdate the stem of these) are likely all derived from the terrestrially modified eutherian that was the stem of the Placentalia.

The Role of Leaping in the Ancestral Euprimate

By 1979, Szalay and Delson noted that the likely breakthrough from an arboreal plesiadapiform ancestry (unequivocally suggested as such by Szalay and Decker, 1974, and corroborated beyond any reasonable expectation by the efforts of Bloch and Boyer) involved "...the establishment of grasp-leaping arboreal adaptations ...necessitated by a particular feeding regime" (p. 99) for the stem of the strepsirhines, considered by them to be the best approximation of the euprimate stem. Szalay and Dagosto (1980) in their extended discussion of what they defined as morphotype locomotor modes (a concept which incorporated a phylogenetic context into the assessments of locomotor behavior/anatomy) have discussed claw-climbing as reflected in the pro-toplesiadapiform condition. They also emphasized in some detail that the interpretation of skeletal features strongly supports grasp-leaping as a mono-phyletic acquisition of the protoeuprimate. They essentially agreed with Le Gros Clark (1959) that arboreal locomotion (but a particular type) was likely part of the causal nexus of the cranial features one observes in the Eocene primates—a foundation on which modern diversity is based. They disagreed with Cartmill's hypothesis, and stated that "The greater importance and more severe selec-tional consequences of judging distances by quadrumanous fast grasp-leapers would clearly put a greater premium on stereoscopy than just running and walking along branches in an arboreal environment. There is no evidence for uniquely associating quadrumanous primate grasp-leaping with arboreal insec-tivory-omnivory. The first euprimate grasp-leaper may or may not have been primarily phytophagous, zoophagous, or ominivorous." (p. 35).

In 1992, Cartmill reviewed, with candor, the differences between the grasp-leaping and the visual predation hypothesis as contributing causal factors in the development of the protoeuprimate cranioskeletal complex, although he continued to think of "arboreality" as some monolithic causal agent. He correctly cites my often-stated view (following those of, e.g., Darwin, Gregory, Matthew, and Simpson, and others'), namely that evolutionary transformations are constrained by history in a highly contingent way, and that the new adaptive solutions mirror that heritage, often to a considerable degree. This view, in light of the prevalence of mosaic evolution (bolstered by an understanding of modularity by students of EvoDevo), demands character level, rather than a taxic, analysis of homologies (the former dubbed as null-group comparison; Szalay, 1994; Szalay and Bock, 1991). In order to arrive at reasonable phylogenetic estimates of character complexes (and subsequently taxon phylogeny hypotheses), the development, functional biology, and adaptation of taxonomic properties need to be considered, in contrast to the declared primacy of algorithm-based rooting with taxic outgroups.

Nevertheless, following our debates of the extant and fossil evidence regarding the ancestral stage of euprimate locomotion, Cartmill has come to consider the issue of phyletics of characters and even the notion of (Darwinian, i.e., evolutionary) homology somewhat moot points (see Cartmill, 1994, on the issue of homology hypotheses; and Cartmill, 1990, for his rejection of historical-narrative explanations as science). In arguing against the grasp-leaping euprimate locomotor mode, Cartmill (1992: 107) noted that "...particular evolutionary events cannot in principle be explained except as instances of some more general regularity," and also stated that "...adaptation to a grasp-leaping habit unique to euprimates, explains nothing." He has professed this belief in a variety of ways, in fact arguing against the very practice of historical-narrative explanations in science. I (and others) completely reject such ahistorical theoretical assumptions about the nature of science.

Cartmill (1992: 107) was correct in stating that other arboreal mammals "...do not look much like euprimates." Of course, few other arboreal mammals (with their independent heritage) do the acrobatic antics of those grasp-leaping lemuriforms whose general skeletal anatomy shows the same derived suite of features that can be reasonably attributed to the protoeuprimates as well. And those skeletal attributes appear to be diagnostic of the order based on the Eocene evidence (i.e., they represent a derived suit of features of the stem). But Cartmill's (1992) discussion of the issue of the euprimate morphotype locomotor mode, including his evaluation of the proposals of Sussman (1995) and Rasmussen (1990) were, in my view, deeply flawed. This was so not only on the theoretical grounds regarding his perspective on how one employs living model species to evaluate fossil animals (e.g., Szalay, 1981a,b; Szalay and Sargis, 2001; and later section). But perhaps more importantly than anything else, Cartmill continued to make only casual, if any, use of the highly specific and functionally well-understood aspects of postcranial morphology for interpreting the fossil postcranial evidence when discussing locomotion in the euprimate stem. This is odd enough by itself, but the postulate that (rapid and frequent) leaping and precise landing by grasping small branches has obvious consequences for both the nervous system and vision should not have been ignored. Habitual great leaping ability in the three-dimensional arboreal environment would certainly suggest a causal relationship to enhanced vision and attendant neurology. And to consider the reduction of the snout, olfaction is far less important for the execution of a leap than visually judging distance and points of landing among variable-sized branches for animals, whose size we cannot be certain of. Nevertheless, the issue remains a particular type of arboreal locomotion (grasp-leaping), not just "arboreality," and testing of that issue resides primarily in the mechanics of the joints of the skeleton of an inferred common ancestor and their near-fossil relatives.

The general area of modeling ecological morphology and its use for fossil species (see later section) is a lot more complex but also far more applicable than Cartmill's (1992: 107) statement that only parallelisms can be explained adaptively. For example, Szalay (1981a) argued against the thesis presented by Kay and Cartmill (1977) that large infraorbital foramina of the plesiadapiform Palaechthon pointed to a terrestrial, hedgehog-like habitus. I pointed out the difficulties of judging habitus (real-time adaptation in a species) based on primitive features because primitive features, while perfectly functional (obviously), do not reflect the most recent shifts in a lineage, unlike their derived attributes. Convergences of complex derived attributes of recognized mechanical consequences, however, are powerful "postdictors" of the habitus of fossil species, and are the most potent tests of historical-narrative explanations. I showed that relatively very large infraorbital foramina persist in some very arboreal species. Therefore, such features simply cannot be very useful in interpreting fossils, "parallelism" aside. Rather instructively, the size of various foramina continues to have a rather checkered history in predicting anything, including scenarios pertaining to the hominid realm.

It is exactly the rejection of the analyzed, ordered, and polarized use of character states of homologous features that is missing from the notion of "parallelism" dictated by Cartmill's views on homology. Is one's assessment of parallelism the result of parsimony analysis? Are we considering some convergent aspects of features, given distinct phylogenetic/taxonomic contexts? For establishing convergence (a tested, and failed homology hypothesis, without the somewhat obfuscating discussion and mixing of levels of organization by Lockwood and Fleagle, 1999), however, one should have some criteria other than the leftover traits expressed as a consequence of "CI" indices of parsimony-derived taxograms. The notion of convergence that Cartmill subscribes to in his pledge to taxic analysis as the arbiter of the nature of similarities is, ipso facto—a residue of a "losing batch of synapomorphies" that one now calls "convergent" (see Szalay, 2000). But beyond how homology is established with some probability, there is the key issue of what particular convergent/parallel properties one is going to employ to explain a particular facet of adaptational history or a fossil species.

R. H. Crompton (1995) has presented a detailed analysis of the literature (albeit with some studied omissions) regarding the origin of feeding and locomotor strategies of the euprimate ancestor. He has paid laudable attention to the connection that must exist between feeding and locomotor strategies. His conflation of the arboreal and scansorial strategy that was suggested for the plesiadapiforms by Szalay (1972) is taken by him as that for the protoeuprimate, one that is a minor lapsus by a primatologist with little practice in sys-tematics or acquaintance with the fossils. What is, however, a recurring pattern in his critique of Cartmill (as well as in Cartmill's own previous contribution) is the consistent lack of attention paid to the details of the fossil dental and postcranial evidence. The circular "chop" diagrams of "total adaptive strategies" of various extant primate species published by Oxnard et al. (1990) are hardly a substitute for the independent assessment of the relevant fossil or even extant evidence. Unfortunately, a remark by Crompton (1995: 19) that a general arboreal form of locomotion "...is typical of many small, primitive mammals...," has less meaning than no statement at all. Within the even conventionally accepted concept of Mammalia, different groups undoubtedly had different primitive locomotor patterns (i.e., morphotype locomotor modes) with their attendant morphological properties that are amenable to specific model-based analysis (see later section).

[One would, in general, hope for the recognition by students of living primate ecology that feeding and locomotor strategies are primarily reflected in the morphologies of the relevant regions of the hard anatomy. Furthermore, it is appropriate to state here that feeding and locomotion can be decoupled not only in terms of morphological mosaic evolution, but also in terms of various solutions for the feeding/locomotion dilemma faced by all lineages. Nothing better exemplifies the mosaic nature of adaptive solutions than the variety of strategies seen within the lorisiforms—a group cited repeatedly by R. H. Crompton.]

Contrary to Crompton's statement (1995: 21), which is relevant here, there is no morphological evidence of any sort that would suggest dwarfism in the ancestry of the living tarsiers, only perhaps if one assumes a large-bodied hap-lorhine ancestry. As I noted earlier, we cannot be certain of the size of either euprimate, strepsirhine, or haplorhine ancestries, even though great antiquity does tend to preserve some aspects of morphology that indicate general functional features. Tarsiers are well within the size range of the group—which they are a relict of—the fossil Tarsiiformes of the Eocene. Their enormous eyes are probably a reflection of the compensation required by secondary nocturnality in a probably diurnal lineage that has shed the tapetum lucidum in its ancestry. But in his conclusions, Crompton (1995) seems to agree with the locomotor mode designation that was proposed by Szalay and Dagosto (1980, 1988) as grasp-leaping, and which was specifically tied to ancestral euprimate postcranial morphology and its inferred biological role (Crompton did not use that term, nor did he cite the 1980 article). Crompton's conclusions certainly corroborate those of Szalay and Dagosto (1980, 1988). Crompton, in spite of his strong disagreement with Cartmill, however, goes on to endorse the dietetic component of Cartmill's visual predation hypothesis. Unfortunately, there is no evidence from the dentition of the earliest euprimates, or from the best estimates of the adaptations of the morphotype of euprimates, that insectivory and predation were the preponderant ancestral dietary strategy. The postcranium and inferred leaping is neutral on that important question. The variety of dental pattern is great, however, so inference as to diet is at best a variety of fruit, flower, nectar, gum, and insect feeding, with no clear-cut emphasis in any reconstructed common ancestor.

It must be stressed that early dietary strategies in the protoeuprimate are not as yet understood, in spite of the often-cited deductive argument of Kay (1984) based on the body weight and diet of living primates, asserting that size is a predictor of diet. Body size is also often inferred from fossil teeth themselves, often a poor measure. According to that view, small fossil primates were, ipso facto, primarily insectivores—a gross oversimplification even on general grounds restricted to living primates as models. Assertions that because some small living lipotyphlans or primates are primarily insectivorous, all small fossil primates had to be as well, are divorced from morphological analysis. Many small fossil primates (as well as marsupials) with the appropriate dental and cranial attributes were probably oblivious to "Kay's rule" (contra Kay and Covert 1984) when it came to their dietary regimes. Similarly, the extant Hapalemur and Lepilemur, or even cheirogaleids, do not adhere to such a rule. Morphological and functional patterns, in light of the appropriate models (but not size alone) supply convincing paleobiological explanations. As argued before (Sussman, 1995; Szalay, 1968, 1969, 1972), the dental evidence leaves little doubt that among early plesiadapiforms and eupri-mates, a mixed feeding strategy, evidenced by relatively low crowned and quasi-bunodont cheek teeth was likely to be both the ancestral and one of the more widespread conditions. One has to look no further than the variety of small rodents who find ample energy and nourishment primarily from seed consumption, ignoring this general "rule." Small fossil primates were not necessarily obligate insectivores unless their morphology corroborates such assessment.

It is also of some importance to note here the relevant point that contrary to Martin (1993), the radiation of a mammalian group is not usually that of an algorithm-based inverted pyramid, and therefore, the living radiation of primates is a poor foundation to model the early story that was driven by ecological context and biogeography, in spite of the putative elegance of such iconography. Given the enormously more extensive favorable habitats for primates in the Paleogene of Holarctica (and probably Africa as well), experimentation of many early lineages among the euprimates probably resulted in a far greater diversity of small omnivorous primates than there is today. An understanding of the fossil record helps in this regard. Massive extinctions with the changing of habitats have resulted in a pattern nearly the opposite of the computer-generated diagram of Martin (1993).

Sussman (1995) and I are in broad agreement on the importance of fru-givory early in primate evolution. Regarding the close relationship of habitat and primate strategies, it is perhaps important to note that primates did not "invent the rainforest," although they certainly carry on the roles started by other clades. At least in South America, where primates did not arrive until relatively late in the Tertiary (and certainly never in Australia and New Guinea until humans ventured there), the radiation of arboreal marsupials was well under way since the Latest Cretaceous or Earliest Paleocene in tropical rainforest environments of South America, and sometime later in the antipodes. And even prior to that, a variety of atribosphenic mammals undoubtedly interacted in a number of ways with the tropical forests and angiosperms of that continent. It should be emphasized that the derived suit of postcranial traits of the stem euprimate certainly does not preclude a reliance on fruits, flowers, gums, or seeds, together with insects as the main items of its dietary regime, although such a diet can be attained by a whole variety of ways other than grasp-leaping. The most corroborated explanation for the morphotype skeletal evidence, however, is a regular practice of bounding leaps and landing with a hindfoot/forefoot grasp ("grasp-leaping"). But such interpretation, of course, does not mean that an animal with such morphology cannot slowly climb, walk, shamble, or in any other way get to its food, or stalk insects. But leaping does make a particular combination of energetic and competitive sense when particulate and discontinuously distributed clumps of food are sought after by many parties, both intra- and interspecific.

It is gratifying indeed that the general idea of grasp-leaping as the eupri-mate morphotype locomotor mode is so thoroughly circumscribed and argued for and advocated in all but name (i.e., without reference to the article by Szalay and Dagosto, 1980, where the hypothesis was first explicitly outlined and supported) by R. H. Crompton (1995).

0 0

Post a comment