Researchers in biology have often directed their efforts toward elucidating the origins of major phyla or classification groups. While we have paid the most attention to the larger questions of transitions between classes, we are also considerably interested in the origins of orders. Cladistic methodology demands that we identify synapomorphies that define different orders but, as Matt Cartmill pointed out in 1972, it has not always been easy to identify the adaptive shift that accompanied the origin of a new order. The class Mammalia has been particularly troublesome in this regard. The interpretation of primate origins is an especially good illustration of the relationship between the characters used to define the taxon and the adaptive zone reconstructed for the order.
Historically, we have considered the primates difficult to define. George Gaylord Simpson stated this explicitly when he wrote that no clear-cut diagnostic adaptation distinguishes primates from other “primitive” placental mammals. The problem is particularly well exemplified by the controversy between those who would assign the treeshrews to the primates, such as Sir Wilfrid E. Le Gros Clark, and those who would not, such as Robert D. Martin and Leigh Van Valen. Workers have sought to define the primates by a distinguishing complex of evolutionary trends instead of defining the order by a single anatomical feature. In The Antecedents of Man, Clark summarizes the evolutionary trends, including generalized limb structure with pentadactyly and retention of skeletal elements including the clavicle that are reduced or lost in other mammalian orders; mobile digits, especially the pollex and hallux, for grasping; presence of flat nails and sensitive tactile pads instead of compressed claws; reduction of the snout and olfactory apparatus; elaboration of the visual apparatus and development of binocular vision; and the enlargement and development of the brain, especially the cerebral cortex. John and Prue Napier suggested additional trends, including the development of truncal uprightness or orthogrady.
Arboreal Theory: Elliot Smith, Wood Jones, Clark
A major paradigm, the arboreal theory of primate origins, defines primates by a complex of characters that adapted them to arboreal life. Indeed, we can find virtually all of the trends listed in the preceding section in the writings of the first exponents of the arboreal theory. It was first formulated by Grafton Elliot Smith and his assistant, Frederic Wood Jones, in the early decades of the 20th century by the study of comparative anatomy.
Elliot Smith, a neuroanatomist, was interested in explaining the distinguishing features of primate brains. In his address at Dundee to the Anthropological Section of the British Association for the Advancement of Science, he stated that it was the evolution of the brain and the ability to learn that led to the origin of the mammals, and that we should pay particular attention to the development of the cerebral cortex in primate evolution. Elliot Smith observed that orders of mammals became successful specialists for modes of life that depended on flight, fast running, or aquatic existence. They lost, however, their primitive simplicity and plasticity of structure. In contrast, the primates did not become narrowly specialized. Elliot Smith considered treeshrews to be insectivores, but he began a tradition in physical anthropology by using them in reconstructing the adaptations that gave rise to the primates. He believed that olfaction was the dominant sense in early mammals, which were essentially terrestrial. Natural selection would favor reduction of the olfactory apparatus and increased development of the neocortex (neopallium) of the cerebrum in arboreal mammals. The arboreal existence favors the development of vision, touch, hearing, agility, and quickness of movement while limiting the utility of smell. Small arboreal early primates could maintain the plasticity of a generalized structure while their brains developed. Elliot Smith pointed to tarsiers as manifesting significant reduction in the size of the olfactory parts of the brain and an increase in the visual cortex of the neo-cortex. He argued that the entire neocortex was affected by the emphasis on vision rather than smell. The sense of touch also became enhanced, and this assisted vision in the conscious appreciation of the environment and in the performance of agile locomotor behaviors. Hearing increased in importance, and the corresponding portion of the cortex expanded. These changes, while increasing the size of the brain and increasing agility, would not modify the primitive characters of the limbs and body. Elliot Smith stated that the interaction of tactile and kinesthetic senses with vision developed the cortex and stimulated the process of specialization of a mechanism for regulating the action of the cerebral cortex (an organ of attention to efficiently manage the nervous centers controlling muscles of the body). He ultimately derived the hominin prefrontal area from reliance on vision rather than smell.
Elliot Smith used the figure reprinted here to support his thesis that the olfactory region became relatively reduced and the visual, tactile, motor, and auditory areas expanded during primate evolution. He related the expansion of the frontal (“prefrontal”) region to learning to perform skilled movements and stereoscopic vision. He stated that the changes are the result of adoption of arboreal habits.
Wood Jones, in his landmark Arboreal Man, drew his evidence from a wider range of anatomical regions than did Elliot Smith. He discussed sensory and cerebral topics in the latter chapters of his book, but was also concerned with skeletal and muscular evidence. His elaboration of Elliot Smith’s thesis reflected his interest in the anatomy of the limbs and emphasizes the functional differentiation of the fore- and hindlimbs. The ancestral primate was arboreal in the late Triassic, and its limbs had not become specialized for pronograde cursorial locomotion. Locomotion in the arboreal context favored use of the forelimb for grasping supports with the palm against the substrate, retention of a clavicle, and use of the hindlimb parallel to the axis of the vertebral column for supporting the animal. Wood Jones used the term emancipation of the forelimb to indicate that the forelimb was no longer used solely for support. Postulated reduction of olfaction, caused by arboreal habits, resulted in the reduction of the facial skeleton. The latter led to a drawing of the eyes toward the midline. The increased specialization of the limbs resulted in a more upright posture with associated changes in the gastrointestinal tract, reproductive organs, and axial skeleton. Indeed, Wood Jones considered arboreal uprightness to be basal to the development of human terrestrial uprightness: He explicitly stated that arboreal uprightness preceded terrestrial uprightness. In Man’s Place Among the Mammals, he further proposed that hominins are descended from an orthograde tarsier-like ancestor. To counter the objection that other orders of mammals include arboreal forms, many of which are dissimilar to primates in their adaptations, Wood Jones used two arguments: The first was that arboreal representatives of other orders are secondarily arboreal, that is, their ancestors were terrestrial, pronograde quadrupeds; in his view, this was not true of the primates. Second, his law of successful minimal specialization stated that the specializations of other forms of arboreal mammals (such as sloths) imperil their further “evolutionary progress”; in contrast, the primates show the “maximum of possibilities.” With his emphasis on the importance of a mobile, grasping forelimb, Wood Jones viewed the use of suspensory postures by the feet as a “pitfall of specialization” and stated that the evolution of a prehensile tail in some New World monkeys has prevented “real [evolutionary] progress.”
In the 1971 The Antecedents of Man, the influential anatomist Wilfred LeGros Clark advocates both the arboreal theory and the inclusion of treeshrews within the primates; these two topics are intimately related within Clark’s framework of primate evolution. He portrays the origin of the primates in the form of arboreal mammals similar to treeshrews during the late Cretaceous or Paleocene. Reflecting earlier arguments by Elliot Smith and Wood Jones, Clark develops the argument that primates are distinguished by increased complexity of general organization, especially in the brain, rather than the elaboration of specializations. The latter was the route taken by more terrestrial mammalian orders in contrast to primates. Clark views primitive insectivores as tree-climbing animals with claws, nonprehensile hands and feet, small eyes and brains, and a well-developed olfactory apparatus. The hindlimbs and (especially) the forelimbs of primates preserve an ancient simplicity of structure and function, although the ability to cling and grasp has been increased by a wider range of movement possible at the shoulder, greater rotatory movements between the radius and ulna, a flexible ankle joint, increased mobility of the digits, and development of digital friction pads. Orbital convergence is considered to be advantageous in the arboreal milieu because it produces overlap of the two visual fields and therefore stereoscopic estimation of distance. Clark also points to the incomplete decussation of optic nerve fivers at the optic chiasm as important in primate stereoscopic vision. He differs from Wood Jones because the latter considers orbital convergence to be a secondary result of snout reduction rather than a feature that would be directly subject to selection. Emphasis on the selective advantage of orbital convergence was consistent with the reconstruction of early primate genera, such as Hemiacodon, Adapis, Notharctus and Nannopithex as vertical clingers and leapers and the conclusion that VCL was the dominant locomotor type in Eocene to Oligocene “prosimians” by John Napier and Alan Walker in the 1970s. Clark relates the evolution of flat nails to increased functional importance of terminal friction pads and believes that pads are more efficeint for grasping than are claws.
Visual Predation Hypothesis: A Challenge to the Theory
A major challenge to the arboreal theory is the visual predation hypothesis, explicated in the doctoral dissertation and series of papers by Matt Cartmill. Cartmill develops the hypothesis by comparing extant arboreal mammals with regard to limb morphology and loco-motion, orbital convergence, and olfactory regression. He believes that it is unclear that primate-like morphology is the most advantageous to arboreal life and states that it may be disadvantageous in certain contexts. He argues, for example, that the degree of orbital convergence manifested in certain primates decreases parallax and would not be valuable in leaping between branches; the slow-climbing, insectivorous slender loris and slow loris actually have more closely approximated orbits than the leaping galagos.
Cartmill observes that we find primate-like character states in marsupial taxa that hunt and manually capture insects among slender branches. He advocates a primate common ancestor that resembled the extant marsupial “dormouse” possum, mountain pygmy possum, and mouse opposum foraging in forest understory. Major components of his model are the deductions that the prehensile hindfoot is the only shared locomotor specialization in primates and the functional differentiation of the fore- and hindlimbs in arboreal mammals often works in the opposite direction than that postulated by Wood Jones. Cartmill then argues that prehensile ability in the hindfoot allows well-controlled and judicious movements on slender branches during hunting. Visual convergence and correlated neurological spe-cializations are predatory adaptations and are similar to adaptations seen in felids and owls, allowing the predators to gauge distance to the prey without moving their heads. Cartmill considers claws to be a hindrance to a bush animal that grasps thin supports by opposition of preaxial and postaxial digits and infrequently climbs on larger surfaces. Simplification of the olfactory apparatus is the result of the approximation of the medial walls of the two orbits; this conclusion is more similar to that of Elliot Smith and Wood Jones than it is to that of Clark. He states that other lineages of visually oriented predators have acquired comparable visual field overlap without significant olfactory reduction.
Both the classic arboreal hypothesis and the visual predation hypothesis examine two basic ideas: that “arboreality” is the niche of early primates and that primate morphology is the best morphology for an arboreal mammal—ideally and uniquely suiting primates for arboreal existence. Elliot Smith and Wood Jones conclude that these statements are true, but Cartmill determines that primate morphology is not the best morphology for an arboreal mammal and therefore seeks to present an alternative explanation for the suite of primate features. Both schools appear to show a limited concept of niche. As G. Evelyn Hutchinson argued in 1958, the niche of an organism is multidimensional. A single tree presents a number of niches for animals depending on factors such as their body size, mobility, and dietary requirements. There is, therefore, no single arboreal niche, and different morphologies may be equally successful in the arboreal milieu and reflect different specific niches. In this context, the visual predation hypothesis may be considered a modification of the classic arboreal theory rather than a completely distinct paradigm. Martin has asserted that a more inclusive answer for primate adaptations is the occupation of a fine-branch niche by small- to medium-bodied, nocturnal, and actively foraging early primates. Robert Sussman also advocates a terminal branch setting for early primates in his mixed diet theory: well developed visual acuity and color vision and grasping feet and hands would be favorable adaptations for primates foraging in the terminal branches of the successful and active radiation of angiosperms.
Many of the fossil undoubted primates that had earlier been identified as vertical clingers and leapers have more recently been reconstructed as quadrupedal (for example, subfamily Adapinae): Although quadrupedal primates often leap between branches, there is not the emphasis on recurrent leaping that would be found in dedicated VCL forms. Selection for leaping ability per se would not figure as prominently in the interpretation of morphological characters.
An important question is whether ancestral primates and other fine-branch mammals (particularly, the ancestors of marsupial taxa discussed by Cartmill) converged in their locomotor adaptations or whether the niche was occupied by their common Jurassic ancestor. Cartmill concludes that there is no evidence that primate-like adaptations were present in Triassic mammalian ancestors. South Africa has produced a complete skeleton of the late Triassic Megazostrodon, interpreted as shrew-like in habits and with a grasping hindlimb. The recent description by Qiang Ji and colleagues of Eomaia scansoria is that of a 125-million-year-old eutherian adapted for climbing. As John Wible and his collaborators point out, Cretaceous eutherians have usually been reconstructed as plantigrade, terrestrial, or scansorial quadrupeds. Based on features of the hands, feet, scapula, wrist, vertebral column, and claws, Eomaia is capable of grasping and branch walking. Ji and colleagues conclude that Eomaia was scansorial (similar to tree shrew Tupaia) or fully arboreal (similar to tree shrew Ptilocercus), that most basal metatherians were scansorial, and that scansorial skeletal factors are primitive for the earliest known eutherians. There is data that primitive marsupials were also climbers. Homoplasy is a persistent problem; for example, there is evidence that the Paleocene multituberculate Ptilodus may have been arboreal with a prehensile tail, but the status of multitu-berculates as Mesozoic mammals is controversial, and Ji and colleagues maintain there was a lot of convergent evolution in the postcrania of early mammals.
A major concern with the models is that the morphological comparisons are often large-scale, that is, between representatives of different genera or orders, rather than between closely related species or populations that differ in anatomical character states and in substrate preference. Even more crucially, the behavioral/ecological information used in formulating models was often anecdotal. Clear understanding of the behavioral/ecological significance of morphological features is best attained by planned behavioral and experimental study that is driven by the goal of testing specific hypotheses. For example, experimental work by B. Demes and colleagues at the State University of New York at Stony Brook on the issue of “hindlimb drive” (recall that Wood Jones emphasized the functional differentiation of the forelimbs and hindlimbs) indicates that primates are variable in their use of hindlimbs for propulsion and that several varieties of nonprimate mammals are “hindlimb driven.” In fact, they conclude that quadrupedal species generally have equal forces on the fore- and hindlimbs.
- Cartmill, M. (1972). Arboreal adaptations and the origin of the order primates. In R. Tuttle (Ed.), Functional and evolutionary biology of primates, pp. 97-122. Chicago: Aldine-Atherton.
- Cartmill, M. (1974). Pads and claws in arboreal locomotion. In F. A. Jenkins (Ed.), Primate locomotion, pp. 45-83. New York: Academic Press.
- Cartmill, M. (1974). Rethinking primate origins. Science, 184, 436-443.
- Clark, W. E. L. G. (1971). The antecedents of man, 3rd edition. Chicago: Quadrangle Books.
- Jones, F. W. (1918). Arboreal man. London: Arnold.
- Jones, F. W. (1929). Man’s place among the mammals.London: Arnold.
- Martin, R. D. (1990). Primate origins and evolution: A phylogenetic reconstruction. Princeton, NJ:Princeton University Press.
- Smith, G. E. (1924). Evolution of man: Essays. London:
- Sussman, R. (1991). Primate origins and the evolution of angiosperms. Journal of Primatology, 23, 209-223.