Atavisms: medical, genetic, and evolutionary implications

Traits expected to be lost in the evolutionary history of a species occasionally reappear apparently out of the blue. Such traits as extra nipples or tails in humans, hind limbs in whales, teeth in birds, or wings in wingless stick insects remind us that certain genetic information is not completely lost, but can be reactivated. Atavisms seem to violate one of the central evolutionary principles, known as Dollo's law, that "an organism is unable to return, even partially, to a previous stage already realized in the ranks of its ancestors." Although it is still not clear what triggers and controls the reactivation of dormant traits, atavisms are a challenge to evolutionary biologists and geneticists. This article presents some of the more striking examples of atavisms, discusses some of the currently controversial issues like human quadrupedalism, and reviews the progress made in explaining some of the mechanisms that can lead to atavistic features.     

Atavisms in homo sapiens: A bolkian heterodoxy revisited

An atavism is the ..reappearance of a lost character (morphology or behaviour) typical of remote ancestors and not seen in the parents or recent ancestors of the organisms displaying the atavistic character (Hall, 1984). In humans, hypertrichosis (extensive body hair), the presence of a tail and supernumerary nipples are often quoted as examples (Hall, 1995). However, Louis Bolk (1866–1930) explained these phenomena in another way. He considered human morphology as an unspecialized expression of the mammalian developmental pattern. The latter also encompasses potentialities for unilateral or propulsive development pathways (specializations) that usually remain latent in humans, but can become expressed in other species. According to Bolk, the appearance of so-called atavisms in humans results from the occasional expression of these latencies in Homo sapiens; they do not recapitulate ancestral conditions.     

Developmental basis of mammalian digit reduction: a case study in pigs

Digit reduction has occurred in parallel in many mammalian lineages. However, despite this pattern's prevalence, the developmental mechanisms underlying mammalian digit reduction remain controversial. We therefore undertook a study of digit development in the pig (Sus scrofa), a mammal with reduced first, second, and fifth digits. Our results indicate that from its earliest formation, the pig limb bud is significantly narrower than that of the model pentadactyl mammal, mouse. Furthermore, the cartilage condensations of the pig's reduced digits are noticeably smaller than those of their nonreduced counterparts from the time of their formation. In addition, growth rates of pig digits are comparable, as are the patterns of cell death in developing pig and mouse limbs. Taken together, results suggest that pig's first, second, and fifth digits are primarily reduced through evolutionary modifications in the early developmental patterning of their limbs. Results of this study, coupled with those from study of limb development in other mammals, suggest that although major developmental reorganizations (e.g., complete digit or limb loss) during early limb development may be selected against, it may be common for more subtle evolutionary modifications in limb development (e.g., changes in relative digit size) to occur at this time.     

CAENOGENESIS,DEVELOPMENTAL VARIABILITY,AND EVOLUTION IN THE CARPUS AND TARSUS OF THE MARBLED NEWT TRITURUS MARMORATUS

In this paper, after a comparative analysis of the development of Triturus marmoratus, we explore the existence of caenogenetic events and their ontogenetic and phylogenetic consequences. The adult morphology of the Triturus marmoratus limb, in terms of number and spatial arrangement of skeletal elements, agrees with the general pattern of urodeles. The congruence in the typical pattern of adult morphology does not hint at the striking differences in embryonic development. These differences can be summarized as follows: 1) Presence of a “central axis” that develops in a distal-to-proximal direction. It originates in the basale commune giving rise to the centrale and the intermedium. Thus, there is no postaxial branching as found in Ambystoma mexicanum. 2) Again, unlike in Ambystoma mexicanum, we find a postaxial structure composed of the ulnare (fibulare)-distal carpal (tarsal) 4-metacarpal (metatarsal) 4 which is independent of the “digital arch.” 3) The (forelimb) digits, in particular, digits 1, 2, and 3, undergo disproportionate elongation. For example, the second digit, composed of a thin continuous, cartilaginous rod, becomes longer than the rest of the limb. Our study of the patterns of embryonic connectivity suggests the coexistence of three directions of growth and morphogenesis in the development of the Triturus marmoratus limb. 1) A proximo-distal one that gives rise to the preaxial axis, “primary axis,” and individual digits. 2) An anterio-posterior axis of development that gives rise to the “digital arch” and determines the number of digits. 3) A disto-proximal central axis that originates in the basale commune and sequentially generates the centrale and the intermedium. We speculate that heterochronic interspecific variation in the time of onset of limb bud formation is related to the degree of precocious digital elongation. Selection for long extremities in early larval stages, associated with functional demands for locomotion and balancing, may be the cause for the above listed changes in developmental pattern. Thus, the reported system is an example of how selection during ontogeny can result in the evolution of the developmental process.     

Ancestral patterns in bird limb development: A new look at Hampé's experiment

The “Archaeopteryx limb” of experimentally treated bird embryos has become a standard quotation in the growing literature on developmental factors in evolution. It has not only been claimed that an early manipulation of the chick limb produces a series of atavistic skeletal features, but the experiment is also frequently interpreted within a genome-centered concept of atavism. The present study provides a morphological, quantitative, and comparative analysis of the skeletal and muscular reactions to the classic barrier insertion experiment of Hampé. The main result of this operation, traditionally seen as a “full length fibula”, is shown to be a relative effect due to tibia shortening, while all the other ancestral skeletal features, which are usually pointed out as being provoked by the elongated fibula, do not appear. The experimentally generated fibula/tibia length ratio and distance, however, mimic the pattern in developing reptilian limbs and are seen to induce secondary effects in the muscular system that are reminiscent of archosaur reptiles. Similar muscle patterns are found as interspecific variations in several bird species. The revised view of the skeletal changes and the additional data on muscular effects allow for a renewed interpretation of the experiment, shifting the emphasis from atavisms to the role of heterochrony, developmental integration, and epigenetic constraint in the evolutionary modification of organismic structures.     

DATA AND DATA INTERPRETATION IN THE STUDY OF LIMB EVOLUTION: A REPLY TO GALIS ET AL. ON THE REEVOLUTION OF DIGITS IN THE LIZARD GENUS BACHIA

Galis and collaborators (2010) claim that our recent paper ( Kohlsdorf and Wagner 2006 ), presenting statistical evidence for the reevolution of digits in the genus Bachia, may be flawed. Their reanalysis of the data does not support the possibility of a reevolution of digits and the authors also argue that such a reevolution would be implausible on functional and developmental grounds. In response, we reanalyzed our data with additional outgroup species. Our results differ from the one published in 2006, but this incongruence is not statistically significant. In contrast, the hypothesis presented by Galis et al. is significantly worse. An analysis of digit number evolution, using novel techniques to test for irreversibility of character loss ( Goldberg and Igic 2008 ), confirmed our original conclusion that there is strong evidence for reevolution of digits in Bachia. We also point out that this result is not in conflict with the hypothesis by Galis and Metz (2001) that mutations affecting the initial digit patterning are associated with strong negative pleiotropic effects and thus unlikely to be fixed in evolution. An important avenue of future research will be to directly test whether reevolved digits develop from conserved digit condensations retained after digit loss.     

EVOLUTIONARY NOVELTY AND ATAVISM IN THE SEMIONOTUS COMPLEX: RELAXED SELECTION DURING COLONIZATION OF AN EXPANDING LAKE

Fishes of the genus Semionotus diversified in the rift lakes of eastern North America during the Mesozoic (Newark Supergroup). Like the well-known cichlid fishes of the African great lakes, diverse complexes of semionotids were apparently endemic to a number of different lakes. Semionotid fishes show considerable morphological diversity in body shape and in a modified row of scales termed “dorsal ridge scales.” A number of distinct dorsal-ridge-scale patterns characterize groups of species from the Newark Supergroup. Interestingly, about 5.5% of individuals examined have anomalous scales mixed in with otherwise stereotypic dorsal-ridge-scale patterns. In this study, I take advantage of nearly annual stratigraphic resolution to determine whether dorsal-ridge-scale anomalies are concentrated stratigraphically in the early phase of lake formation and colonization by semionotids. More than 1,700 specimens of semionotid fish were collected from a single lake deposit (cycle P4), representing approximately 21,000 years, in the Early-Jurassic Towaco Formation of the Newark Basin. Dorsal-ridge-scale anomalies are significantly more frequent in older than in younger lake sediments, which I interpret as being the result of relaxed selection during the early colonization of the lake. Anomalous variation parallels variation in dorsal ridge scales between species-groups. Some anomalies are atavisms, while others are unique or foreshadow future evolutionary events. One type of anomaly is incorporated into the dorsal-ridge-scale series of two new species that gave rise to a radiation in a subsequent lake filling the same topographic basin. Because both novelties and atavisms occur in the dorsal-ridge-scale series of single individuals, I argue that the disruptions of the same “developmental program” produced both atavistic and novel traits.     

Restoring avian wing digits

The precise identification of the digits of the avian wing is of importance in evolutionary studies. If the digits are numbered two, three and four, this has been taken to suggest that birds are not descended directly from dinosaurs. If the digits are numbered one, two and three, dinosaur origins become more plausible. Studies of the development of the avian wing have failed to resolve this dilemma. However, in some instances, it is possible to deduce information about evolutionary morphologies by manipulating development experimentally. We grafted beads loaded with fibroblast growth factor 4 into the distal tip of chick wing buds at times when the apical ectodermal ridge is regressing. The consequence was that the cartilage structure conventionally labelled ''element 5'' increased dramatically in size and acquired a digit-like morphology in some instances. Corresponding changes in soft tissue morphology were also observed. We conclude that it may be possible to resolve the issue of avian digit homology by the induction of experimental atavisms of this kind.     

On atavisms and atavistic genes

The authors propose the term atavistic to designate a gene producing an ancestral phenotype (atavism). Several examples are presented, and the possible origin of atavistic genes, as well as their pathological implications discussed.     

Atavisms and the homology of hyobranchial elements in lower vertebrates

The homology of branchial arch segments in salamanders has been a matter of controversy since the last century. Many investigators term the most medial paired elements of salamander branchial arches “ceratobranchials” and the next distal paired elements “epibranchials.” This suggests that the first two segmental elements of the salamander branchial arch are not homologous with elements occupying the same position in ray-finned fishes, Latimeria, “rhipidistians,” and lungfishes, in which these bones are called hypobranchials and ceratobranchials, respectively. Three lines of evidence suggest that it is more parsimonious to interpret urodele branchial arch segments as being homologous with those of other vertebrate clades?(1) comparative osteology, (2) comparative myology, and (3) the discovery of cartilaginous structures forming a third segmental unit that we interpret as atavistic epibranchials of the branchial arch in one population of the salamander Notophthalmus viridescens. These structures possess all the defining attributes of atavisms, and illustrate the special role that atavistic features play in resolving questions of homology recognition.     

Developmental constraints and the evolution of vertebrate digit patterns

The skeletal makeup of the digits of 145 hands and feet from species from the four classes of tetrapod vertebrates is analysed. The analysis leads to the conclusion that developmental constraints are very influential in the evolution of vertebrate digit patterns. Furthermore, 98% of the patterns analysed fall within the specific constraints of the Stock & Bryant (1981) version of the polar coordinate model for the formation of digit patterns during development and pattern regulation. Three cases of apparently forbidden morphology are discussed in terms of the phenomenon of differential growth during development.     

Patterns of correlation and covariation of anthropoid distal forelimb segments correspond to Hoxd expression territories

Anthropoids in general and hominoids in particular exhibit differential adaptations in forearm and digital skeletal proportions to a diverse array of locomotor modes. Hox genes act as selector genes with spatially regulated expression patterns during development. Their expression in the forelimb appears to define modules that specify differential skeletal growth. Here we explore forelimb skeletal proportions in a large sample of anthropoids from a background provided by Hoxd expression patterns in late-stage murine embryonic forelimbs. Interspecific correlation and principal components analyses of primate forelimb data indicate that morphological variation in anthropoids reflects well-defined developmental modules downstream of Hoxd expression. The phalanges of digit one appear to represent a single growth module, whereas the metacarpals and manual phalanges of the posterior digits correspond to a second, independent, expression territory that extends proximally into the distal zeugopod. In particular, hominoids show very high correlations among the posterior digits and the independence of digit one. In addition, the distal radius is generally highly correlated with the posterior digits and not digit one. Relying on established functional differences among Hox paralogs, we present a model that parsimoniously explains hominoid forearm and digital proportions as a consequence of downstream effects of Hox. We, therefore, suggest that Hox-defined developmental modules have served as evolutionary modules during manual evolution in anthropoids.     

Autopodial skeleton evolution in side-necked turtles (Pleurodira)

Carpal and tarsal anatomy was documented based on the observation of dry skeletons of adult specimens representing 25 species in 15 genera and on data taken from the literature. In addition, histological sections and cleared and double‐stained autopodia of recently hatched and juvenile specimens representing seven chelid and pelomedusoid species were studied. There is much more morphological diversity in the manus than in the pes. Variation in autopodial skeletons includes: the astragalus and calcaneum are either separated or fused; fusion of distal carpals 3–4−5 or just 4–5; number of centralia in the carpus; and presence/absence of a pisiform and of an accessory radial element. The widespread and probably basal phalangeal formula for Pleurodira is 2.3.3.3.3. Deviations are Pelomedusa subrufa, exhibiting a reduction to 2.2.2.2.2, Pelusios spp. with one phalanx less in digit I and for one species in digit V as well, and Acanthochelys pallidipectoris with an additional phalanx in the fourth finger. Six discrete characters itemizing some of the morphological variation observed were plotted on a composite pleurodire phylogeny, revealing not only homoplastic patterns but also the utility of some characters in supporting the monophyly of several clades. The pisiform is the last carpal element to ossify in Chelus fimbriatus. We hypothesize that the so‐called fifth hooked metatarsal represents the fusion of distal tarsal 5 with metatarsal V. The accessory radial element that was occasionally present in the turtles examined may represent an atavism of the otherwise lost radiale of turtles.     

Bambi and Sp8 Expression Mark Digit Tips and Their Absence Shows That Chick Wing Digits 2 and 3 Are Truncated

An often overlooked aspect of digit development is the special nature of the terminal phalanx, a specialized structure with characteristics distinct from other phalanges, for example the presence of ectodermal derivatives such as nails and claws. Here, we describe the unique ossification pattern of distal phalanges and characteristic gene expression in the digit tips of chick and duck embryos. Our results show that the distal phalanx of chick wing digit 1 is a genuine tip with a characteristic ossification pattern and expression of Bambi and Sp8; however, the terminal phalanx of digits 2* and 3 is not a genuine tip, and these are therefore truncated digits. Bambi and Sp8 expression in the chick wing provides a direct molecular assessment of digit identity changes after experimental manipulations of digit primordia. In contrast, digits 1 and 2 of the duck wing both possess true tips. Although chick wing-tip development was not rescued by application of Fgf8, this treatment induced the development of extra phalanges. Grafting experiments show that competence for tip formation, including nails, is latent in the interdigital tissue. Our results deepen understanding of the mechanisms of digit tip formation, highlighting its developmental autonomy and modular nature, with implications for digit reduction or loss during evolution. * Numbering of wing digits is 1, 2, 3 from anterior to posterior.     

Development and variation of the anuran webbed feet (Amphibia, Anura)

Webbed feet evolved convergently in most groups of aquatic tetrapods. However, extensive webbing is not always limited to an aquatic life style. In Anurans, hind limbs display great variation, including absence, of interdigital membranes, which is explained by differential growth rates of digital and interdigital tissues during early limb development. In order to explore web diversification in anurans, this paper presents analyses of: (1) hind limb early development and its relationship to the expression of interdigital membranes; (2) intraordinal variation of interdigital membranes in adult feet; and (3) intraordinal variation of metatarsal and digit lengths, including comments on metatarsal development. Study of limb development is carried out in larval series of 12 anuran species. Analysis of intraordinal variation comprises a sample of adults of 111 species. We recognize two configurations in the autopodium bud: (1) paddle-like shape with digits differentiated within the confines of interdigital tissues, and (2) pointed autopodium with digits differentiated beyond interdigital tissues. These early differences are conserved in adult morphology, in which allometry and isometry of digit IV (and metatarsal IV) with respect to other digits (and metatarsals) result in asymmetrical and paddle-like autopodium, respectively. The paddle-like autopodium is restricted to fossil and extant pipids and the hylids Pseudis and Lysapsus , whereas the asymmetrical one is present in most anurans. Both configurations seem to represent an early divergence of the autopodium shape. The paddle-like configuration observed in hylids appears as a reversion to an ancient condition that results from a conserved program of limb development.  © 2008 The Linnean Society of London, Zoological Journal of the Linnean Society , 2008, 152 , 39–58.     

Even-toed but uneven in length: the digits of artiodactyls

In captive ruminants housed in small enclosures, hypertrophy of the outer hooves of the hindlimbs is often observed. We hypothesised that the underlying cause is overload attributable to an asymmetry of the digits, especially with respect to their length. To test this hypothesis, the bones of the digits of four species of artiodactyls, which included 11 wild chamois (Rupicapra rupicapra), 11 captive fallow deer (Dama dama), 11 captive bison (Bison bison) and 11 European moose (Alces alces; 9 wild, 2 captive), were radiographed post mortem and measured using a computer programme. In addition, the dimensions of the outer and inner hooves were measured directly with a calliper. The mean lengths of the epiphysis of the fourth metacarpal/metatarsal bone and the first and second phalanges of the fourth digit were greater than that of the third digit, whereas the third phalanx of the third digit had a greater mean length. The mean total length of the fourth digit of the forelimbs was greater than that of the third digit in 73–95% of specimens, depending on species. In the hindlimbs, the fourth digit was longer in 91–100% of the specimens. The hooves of the fourth digit were significantly broader than the hooves of the third digit, whereas the inner hooves of the third digits had a greater toe length than those of the fourth digit. The paired digits of artiodactyls are uneven in length, which suggests a different function during stance and weight bearing. It is conceivable that this asymmetry is the result of selection processes that favoured locomotion on soft ground.     

The morphological basis of hallucal orientation in extant birds

The perching foot of living birds is commonly characterized by a reversed or opposable digit I (hallux). Primitively, the hallux of nonavian theropod dinosaurs was unreversed and lay parallel to digits II-IV. Among basal birds, a unique digital innovation evolved in which the hallux opposes digits II-IV. This digital configuration is critical for grasping and perching. I studied skeletons of modern birds with a range of hallucal designs, from unreversed (anteromedially directed) to fully reversed (posteriorly directed). Two primary correlates of hallucal orientation were revealed. First, the fossa into which metatarsal I articulates is oriented slightly more posteriorly on the tarsometatarsus, rotating the digit as a unit. Second, metatarsal I exhibits a distinctive torsion of its distal shaft relative to its proximal articulation with the tarsometatarsus, reorienting the distal condyles and phalanges of digit I. Herein, I present a method that facilitates the re-evaluation of hallucal orientation in fossil avians based on morphology alone. This method also avoids potential misinterpretations of hallucal orientation in fossil birds that could result from preserved appearance alone.     

Anatomical features associated with preaxial duplication (pd): a recessive mutation in the rat

Preaxial polydactyly of the fore- and hindlimbs was found in Wistar-derived rats in 1978. Genetic analysis indicated that the polydactyly was due to the effects of an autosomal recessive gene (gene symbol; pd). Polydactylous homozygous rats had two or three pollices (six or seven digits) in the forelimbs and one to three preaxial extra digits (six to eight digits) in the hindlimbs. Skeletal examination revealed the presence of the extra carpal, metacarpal, and phalangeal bones that seemed to be complete or incomplete duplication of the navicular, greater multangular, first metacarpal, and phalanges of digit I in the forelimbs. In the hindlimbs, extra tarsal, metatarsal, and phalangeal bones were also observed preaxially. These extra elements seemed to be mirror-image duplications of the talus, navicular, second cuneiform, third cuneiform, cuboid, and metatarsals and phalanges of digits II-V with the absence of the first cuneiform, tibiale, first metatarsal, and phalanges of digit I. In addition, morphological changes were observed in the humerus, radius, and ulna in the forelimbs and femur, tibia, and fibula in the hindlimbs. Especially in the radius and tibia, thickening and bifurcation were found, indicating incomplete duplication of these bones. Based on these findings, the limb anomaly was classified as preaxial carpometacarpal/tarsometatarsal-type polydactyly with incomplete duplication of the radius and tibia. The mutant rats had other associated anomalies such as accessory spleens and cryptorchism. The males are sterile, whereas the females breed normally.     

Phylogenetic memory of developing mammalian dentition

Structures suppressed during evolution can be retraced due to atavisms and vestiges. Atavism is an exceptional emergence of an ancestral form in a living individual. In contrast, ancestral vestige regularly occurs in all members of an actual species. We surveyed data about the vestigial and atavistic teeth in mammals, updated them by recent findings in mouse and human embryos, and discussed their ontogenetic and evolutionary implications. In the mouse incisor and diastema regions, dental placodes are transiently distinct being morphologically similar to the early tooth primordia in reptiles. Two large vestigial buds emerge in front of the prospective first molar and presumably correspond to the premolars eliminated during mouse evolution. The incorporation of the posterior premolar vestige into the lower first molar illustrates the putative mechanism of evolutionary disappearance of the last premolar in the mice. In mutant mice, devious development of the ancestral tooth primordia might lead to their revivification and origin of atavistic supernumerary teeth. Similarity in the developmental schedule between three molars in mice and the respective third and fourth deciduous premolar and the first molar in humans raises a question about putative homology of these teeth. The complex patterning of the vestibular and dental epithelium in human embryos is reminiscent of the pattern of "Zahnreihen" in lower vertebrates. A hypothesis was presented about the developmental relationship between the structures at the external aspect of the dentition in mammals (oral vestibule, pre-lacteal teeth, paramolar cusps/teeth), the tooth glands in reptiles, and the earliest teeth in lower vertebrates.