The problem of the skin derivatives' origin in the amniote evolution. The skin appendages--scale, feather, and hair

Overview of modern data on morphology of the skin derivatives in the higher vertebrates is given. Analysis of convergent similarities between the hair and feathers themselves as well as between their follicles makes it possible to forward a "generative" concept of the evolutionary origin of various ecto-mesodermal derivatives, such as keratinized dermal appendages (scales, feathers, hair). This concept appeared as a result of the author's studies on the skin derivatives, as well as of the data on molecular biology and the tissue engineering showing similar mechanisms of morphogenesis of the dermal appendages. Recurrently published ideas on various heterochronies in generations of the skin derivatives both in the onto- and the phylogeneses are also taken into acount. Various dermal appendages have appeared in the evolution of the higher vertebrates as independent generations of the ecto- and mesodermal tissues. Their parallel origin was caused by similar changes in the metabolism and molecular regulation of morphogenesis.     

Teeth outside the mouth: The evolution and development of shark denticles

Vertebrate skin appendages are incredibly diverse. This diversity, which includes structures such as scales, feathers, and hair, likely evolved from a shared anatomical placode, suggesting broad conservation of the early development of these organs. Some of the earliest known skin appendages are dentine and enamel-rich tooth-like structures, collectively known as odontodes. These appendages evolved over 450 million years ago. Elasmobranchs (sharks, skates, and rays) have retained these ancient skin appendages in the form of both dermal denticles (scales) and oral teeth. Despite our knowledge of denticle function in adult sharks, our understanding of their development and morphogenesis is less advanced. Even though denticles in sharks appear structurally similar to oral teeth, there has been limited data directly comparing the molecular development of these distinct elements. Here, we chart the development of denticles in the embryonic small-spotted catshark (Scyliorhinus canicula) and characterize the expression of conserved genes known to mediate dental development. We find that shark denticle development shares a vast gene expression signature with developing teeth. However, denticles have restricted regenerative potential, as they lack a sox2+ stem cell niche associated with the maintenance of a dental lamina, an essential requirement for continuous tooth replacement. We compare developing denticles to other skin appendages, including both sensory skin appendages and avian feathers. This reveals that denticles are not only tooth-like in structure, but that they also share an ancient developmental gene set that is likely common to all epidermal appendages.     

Rooster feathering, androgenic alopecia, and hormone-dependent tumor growth: what is in common?

Different epithelial organs form as a result of epithelial-mesenchymal interactions and share a common theme modulated by variations (Chuong ed. In Molecular Basis of Epithelial Appendage Morphogenesis, 1998). One of the major modulators is the sex hormone pathway that acts on the prototype signaling pathway to alter organ phenotypes. Here, we focus on how the sex hormone pathway may interface with epithelia morphogenesis-related signaling pathways. We first survey these sex hormone-regulated morphogenetic processes in various epithelial organs. Sexual dimorphism of hairs and feathers has implications in sexual selection. Diseases of these pathways result in androgenic alopecia, hirsutism, henny feathering, etc. The growth and development of mammary glands, prostate glands, and external genitalia essential for reproductive function are also dependent on sex hormones. Diseases affecting these organs include congenital anomalies and hormone-dependent breast and prostate cancers. To study the role of sex hormones in new growth in the context of system biology/pathology, an in vivo model in which organ formation starts from stem cells is essential. With recent developments (Yu et al. (2002) The morphogenesis of feathers. Nature 420:308-312), the growth of tail feathers in roosters and hens has become a testable model in which experimental manipulations are possible. We show exemplary data of differences in their growth rate, proliferative cell population, and signaling molecule expression. Working hypotheses are proposed on how the sex hormone pathways may interact with growth pathways. It is now possible to test these hypotheses using the chicken model to learn fundamental mechanisms on how sex hormones affect organogenesis, epithelial organ cycling, and growth-related tumorigenesis.     

Wnt signaling in skin organogenesis

While serving as the interface between an organism and its environment, the skin also can elaborate a wide range of skin appendages to service specific purposes in a region-specific fashion. As in other organs, Wnt signaling plays a key role in regulating the proliferation, differentiation and motility of skin cells during their morphogenesis. Here I will review some of the recent work that has been done on skin organogenesis. I will cover dermis formation, the development of skin appendages, cycling of appendages in the adult, stem cell regulation, patterning, orientation, regional specificity and modulation by sex hormone nuclear receptors. I will also cover their roles in wound healing, hair regeneration and skin related diseases. It appears that Wnt signaling plays essential but distinct roles in different hierarchical levels of morphogenesis and organogenesis. Many of these areas have not yet been fully explored but are certainly promising areas of future research.Key words: morphogenesis, hair, feathers, tracts, epithelium-mesenchyme interactions, Wnt signaling pathwayThe integument forms the interface between an organism and its environment.^1,2 As such it protects against dehydration, infection, temperature extremes, etc while providing a means for display, camouflage and other functions.³ The skin can elaborate remarkable structural diversity producing specialized functions in a region-specific fashion to provide organisms with a selective advantage. For example, the development of feathers led to the acquisition of flight in birds and the formation of mammary glands enabled mammals to nurse their young.⁴ The advantage of these evolutionary developments can be seen by the number of birds and mammals present today.Skin appendages, such as skin, hairs, feathers, scales, glands and teeth grow from the epithelium as a result of epithelial-mesenchymal interactions,⁵ largely in response to common molecular signals with slight variations in their placement and timing during tissue morphogenesis.⁶ Theoretically, stem cells are totipotent and progressively can be guided toward their specific fates by exposure to specific regulatory signals. The juxtaposition of molecular signals or lack thereof may have a tremendous impact on cell fate decisions. Hence, the difference between skin appendages is due to the topological arrangement of the epithelia during developmental processes. These are presumably regulated by adhesion molecules whose expression is controlled by signaling molecules as well as by physical constraints.Hairs and feathers are attractive model systems for experimental research because of their ability for seasonal or periodic renewal. Obviously not all hairs or feathers are replaced at one time or birds would lose all of their feathers at once and fall from the sky in mid-flight; rather hairs and feathers are replaced over a period of time in a wave-like pattern.⁷ Yet this cycling behavior enables thousands of entirely new organs to be regenerated again and again throughout these animal''s lives. Hairs and feathers demonstrate an incredible diversity of forms arising in different locations over the body surface. For instance, hairs on the scalp, face and body differ in size, coarseness, color, etc. This regional specificity indicates that in each cycle skin stem cells are directed to form distinct structures through a series of molecular and cellular interactions.     

Evolution of the morphological innovations of feathers

Feathers are complex assemblages of multiple morphological innovations. Recent research on the development and evolution of feathers has produced new insights into the origin and diversification of the morphological innovations in feathers. In this article, I review and discuss the contribution of three different factors to the evolution of morphological innovations in feathers: feather tubularity, hierarchical morphological modularity, and the co-option molecular signaling modules. The developing feather germ is a tube of epidermis with a central dermal pulp. The tubular organization of the feather germ and follicle produces multiple axes over which morphological differentiation can be organized. Feather complexity is organized into a hierarchy of morphological modules. These morphological modules evolved through the innovative differentiation along multiple different morphological axes created by the tubular feather germ. Concurrently, many of the morphological innovations of feathers evolved through the evolutionary co-option of plesiomorphic molecular signaling modules. Gene co-option also reveals a role for contingency in the evolution of hierarchical morphological innovations.     

Shh-Bmp2 signaling module and the evolutionary origin and diversification of feathers

To examine the role of development in the origin of evolutionary novelties, we investigated the developmental mechanisms involved in the formation of a complex morphological novelty-branched feathers. We demonstrate that the anterior-posterior expression polarity of Sonic hedgehog (Shh) and Bone morphogenetic protein 2 (Bmp2) in the primordia of feathers, avian scales, and alligator scales is conserved and phylogenetically primitive to archosaurian integumentary appendages. In feather development, derived patterns of Shh-Bmp2 signaling are associated with the development of evolutionarily novel feather structures. Longitudinal Shh-Bmp2 expression domains in the marginal plate epithelium between barb ridges provide a prepattern of the barbs and rachis. Thus, control of Shh-Bmp2 signaling is a fundamental component of the mechanism determining feather form (i.e., plumulaceous vs. pennaceous structure). We show that Shh signaling is necessary for the formation and proper differentiation of a barb ridge and that it is mediated by Bmp signaling. BMP signaling is necessary and sufficient to negatively regulate Shh expression within forming feather germs and this epistatic relationship is conserved in scale morphogenesis. Ectopic SHH and BMP2 signaling leads to opposing effects on proliferation and differentiation within the feather germ, suggesting that the integrative signaling between Shh and Bmp2 is a means to regulate controlled growth and differentiation of forming skin appendages. We conclude that Shh and Bmp signaling is necessary for the formation of barb ridges in feathers and that Shh and Bmp2 signaling constitutes a functionally conserved developmental signaling module in archosaur epidermal appendage development. We propose a model in which branched feather form evolved by repeated, evolutionary re-utilization of a Shh-Bmp2 signaling module in new developmental contexts. Feather animation Quicktime movies can be viewed at http://fallon.anatomy.wisc.edu/feather.html.     

Stimulation of ectodermal organ development by Ectodysplasin-A1

Organs developing as ectodermal appendages share similar early morphogenesis and molecular mechanisms. Ectodysplasin, a signaling molecule belonging to the tumor necrosis factor family, and its receptor Edar are required for normal development of several ectodermal organs in humans and mice. We have overexpressed two splice forms of ectodysplasin, Eda-A1 and Eda-A2, binding to Edar and another TNF receptor, Xedar, respectively, under the keratin 14 (K14) promoter in the ectoderm of transgenic mice. Eda-A2 overexpression did not cause a detectable phenotype. On the contrary, overexpression of Eda-A1 resulted in alterations in a variety of ectodermal organs, most notably in extra organs. Hair development was initiated continuously from E14 until birth, and in addition, the transgenic mice had supernumerary teeth and mammary glands, phenotypes not reported previously in transgenic mice. Also, hair composition and structure was abnormal, and the cycling of hairs was altered so that the growth phase (anagen) was prolonged. Both hairs and nails grew longer than normal. Molar teeth were of abnormal shape, and enamel formation was severely disturbed in incisors. Furthermore, sweat gland function was stimulated and sebaceous glands were enlarged. We conclude that ectodysplasin-Edar signaling has several roles in ectodermal organ development controlling their initiation, as well as morphogenesis and differentiation.     

Mechanisms of ectodermal organogenesis

All ectodermal organs, e.g. hair, teeth, and many exocrine glands, originate from two adjacent tissue layers: the epithelium and the mesenchyme. Similar sequential and reciprocal interactions between the epithelium and mesenchyme regulate the early steps of development in all ectodermal organs. Generally, the mesenchyme provides the first instructive signal, which is followed by the formation of the epithelial placode, an early signaling center. The placode buds into or out of the mesenchyme, and subsequent proliferation, cell movements, and differentiation of the epithelium and mesenchyme contribute to morphogenesis. The molecular signals regulating organogenesis, such as molecules in the FGF, TGFbeta, Wnt, and hedgehog families, regulate the development of all ectodermal appendages repeatedly during advancing morphogenesis and differentiation. In addition, signaling by ectodysplasin, a recently identified member of the TNF family, and its receptor Edar is required for ectodermal organ development across vertebrate species. Here the current knowledge on the molecular regulation of the initiation, placode formation, and morphogenesis of ectodermal organs is discussed with emphasis on feathers, hair, and teeth.     

TNF superfamily in skin appendage development

The development of skin appendages such as hairs, teeth, and mammary glands is regulated by signaling molecules of the Wnt, FGF, TGFbeta, and Hedgehog pathways. Last decade has also revealed a pivotal role for the TNF family ligand ectodysplasin (Eda) in multiple steps of epithelial appendage morphogenesis, from initiation to differentiation. Surprisingly, other members of the TNF superfamily such as Rank ligand, lymphotoxins, and TNF have recently been linked with specific aspects of skin appendage biology including branching of the mammary gland, hair shaft formation, and hair follicle cycling. This review focuses on the novel discoveries of Eda and other TNF related cytokines in skin appendage development made since the previous review on this topic in Cytokine and Growth Factor reviews in 2003.     

FGF signaling is required for initiation of feather placode development

Morphogenesis of hairs and feathers is initiated by an as yet unknown dermal signal that induces placode formation in the overlying ectoderm. To determine whether FGF signals are required for this process we over-expressed soluble versions of FGFR1 or FGFR2 in the skin of chicken embryos. This produced a complete failure of feather formation prior to any morphological or molecular signs of placode development. We further show that Fgf10 is expressed in the dermis of nascent feather primordia, and that anti-FGF10 antibodies block feather placode development in skin explants. In addition we show that FGF10 can induce expression of positive and negative regulators of feather development and can induce its own expression under conditions of low BMP signaling. Together these results demonstrate that FGF signaling is required for the initiation of feather placode development and implicate FGF10 as an early dermal signal involved in this process.     

Evo-Devo of amniote integuments and appendages

Integuments form the boundary between an organism and the environment. The evolution of novel developmental mechanisms in integuments and appendages allows animals to live in diverse ecological environments. Here we focus on amniotes. The major achievement for reptile skin is an adaptation to the land with the formation of a successful barrier. The stratum corneum enables this barrier to prevent water loss from the skin and allowed amphibian / reptile ancestors to go onto the land. Overlapping scales and production of beta-keratins provide strong protection. Epidermal invagination led to the formation of avian feather and mammalian hair follicles in the dermis. Both adopted a proximal - distal growth mode which maintains endothermy. Feathers form hierarchical branches which produce the vane that makes flight possible. Recent discoveries of feathered dinosaurs in China inspire new thinking on the origin of feathers. In the laboratory, epithelial - mesenchymal recombinations and molecular mis-expressions were carried out to test the plasticity of epithelial organ formation. We review the work on the transformation of scales into feathers, conversion between barbs and rachis and the production of "chicken teeth". In mammals, tilting the balance of the BMP pathway in K14 noggin transgenic mice alters the number, size and phenotypes of different ectodermal organs, making investigators rethink the distinction between morpho-regulation and pathological changes. Models on the evolution of feathers and hairs from reptile integuments are discussed. A hypothetical Evo-Devo space where diverse integument appendages can be placed according to complex phenotypes and novel developmental mechanisms is presented.     

Conserved features and evolutionary shifts of the EDA signaling pathway involved in vertebrate skin appendage development

It is widely accepted that evolutionary changes in conserved developmental signaling pathways play an important role in morphological evolution. However, few in silico studies were interested in tracking such changes in a signaling pathway. The Ectodysplasin (EDA) pathway provides an opportunity to fill this gap because it is involved in vertebrate skin appendage development such as scales, teeth, hair, and feathers that take an obvious part in the adaptation of species to their environment. We benefited from the large amount of genomic data now available to explore the evolution of the upstream genes of the EDA pathway. In mammals, these genes are eda (encoding 2 ligands, EDA-A1 and EDA-A2), edar (EDA-A1 receptor), edaradd (EDA receptor [EDAR] adapter), xedar (EDA-A2 receptor), and troy (a XEDAR-related receptor). We show that the evolution of EDA pathway genes combines both strongly conserved features and evolutionary shifts. These shifts are found at different signaling levels (from the ligand to intracellular signaling) and at different taxonomic levels (class, suborder, and genera). Although conserved features likely participate to the similarities found in the early development of vertebrate skin appendages, these shifts might account for innovations and specializations. Moreover, our study demonstrates that we can now benefit from the large number of sequenced vertebrate genomes to explore the evolution of specific signaling pathways and thereby to open new perspectives for developmental biology and evolutionary developmental biology.     

Molecular biology of hair morphogenesis: development and cycling

In mammals, hair follicles produce hairs that fulfill a number of functions including thermoregulation, collecting sensory information, protection against environmental trauma, social communication, and mimicry. Hair follicles develop as a result of epithelial-mesenchymal interactions between epidermal keratinocytes committed to hair-specific differentiation and cluster of dermal fibroblasts that form follicular papilla. During postnatal life, hair follicles show patterns of cyclic activity with periods of active growth and hair production (anagen), apoptosis-driven involution (catagen), and relative resting (telogen). During last decade, substantial progress has been achieved in delineating molecular mechanisms that control hair follicle development and cyclic activity. In this review, we summarize the data demonstrating that regulation of hair follicle development in the embryo and control of hair follicle growth during postnatal life are highly conserved and both require involvement of similar molecular mechanisms. Since many of the molecules that control hair follicle development and cycling are also involved in regulating morphogenesis and postnatal biology of other ectodermal derivatives, such as teeth, feathers, and mammary glands, basic principles and molecular mechanisms that govern hair follicle development and growth may also be applicable for other developmental systems.     

Ectodysplasin signaling in development

Ectodysplasin (Eda), a signaling molecule belonging to the tumor necrosis factor family, is required for normal development of several ectodermally derived organs in humans and mice. Two closely related isoforms of ectodysplasin, Eda-A1 and Eda-A2, have been described which bind to and activate two different receptors, Edar and X-linked Eda-A2 receptor (Xedar), respectively. Mutations in Eda, Edar or other molecules of this signaling pathway cause ectodermal dysplasias characterized by defective development of teeth, hairs, and several exocrine glands such as sweat glands presumably due to impaired NF-kappaB response. Studies with mice either lacking the functional proteins of Edar pathway or overexpressing the ligand or receptor suggest that Eda-A1-Edar signaling has multiple roles in ectodermal organ development regulating their initiation, morphogenesis, and differentiation.     

Cryptic patterning of avian skin confers a developmental facility for loss of neck feathering

Vertebrate skin is characterized by its patterned array of appendages, whether feathers, hairs, or scales. In avian skin the distribution of feathers occurs on two distinct spatial levels. Grouping of feathers within discrete tracts, with bare skin lying between the tracts, is termed the macropattern, while the smaller scale periodic spacing between individual feathers is referred to as the micropattern. The degree of integration between the patterning mechanisms that operate on these two scales during development and the mechanisms underlying the remarkable evolvability of skin macropatterns are unknown. A striking example of macropattern variation is the convergent loss of neck feathering in multiple species, a trait associated with heat tolerance in both wild and domestic birds. In chicken, a mutation called Naked neck is characterized by a reduction of body feathering and completely bare neck. Here we perform genetic fine mapping of the causative region and identify a large insertion associated with the Naked neck trait. A strong candidate gene in the critical interval, BMP12/GDF7, displays markedly elevated expression in Naked neck embryonic skin due to a cis-regulatory effect of the causative mutation. BMP family members inhibit embryonic feather formation by acting in a reaction-diffusion mechanism, and we find that selective production of retinoic acid by neck skin potentiates BMP signaling, making neck skin more sensitive than body skin to suppression of feather development. This selective production of retinoic acid by neck skin constitutes a cryptic pattern as its effects on feathering are not revealed until gross BMP levels are altered. This developmental modularity of neck and body skin allows simple quantitative changes in BMP levels to produce a sparsely feathered or bare neck while maintaining robust feather patterning on the body.     

Regulation of hair follicle development by the TNF signal ectodysplasin and its receptor Edar

X-linked and autosomal forms of anhidrotic ectodermal dysplasia syndromes (HED) are characterized by deficient development of several ectodermal organs, including hair, teeth and exocrine glands. The recent cloning of the genes that underlie these syndromes, ectodysplasin (ED1) and the ectodysplasin A receptor (EDAR), and their identification as a novel TNF ligand-receptor pair suggested a role for TNF signaling in embryonic morphogenesis. In the mouse, the genes of the spontaneous mutations Tabby (Ta) and downless (dl) were identified as homologs of ED1 and EDAR, respectively. To gain insight into the function of this signaling pathway in development of skin and hair follicles, we analyzed the expression and regulation of Eda and Edar in wild type as well as Tabby and Lef1 mutant mouse embryos. We show that Eda and Edar expression is confined to the ectoderm and occurs in a pattern that suggests a role of ectodysplasin/Edar signaling in the interactions between the ectodermal compartments and the formation and function of hair placodes. By using skin explant cultures, we further show that this signaling pathway is intimately associated with interactions between the epithelial and mesenchymal tissues. We also find that Ta mutants lack completely the placodes of the first developing tylotrich hairs, and that they do not show patterned expression of placodal genes, including Bmp4, Lef1, Shh, Ptch and Edar, and the genes for beta-catenin and activin A. Finally, we identified activin as a mesenchymal signal that stimulates Edar expression and WNT as a signal that induces Eda expression, suggesting a hierarchy of distinct signaling pathways in the development of skin and hair follicles. In conclusion, we suggest that Eda and Edar are associated with the onset of ectodermal patterning and that ectodysplasin/edar signaling also regulates the morphogenesis of hair follicles.     

Involvement of selective epithelial cell death in the formation of feather buds on a bioengineered skin

Selective cell death by apoptosis plays important roles in organogenesis. Apoptotic cells are observed in the developmental and homeostatic processes of several ectodermal organs, such as hairs, feathers, and mammary glands. In chick feather development, apoptotic events have been observed during feather morphogenesis, but have not been investigated during early feather bud formation. Previously, we have reported a method for generating feather buds on a bioengineered skin from dissociated skin epithelial and mesenchymal cells in three-dimensional culture. During the development of the bioengineered skin, epithelial cavity formation by apoptosis was observed in the epithelial tissue. In this study, we examined the selective epithelial cell death during the bioengineered skin development. Histological analyses suggest that the selective epithelial cell death in the bioengineered skin was induced by caspase-3-related apoptosis. The formation of feather buds of the bioengineered skin was disturbed by the treatment with a pan-caspase inhibitor. The pan-caspase inhibitor treatment suppressed the rearrangement of the epithelial layer and the formation of dermal condensation, which are thought to be essential step to form feather buds. The suppression of the formation of feather buds on the pan-caspase inhibitor-treated skin was partially compensated by the addition of a GSK-3β inhibitor, which activates Wnt/β-catenin signaling. These results suggest that the epithelial cell death is involved in the formation of feather buds of the bioengineered skin. These observations also suggest that caspase activities and Wnt/β-catenin signaling may contribute to the formation of epithelial and mesenchymal components in the bioengineered skin.     

Cytochemical and molecular characteristics of the process of cornification during feather morphogenesis

Feathers are the most complex epidermal derivatives among vertebrates. The present review deals with the origin of feathers from archosaurian reptiles, the cellular and molecular aspects of feather morphogenesis, and focus on the synthesis of keratins and associated proteins. Feathers consist of different proteins among which exists a specialized group of small proteins called beta-keratins. Genes encoding these proteins in the chick genome are distributed in different chromosomes, and most genes encode for feather keratins. The latter are here recognized as proteins associated with the keratins of intermediate filaments, and functionally correspond to keratin-associated proteins of hairs, nails and horns in mammals. These small proteins possess unique properties, including resistance and scarce elasticity, and were inherited and modified in feathers from ancestral proteins present in the scales of archosaurian progenitors of birds. The proteins share a common structural motif, the core box, which was present in the proteins of the reptilian ancestors of birds. The core box allows the formation of filaments with a different molecular mechanism of polymerization from that of alpha-keratins. Feathers evolved after the establishment of a special morphogenetic mechanism gave rise to barb ridges. During development, the epidermal layers of feathers fold to produce barb ridges that produce the ramified structure of feathers. Among barb ridge cells, those of barb and barbules initially accumulate small amounts of alpha-keratins that are rapidly replaced by a small protein indicated as “feather keratin”. This 10 kDa protein becomes the predominant form of corneous material of feathers. The main characteristics of feather keratins, their gene organization and biosynthesis are similar to those of their reptilian ancestors. Feather keratins allow elongation of feather cells among supportive cells that later degenerate and leave the ramified microstructure of barbs. In downfeathers, barbs are initially independent and form plumulaceous feathers that rest inside a follicle. Stem cells remain in the follicle and are responsible for the regeneration of pennaceous feathers. New barb ridges are produced and they merge to produce a rachis and a flat vane. The modulation of the growth pattern of barb ridges and their fusion into a rachis give rise to a broad variety of feather types, including asymmetric feathers for flight. Feather morphogenesis suggests possible stages for feather evolution and diversification from hair-like outgrowths of the skin found in fossils of pro-avian archosaurians.     

The evolutionary origin and diversification of feathers

Progress on the evolutionary origin and diversification of feathers has been hampered by conceptual problems and by the lack of plesiomorphic feather fossils. Recently, both of these limitations have been overcome by the proposal of the developmental theory of the origin of feathers, and the discovery of primitive feather fossils on nonavian theropod dinosaurs. The conceptual problems of previous theories of the origin of feathers are reviewed, and the alternative developmental theory is presented and discussed. The developmental theory proposes that feathers evolved through a series of evolutionary novelties in developmental mechanisms of the follicle and feather germ. The discovery of primitive and derived fossil feathers on a diversity of coelurosaurian theropod dinosaurs documents that feathers evolved and diversified in nonavian theropods before the origin of birds and before the origin of flight. The morphologies of these primitive feathers are congruent with the predictions of the developmental theory. Alternatives to the theropod origin of feathers are critique and rejected. Hypotheses for the initial function of feathers are reviewed. The aerodynamic theory of feather origins is falsified, but many other functions remain developmentally and phylogenetically plausible. Whatever their function, feathers evolved by selection for a follicle that would grow an emergent tubular appendage. Feathers are inherently tubular structures. The homology of feathers and scales is weakly supported. Feathers are composed of a suite of evolutionary novelties that evolved by the duplication, hierarchical organization, interaction, dissociation, and differentiation of morphological modules. The unique capacity for modular subdivision of the tubular feather follicle and germ has fostered the evolution of numerous innovations that characterize feathers. The evolution of feather keratin and the molecular basis of feather development are also discussed.