Microscopic analysis of lizard claw morphogenesis and hypothesis on its evolution

Morphogenesis of claws in the lizard Lampropholis guichenoti has been studied by light and electron microscopy. Claws originate from a thickening of the epidermis covering the tips of digits under which mesenchymal cells aggregate. Mesenchymal cells are in continuity with perichondrial cells of the last phalange, and are connected to the epidermis through numerous cell bridges that cross an incomplete basement membrane. The dense lamella is completed in non‐apical regions of the digit where also collagen fibrils increase. The dorsal side of the developing claw derives from the growth of the outer scale surface of the last scale of the digit. The corneous layer, made of beta‐keratin cells, curves downward by the tip of the growing claw. The epidermis of the ventral side of the claw contains keratohyaline‐like granules and alpha‐keratinocytes like an inner scale surface. The thickness of the horny layer increases in the elongating unguis while a thinner and softer corneous layer remains in the subunguis. These observations show that lizard claws derive from the modification of the last scale or scales of the digit, probably under the influence of the growing terminal phalanx. Some hypotheses on the evolution of claws in reptiles are presented.     

Ultrastructural localization of hair keratin homologs in the claw of the lizard Anolis carolinensis

The claw of lizards is largely composed of beta‐keratins, also referred to as keratin‐associated beta‐proteins. Recently, we have reported that the genome of the lizard Anolis carolinensis contains alpha keratin genes homologous to hair keratins typical of hairs and claws of mammals. Molecular and immunohistochemical studies demonstrated that two hair keratin homologs named hard acid keratin 1 (HA1) and hard basic keratin 1 (HB1) are expressed in keratinocytes forming the claws of A. carolinensis. Here, we extended the immunocytochemical localization of the novel reptilian keratins to the ultrastructural level. After sectioning, claws were subjected to immunogold labeling using antibodies against HA1, HB1, and, for comparison, beta‐keratins. Electron microscopy showed that the randomly organized network of tonofilaments in basal and suprabasal keratinocytes becomes organized in long and parallel bundles of keratin in precorneous layers, resembling cortical cells of hairs. Entering the cornified part of the claw, the elongated corneous cells fuse and accumulate corneous material. HA1 and HB1 are absent in the basal layer and lower spinosus layers of the claw and are expressed in the upper and precorneous layers, including the elongating corneocytes. The labeling for alpha‐keratin was loosely associated with filament structures forming the fibrous framework of the claws. The ultrastructural distribution pattern of hard alpha‐keratins resembled that of beta‐keratins, which is compatible with the hypothesis of an interaction during claw morphogenesis. The data on the ultrastructural localization of hair keratin homologs facilitate a comparison of lizard claws and mammalian hard epidermal appendages containing hair keratins. J. Morphol., 2011. © 2010 Wiley‐Liss, Inc.     

Autoradiographic observations on developing and growing claws of reptiles

Alibardi, L. 2010. Autoradiographic observations on developing and growing claws of reptiles. —Acta Zoologica (Stockholm) 91 : 233–241 The present qualitative autoradiographic analysis aims to present the main features of morphogenesis and growth of claws in reptiles. Lizard embryos treated with tritiated thymidine reveal that epidermal cell proliferation in terminal digits is prevalent in the dorsal side and gives origin to the curved unguis of the claw. Less proliferation occurs in the ventral side of the digit tip where the concave sub‐unguis is derived. Adult claws of a turtle show that thymidine‐labelled cells are present along most of the epidermis of the claw, especially at the claw tip. Also, injection of tritiated histidine and proline, indicating active protein synthesis, confirm autoradiographic labelling along most of the epidermis of claws, in particular at the apical tip. The present study indicates that proximal matrix regions, as have been described in mammalian nails, are absent in reptiles. This pattern of claw growth probably derives from that of terminal digital scales. In fact reptilian (and avian) claws are formed from a modification of scales, a different condition from that present in mammals.     

Microscopic and immunohistochemical study on the cornification of the developing beak in the turtle Emydura macquarii

The development and cornification of the ramphoteca (beak) in turtles are not known. The microscopic aspects of beak formation have been analyzed in the pleurodirian turtle Emydura macquarii using histological, immunocytochemical and ultrastructural methods. At embryonic Stage 15 the maxillar beak is originated from discontinuous placodes (one frontal and two oral) formed in the epidermis above and below the mouth that later merge into the epidermis of the beak. The mandibular beak is formed by two lateral placodes. In the placodes, basal keratinocytes in contact with local mesenchymal condensations become columnar, and generate suprabasal cells forming 5–6 layers of embryonic epidermis at Stages 17–20 and a compact shedding alpha‐layer at the base of the embryonic epidermis. These keratinocytes contain irregular or aggregated reticular bodies made of 30–40 nm thick strands of coarse filaments, mixed with tonofilaments and sparse lipid droplets. Beneath the shedding layer are present 3–4 layers of keratinocytes accumulating coarse filaments mixed with beta‐corneous packets, and underneath spindle‐shaped beta‐cells differentiate where beta‐corneous packets completely replace the reticulate bodies. Differently from scales where corneocytes partially merge, beak corneocytes remain separated but they are joined by numerous interlocking spines. The production of beta‐cells in the thick corneous layer of the developing beak, like in claws, occurs before the differentiation of beta‐cells in the body scutes. This indicates that a massive mesenchymal condensation triggers beta‐differentiation before this process is later activated in most of body scutes of the carapace and plastron. J. Morphol. 277:1309–1319, 2016. © 2016 Wiley Periodicals, Inc.     

Microanatomy of passerine hard‐cornified tissues: Beak and claw structure of the black‐capped chickadee (Poecile atricapillus)

The microanatomy of healthy beaks and claws in passerine birds has not been well described in the literature, despite the importance of these structures in avian life. Histological processing of hard‐cornified tissues is notoriously challenging and only a few reports on effective techniques have been published. An emerging epizootic of beak deformities among wild birds in Alaska and the Pacific Northwest region of North America recently highlighted the need for additional baseline information about avian hard‐cornified structures. In this study, we examine the beak and claw of the Black‐capped Chickadee (Poecile atricapillus), a common North American passerine that is affected by what has been described as “avian keratin disorder.” We use light and scanning electron microscopy and high‐magnification radiography to document the healthy microanatomy of these tissues and identify features of functional importance. We also describe detailed methods for histological processing of avian hard‐cornified structures and discuss the utility of special stains. Results from this study will assist in future research on the functional anatomy and pathology of hard‐cornified structures and will provide a necessary reference for ongoing investigations of avian keratin disorder in Black‐capped Chickadees and other wild passerine species. © 2011 Wiley Periodicals, Inc.     

Immunocytochemistry and protein analysis suggest that reptilian claws contain small high cysteine-glycine proteins

The present study analyzes the structure and the main proteins of reptilian claws. Mature claws are formed by two to four layers of keratinocytes, a transitional layer of spindle-shaped cells and a thick corneous layer. Transitional cells elongate and merge into a compact corneous layer that is immunoreactive for beta-keratins, now indicated as sauropsid keratin-associated proteins (sKAPs). Most proteins extracted from claws in representative reptiles have a molecular weight of 13-20 kDa, an acidic to basic isoelectric point, and are identified from the positive immunoreactivity to beta-keratin antibodies. The comparative analysis between lizard and avian claw beta-keratins shows the presence of an internal region of 20 amino acids with the highest identity, indicated as core-box, within an extended 32-amino acid region with a prevalent beta-sheet secondary conformation. This region is structurally equivalent to a 32-amino acid region present in scale beta-keratins of most reptiles. Both reptilian and avian keratins contain glycine-rich regions for stabilization of the beta-keratin polymer. The N- and C-regions contain most cysteine for disulphide-bonds formation. Claw proteins contain higher amount of cysteine and glycine than other scale proteins, suggesting that claw proteins are specialized cysteine-glycine-rich proteins suited to produce a very hard corneous material.     

Ontogenetic changes in diet and related morphological adaptations in Ocypode gaudichaudii

There are conflicting reports as to whether Ocypode gaudichaudii individuals switch from carnivory as juveniles to deposit‐feeding primarily on diatoms as adults, or whether they expand diet range and become omnivorous with maturity. At the onset of deposit‐feeding, crabs develop specialized claws with truncated ends that they use to shovel sediment during foraging. Eighty‐eight crabs were collected from Culebra Island (Republic of Panama) to study how the diet of this crab shifts with changes in claw shape, mouthpart proportions, and setation, as well as gastric mill width. Forty‐four crabs had identifiable material in their foreguts: 30 had animal material, 12 had diatoms, and two had a mix of both. There were no differences between the gastric mill, mandibles, and the proximal endites of the first maxillipeds of predators and deposit‐feeders, but extra rows of plumose setae were present on the second maxilliped of deposit‐feeding crabs with carapace length (CL) >10.6 mm. All individuals with CL <12.3 mm and non‐truncated claws ate animals, but those with larger CL and truncated claws had animal, diatom, or mixed diets; hence, claw truncation does not restrict the crab's diet to diatoms but, instead, broadens the diet to include both animals and diatoms. Perhaps this is a strategy to balance the economics of foraging on animals and diatoms on medium to low‐energy beaches that lack the larger invertebrates that adults of other species of ghost crabs eat. More generally, our study shows that specialized feeding structures need not imply a narrow or specialist diet.     

Differentiation of the epidermis of scutes in embryos and juveniles of the tortoise Testudo hermanni with emphasis on beta-keratinization

The sequence of differentiation of the epidermis of scutes during embryogenesis in the tortoise Testudo hermanni was studied using autoradiography, electron microscopy and immunocytochemistry. The study was mainly conducted on the epidermis of the carapace, plastron and nail. Epidermal differentiation resembles that described for other reptiles, and the embryonic epidermis is composed of numerous cell layers. In the early stages of differentiation of the carapacial ridge, cytoplasmic blebs of epidermal cells are in direct contact with the extracellular matrix and mesenchymal cells. The influence of the dermis on the formation of the beta‐layer is discussed. The dermis becomes rich in collagen bundles at later stages of development. The embryonic epidermis is formed by a flat periderm and four to six layers of subperidermal cells, storing 40–70‐nm‐thick coarse filaments that may represent interkeratin or matrix material. Beta‐keratin is associated with the coarse filaments, suggesting that the protein may be polymerized on their surface. The presence of beta‐keratin in embryonic epidermis suggests that this keratin might have been produced at the beginning of chelonian evolution. The embryonic epidermis of the scutes is lost around hatching and leaves underneath the definitive corneous beta‐layer. Beneath the embryonic epidermis, cells that accumulate typical large bundles of beta‐keratin appear at stage 23 and at hatching a compact beta‐layer is present. The differentiation of these cells shows the progressive replacement of alpha‐keratin bundles with bundles immunolabelled for beta‐keratin. The nucleus is degraded and electron‐dense nuclear material mixes with beta‐keratin. In general, changes in tortoise skin when approaching terrestrial life resemble those of other reptiles. Lepidosaurian reptiles form an embryonic shedding layer and crocodilians have a thin embryonic epidermis that is rapidly lost near hacthing. Chelonians have a thicker embryonic epidermis that accumulates beta‐keratin, a protein later used to make a thick corneous layer.     

Embryonic keratinization in vertebrates in relation to land colonization

The embryogenesis and cytology of the epidermis in different vertebrates is variable in relation to the formation of a stratum corneum of different complexity. The latter process was essential for land colonization during vertebrate evolution and produced an efficient barrier in amniotes. Keratinocytes are made of cross‐linked keratins associated with specific proteins and lipids that are produced at advanced stages of embryogenesis when the epidermis becomes stratified. In these stages the epidermis changes from an aquatic to a terrestrial type, preadapted in preparation for the impact with the dry terrestrial environment that occurs at hatching or parturition. The epidermal barrier against water‐loss, mechanical and chemical stress, and microbe penetration is completely formed shortly before birth. Beneath the outer periderm, variably stratified embryonic layers containing glycine‐rich alpha‐keratins are formed in preparation for adult life. The following layers of the epidermis produce proteins for the formation of the cornified cell membrane and of the cornified material present in keratinocytes of the adult epidermis in reptiles, birds and mammals. The general features of the process of soft cornification in the embryonic epidermis of vertebrates are presented. Cornification in developing scales in reptiles, avian feathers and mammalian hairs is mainly related to the evolution of keratin‐associated proteins. The latter proteins form the resistant matrix of hard skin derivatives such as claws, beaks, nails and horns.     

Points on the curve: An analysis of methods for assessing the shape of vertebrate claws

The form of amniote claws has been extensively investigated, often with inferences about ecological association being drawn from studies of their geometry. Various methods have been used to quantify differences in the geometry of claws, but rarely have the underlying assumptions of such methods been addressed. Here, we use one set of bird claws and apply six methods (five that have been previously used, and a new one) that are tasked with comparing their shape. In doing so, we compare the (1) ability of these methods to represent the shape of the claw; (2) validity of the assumptions made about underlying claw geometry; (3) their ability to be applied unambiguously; and (4) their ability to differentiate between predetermined functional clusters. We find that of the six methods considered only the geometric morphometric approach reveals differences in the shapes of bird claws. Our comparison shows that geometry‐based methods can provide a general estimate of the degree of curvature of claw arcs, but are unable to differentiate between shapes. Of all of the geometry‐based approaches, we conclude that the adjusted version of the Zani (2000) method is the most useful because it can be applied without ambiguity, and provides a reliable estimate of claw curvature. The three landmarks that define that method (tip and base of the claw arc, plus the intersection between said claw arc and a line drawn perpendicular from the midpoint of tip and claw base) do not all bear biological significance, but relatively clearly circumscribe the length‐to‐height ratio of the claw, which relates to its curvature. Overall, our comparisons reveal that the shape of avian claws does not differ significantly between climbing and perching birds, and that the utilization of preordained functional clusters in comparative data analysis can hinder the discovery of meaningful differences in claw shape. J. Morphol. 278:150–169, 2017. © 2016 Wiley Periodicals,Inc.     

Post-autotomy claw regrowth and functional recovery in the snapping shrimp Alpheus angulosus

The ability to regenerate lost tissues, organs or whole body parts is widespread across animal taxa; in some animals, regeneration includes transforming a remaining structure to replace the one that was lost. The transformation of one limb into another involves considerable plasticity in morphology, physiology and behavior, and snapping shrimp offer excellent opportunities for studying this process. We examined the changes required for the transformation of the small pincer to a mature snapping claw in Alpheus angulosus. First molt claws differ from mature claws in overall shape as well as in morphology related to snapping function; nonetheless, shrimp with first molt claws do produce snaps. While most shape variables of second molt claws do not differ significantly from mature claws, the plunger (structure required for snap production) does not reach mature size until the third molt for females, or later for males. Thus, the pincer claw can be transformed into a functional snapping claw in one molt, although both the underlying morphology and superficial shape are not fully regenerated at this stage. The rapid production of a functional snapping claw that we observe in this study suggests that this particular function is of significant importance to snapping shrimp behavior and survival.     

Synthesis of interkeratin matrix in differentiating lizard epidermis: an ultrastructural autoradiographic study after injection of tritiated proline and histidine

During epidermal differentiation in mammals, keratins and keratin-associated matrix proteins rich in histidine are synthesized to produce a corneous layer. Little is known about interkeratin proteins in nonmammalian vertebrates, especially in reptiles. Using ultrastructural autoradiography after injection of tritiated proline or histidine, the cytological process of synthesis of beta-keratin and interkeratin material was studied during differentiation of the epidermis of lizards. Proline is mainly incorporated in newly synthesized beta-keratin in beta-cells, and less in oberhautchen cells. Labeling is mainly seen among ribosomes within 30 min postinjection and appears in beta-keratin packets or long filaments 1-3 h later. Beta-keratin appears as an electron-pale matrix material that completely replaces alpha-keratin filaments in cells of the beta-layer. Tritiated histidine is mainly incorporated into keratohyalin-like granules of the clear layer, in dense keratin bundles of the oberhautchen layer, and also in dense keratin filaments of the alpha and lacunar layer. The detailed ultrastructural study shows that histidine-labeling is localized over a dense amorphous material associated with keratin filaments or in keratohyalin-like granules. Large keratohyalin-like granules take up labeled material at 5-22 h postinjection of tritiated histidine. This suggests that histidine is utilized for the synthesis of keratins and keratin-associated matrix material in alpha-keratinizing cells and in oberhautchen cells. As oberhautchen cells fuse with subjacent beta-cells to form a syncytium, two changes occur : incorporation of tritiated histidine, but uptake of proline increases. The incorporation of tritiated histidine in oberhautchen cells lowers after merging with cells of the beta-layer, whereas instead proline uptake increases. In beta-cells histidine-labeling is lower and randomly distributed over the cytoplasm and beta-keratin filaments. Thus, change in histidine uptake somehow indicates the transition from alpha- to beta-keratogenesis. This study indicates that a functional stratum corneum in the epidermis of amniotes originates only after the association of matrix and corneous cell envelope proteins with the original keratin scaffold of keratinocytes.     

Stratum corneum hydration: Phase transformations and mobility in stratum corneum, extracted lipids and isolated corneocytes

The outermost layer of skin, stratum corneum (SC), functions as the major barrier to diffusion. SC has the architecture of dead keratin filled cells embedded in a lipid matrix. This work presents a detailed study of the hydration process in extracted SC lipids, isolated corneocytes and intact SC. Using isothermal sorption microcalorimetry and relaxation and wideline ¹H NMR, we study these systems at varying degrees of hydration/relative humidities (RH) at 25 °C. The basic findings are (i) there is a substantial swelling both of SC lipids, the corneocytes and the intact SC at high RH. At low RHs corneocytes take up more water than SC lipids do, while at high RHs swelling of SC lipids is more pronounced than that of corneocytes. (ii) Lipids in a fluid state are present in both extracted SC lipids and in the intact SC. (iii) The fraction of fluid lipids is lower at 1.4% water content than at 15% but remains virtually constant as the water content is further increased. (iv) Three exothermic phase transitions are detected in the SC lipids at RH = 91-94%, and we speculate that the lipid re-organization is responsible for the hydration-induced variations in SC permeability. (v) The hydration causes swelling in the corneocytes, while it does not affect the mobility of solid components (keratin filaments).     

Stratum corneum hydration: phase transformations and mobility in stratum corneum, extracted lipids and isolated corneocytes

The outermost layer of skin, stratum corneum (SC), functions as the major barrier to diffusion. SC has the architecture of dead keratin filled cells embedded in a lipid matrix. This work presents a detailed study of the hydration process in extracted SC lipids, isolated corneocytes and intact SC. Using isothermal sorption microcalorimetry and relaxation and wideline (1)H NMR, we study these systems at varying degrees of hydration/relative humidities (RH) at 25 degrees C. The basic findings are (i) there is a substantial swelling both of SC lipids, the corneocytes and the intact SC at high RH. At low RHs corneocytes take up more water than SC lipids do, while at high RHs swelling of SC lipids is more pronounced than that of corneocytes. (ii) Lipids in a fluid state are present in both extracted SC lipids and in the intact SC. (iii) The fraction of fluid lipids is lower at 1.4% water content than at 15% but remains virtually constant as the water content is further increased. (iv) Three exothermic phase transitions are detected in the SC lipids at RH=91-94%, and we speculate that the lipid re-organization is responsible for the hydration-induced variations in SC permeability. (v) The hydration causes swelling in the corneocytes, while it does not affect the mobility of solid components (keratin filaments).     

Keratin expression in cultures of adult human epidermal cells.

Keratins are complex fibrous proteins characteristic of epithelial cells. We have developed a procedure that allows us to culture and passage adult human dermal keratinocytes in the absence of mesenchymal substrates. Electron microscopic examination of stratifying cultures showed the presence of numerous filament bundles, desmosomes and electron dense granules. The expression of different classes of keratin was examined by immunofluorescence, SDS-PAGE and immunoblots using monoclonal antibodies. The analysis of water-insoluble proteins revealed the presence of keratins of molecular weights 40 Kda, 50-52 Kda, 56 Kda and 65-67 Kda. Our results indicate that the terminal differentiation of keratinocytes may not require dermal factors.     

Changes in nano‐mechanical properties of human epidermal cornified cells depending on their proximity to the skin surface

During formation of the stratum corneum (SC) barrier, terminally differentiated keratinocytes continue their maturation process within the dead superficial epidermal layer. Morphological studies of isolated human corneocytes have revealed differences between cornified envelopes purified from the deep and superficial SC. We used atomic force microscopy to measure the mechanical properties of native human corneocytes harvested by tape‐stripping from different SC depths. Various conditions of data acquisition have been tested and optimized, in order to obtain exploitable and reproducible results. Probing at 200 nN allowed us to investigate the total stiffness of the cells (at 50 nm indentation) and that of the cornified envelopes (at 10 to15 nm), and lipid envelopes (at 5 to 10 nm). The obtained data indicated statistically significant differences between the superficial (more rigid) and deep (softer) corneocytes, thus confirming the existence of depth and maturation‐related morphological changes within the SC. The proposed approach can be potentially used for minimally invasive evaluation of various skin conditions such as aging, skin hydration, and pathologies linked to SC.     

白边油蝠表皮角质化细胞末端分化的超微结构

蝙蝠是一种唯一能够飞行的哺乳动物,其皮肤的超微结构尚未见报道。在电镜下观察了白边油蝠(Pipistrellus kuhlii)背部和翼膜皮肤的超微结构。表皮的厚度较低(10~12μm),角质层下有1~2层的刺细胞,该刺细胞由相似于鸟类无羽表皮的纤细角化细胞形成。颗粒层不连续且仅有少量小型透明角质颗粒(<0.3μm)。在翼膜的若干区域,表皮简化为一层与角质层相连的基底层。过渡期的角化细胞几乎不存在,提示其角质化过程非常迅速。基底膜上的无数半桥粒在真皮下面形成密集的附着点。大量胶原纤维直接维系在半桥粒和基底膜的致密层上,稀疏的弹性纤维使得蝙蝠表皮在飞行时易于伸展、在飞行后易于迅速折叠而不会受到损伤。与鸟类的表皮相似,蝙蝠角化细胞富有大量的脂质。由于脂质有助于蝙蝠皮肤在飞行中与冷空气流的传热绝缘,大量脂质的存在可能是为补偿蝙蝠翼膜的真皮缺乏厚的脂肪层。研究还表明,毛发较薄(4~7μm),并具有与皮层相似的突状物组成的精细表皮,其表皮细胞形成钩状抓握点使毛发紧紧粘结在一起,通过这种方式毛皮保持紧凑以恒定体温。     

Deleterious Mutations of a Claw Keratin in Multiple Taxa of Reptiles

We have recently shown that homologs of mammalian hair keratins are expressed in the claws of the green anole lizard, Anolis carolinensis. To test whether reptilian hair keratin homologs are functionally associated with claws, we investigated the conservation of the prototypical reptilian hair keratin homolog, hard acidic keratin 1 (HA1), in representative species from all main clades of reptiles. A complete cDNA of HA1 was cloned from the claw-forming epidermis of the lacertid lizard Podarcis sicula, and partial HA1 gene sequences could be amplified from genomic DNA of tuatara, lizards, gekkos, turtles, and crocodiles. In contrast, the HA1 gene of the limbless slow worm, Anguis fragilis, and of two species of turtles contained at least one deleterious mutation. Moreover, an HA1 gene was undetectable in the softshell turtle, snakes, and birds. Mapping the presence and absence of HA1 onto the phylogenetic tree of sauropsids suggested that the HA1 gene has been lost independently in several lineages of reptiles. The species distribution of HA1 is compatible with the hypothesis of a primary function of HA1 in claws but also shows that the formation of reptilian claws does not strictly depend on this keratin.     

The anatomical basis of sexual dichromatism in non-iridescent ultraviolet-blue structural coloration of feathers

Despite extensive research on the evolution of avian dichromatism, the anatomical bases for differences between the sexes in species with structurally coloured plumage remain largely unknown. Using full‐spectrum spectrometry and transmission electron microscopy, we compared the colour and morphology of rump feathers of male and female eastern bluebirds (Sialia sialis). The ultraviolet (UV)‐blue feather colour in this species is caused by coherent scattering of light within the medullary ‘spongy layer’ of feather barbs. This spongy layer lies beneath a keratin cortex and on top of a layer of melanin granules that surround a hollow central vacuole. Irregularly shaped electron‐dense regions are present within the cortex. Male and female S. sialis differed substantially in their plumage colour and feather structure. A backwards logistic regression predicted sex with 100% accuracy using the colour variables brightness, UV‐violet (UV‐V) chroma and spectral saturation. A second backwards logistical regression predicted sex with 100% accuracy using relative cortex area and size of air spaces. Thus, S. sialis are dimorphic both in colour and in the structures causing this colour. Multiple regression analyses using data pooled from both sexes indicated that multiple features of feather barb structure contributed to colour variation in complex ways. Brightness was negatively related to the relative surface area of cortex in barb cross‐sections. Hue was positively related and UV‐V chroma was negatively related to the distance between scattering elements (i.e. keratin rods and air spaces) in the spongy layer. In contrast, hue was negatively related and UV‐V chroma was positively related to the thickness of the spongy layer. UV‐V chroma was also negatively related to the relative area of electron‐dense regions in the cortex. Spectral saturation was negatively related to the distance between scatterers and the standard error of the size of air spaces. These results suggest that the dimensions of spongy‐layer elements are critical to colour production, but that UV‐blue coloration can also be modified by the cortex and the thickness of the spongy layer. © 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 84 , 259–271.