SEARCH

SEARCH BY CITATION

Keywords:

  • electron microscopy;
  • birds;
  • ultraviolet;
  • structural colors

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

The visual perception of many birds extends into the near-ultraviolet (UV) spectrum and ultraviolet is used by some to communicate. The beak horn of the King Penguin (Aptenodytes patagonicus) intensely reflects in the ultraviolet and this appears to be implicated in partner choice. In a preliminary study, we recently demonstrated that this ultraviolet reflectance has a structural basis, resulting from crystal-like photonic structures, capable of reflecting in the near-UV. The present study attempted to define the origin of the photonic elements that produce the UV reflectance and to better understand how the UV signal is optimized by their fine structure. Using light and electron microscopic analysis combined with new spectrophotometric data, we describe here in detail the fine structure of the entire King Penguin beak horn in addition to that of its photonic crystals. The data obtained reveal a one-dimensional structural periodicity within this tissue and demonstrate a direct relationship between its fine structure and its function. In addition, they suggest how the photonic structures are produced and how they are stabilized. The measured lattice dimensions of the photonic crystals, together with morphological data on its composition, permit predictions of the wavelength of reflected light. These correlate well with experimentally observed values. The way the UV signal is optimized by the fine structure of the beak tissue is discussed with regard to its putative biological role. © 2006 Wiley-Liss, Inc.

Bird vision is highly specialized in terms of both acuity and sensitivity, many birds being capable of perceiving near-ultraviolet (UV) light (Cuthill et al.,2000). Several reports have emphasized the potential biological role of both UV reflectance and fluorescence in avian communication (Andersson and Amundsen,1997; Hunt et al.,1999; Örnborg et al.,2002; Siitari et al.,2002; Pearn et al.,2003). In particular, the capacity of avian ornaments to reflect UV, in addition to their being colored, has been shown to play an important role during sexual displays (Hausmann et al.,2003). King Penguins (Aptenodytes patagonicus) are endowed with several highly colored ornaments (Fig. 1A), certain of which have been suggested to be implicated in mate choice in this species (Dresp et al.,2005). These include yellow/orange breast and auricular feathers and two orange/pink beak horns, which are located on each side of the beak and stand out against its black background. Of these ornaments, only the beak horn reflects in the UV spectral range.

thumbnail image

Figure 1. A: Two King Penguins in a face-to-face encounter, illustrating the orange/yellow-colored breast and auricular plumage and the orange/pink beak horn on the black beak. B: Isolated beak horn specimen collected after molting. Scale bar = 1.5 cm.

Download figure to PowerPoint

Color is produced either chemically from pigments or by physical interaction between tissue structures and incident light. Although colored pigments such as carotenoids and psittacofulvins are well documented in the feathers or bills of certain birds (McGraw and Nogare,2004; Peters et al.,2004), no avian pigments that significantly enhance UV reflectance are currently known. The so-called structural colors result from reflectance of incident light due to the physical nature of the reflecting material (Dyck,1976; Parker,1998; Prum et al.,1999a,1999b; Vukusic et al.,2001). Examples of structural colors are widespread in the natural world (Parker,2000; Vukusic and Sambles,2003) and have been investigated in birds for over a century (for reviews, see Auber,1957; Dyck,1976; Prum et al.,1999a). Since most birds detect near-UV light, photonic structures in birds reflecting in the UV range have aroused considerable interest in recent years, particularly with regard to their potential biological function (Finger and Burkhardt,1994). They were first discovered in the plumage of several avian species (Prum et al.,1999b; Hausmann et al.,2003) but have also more recently been reported to be present in other avian tissues, including skin (Prum et al.,1999b; Prum and Torres,2003), mouth tissue (Hunt et al.,2003), and comb tissue. In a recent preliminary report, we provided evidence for the structural basis of UV reflectance from the King Penguin beak horn (Dresp et al.,2005). In the present study, we provide additional spectrophotometric data supporting the structural basis of this reflectance and describe in detail for the first time the fine structure of the entire King Penguin horn, including that of the vast numbers of interconnected photonic microstructures responsible for UV reflectance, situated in the upper region of the horn. These data suggest the likely origin of these microstructures, how they are produced, and how their structure is stabilized in this tissue. They also reveal how they function to produce a biological signal for this species and how the signal is optimized.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

Beak Horns

The orange/pink beak horns of adult king penguins are keratinous structures, asymmetrically ellipsoidal in shape, ca. 8 cm long and ca. 1.5 cm wide at the broadest part, which is nearest to the eye in situ (Fig. 1B), located on each side of the black beak. They molt annually, exposing a new horn on the beak. Their thickness, measured with a micrometer, varies from ca. 0.3 to 0.4 mm between individuals. Naturally air-dried specimens are hard, with a smooth matt outer surface, curled at their edges. Thin translucent layers often peel away from their inner surface. Specimens were collected after molting on the beach at the Baie du Marin of Possession Island (46°25′ S, 51°45′ E) in the Crozet Archipelago, where large colonies of King Penguins gather to breed. They were stored dry at ambient temperature for several weeks before analysis for UV reflectance and histological structure.

Microscopy

Beak horns were sectioned laterally into strips 2 mm wide, approximately 2.5 cm from the broadest end. Strips were subsequently soaked in water for at least 2 hr, before fixation overnight in 2.5% glutaraldehyde in 0.2 M phosphate buffer, pH 7.2. Some strips were initially pretreated for 2 hr in sodium hydroxide before fixation, as previously described (Dyck,1976). However, since this procedure, designed to facilitate subsequent sectioning of bird feathers, appeared to have no significant influence either on sectioning or histology of beak horns, it was abandoned as a routine preparative procedure. After extensive rinsing in water, strips were postfixed for 5 hr in 2% osmium tetroxide and washed again in water before progressive dehydration in ethanol. Specimens were subsequently infiltrated with and embedded in Spurr epoxy resin (Spurr,1969). For light microscopy, after polymerization for 48 hr at 60°C, blocks were sectioned with an ultramicrotome at 0.7 μm, transverse to the longitudinal axis of the beak horn, stained with Toluidine blue, and viewed with bright-field optics with a Zeiss axioscope microscope equipped with Neofluar objectives. Digital photographs were acquired as Jpeg files with Nikon ACT-1 software using a Nikon DXM 1200 digital camera. For electron microscopy, 70 nm sections were poststained with uranyl acetate and lead citrate and viewed in a Zeiss 900 microscope at 80 kV. Electron micrographs were scanned to produce digital images and structural parameters were measured on prints at final magnifications of 20–80,000×.

Spectral Reflectance

Reflectance spectra were measured on five air-dried beak horns at a point approximately 2.5 cm from the end nearest the eye in situ, which is the flattest region of the horn, as previously described (Prum et al.,1999a; Dresp et al.,2005). Briefly, they were obtained using an S2000 fiber optic diode array spectrophotometer with a PX-2 pulsed xenon light source (Ocean Optics, Dunedin, FL). Spectra were measured at least three times, with the probe perpendicular to (incident light angle, 90°) and 2–3 mm away from the reflecting surface for an illuminated field of ca. 3 mm2. A Spectralon diffuse white reference standard was used for calibration and percentage reflectance values were calculated automatically by software provided with the instrument, the details of which have been published (Prum et al.,1999a). The reflectance spectrum was also obtained on one beak horn specimen after soaking in water overnight. The surface of this horn was blotted dry just prior to spectrophotometry. For comparison, spectra were also measured on beak horns of live penguins.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

Histology

Toluidine blue-stained sections cut in the plane perpendicular to the long axis and parallel to the short axis of the beak horn displayed three superposed regions with distinct morphological characteristics. An upper region, extending inward from the outer surface to a depth of ca. 40% of the total thickness of the horn, consisted of a relatively weakly stained matrix containing densely stained, interconnected profiles (U; Fig. 2A). Most of these were elongated in shape, varying from 3.5 to 13 μm in length and from 0.5 to 2.0 μm in width, with the majority between 5 and 10 μm long and 1 and 2 μm wide (Fig. 2A). Some of the interconnected dense profiles were more rounded in shape, varying from 2.0 to 4.5 μm in diameter. The most superficial part of this region of the beak horn appeared to contain air pockets. Cell nuclei were not observed within the upper region. Immediately beneath this upper region was a narrow, more intensely stained, central region (C; Fig. 2A) occupying ca. 10% of the horn thickness, which consisted of seven to eight layers of closely associated clear cells, with elongated central nuclei. The borders of cells in this region were relatively broad and densely stained; in the uppermost part of this region, they appeared to be contiguous with the dense profiles in the upper region. Beneath this central region, clear flattened cellular profiles occupied the lower half of the depth of the horn (L; Fig. 2A). These cells appeared to be increasingly compact toward the lower surface. Elongated nuclei were observed within this lower region.

thumbnail image

Figure 2. A: Light micrograph of Toluidine blue-stained transverse section of intact beak horn. Note the three regions. The upper region (U) is filled with interconnected densely stained profiles of variable shapes and sizes; the central region (C) is constituted of layers of cells with broad densely stained borders; the lower region (L) contains compact, clear, flat cellular profiles. B: Toluidine blue-stained transverse section of beak horn processed for histology after scraping successive layers from the upper surface until the ultraviolet reflection was eliminated. Note the absence of the upper region, U. Scale bar = 25 μm.

Download figure to PowerPoint

Electron Microscopy

Transmission electron microscopy of transverse sections cut perpendicular to the long axis of the beak horn showed the upper region of the horn to be filled with vast numbers of interconnected structures, subsequently referred to here as “microstructures” (Fig. 3, arrows). Each of these was constituted of a membrane doublet (dm; Fig. 3) folded backward and forward to form multiple layers, a complete layer comprising the membrane doublet plus the material filling the space between successive membrane doublets. These microstructures differed considerably in shape and size (Fig. 4). Their lengths (distance between the two extremities that connect with adjacent microstructures, e.g., *–*, Fig. 3) varied from 1.9 to 12.5 μm (mean, 5.8 ± 2.7 μm) and their widths (distance perpendicular to the axis between extremities) from 0.8 to 6.6 μm (mean, 2.0 ± 1.2 μm). The length of the widest layers (e.g., #–#, Fig. 3) varied from 2.0 to 9.7 μm between different microstructures (mean, 3.9 ± 1.6 μm). This is almost twice that of the average width of microstructures and is due to the fact that the layers were not in general perpendicular to the long axis of the microstructures. The lengths of individual layers of a given microstructure were in general smaller at the extremities, i.e., the microstructures tapered at the extremities. Since the size range of these multiple layer microstructures identified by electron microscopy closely correlates with that of the densely Toluidine blue-stained profiles observed in the histological sections by light microscopy, it may be concluded that they are indeed the same structures and that the space between them is not stained significantly by this dye. The layers within a given microstructure were parallel to each other, but they were not always orientated in the same direction in different microstructures: their orientations differed by up to 30–45° (compare microstructures I and II in Fig. 3). The number of layers, i.e., folds, in microstructures varied considerably. The smallest microstructure seen contained only six layers, while the biggest comprised up to 40 layers. Occasionally, within a given microstructure, certain layers were considerably shorter than adjacent layers and did not extend across the entire width of the microstructure (large arrow; Fig. 5). In such cases, adjacent layers compacted together, maintaining a relatively constant space between successive membrane doublets. Each complete layer in the microstructures comprised a membrane doublet consisting of two dense membranes, each approximately 9.5 nm thick, separated by a narrow electron lucent space, which was approximately 14.5 ± 3 nm wide for most of its length, although elongated dense profiles of variable sizes were frequently observed (e.g., small arrow; Fig. 5) within slightly enlarged portions of this intra double-membrane space. Membrane doublets in successive layers were separated by a moderately electron lucent space approximately 97 ± 11 nm wide, containing a few fine filaments approximately 3.5 nm in diameter (Figs. 5 and 6). These filaments occasionally appeared to be branched, forming a loose skeletal-like network, and individual filaments often appeared to be attached to one of the membranes forming the doublet (arrow; Fig. 6). These filaments extended from irregular large bundles (F; Figs. 5 and 6) interspersed between the multiple-layered microstructures. These filament bundles, which were the only other structural element apart from the microstructures observed in this region, were found to occupy approximately 44% of the total volume of this upper region and thus represent a significant component of its composition. The perpendicular distance between the surfaces of successive doublet membranes measured on micrographs at a final magnification of 80,000× was found to be 130.5 ± 11 nm. These microstructures thus approximate to crystals with one-dimensional periodicity of 130.5 ± 11 nm. Each multiple-layered microstructure was interconnected to four others, two at each extremity. They were thus interconnected in triads at each extremity. Such connections were made by one of the juxtaposed membranes, which constituted part of the doublet of the folded layer at the extremity of the microstructure, pairing up with one of the membranes of the doublet of one of the neighboring microstructures to form a new doublet, while the other juxtaposed membrane formed the doublet of the other neighboring microstructure with this same membrane (Fig. 7).

thumbnail image

Figure 3. Electron micrograph of the upper region of the beak horn. This region is filled with interconnected microstructures (e.g., arrows), constituted of multiple layers of folded membrane doublets (dm). The microstructures vary in shape and size. Their lengths (e.g., *–*), widths (perpendicular to long axis), and lengths of widest folds (e.g., #–#) were measured to obtain the histograms shown in Figure 4. The orientation of the layers often differs between individual microstructures, e.g., I and II. Irregular bundles of filaments (e.g., F) fill the space between microstructures. Scale bar = 2.0 μm.

Download figure to PowerPoint

thumbnail image

Figure 4. Distribution of the lengths (A) and widths (B) of the microstructures. Dimensions (in μm) of 47 microstructures were measured directly on electron micrographs, as indicated in legend to Figure 3, and classed into groups differing by 2 μm (A) or 1 μm (B).

Download figure to PowerPoint

thumbnail image

Figure 5. Higher magnification of part of a microstructure. Note the small dense profiles (e.g., small arrows) frequently observed between the juxtaposed membrane doublets, which form the folded layers. Note also that one of the folds (large arrow) is shorter than the others. Membrane doublets adjacent to this compact together reducing the distance between them. Fine filaments are present in the space between successive membrane doublets. A bundle of fine filaments (F) fills the space adjacent to the microstructure. Scale bar = 0.5 μm.

Download figure to PowerPoint

thumbnail image

Figure 6. Electron micrograph illustrating the fine keratin filaments extending from large bundles (F) into the spaces between the folds of the microstructures, where they appear to be attached to the membranes to form a skeletal-like structure (arrow). Scale bar = 0.5 μm.

Download figure to PowerPoint

thumbnail image

Figure 7. Electron micrograph illustrating how microstructures are interconnected. Parts of three adjacent microstructures are shown. Note that the two juxtaposed membranes, which constitute the doublet forming the terminal fold of the uppermost of these, originate from different adjacent microstructures. Scale bar = 1.0 μm.

Download figure to PowerPoint

No cell nuclei were ever observed in this upper region, which extended from the surface to a depth of approximately 120 μm (i.e., ca. 40% of the horn thickness). In order to obtain a three-dimensional perception of these microstructures, transverse sections of the beak horn were cut in a plane parallel to its long axis, i.e., perpendicular to that used for Figures 2, 3, 5, 6, and 7. Identical microstructures with similar dimensions were observed in sections cut in this plane. Thus, it may be concluded that these microstructures, 5.8 ± 2.7 μm long and 2.0 ± 1.2 μm wide, consist of three-dimensional tilted stacks of 6–40 layers of superposed flat double-membrane folds.

Immediately beneath the upper region filled with the multiple-layer folded membrane microstructures, the central beak horn region was found to contain large cells with copious cytoplasm and elongated, frequently convoluted, nuclei (Fig. 8). A well-defined nucleolus was generally observed within these nuclei, which also contained condensed heterochromatin. Although the morphological preservation was not optimal, probably due to the fact that the beak horns were naturally air-dried when collected, considerable quantities of well-preserved, compact bundles of fine filaments, with dimensions similar to those in the upper region, were observed within the cytoplasm, but few other organelles were identified (Fig. 8). The filaments were sometimes aligned approximately along the same axis and sometimes arranged irregularly, when they had the appearance of eddy currents in a river. In addition, well-defined interdigitating folded membrane systems were present in these cells, although their complexity differed from one cell to another. They were constituted of juxtaposed double membranes, although the membranes often appeared in negative contrast. They bore some resemblance to the folded membranes constituting the microstructures present in the upper region described above, but their fine structure differed radically in several aspects (Fig. 9). They were constituted of fewer folds and these folds were not compacted together, although some were more compact than others (Fig. 8). Thus, much wider spaces separated successive folds and the spaces were irregular compared to those present in the upper region microstructures. Moreover, the narrow space between juxtaposed membranes was less regular and slightly larger, 17.5 ± 6 nm, instead of 14.5 ± 3 nm, and it was completely filled with electron-dense material, while in the microstructures in the upper region of the horn, this intermembrane space was mostly clear, with only occasional electron-dense inclusions. In addition, the relatively broad irregular spaces between these loosely folded membranes were filled with large numbers of fine filaments, rather than just a few individual filaments, as was found in the microstructures of the upper region. Interconnections between folded doublet membrane systems were observed and these were similar to those found between microstructures in the upper horn region.

thumbnail image

Figure 8. Electron micrograph of the central cellular layer. The cell has an elongated nucleus (N) and is filled with fine filaments, which are either orientated in bundles or are disorganized. Note the large numbers of membranes that form complex, interdigitated, double-membrane systems. Scale bar = 1.5 μm.

Download figure to PowerPoint

thumbnail image

Figure 9. Higher magnification of complex membrane systems in a cell in the central region of the beak horn. Note that the space between the two membranes that form the doublet is completely filled with electron-dense material. Scale bar = 0.5 μm.

Download figure to PowerPoint

In the lower region of the beak horn, beneath this well-defined layer of large cells, compact layers of flattened cells with elongated cell nuclei were observed, the cell membranes of adjacent cells running parallel to the surface of the beak horn. The cytoplasm of these cells was filled with fine filaments essentially aligned parallel to the long axis of these cells, parallel to the surface of the beak horn. No elaborate membrane systems were observed in these cells (Fig. 10).

thumbnail image

Figure 10. Cells in lower region of the beak horn. These cells are filled with fine filaments and their mainly rectilinear cell membranes are closely juxtaposed. Scale bar = 0.5 μm.

Download figure to PowerPoint

Thus, in summary, the lower region of the beak horn tissue contained cells filled with fine filaments, but no complex interdigitating membrane systems. The cell membranes of these cells were closely juxtaposed to each other and not interdigitated. In the central region, cells filled with filaments were present and interdigitated membrane systems were abundant. These systems may be connected to or contiguous with their cell membranes, which also appeared to be interdigitated. The upper region contained only well-organized, interconnected, compact, interdigitated membrane systems, termed microstructures, with bundles of filaments filling the intervening space.

Spectral Reflectance

On visual inspection, dried beak horn specimens have a yellow-orange color with a pinkish-violet tint. Reflectance from the upper surface of all specimens of dried beak horns, measured by spectrophotometry, was characterized by a pronounced peak in the near-UV region (Fig. 11A). The wavelength of maximum reflectance (λmax) and the maximum percentage reflectance varied slightly between different specimens and also between different positions on the same beak horn. λmax ranged from 360 to 400 nm, depending on the specimen (mean, ca. 370 nm). A second, very broad reflectance peak in the visible spectrum (yellow, orange, red region) with a λmax around 630 nm was also observed on all specimens. In contrast, no UV reflectance was observed from the opposite (under) surface of the beak horn. Spectra obtained from live birds were qualitatively similar. However, the average λmax was ca. 390 nm instead of ca. 370 nm. In addition, the maximum reflectance was 50–100% higher. When beak horn specimens were rehydrated for 24 hr by immersion in water, they became supple, their pinkish-violet tint increased, and the UV peak of their reflectance spectrum was not symmetrical, as with dried specimens. λmax shifted from ca. 370 to 420 nm and the percentage of maximum reflectance approximately doubled (Fig. 11B).

thumbnail image

Figure 11. Reflectance spectra obtained from an intact dried beak horn (A), and from the same beak horn after soaking in water overnight (B). Note the increase in maximal reflectance, the shift in wavelength of maximal reflectance, and the broadening of the reflectance peak.

Download figure to PowerPoint

In previously reported scraping experiments in which successive layers were removed from the upper surface of beak horns with a scalpel blade, UV reflectance was found to be little affected until the upper 38% of the horn had been removed, after which it was suddenly abolished (Dresp et al.,2005). The broad peak in the visible spectrum was found to persist, however, and its percentage reflectance increased slightly. Figure 2B shows the histology of the scraped beak horn at this stage. Almost the entire upper region had been removed by the successive scrapings, but all of the central cellular and lower regions were still present. This demonstrates that the cellular and lower regions are incapable of reflecting in the UV range and that the structures responsible for UV reflectance in this tissue are situated within the upper region of the beak horn.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

In a preliminary report, we recently presented spectrophotometric and morphological evidence that the ultraviolet reflectance of the King Penguin beak horn is likely to be structural in origin and suggested that it is produced according to the laws of optical physics from photonic structures present in its upper region (Dresp et al.,2005). The ultrastructural study of the entire beak horn we present here provides a detailed account of fine structure of these photonic structures and permits a more complete understanding of how they are likely to be formed, how they are stabilized, and the way in which they could produce a biologically functional, appropriate signal. Light and electron microscopy data presented here show that the part of the King Penguin beak horn responsible for UV reflectance, the uppermost region, is a homogeneous tissue, filled throughout its depth with only two types of structure. One of these consisted of interconnected microstructures that stained intensely with Toluidine blue in histological sections. Electron microscopy revealed these to consist of a multiply folded membrane doublet that produces up to 40 quasiparallel layers in individual microstructures. The other morphological element consisted of bundles of fine filaments that filled the space between microstructures.

Although detailed biochemical data on this tissue are not available at the present time, it may be concluded that these filaments consist of keratin, as in all other hard integument structures of birds, including feathers, ramphotheca, and scales. Mammalian keratins consist essentially of α-keratin, which has an α-helical structure forming typical intermediate filaments. In contrast, avian keratins are mostly composed of β-keratins, which form fine filaments with a twisted β-sheet structure. The present ultrastructural study clearly demonstrates that the filaments present in this upper beak horn region are approximately 3.5 nm in diameter, which correlates with that of β-keratin filaments.

β-keratin is synthesized by keratinocytes. In the penguin beak horn, these cells are situated in its central and lower regions. The upper region of the beak horn constitutes a corny layer (stratum corneum) of this specialized skin tissue. As in avian skin, keratinocytes in this tissue migrate from deeper layers toward the surface, while at the same time producing large amounts of keratin. They subsequently die, losing both nuclei and cytoplasmic organelles, as they assume a particular functional role in the integument. The observation that no nuclei were observed in the entire upper region, which is approximately 120–150 μm deep, depending on the specimen, is in keeping with this conclusion and reflects the terminal status of keratinocytes in this region. The filaments produced by these cells serve an evident mechanical function in the King Penguin beak horn by adding overall rigidity to it. In its upper region, since individual filaments separate from compact bundles and extend into the spaces between membrane folds within the multiple-layered microstructures, apparently forming a skeletal like network, it can be concluded that keratin filaments also serve to confer stability to these structures. The bundles of keratin filaments present in the upper region of the horn completely fill the space between microstructures and thus may also play a role in physically forcing the membrane doublet layers together as they are formed to produce the regular parallel layers characteristic of the microstructures.

Precisely how are the microstructures formed and from what? Electron microscopy demonstrated that cells in the lower region have a structural composition similar to that of cells in the central region, i.e., are filled with large quantities of fine keratin filaments. They would thus represent the precursors of the cells in the central region. The beak horn, like skin and its derivatives, appears to grow from the bottom toward the surface. The light microscope images show contiguity between densely stained structures from the central layers and those in the upper layer, suggesting that the membrane layers found in these microstructures are likely to be formed by cells in the central layer as they migrate toward the surface. Electron microscopy showed that the specialized keratinocytes in the central region of the beak horn, i.e., just beneath the region containing the multiple-layered membrane microstructures, contain interdigitated membrane systems, suggesting that while continuing the mass production of keratin, they are also responsible for large-scale membrane synthesis. The membrane systems they contain share some aspects with the multiple-layered microstructures in the upper region. Although some of these folded membranes were more compact than others, they were not as compact as in the microstructures, nor were they as elaborate. However, given the similarity between the more compact interdigitated membrane systems in these cells in the central region of the horn and the microstructures present in the upper region, it is highly probably that these represent the initial stages in the formation of the folded membrane doublets in the microstructures.

Two possibilities exist for their formation. They could be an unusual, excessive elaboration of the endoplasmic reticulum of these cells in the central region. However, it seems much more likely that they are formed from cell plasma membranes, which under the pressure of massive synthesis produce folds that interdigitate with membranes from adjacent cells, i.e., they are intracellular invaginations of the cell wall. Interestingly, mammalian epithelial cells also form complex interdigitated patterns with neighboring cells. It is likely that they are compressed into compact, well-organized folds in the microstructures in the upper region by the action of continued growth of cells in the central layer and by compression from filament bundles during terminal differentiation of these specialized cells, a process that is accompanied by nuclear loss. The electron-dense material that entirely fills the intermembrane space in interdigitated membrane systems in the central region of the horn probably consists of extracellular material, which is squeezed into the small elongated profiles observed in the upper region, as membrane synthesis and compaction of the doublet continues.

Although several physical mechanisms have been evoked to explain structural color in birds, that in avian plumage is now considered to result from coherent light scattering either from a spongy keratin-air matrix (Prum et al.,1999b) or from photonic crystals constituted of melanin rods within a keratin-air matrix (Zi et al.,2003). Avian skin structural color is produced by coherent light scattering from organized collagen fibrils within the dermis (Prum and Torres,2003). Ordered collagen arrays in the facial skin of certain birds can also reflect strongly in the UV range (Prum et al.,1999a). The new spectrophotometric data presented here comparing the reflectance spectra from naturally air-dried and rehydrated specimens, which show an increase in wavelength maxima, λmax, after rehydration (Fig. 11B), support the conclusion that the UV reflectance from the beak horn is indeed produced by manipulation of incident light by photonic structures within the tissue, further justifying the use of Bragg's law to predict reflectance λmax (Bragg and Bragg,1915) as follows: λmax = n2d sinθ, where λmax is the peak wavelength of reflected light, n is the average refractive index of the tissue, d is the separation of the layers (lattice dimension), and θ is the angle of incidence of the light (here 90°). Rehydration of specimens could either increase dimensions of structural parameters within the reflecting elements or increase the average refractive index, both of which would lead to an increase in λmax. Since no increase in overall thickness of the beak horn was observed by micrometry, it may be concluded that the change in λmax is due to an increase in the average refractive index, with water (refractive index 1.33) either filling up air spaces (refractive index 1.00) or dissolving dried salt deposits within the tissue, thus increasing its refractive index.

Since the bundles of filaments in the upper region were not regular in size, nor systematically orientated in the same direction, nor organized into a regular or repetitive pattern, they can be excluded as a possible origin of coherent light scattering. Thus, of the two types of structure present in this UV-reflective region, only the multiple-folded membrane microstructures, with their well-ordered, repetitive, crystal-like fine structure with one-dimensional periodicity, have the physical characteristics capable of producing structural color, by interference between incident light and that reflected from successive folds. The spacing between successive folds, the overall refractive index of the material, and the incident angle are the critical parameters that determine the wavelength of the reflected light according to Bragg's law.

A major unknown here is the overall refractive index of the beak horn tissue. At the present time, biochemical analysis on this tissue is lacking and it is only possible to predict a value for its refractive index. We previously used a value of 1.4 (Dresp et al.,2005), which has been employed to predict reflectance wavelengths in avian and mammalian skin, based on the refractive indexes of collagen (1.42) and mucopolysaccharide (1.35) (Prum and Torres,2004). The filamentous structures found by electron microscopy of the upper region here are typical of β-keratin in avian tissues. The bundles fill at least 44% of the volume of the upper region and individual filaments extend into the spaces between the membrane layers of the microstructures. It may thus be assumed that keratin is a major component. The reflecting region of the dried beak horn thus consists of keratin, the membrane folds, the refractive index of which is uncertain, and the fluid or air that fills intervening spaces. The refractive index of keratin (1.54) is slightly higher than that of collagen (Prum et al.,1999b) and we suggest that a more realistic value of the average refractive index of the reflecting region on the beak horn is at least 1.45, rather than the value of 1.4 we previously employed.

A second major parameter in Bragg's law is the lattice dimension of the photonic crystals. Preparative procedures used for electron microscopy are well known to provoke possible shrinkage effects. However, since the horns employed for this study were previously air-dried before collection, it is probable that additional shrinkage artifacts due to dehydration of specimens with alcohol would not have been produced.

Using a value of 1.45 for the refractive index of the upper region of the dried specimen, an average lattice separation (interfold distance) of 130.5 ± 11 nm, and an incident light angle of 90°, Bragg's law predicts the maximum wavelength of reflected light to be 378.5 ± 31.9 nm. This predicted value is remarkably close to that experimentally measured, demonstrating that Bragg's law provides a good estimate when the structure responsible for color is characterized by one-dimensional periodicity. It may thus be concluded that ultraviolet reflectance in the King Penguin results from the multiple-layered microstructures described here. λmax was found to range from 365 to 390 nm for dried specimens and from 367 to 420 nm on live penguins. These interspecimen variations could result from slight differences in lattice separations in different specimens. The generally higher λmax recorded with live birds probably reflects the degree of hydration of beak horns in vivo, as rehydration of isolated dried specimens resulted in an increase in λmax.

When the region containing the photonic structures had been removed experimentally by scraping, the reflectance spectrum of beak horns was characterized by a broad plateau, corresponding to a yellow/orange hue, with increased amplitude compared with intact horns. This increase is probably due to the fact that the upper region acts as a diffuse filter. The fact that this peak was unaltered by rehydrating intact beak horns suggests that it results from pigments situated in the central cellular layer of the beak horn, or in the region beneath. This reflectance spectrum is typical of carotenoid pigments.

The biological significance of the multiple-layered photonic crystals found in this study is undoubtedly linked to the visual signal they send to other individuals of the same species. As has been shown in other avian species (Hausmann et al.,2003), the most important biological function of the visual signals produced by the King Penguin beak described here is likely to be associated with choosing a mate (Parker,1995). In this context, it is of interest that during courtship displays, King Penguins flaunt their beak ornaments when encountering potential partners. The fact that the horn is both ultraviolet and also orange-pink in color would provoke an increased signal, as more than a single type of photoreceptor would be activated. The perception of the beak horn signal would also be heightened by the fact that the beak tissue surrounding the horn is black.

The microstructures are extremely numerous and of different sizes and shapes. However, not all appear to be necessary to produce a maximal signal. The percentage reflectance amplitude was not found to be significantly reduced by experimental removal of a substantial proportion of the upper region of the beak horn, indicating that the amplitude of reflectance does not result from an additive effect of all the reflective structures present throughout the upper region, but arises from reflectance from photonic structures present within a relatively limited depth of tissue, probably no greater than 10 μm thick. Although the presence of many microstructures distributed throughout the depth of the horn does not appear to reinforce the signal, this may be important in case of partial damage. In addition, the multiplicity of microstructures with slightly different orientations would produce a wider angle of reflectance. The fact that the planes of the membrane layers in a given microstructure were not oriented parallel to the beak horn surface, but varied with regard to those in other microstructures by about ±20°, would affect the visual signal from the penguin beak. According to Bragg's law, orientation of photonic crystals with regard to incident light determines both the wavelength of the reflectance and also the reflectance angle. The presence of many individual microstructures with different orientations with respect to the surface of the beak horn would result in a spread of the wavelength and also a spread in the angle over which it is reflected. The UV signal would thus not be reflected over a very narrow angle, quasiperpendicular to the beak, but would rather be perceptible over an angle of about 45°. This represents an additional factor with a bearing on the biological function of these microstructures, since it would produce a more easily perceptible signal.

In summary, the physical mechanism of UV reflection (i.e., coherent light scattering) from the King Penguin beak horn that may be concluded from the present data is similar to that proposed for feathers or avian skin. However, the general morphology and the fine structure of the UV reflecting elements present in the penguin beak differ radically from those of all previously described photonic structures. Moreover, this is the first time that UV reflecting microstructures have been characterized in the beak tissue of any bird. Several questions remain unanswered at the present time concerning the development of these microstructures. Nothing at present is known of possible morphological changes that might occur within the penguin beak horn during the months before molting occurs. Since the King Penguin is a protected species, precluding the possibility of studying the fine structure of beak horns in situ, it can not be known if beak horns increase in thickness during the year before molting and whether more photonic microstructures are produced throughout this period. Studies in progress on beak horns obtained from immature penguins could shed further light on these questions.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED

This study was performed while B.D. was on study leave in UMR 5175, CNRS, Montpellier, France.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED