The relationship between IGF-I and changes in muscle fibre phenotype in response to 6 d of stretch or disuse of the lower limb muscles of the rabbit was studied by combining in situ hybridisation and immunohistochemistry procedures. Passive stretch by plaster cast immobilisation of the muscle in its lengthened position not only induced an increase in IGF-I mRNA expression within the individual muscle fibres but also an increase in the percentage of fibres expressing neonatal and slow myosin. This change in phenotype was also found to be accompanied by a rapid and marked increase of muscle mass, total RNA content as well as IGF-I gene expression. In contrast, IGF-I appears not to be involved in muscle atrophy induced by immobilisation in the shortened position and the inactivity which results from this procedure. The level of increase in expression of IGF-I mRNA varied from fibre to fibre. By using adjacent serial sections, the fibres which expressed IGF-I mRNA at the highest levels were identified as expressing neonatal and the slow type 1 myosin. These data suggest that the expression of IGF-I within individual muscle fibres is correlated not only with hypertrophy but also with the muscle phenotypic adaptation that results from stretch and overload.

Qualitative and quantitative information is reported on the morphological changes that occur in nerve fibres and nonneuronal cells of peripheral nerve during the lifetime of the mouse. Tibial nerves of mice aged 6–33 mo were studied. With ageing, collagen accumulates in the perineurium and lipid droplets in the perineurial cells. Macrophages and mast cells increase in number, and onion bulbs and collagen pockets are frequently present. Schwann cells associated with myelinated fibres (MF) slightly decrease in number in parallel with an increase of the internodal length from 6 to 12 mo, but increase in older nerves when demyelination and remyelination are common. The unmyelinated axon to myelinated fibre (UA/MF) ratio was about 2 until 12 mo, decreasing to 1.6 by 27 mo. In older mice, the loss of nerve fibres involves UA (50% loss of 27–33 mo cf. 6 mo) more markedly than MF (35%). In aged nerves wide incisures and infolded or outfolded myelin loops are frequent, resulting in an increased irregularity in the morphology of fibres along the internodes. In the mouse there is an adult time period, 12–20 mo, during which several features of degeneration progressively appear, and an ageing period from 20 mo upwards when the nerve suffers a general disorganisation and marked fibre loss.

The discovery of fossil hand bones from an early human ancestor at Olduvai Gorge in 1960, at the same level as primitive stone tools, generated a debate about the role of tools in the evolution of the human hand that has raged to the present day. Could the Olduvai hand have made the tools? Did the human hand evolve as an adaptation to tool making and tool use? The debate has been fueled by anatomical studies comparing living and fossil human and nonhuman primate hands, and by experimental observations. These have assessed the relative abilities of apes and humans to manufacture the Oldowan tools, but consensus has been hampered by disagreements about how to translate experimental data from living species into quantitative models for predicting the performance of fossil hands. Such models are now beginning to take shape as new techniques are applied to the capture, management and analysis of data on kinetic and kinematic variables ranging from hand joint structure, muscle mechanics, and the distribution and density of bone to joint movements and muscle recruitment during manipulative behaviour. The systematic comparative studies are highlighting a functional complex of features in the human hand facilitating a distinctive repertoire of grips that are apparently more effective for stone tool making than grips characterising various nonhuman primate species. The new techniques are identifying skeletal variables whose form may provide clues to the potential of fossil hominid hands for one-handed firm precision grips and fine precision manoeuvering movements, both of which are essential for habitual and effective tool making and tool use.

The structure of porcine skin as examined by light microscopy is reviewed and its similarities to and differences from human skin are highlighted. Special imaging techniques and staining procedures are described and their use in gathering morphological information in porcine skin is discussed. Confocal laser scanning microscopy (CLSM) was employed to examine the structure of porcine skin and the findings are presented as an adjunct to the information already available in the literature. It is concluded that CLSM provides valuable additional morphological information to material examined by conventional microscopy and is useful for wound healing studies in the porcine model.

After the unilateral destruction of the dopamine input to the neostriatum there are enduring changes in rat behaviour. These have been ascribed to the loss of dopamine and the animals are often referred to as ‘hemiparkinsonian’. In the denervated neostriatum, we have shown that not only are the tyrosine hydroxylase positive boutons missing, but also the medium sized densely spiny output cells have fewer spines. Spines usually have asymmetric synapses on their heads. In a recent stereological study we were able to show that there is a loss of approximately 20% of asymmetric synapses in the lesioned neostriatum by 1 mo after the lesion. Current experiments are trying to establish the specificity of this loss. So far we have evidence suggesting that there is no obvious preferential loss of synapses from either D1 or D2 receptor immunostained dendrites in the neostriatum with damaged dopamine innervation. These experiments suggest that dopamine is somehow necessary for the maintenance of corticostriatal synapses in the neostriatum. In a different series of experiments slices of cortex and neostriatum were maintained in vitro in such a way as to preserve at least some of the corticostriatal connections. In this preparation we have been able to show that cortical stimulation results in robust excitatory postsynaptic potentials (EPSPs) recorded from inside striatal neurons. Using stimulation protocols derived from the experiments on hippocampal synaptic plasticity we have shown that the usual consequence of trains of high frequency stimulation of the cortex is the depression of the size of EPSPs in the striatal cell. In agreement with similar experiments by others, the effect seems to be influenced by NMDA receptors since the unblocking of these receptors with low Mg++ concentrations in the perfusate uncovers a potentiation of the EPSPs after trains of stimulation. Dopamine applied in the perfusion fluid round the slices has no effect but pulsatile application of dopamine, close to the striatal cell being recorded from, and in temporal association with the cortical trains, leads to a similar LTP like effect. The reduction of K+ channel conductance in the bath with TEA also has the effect of making cortical trains induce potentiation of corticostriatal transmission. TEA applied only to the cell being recorded from has no similar effect; the cortical stimulation again depresses the EPSP amplitude, so the site of action of TEA may well be presynaptic to the striatal cell. The morphological and physiological experiments may not necessarily be related but it is tempting to suggest that dopamine protects some corticostriatal synapses by potentiating them but that in the absence of dopamine others simply disconnect and are no longer detectable on electron microscopy.

In numerical simulations of skeletal muscle contractions, geometric information is of major importance. The aim of the present study was to determine whether the diffusion tensor imaging (DTI) technique is suitable to obtain valid input with regard to skeletal muscle fibre direction. The accuracy of the DTI method was therefore studied by comparison of DTI fibre directions in the rat tibialis anterior muscle with fascicle striation patterns visible on high-resolution magnetic resonance imaging (MRI) and with fibre directions in an actual longitudinal section (ALS) through the same muscle. The results showed an excellent qualitative agreement between high-resolution MRI and DTI. Despite less accurate quantitative comparison with ALS, it was concluded that DTI does indeed measure skeletal muscle fibre direction. After the experiment, it was possible to determine an appropriate voxel size (0.9 mm3) that provided enough resolution and acceptable accuracy (5°) to use DTI fibre directions in biomechanical analyses. Muscle deformation during contraction, resulting from a finite element simulation with a mesh that was directly generated from the experimental data, has been presented.

The basal ganglia (BG) have received much attention during the last 3 decades mainly because of their clinical relevance. Our understanding of their structure, organisation and function in terms of chemoarchitecture, compartmentalisation, connections and receptor localisation has increased equally. Most of the research has been focused on the mammalian BG, but a considerable number of studies have been carried out in nonmammalian vertebrates, in particular reptiles and birds. The BG of the latter 2 classes of vertebrates, which together with mammals constitute the amniotic vertebrates, have been thoroughly studied by means of tract-tracing and immunohistochemical techniques. The terminology used for amniotic BG structures has frequently been adopted to indicate putative corresponding structures in the brain of anamniotes, i.e. amphibians and fishes, but data for such a comparison were, until recently, almost totally lacking. It has been proposed several times that the occurrence of well developed BG structures probably constitutes a landmark in the anamniote-amniote transition. However, our recent studies of connections, chemoarchitecture and development of the basal forebrain of amphibians have revealed that tetrapod vertebrates share a common pattern of BG organisation. This pattern includes the existence of dorsal and ventral striatopallidal systems, reciprocal connections between the striatopallidal complex and the diencephalic and mesencephalic basal plate (striatonigral and nigrostriatal projections), and descending pathways from the striatopallidal system to the midbrain tectum and reticular formation. The connectional similarities are paralleled by similarities in the distribution of chemical markers of striatal and pallidal structures such as dopamine, substance P and enkephalin, as well as by similarities in development and expression of homeobox genes. On the other hand, a major evolutionary trend is the progressive involvement of the cortex in the processing of the thalamic sensory information relayed to the BG of tetrapods. By using the comparative approach, new insights have been gained with respect to certain features of the BG of vertebrates in general, such as the segmental organisation of the midbrain dopaminergic cell groups, the occurrence of large numbers of dopaminergic cell bodies within the telencephalon itself and the variability in, among others, connectivity and chemoarchitecture. However, the intriguing question whether the basal forebrain organisation of nontetrapods differs essentially from that observed in tetrapods still needs to be answered.

Microdamage in bone contributes to the loss of bone quality in osteoporosis and is thought to play a major role in both fragility and stress fractures (Schaffler et al. 1995). In this study, in vivo microcracks in human ribs were bulk-stained in basic fuchsin and viewed in longitudinal section and in 3 dimensions using 2 different computer-based methods of reconstruction: (1) serial sectioning of methylmethacrylate embedded sections using a sledge macrotome and identification of microcracks using UV epifluorescence followed by computerised reconstruction of microcracks using software and (2) laser scanning confocal microscopy of thick sections followed by reconstruction of microcracks into a 3-D image. The size and shape of microcracks were found to be similar using both techniques. Both techniques of reconstruction showed microcracks to be approximately elliptical in shape. From the serial sectioning reconstructions (n = 9), microcracks were found to have a mean length of 404±145 μm (mean±S.D.) (in the longitudinal direction) and mean width of 97±38 μm (in the transverse direction). Using epifluorescence microscopy, 92 microcracks were identified; mean microcrack length was 349±100 μm in the longitudinal direction. This was consistent with other results (Burr & Martin, 1993) and with the theoretical prediction of an elliptical crack shape with aspect ratio (longitudinal[ratio ]transverse) of 5[ratio ]1 deduced from analysis of random 2-D sections (Taylor & Lee, 1998). The results obtained provide new data on the nature of microcracks in bone and the method has the potential to become a useful tool in the calculation of stress intensity values which indicate the probability of an individual microcrack propagating to cause a stress or fragility fracture.

This study provides a comparative analysis of the temporal and spatial distribution of 5 intervertebral disc (IVD) proteoglycans (PGs) in sheep. The main PGs in the 2 and 10 y old sheep groups were polydisperse chondroitin sulphate and keratan sulphate substituted species. Their proportions did not differ markedly either with spinal level or disc zone. In contrast, the fetal discs contained 2 slow migrating (by composite agarose polyacrylamide gel electrophoresis, CAPAGE), relatively monodisperse chondroitin sulphate-rich aggrecan species which were also identified by monoclonal antibody 7-D-4 to an atypical chondroitin sulphate isomer presentation previously found in chick limb bud, and shark cartilage. The main small PG detectable in the fetal discs was biglycan, whereas decorin predominated in the 2 and 10 y old IVD samples; its levels were highest in the outer annulus fibrosus (AF). Versican was most abundant in the AF of the fetal sheep group; it was significantly less abundant in the 2 and 10 y old groups. Furthermore, versican was immunolocalised between adjacent layers of annular lamellae suggesting that it may have some role in the provision of the viscoelastic properties to this tissue. Versican was also diffusely distributed throughout the nucleus pulposus of fetal IVDs, and its levels were significantly lower in adult IVD specimens. This is the first study to identify versican in ovine IVD tissue sections and confirmed an earlier study which demonstrated that ovine IVD cells synthesised versican in culture (Melrose et al. 2000). The variable distribution of the PGs identified in this study provides further evidence of differences in phenotypic expression of IVD cell populations during growth and development and further demonstrates the complexity of the PGs in this heterogeneous but intricately organised connective tissue.

The management of peripheral nerve injury remains a major clinical problem. Progress in this field will almost certainly depend upon manipulating the pathophysiological processes which are triggered by traumatic injuries. One of the most important determinants of functional outcome after the reconstruction of a transected peripheral nerve is the length of the gap between proximal and distal nerve stumps. Long defects (> 2 cm) must be bridged by a suitable conduit in order to support axonal regrowth. This review examines the cellular and acellular elements which facilitate axonal regrowth and the use of acellular muscle grafts in the repair of injuries in the peripheral nervous system.

Collagen fibrils in extracellular matrices of connective tissues (tendon, cornea, etc.) are bridged and linked by the anionic glycosaminoglycans (AGAGs) of the small proteoglycans (decoron, etc.). It was proposed that these bridges and ties maintain the collagen fibril dispositions in relation to each other, helping to define tissue shape, and hence called shape modules. This investigation describes chemical and physicochemical conditions in which these structures are stable and what treatments cause their disruption. The effects on fixed and unfixed sections of tendon, cornea, lung and ear from rat, mouse and rabbit of pH, electrolyte concentration, EDTA, mercaptoethanol, hydrogen peroxide, free radicals, periodate, acetylation, urea, nonionic detergent and organic solvents were assessed by staining with Cupromeronic blue or Alcec blue in CEC techniques to localise AGAG bridges or their disintegration products. Ca2+ was not involved in the structures, oxidation/reduction had no effect and Triton X100, a nonionic detergent did not damage them. They were stable between pH 4.5 and 9.5. Periodate as a glycol-cleaving reagent did not affect them. High concentrations of urea (>2.0 m) and MgCl2 (0.5 m) disrupted the tissues. The combination of Triton and urea at concentrations too low to cause damage separately was disruptive. Free radicals in periodate solutions were damaging. Organic solvents caused collapse and rearrangements of the AGAG filaments. Acetylation caused considerable disruption of shape modules. Dermochondan but not keratan sulphate AGAGs were removed by treatment with NaOH. After fixing with glutaraldehyde only free radical and NaOH treatments were severely disruptive of shape modules. The results are compatible with a previously proposed structure for the shape modules, stabilised by hydrophobic and hydrogen bonding.

The retinal vasculature of postmortem normal human and diabetic eyes was studied using an immunohistochemical technique in conjunction with confocal laser scanning microscopy. The technique, which stained for von Willebrand factor, allowed both large areas of the retinal vasculature to be visualised and abnormalities to be studied in detail without disturbing the tissue architecture. Only one microaneurysm, defined as any focal capillary dilation, was observed in 10 normal eyes but numerous microaneurysms were seen in 4 out of 5 diabetic retinas; counts varied between 0 and 26 per 0.41 mm2 sample area. Microaneurysms were classified into 3 categories according to morphology: saccular, fusiform and focal bulges. Most were saccular, these having no preferred orientation. The majority of microaneurysms were associated with just 2 vessels suggesting they were unlikely to develop at vascular junctions. The majority were observed to originate from the inner nuclear layer and were therefore in the deeper part of the inner retinal capillary plexus. Variation in the staining of microaneurysms may correlate with endothelial dysfunction seen clinically as dye leakage during fluorescein angiography.

The pain of a migraine attack is often described as unilateral, with a throbbing or pulsating quality. The middle meningeal artery (MMA) is the largest artery supplying the dura mater, is paired, and pain-producing in humans. This artery, or its branches, and other large intracranial extracerebral vessels have been implicated in the pathophysiology of migraine by theories suggesting neurogenic inflammation or cranial vasodilatation, or both, as explanations for the pain of migraine. Having previously studied in detail the distribution of the second order neurons that are involved in the transmission of nociceptive signals from intracranial venous sinuses, we sought to compare the distribution of second order neurons from a pain-producing intracranial artery in both monkey and cat. By electrically stimulating the middle meningeal artery in these species and using immunohistochemical detection of the proto-oncogene Fos as a marker of neuronal activation, we have mapped the sites of the central trigeminal neurons which may be involved in transmission of nociception from intracranial extracerebral arteries. Ten cats and 3 monkeys were anaesthetised with α-chloralose and the middle meningeal artery was isolated following a temporal craniotomy. The animals were maintained under stable anaesthesia for 24 h to allow Fos expression due to the initial surgery to dissipate. Following the rest period, the vessel was carefully lifted onto hook electrodes, and then left alone in control animals (cat n = 3), or stimulated (cat n = 6, monkey n = 3). Stimulation of the left middle meningeal artery evoked Fos expression in the trigeminocervical nucleus, consisting of the dorsal horn of the caudal medulla and upper 2 divisions of the cervical spinal cord, on both the ipsilateral and contralateral sides. Cats had larger amounts of Fos expressed on the ipsilateral than on the contralateral side. Fos expression in the caudal nucleus tractus solitarius and its caudal extension in lamina X of the spinal cord was seen bilaterally in response to middle meningeal artery stimulation. This study demonstrates a comparable anatomical distribution of Fos activation between cat and monkey and, when compared with previous studies, between this arterial structure and the superior sagittal sinus. These data add to the overall picture of the trigeminovascular innervation of the intracranial pain-producing vessels showing marked anatomical overlap which is consistent with the often poorly localised pain of migraine.

Articular cartilage undergoes cycles of compressive loading during joint movement, leading to its cyclical deformation and recovery. This loading is essential for chondrocytes to perform their normal function of maintenance of the extracellular matrix. Various lines of evidence suggest the involvement of the cytoskeleton in load sensing and response. The purpose of the present study is to describe the 3-dimensional (3D) architecture of the cytoskeleton of chondrocytes within their extracellular matrix, and to examine cytoskeletal responses to experimentally varied mechanical conditions. Uniformly sized explants of articular cartilage were dissected from adult rat femoral heads. Some were immediately frozen, cryosectioned and labelled for filamentous actin using phalloidin, and for the focal contact component vinculin or for vimentin by indirect immunofluorescence. Sections were examined by confocal microscopy and 3D modelling. Actin occurred in all chondrocytes, appearing as bright foci at the cell surface linked to an irregular network beneath the surface. Cell surface foci colocalised with vinculin, suggesting the presence of focal contacts between the chondrocyte and its pericellular matrix. Vimentin label occurred mainly in cells of the deep zone. It had a complex intracellular distribution, with linked networks of fibres surrounding the nucleus and beneath the plasma membrane. When cartilage explants were placed into organ culture, where in the absence of further treatments cartilage imbibes fluid from the culture medium and swells, cytoskeletal changes were observed. After 1 h in culture the vimentin cytoskeleton was disassembled, leading to diffuse labelling of cells. After a further hour in culture filamentous vimentin label reappeared in deep zone chondrocytes, and then over the next 48 h became more widespread in cells of the explants. Actin distribution was unaffected by culture. Further experiments were performed to test the effects of load on the cytoskeleton. Explants were placed in culture and immediately subjected to static uniaxial radially unconfined compressive loads of 0.5, 1, 2 or 4 MPa for 1 h using a pneumatic loading device. Loads greater than 0.5 MPa maintained the vimentin organisation over the culture period. At 0.5 MPa, the chondrocytes within the explant behaved as in free-swelling culture. The rapid change in vimentin organisation probably relates to rapid swelling of the explants—under free-swelling conditions, these reached their maximum swollen size in just 15 min of culture. The chondrocytes' response to change in tissue dimensions, and thus to their relationship to their immediate environment, was to disassemble their vimentin networks. Loading probably counteracts the swelling pressure of the tissue. Overall, this work suggests that chondrocytes maintain their actin cytoskeleton and modify their vimentin cytoskeleton in response to changing mechanical conditions.

Complete cross-sections from the femoral midshaft of 180 individuals of known height and weight, aged 21–97 y, from a modern Australian population were examined using automatic video image analysis to quantify total subperiosteal porosity (TSPP). More specifically, the aim was to investigate whether age changes were similar in both sexes in (1) total subperiosteal area (TSPA), cortical area (CA) and medullary area (MA), (2) intracortical porosity (ICP), and (3) the respective contributions to TSPP made by MA and intracortical void area (ICVA). Our findings indicated that both sexes showed a significantly greater height normalised TSPA in the 70s as compared with the 20s. Males had consistently larger bones with a greater height normalised CA. In both sexes CA showed a tendency to increase till the 7th decade and then to decline, more so in females. MA approximately trebled in females and doubled in males over the age range studied. Although ICP also increased, from 4–6% in young adults to over 9% in the elderly, it showed a significant difference between the sexes only in the 3rd decade, being greater in males at this stage. By contrast, TSPP became significantly greater in females (from that recorded in the 3rd decade) by the time they reached the 50s, while in males this did not occur till the 80s. It increased from ∼25% in young adults of both sexes to ∼50% in females and ∼37% in males in their 80s. However, in the elderly there was great variability in both sexes in the appearance of bones from individuals of similar chronological age. Some bones differed little from those in younger subjects, others showed greatly increased ICP, still others displayed reduced cortical widths with low ICP. The femoral midshaft resembles other skeletal sites in that age changes in TSPP are more marked in females than males.

Regeneration poses a distinctive set of problems for evolutionary biologists, but there has been little substantive progress since these issues were clearly outlined in the monograph of T. H. Morgan (1901). The champions at regeneration among vertebrates are the urodele amphibians such as the newt, and we tend to regard urodele regeneration as an exceptional attribute. The ability to regenerate large sections of the body plan is widespread in metazoan phylogeny, although it is not universal. It is striking that in phylogenetic contexts where regeneration occurs, closely related species are observed which do not possess this ability. It is a challenge to reconcile such variation between species with a conventional selective interpretation of regeneration. The critical hypothesis from phylogenetic analysis is that regeneration is a basic, primordial attribute of metazoans rather than a mechanism which has evolved independently in a variety of contexts. In order to explain its absence in closely related species, it is postulated to be lost secondarily for reasons which are not understood. Our approach to this question is to compare a differentiated newt cell with its mammalian counterpart in respect of the plasticity of differentiation.

Research studies clearly indicate that age-related changes in cellular and tissue function are linked to decreases in the anabolic hormones, growth hormone and insulin-like growth factor (IGF)-1. Although there has been extensive research on the effects of these hormones on bone and muscle mass, their effect on cerebrovascular and brain ageing has received little attention. We have also observed that in response to moderate calorie restriction (a treatment that increases mean and maximal lifespan by 30–40%), age-related decreases in growth hormone secretion are ameliorated (despite a decline in plasma levels of IGF-1) suggesting that some of the effects of calorie restriction are mediated by modifying the regulation of the growth hormone/IGF-1 axis. Recently, we have observed that microvascular density on the surface of the brain decreases with age and that these vascular changes are ameliorated by moderate calorie restriction. Analysis of cerebral blood flow paralleled the changes in vasculature in both groups. Administration of growth hormone for 28 d was also found to increase microvascular density in aged animals and further analysis indicated that the cerebral vasculature is an important paracrine source of IGF-1 for the brain. In subsequent studies, administration of GHRH (to increase endogenous release of growth hormone) or direct administration of IGF-1 was shown to reverse the age-related decline in spatial working and reference memory. Similarly, antagonism of IGF-1 action in the brains of young animals impaired both learning and reference memory. Investigation of the mechanisms of action of IGF-1 suggested that this hormone regulates age-related alterations in NMDA receptor subtypes (e.g. NMDAR2A and R2B). The beneficial role of growth hormone and IGF-1 in ameliorating vascular and brain ageing are counterbalanced by their well-recognised roles in age-related pathogenesis. Although research in this area is still evolving, our results suggest that decreases in growth hormone and IGF-1 with age have both beneficial and deleterious effects. Furthermore, part of the actions of moderate calorie restriction on tissue function and lifespan may be mediated through alterations in the growth hormone/IGF-1 axis.

We examined cells from the nucleus pulposus and annulus fibrosus of adult bovine intervertebral discs, using confocal laser scanning microscopy on living unfixed tissue. These cells were visualised using chloromethyl fluorescein diacetate, a membrane-impermeant fluorescent dye. The organisation of cells from the outer annulus was also determined using confocal microscopy after fixation and staining the actin-filaments with FITC-phalloidin. We found that cellular processes were a dominant feature of cells from all regions of the disc including the cells of the nucleus pulposus and inner annulus. These processes were also visible in histological sections of disc examined both at the light and electron microscope level, even though cells from the nucleus and inner annulus appeared chondrocyte-like, being rounded and enclosed in a capsule. The function of these processes is at present unknown. We suggest that they may serve to sense mechanical strain.

This review describes the evolutionary history of the mammalian skull vault as a basis for understanding its complex structure. Current information on the developmental tissue origins of the skull vault bones (mesoderm and neural crest) is assessed for mammals and other tetrapods. This information is discussed in the context of evolutionary changes in the proportions of the skull vault bones at the sarcopterygian-tetrapod transition. The dual tissue origin of the skull vault is considered in relation to the molecular mechanisms underlying osteogenic cell proliferation and differentiation in the sutural growth centres and in the proportionate contributions of different sutures to skull growth.

In the course of an extensive comparative, structural and developmental study of the cranial and postcranial dermal skeleton (teeth and scales) in osteichthyan fishes, we have undertaken investigations on scale development in zebrafish (Danio (Brachydanio) rerio) using alizarin red staining, and light and transmission electron microscopy. The main goal was to know whether zebrafish scales can be used as a model for further research on the processes controlling the development of the dermal skeleton in general, especially epithelial–mesenchymal interactions. Growth series of laboratory bred specimens were used to study in detail: (1) the relationship of scale appearance with size and age; (2) the squamation pattern; and (3) the events taking place in the epidermis and in the dermis, before and during scale initiation and formation, with the aim of searching for morphological indications of epithelial-mesenchymal interactions. Scales form late in ontogeny, generally when zebrafish are more than 8.0 mm in standard length. Within a population of zebrafish of the same age scale appearance is related to standard length, but when comparing populations of different age the size of the fish at scale appearance is also related to age. Scales always appear first in the posterior region of the body and the squamation then extends anteriorly. Scales develop in the dermis but closely apposed to the epidermal–dermal boundary. Cellular modifications occurring in the basal layer of the epidermis and in the dermis before scale formation clearly indicate that the basal epidermal cells differentiate first, before any evidence of differentiation of the progenitors of the scale-forming cells in the dermis. This strongly suggests that scale differentiation could be initiated by the epidermal basal layer cells which probably produce a molecular signal towards the dermis below. Subsequently dermal cells accumulate close to the epidermis, and differentiate to form scale papillae. The late formation of the scales during ontogeny is due to a late colonisation of the dermis by the progenitors of the scale-forming cells. Because of their late formation during ontogeny and of their regular pattern of development, scales in zebrafish represent a good model for further investigations on the general mechanisms of epithelial-mesenchymal interactions during dermal skeleton development, and in particular for the study of the gene expression patterns.