Mandible mechanical properties and composition of the larval Glossosoma boltoni (Trichoptera, Insecta)

Insect feeding structures, such as mandibles, interact with the ingesta (food or/and substrate) and can be adapted in morphology, composition of material and mechanical properties. The foraging on abrasive ingesta, as on algae covering rocks, is particularly challenging because the mandibles will be prone to wear and structural failure, thus suggesting the presence of mandibular adaptations to accompany this feeding behavior. Adaptations to this are well studied in the mouthparts of molluscs and sea urchins, but for insects there are large gaps in our knowledge. In this study, we investigated the mandibles of a grazing insect, the larvae of the trichopteran Glossosoma boltoni. Using scanning electron microscopy, wear was documented on the mandibles. The highest degree was identified on the medial surface of the sharp mandible tip. Using nanoindentation, the mechanical properties, such as hardness and Young’s modulus, of the medial and lateral mandible cuticles were tested. We found, that the medial cuticle of the tip was significantly softer and more flexible than the lateral one. These findings indicate that a self-sharpening mechanism is present in the mandibles of this species, since the softer medial cuticle is probably abraded faster than the harder lateral one, leading to sharp mandible tips. To investigate the origins of these properties, we visualized the degree of tanning by confocal laser scanning microscopy. The autofluorescence signal related to the mechanical property gradients. The presence of transition and alkaline earth metals by energy dispersive X-ray spectroscopy was also tested. We found Ca, Cl, Cu, Fe, K, Mg, Mn, P, S, Si, and Zn in the cuticle, but the content was very low and did not correlate with the mechanical property values.


Specimens and preparation
Individuals of Glossosoma boltoni were collected by Prof. A. Thienemann at Plön, Germany, in 1930, and fixed in 70% EtOH (inventory number: 210, Leibniz Institute for the Analysis of Biodiversity Change, former Zoological Museum Hamburg).To perform this study, ten specimens of the ultimate larval instar were dissected.The material used is old, but we think that the material can still be used since the here presented results are part of a larger study on trichopteran mandibles.We tested the mandibles of six species; most of the material used was older, but we also studied samples that were preserved for only few years in 70% EtOH.The results from our EDX and nanoindentation measurements on G. boltoni are within the results from the fresher specimens.In these latter, no relationship between the elements and the mechanical properties could be detected.In addition, the autofluorescence signal received from G. boltoni was similarly strong to the signal received from the fresher cuticle.Surely, the drying and embedding procedure alters material properties of cuticle to some extent, but structures with a high 3dimensionality cannot be measured by nanoindentation and EDX.

Light microscopy
All specimens were first documented by light microscopy, employing a Keyence Digital Microscope VHX-7000 (KEYENCE, Neu-Isenburg, Germany) equipped with automatic stacking software.

Scanning electron microscopy (SEM)
Two heads with the first thoracic segments were deposited in small tubes that were filled with 70% EtOH and cleaned by a short ultrasonic bath for 20 s.The morphology was documented in SEM.First, heads were mounted on SEM specimen holders by double-sided adhesive carbon tape.After drying at room temperature for 30 min, samples were then sputter-coated with platinum (5 nm layer) and visualized with the Zeiss LEO 1525 (One Zeiss Drive, Thornwood, NY, USA).Afterwards, samples were rewetted by 70% EtOH.Then, the heads were dissected, the four mandibles carefully extracted, cleaned by a short ultrasonic bath for 20 s, to remove the previous sputter-coating, arranged on SEM sample holders, sputter-coated and visualized with the Zeiss Leo again (15 kV, magnifications were altered between the images).

Staining
To test whether the localities on the mandibles, which appeared worn under SEM, really experienced abrasion of the epicuticle (which reduces the intrusion of fluids into the cuticle), two mandibles of one additional specimen were stained.For this purpose, mandibles were deposited in toluidine blue (Sigma-Aldrich, St. Louis, MO, USA; 1 g in 100 ml 70% EtOH) for 2 min.The toluidine blue binds strongly to chitin and can this visualize areas without epicuticle.Afterwards, the whole mounts were documented using light microscopy.

Confocal laser scanning microscopy (CLSM)
For this step, three mandibles of two additional specimens were cleaned by a short ultrasonic bath for 20 s.Before visualization of the autofluorescence, each mandible was documented from both sides using light microscopy, to identify potential pigmentation of cuticle.Two unbroken mandibles (one in ventral and one in dorsal view) and one mandible with cut-off tip (in medial view) were arranged on object glass slides, following the procedure of 49 .Each mandible was surrounded by a stack of reinforcement rings, filled with glycerine (greater than or equal to 99.5%, free of water, Carl Roth GmbH & Co. KG, Karlsruhe, Germany) and covered with a cover slip.Following previous protocol 41,49 , samples were visualized with a Zeiss LSM 700 confocal laser scanning microscope (Carl Zeiss Microscopy GmbH, Jena, Germany), equipped with four stable solid-state lasers with wavelengths of 405 nm, 488 nm, 555 nm, and 639 nm.Bandpass or longpass emission filters (420-480 nm, greater than or equal to 490 nm, greater than or equal to 560 nm, or greater than or equal to 640 nm) were used.After scanning, images of autofluorescence were superimposed (with maximum intensity projection) using the software Zeiss Efficient Navigation (Zen) (Carl Zeiss MicroImaging GmbH).Finally, the color blue was assigned to the autofluorescence signal received from the laser with wavelength 405 nm, green to 488 nm, red (50% saturation) to 555 nm and red (50% saturation) to 639 nm.

Energy dispersive X-ray spectroscopy (EDX)
With EDX we tested two mandibles (one left mandible of one specimen and one right mandible of another specimen to compare for differences between individuals and between right and left side).The same localities of the counter mandibles from the same specimens were later tested by nanoindentation.
For this purpose, the two clean (20 s in ultrasonic bath) and dry mandibles were attached to glass object slides by double-sided adhesive carbon tape, following previous protocols 9,41,50 .Then, each mandible was surrounded by a small metallic ring.This ring was afterwards filled with epoxy resin (Reckli Epoxy WST, RECKLI GmbH, Herne, Germany) to cover the mandibles completely.Polymerization lasted for three days at room temperature.
Glass object slide and carbon tape were removed and each sample polished with different sand papers until a region of interest was on display.Since the different regions of interest (e.g., ventral and dorsal condyle) are at different levels of the sample, a sample was always polished until one target region was on display, then EDX and nanoindentation were performed.Afterwards the sample was polished until the next target region was on display.Then each step of the protocol was repeated (see below) and the next regions were tested.
To prepare samples for the EDX, the surface was, after polishing, smoothened with aluminium oxide polishing powder suspension of 0.3 μm grainsize (PRESI GmbH, Hagen, Germany) on a polishing machine (Minitech 233/333, PRESI GmbH, Hagen, Germany) to receive a plain sample surface.Before measurements, embedded mandibles were cleaned in an ultrasonic bath for 5 min to remove the polishing powder.We used sections of the mandibles to test the exocuticle by nanoindentation and EDX, but not the thin epicuticle.
Samples were mounted on SEM sample holders and sputter-coated with platinum (5 nm layer).We chose platinum for coating to have one element that could be used as marker to check if each individual EDX measurement was correct and not corrupted by e.g., surface roughness.For this, we measured later 20 areas of pure epoxy to receive values on the Pt content (mean ± SD; 0.14 ± 0.02 atomic %).After all analyses were completed, we excluded the measurements (N = 3) with very high proportions of Pt (e.g., more than 5 atomic %).
Elemental composition was determined with the SEM Zeiss LEO 1525 equipped with an Octane Silicon Drift Detector (SDD) (micro analyses system TEAM, EDAX Inc., New Jersey, USA).For each sample, the same settings were used (i.e. an acceleration voltage of 20 kV, working distance, lens opening, etc.).Before analysis, the detector was calibrated with copper.We did not perform mappings, but instead analysed small areas of the cuticle sections.
We also did not discuss the following elements, as they are the elemental basis of chitin and proteins (H, C, N, O), the coating (Pt), or the polishing powder (Al, O).For test purposes, we performed 10 EDX point measurements on the epoxy, but could not detect Si (which could be part of the sandpaper), or of other elements that we further discuss-which shows, that these elements are not an artifact of embedding and polishing, but are part of the mandibles.
The single peak of P overlaps with the one of Pt.Due of this, the software could not discriminate between these two elements and P content could not be reliably determined.Therefore, P and Pt were discussed together (P + Pt).As mentioned above, we measured 20 areas of pure epoxy to receive values on the Pt content (mean ± SD; 0.14 ± 0.02 atomic %) to further estimate the proportions of P.

Nanoindentation
Nanoindentation was performed on the two counter mandibles from the same specimens that were studied by EDX.Additionally, to increase the sample size, six mandibles of three more specimens were used for mechanical property tests (thus, overall four individuals were studied).Samples for nanoindentation were prepared following the same protocol as used for EDX: mandibles were embedded in epoxy, samples polished and smoothened (for detailed protocol, see 9,50 ).
Samples were then attached onto the nanoindenter sample holder.Indentation was performed with a nanoindenter SA2 (MTS Nano Instruments, Oak Ridge, Tennessee, USA), equipped with a Berkovich indenter tip and a dynamic contact module (DCM) head.The mechanical properties hardness (H) and Young's modulus (E) were determined from force-distance curves by applying the continuous stiffness mode.E and H were determined at penetration depths of 800-1000 nm.For each site indented, we received ~ 30 values obtained at different

Autofluorescence signals
Most regions of the mandible appeared red (showed a high autofluorescence from the lasers of 555 nm and 639 nm wavelength) (Fig. 4).A gradient within this structure could be determined: the condyles were of darker reddish colour, which became lighter towards the tip.The tip appeared green (showed a strong autofluorescence signal from the laser of 488 nm wavelength) (Fig. 4A,C).In dorsal view, the medial surface of the mandible was green in contrast to the lateral surface, which emitted a red signal (Fig. 4A).In the mandible with the cut-off tip, we found that the lateral surface emitted a red signal and the medial surface a green one (Fig. 4B).
The tissues surrounding the mandibles (i.e., muscles, etc.) appeared blue (showed a strong autofluorescence signal from the laser of 405 nm) to green (showed a strong autofluorescence signal from the laser of 488 nm wavelength) (Fig. 4A,C) or, when they covered part of the mandibles, were of purple colour as result from blue (tissue) and red colour (mandible) in overlay (Fig. 4A).

Elemental analysis by EDX
Based on the analysis of 141 point measurements, we determined Ca, Cl, Cu, Fe, K, Mg, Mn, P + Pt, S, Si, and Zn in the mandibles.All these elements were present in small proportions (< 0.30 atomic %) (see Figs. 5 and 6 and Supplementary Table 1).P + Pt (mean and standard deviation in atomic %: 0.21 ± 0.09) was detected with highest proportions, followed by Ca (0.11 ± 0.06), Mg (0.10 ± 0.05), Cu (0.07 ± 0.06), Fe (0.06 ± 0.04), Zn (0.06 ± 0.01), Si (0.03 ± 0.04), S (0.03 ± 0.02), Mn (0.03 ± 0.02), Cl (0.02 ± 0.02), and finally K (0.02 ± 0.01) with lowest proportions.Some differences between the tested regions could be determined, but we could not find a general pattern (see Supplementary Table 2 for results from pairwise comparison).When the whole content of the mandibles was compared between the two tested specimens, we could detect some differences for the proportions of Ca, Cl, Cu, Fe, K and S, but, in general, differences in proportions were rather small (see Supplementary Table 3 for means, SD and p-values).When the elements were sorted to the region and then compared between the two tested specimens, differences between the individuals were not significant (see Supplementary Table 4 for p-values).When the proportions were sorted to the tested localities, we could not detect a general pattern in the distribution as well (see Fig. 6 and Supplementary Table 5 for means and SDs).

Mechanical properties
The Young's modulus (E) describes the stiffness of a solid material and describes the relationship between axial strain and tensile stress.The hardness (H) is the measure of the resistance to local plastic deformation induced by indentation.
In the mandibles, the condyles were the stiffest and hardest regions (dorsal condyle: E mean ± SD = 7.97 ± 0.25, H = 0.37 ± 0.06 GPa; ventral condyle: E = 8.06 ± 0.31 GPa, H = 0.37 ± 0.07 GPa), followed by the lateral surface of the tip (E = 6.35 ± 0.21 GPa; H = 0.24 ± 0.05 GPa), the lateral mandible surface (E means ranged from 6.07 to 7.41 GPa and H from 0.27 to 0.37 GPa), the medial mandible surface (E means ranged from 2.54 to 7.21 GPa; H from 0.10 to 0.33 GPa) and finally the medial surface of the tip as the softest and most flexible region (E = 2.20 ± 0.46 GPa; H = 0.05 ± 0.02 GPa) (see Fig. 7 and Supplementary Table 6).For the medial side of the mandibles, we determined mechanical property gradients with the medial locality (tip) as the softest and most flexible one.Towards proximal, values of E and H increased (see Fig. 7 and Supplementary Table 6 for all values).Pairwise comparison by Wilcoxon method revealed significant differences between the different regions (see Supplementary Tables 7 and 8 for p-values).For the lateral surface of the mandibles we, however, found a different pattern: values of E and H were rather similar as revealed by Wilcoxon method pairwise comparison (see Supplementary Tables 7 and 8 for p-values).
Between the individuals, H and E values, sorted to the distinct regions, were not different (see Supplementary Table 4 for p-values).Between left and right mandible, no differences with regard to E and H could be determined (see Supplementary Table 9 for p-values).

Relationship between parameters
We determined that the following parameters were very strongly correlated (see Supplementary Table   mechanical properties and the elemental compositions were detected (see Supplementary Table 10 for correlations coefficients).
The mechanical properties were reflected by the emitted autofluorescence signal.All structures, that were harder and stiffer (e.g., the condyles) showed a strong red signal in CLSM, whereas the structures, that were softer and more flexible (e.g., the medial cuticle of the mandible tip) appeared green in CLSM.

Discussion
Insects represent the most species-rich and diverse animal group (see 51 ), are abundant in nearly all habitats and provide essential ecological functions by their actions (for review, see 52 ).One reason for this diversity can be found, among other factors, in the strong variety of life history traits, which includes the foraging on different food sources or from different surfaces.Since mouthparts are interfaces between the organisms and their preferred food, research on them contributes to our understanding of insect evolution, behaviour and ecology (for review, see 53 ).Insect mouthpart morphologies were intensively investigated in the past, giving important insights into functional principles and ecology (for reviews, see 52,53 ).With regard to the mechanical properties (e.g., hardness and Young's modulus), which determine the functionality of mouthparts, there are large gaps in knowledge, as only few studies address them and their origins (i.e., the degree of tanning or the inorganic content) (see [24][25][26]32,41 ).
When investigating the adaptations in mouthparts to the food, Trichoptera larvae can be regarded as a good model system, since the larvae are ecologically highly diverse and show adaptations to grazing, shredding, filtering, or predatory behaviour (see 42,[54][55][56][57][58][59] ).We here investigated the mechanical properties of just one species, but, however, hope, that mechanical property tests on more species with distinct ecological niches will follow.
In mandibles, E values from ~ 6 to ~ 11 GPa and H values from 0.4 to 1.2 GPa were documented in termites 73 , E values from ~ 7 to ~ 11 GPa and H values from ~ 0.7 to ~ 1.0 GPa in dragonfly nymphs 76 , E values from ~ 7 to ~ 11 GPa and H values from ~ 0.3 to ~ 1.0 GPa in beetle larvae 77 , and E values from ~ 3.5 to ~ 20.9 GPa and H values from ~ 0.2 to ~ 2.0 GPa in antlion larvae 41 .The maximum values of E and H of Glossosoma boltoni (see Fig. 8 for habitus) are in general within the range of ant, dragonfly nymph and beetle larvae mouthparts.However, in G. boltoni the mandible tip, which interacts with the ingesta, was softer and more flexible inside and harder and stiffer outside, which was not detected for the other insect mandibles tested by nanoindentation; they show decreasing mechanical property values from the interacting mouthpart regions towards the mouthpart basis.

Origin of mechanical property gradients
The heterogeneity in mechanical properties of insect cuticle (see review 32 ) can come from the degree of sclerotization by quinone reactions 24,78 , the chitin microstructure, the abundance of the mineral content (abundant in small proportions), proteins, transition metals and halogens ( 28,29 ; for review on mechanical property gradients and their various origins, see 31 ).
The degree of sclerotisation and its relationship with the mechanical properties were well investigated in cuticle structures.Autofluorescence signals, received after laser excitation via CLSM according to the protocol of 49 , are commonly studied to identify cuticle regions with certain dominating material compositions.Sclerotized and stiff cuticle is related to the signal received after excitation with lasers of 555 nm and 639 nm wavelength (here, the red colour was assigned to the signal).Less sclerotized chitin-rich cuticle, which is flexible and relatively tough, emits signals from the laser of 488 nm wavelength (here, the green colour was assigned to the signal).Regions with high proportions of resilin, an elastic and flexible protein (see 107 ; for review see 108 ), or other proteins 109,110 emit a strong signal after excitement with the laser of 405 nm wavelength (here, the blue colour was assigned to the signal).Regions of mixed material compositions can appear brown, yellow, or pink from overlay of different signals.CLSM protocols, mostly following 49 , were previously applied to chitinous structures, such as wings 111 , foot attachment devices 112 , antennae 113 , genitals 114 , and also mouthparts ( [115][116][117][118][119][120] ; for Trichoptera, see 121 ).www.nature.com/scientificreports/A relationship between the autofluorescence signals and the material properties (see 122 ) was previously crossvalidated via AFM-nanoindentation for leg attachment devices in lady bird beetles 112 , antlion larvae mandibles 41  6 for EDX results sorted to each locality).Please note that the EDX analyses were performed on embedded and polished samples (sections of mandibles), this SEM image just illustrates the localities of the studied regions.Scale bar: 80 µm.and for chitinous radular teeth of gastropod molluscs 50 .The same relationship is also detected here, as we determined that blue to green areas are softer and more flexible (e.g., the medial surface of the mandible tip) than regions of red colour (e.g., the bases of the mandible or the lateral surface of the mandible tip), which suggests that the mechanical property gradients in Glossosoma boltoni have their origin in the degree of tanning.This relationship was previously also suggested by 11 for the mandibles of this species.

Interaction with hard substrate
Interaction with hard substrate is challenging due to the risk of structural failure or wear.The chitinous molluscan teeth are among the best studied structures that interact directly with hard and abrasive ingesta as e.g.stones.
Here, the teeth can be heavily mineralized with iron oxides and/or silica (Polyplacophora and Patellogastropoda), which are located between the chitin fibres and increase the hardness and elasticity dramatically (for reviews, see 13,14,123 ).However, in some gastropod taxa (e.g.Paludomidae) foraging from stones, teeth are not heavily mineralized 50,124,125 .Here, teeth can withstand high forces by their clever geometry and mechanical property gradients, which allows the teeth of one row to bend and to gain support from the adjacent row distributing the stress 8,63 .Additionally, the tooth tips show a high degree of cross-linking of the chitin fibres (sclerotization), which seems to reduce the degree of wear 50 .In G. boltoni, we did not determine high proportions of metals, but found by CLSM that the mechanical property gradients probably originate from the degree of tanning.This system seems to be analogous to the not-mineralized radular teeth.
From the wear distribution we could receive data on the regions that interact with the hard substrate during foraging.As the medial surface of the mandible tip showed sign of heavy wear (Fig. 2F), we propose that this species performs a scooping motion on the rocks to remove algae.The wear on the outer mandible surface (the lateral side; see Fig. 2A-E) could potentially come from interactions with adjacent sand particles or adjacent smaller stones.We could not identify wear on the mandible teeth (Fig. 1H), which indicates that they are not used for loosening algae from stones; they are potentially involved in cutting larger algae.www.nature.com/scientificreports/

Self-sharpening and function
Due to the abrasive particles or substrates (ingesta), wear on a cutting edge of tools or structures can affect its sharpness and hence functionality.Adaptations to prevent or reduce dulling can be found in teeth of beavers 126,127 , Triceratops 128 , sea urchins 3,5,7 and molluscs 1,4,6 .Additionally, self-sharpening mechanisms were previously determined for the mandibles of the locust Locusta migratoria by mechanical property tests 96 and proposed for the mandibles of beetles by documenting the geometry 77,84 .Based on the data from 96 , insect mandibles were already on focus in the development of self-sharpening cutting tools 129 .
In general, self-sharpening mechanisms, which are also applied to bionic tribological systems (see 130,131 ), result from the following: the hardness is not uniform within structure.One part (e.g., the layers of enamel on the beaver incisor surface, see 127 ), which is hard and wear-resistant, interacts with the food or substrate.The second part, which is rather soft, can withstand possible impacts, compressive and tensile stresses (e.g., the dentine underneath the enamel in beaver incisor, see 127 ).However, this region shows a high degree of wear.Due to weaker interfaces between the harder and the softer region, cracks produced during feeding can propagate along the interfaces more easily than perpendicular to them (see 3 ).This heterogeneity can be achieved by the microstructure of the organic matrix and the distribution of inorganics and results in a sharpening of the tip profile.
In Glossosoma boltoni, we could not test the interfacial strengthening, but determined mechanical property heterogeneities between the medial and the lateral mandible surface (cuticle).The mechanical property distribution of G. boltoni is different to the one determined in the self-sharpening mandibles of the locust Locusta migratoria; here the lateral surface of one mandible tip was harder than the other and the medial surface of the counter mandible tip was harder than the other 96 .
However, the medial surface of G. boltoni was significantly softer, more flexible and showed pronounced signs of wear, similar to the situation in beavers.In contrast, the lateral surface, which presumably interacts with the stone during feeding, was significantly harder, stiffer and showed less wear, similar to the situation in beavers as well.Potentially, when the mandible tip scratches across the stone surface to loosen algae, the soft medial cuticle abrades faster than the harder lateral cuticle, which would lead to a sharp cutting edge at the mandible tip.By SEM we found that mandible tips of G. boltoni appeared rather sharp in profile, similar to the situation found in some beetle mandibles that are thought to be self-sharpened as well 77,84 .Due to these indications, we suggest that a self-sharpening mechanism is present in the mandibles of these trichopteran larvae.However, this should be verified in the future, potentially in the form of in situ wear studies, similar to studies on sea urchin teeth 7 .

Conclusion
We studied the mandibles of a trichopteran larvae foraging from stones by scanning electron microscopy, confocal laser scanning microscopy, energy dispersive X-ray spectroscopy and nanoindentation.We found regional heterogeneities in mechanical properties: the softest and most flexible region was the medial surface of the mandible tip, which showed high degrees of wear, whereas the lateral region was hard and stiff.Due to the presence of these gradients, which were found to relate to the degree of tanning, we suggest that potentially a self-sharpening mechanism is present in these mouthparts facilitating the foraging on challenging surfaces.

Figure 1 .
Figure 1.SEM images of the head and mandibles of larval Glossosoma boltoni.(A) Ventral view on the head and the first thoracic segment of one specimen.(B) Dorsal view on the left mandible.(C) Magnification of the trichoid sensilla.(D) Magnification of the setae surfaces.(E) Ventral view on the left mandible with wear on the medial surface of the tip.(F) Magnification of the prostheca.(G,H) Mandibular teeth on the dorsal edge of the mandible with magnification (H).Cd dorsal condyle, Cv ventral condyle, di distal, la lateral, Li labium, Lr labrum, me medial, Md mandible, Ms medial surface, Mt mandibular tooth, Mx maxilla, pr proximal, Pr prostheca, Se setae, Tp tip, ve ventral.Scale bars: (A) 400 µm; (B,E) 80 µm; (C,F,G) 20 µm; (D,H) 4 µm.

Figure 2 .
Figure 2. SEM images of wear on the mandibles of larval Glossosoma boltoni.(A) Ventral view on the head, showing the interaction between both mandibles.(B-D) Ventral view on the right mandible (B) with magnification of the scratches on the tip (C) and the wear on the lateral surface exposing the chitin fibres (D).(E) Magnification of scratches on lateral surface of the left mandible.(F) Ventral view on the left mandible with high degrees of wear on the medial surface of the tip exposing the chitin fibres.The dashed line represents the locality of the section (on the embedded sample) to test the mechanical properties and the elemental composition of the medial and lateral mandible surface.(G,H) Medial view on the right mandible at a high magnification (H) showing the sharp cutting edge of the mandible tip.di distal, do dorsal, la lateral, Ls lateral surface (the surface of the mandible facing towards lateral), me medial, Ms medial surface (the surface of the mandible facing towards medial), pr proximal, Tp tip, ve ventral.Scale bars: (A,G) 80 µm; (B,H) 20 µm; (C-E) 4 µm, (F) 40 µm.

Figure 3 .
Figure 3. (A) Light microscopy image of an unstained mandible.(B) Mandible stained with toluidine blue.The setae, internal tissues and the mandible tip are blue.Ar artifact, a loose piece of tissue, Cv ventral condyle, Ti tip.

Figure 4 .
Figure 4. CLSM images of three different mandibles.All images were taken separately, thus the autofluorescence is not directly comparable between the images.(A) Dorsal view on left mandible (same specimen as in C). (B) Medial view on the left mandible with the cut-off tip.(C) View from medial on the right mandible (same specimen as in A).Cd dorsal condyle, di distal, do dorsal, Ls lateral surface, la lateral, me medial, Mt mandibular tooth, pr proximal, Se setae, Tp tip, Os outer surface, ve ventral.

Figure 5 .
Figure 5. Results from EDX analysis, given in atomic %, based on the testing of two mandibles.(A) Elemental content of the tip, the lateral and medial surfaces and the condyles.(B) Results, summarized from all regions, sorted to the body side (left and right mandible are from two different specimens.(C) Dorsal view on one right mandible.Letters represent the localities tested (see Fig.6for EDX results sorted to each locality).Please note that the EDX analyses were performed on embedded and polished samples (sections of mandibles), this SEM image just illustrates the localities of the studied regions.Scale bar: 80 µm.

Figure 6 .
Figure 6.(A) Results from EDX analysis, given in atomic %, based on the testing of two mandibles.The results are sorted to the individual localities tested (see SEM image for the localities).(B) Dorsal view on one right mandible.Letters represent the localities tested.Please note that the EDX analyses were performed on embedded and polished samples (sections of mandibles).The SEM image illustrates the localities of the studied regions.Scale bar: 80 µm.

Figure 7 .
Figure 7. (A) SEM image shows the localities tested with nanoindentation.Please note that the nanoindentation was performed on embedded and polished samples (sections of mandibles); this SEM should rather illustrate the localities of the studied regions.(B) Light microscopy image of an embedded mandible.(C) Results from nanoindentation: hardness (H) and Young's modulus (E), both given in GPa, for the tested regions on the mandible tip, medial and lateral surfaces, and both condyles.These results are based on the testing of eight mandibles.Scale bar: 80 µm.