Arthropod Cuticle: Time-Lapse 3D Imaging to Assess Toughening and Failure Mechanisms

Arthropod cuticle: Time-lapse 3D imaging to assess toughening and failure mechanisms

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Science and Engineering.

2019

Daniel Sykes

School of Natural Sciences

Department of Materials TABLE OF CONTENTS

INTRODUCTION: 11 Fracture mechanics of arthropod cuticle 13 Toughening mechanisms in composite materials 15 The use of X-ray computed tomography to investigate microstructure 19 Aims of Papers 1, 2, 3 & 4 22 Alternative format 23 References 25

PAPER 1: Arthropod cuticle: a biological material with diverse mechanical properties 31

PAPER 2: Time-lapse three-dimensional imaging of crack propagation in beetle cuticle 84

PAPER 3: Preservation of mechanical properties in locust tibiae for in situ time-lapse three-dimensional imaging 93

PAPER 4: Effect of hydration on crack propagation in beetle elytra using time-lapse three-dimensional imaging 112

DISCUSSION: 140 Development of a methodology for in situ mechanical testing of hydration-sensitive biological materials 145 Toughness in arthropod cuticle 146 Future work 148 Conclusions 152 References 153

APPENDIX A: Supplementary information 159

Word count: 46,267

ABSTRACT

Thesis title: Arthropod cuticle: Time-lapse 3D imaging to assess toughening and failure mechanisms

Name: Daniel Sykes Institution: University of Manchester Degree Title: Doctor of Philosophy Date: June 2019

This thesis concentrates on one of the least studied mechanical properties of cuticle – toughness. In particular, it investigates how cuticle microstructure impacts crack propagation, and the toughening mechanisms cuticle possesses to protect the vulnerable soft tissues contained within the exoskeleton. I use time-lapse 3D nCT imaging with in situ mechanical tests to investigate toughness and damage progression in arthropod cuticle. In addition, I develop new methodologies of standardised sample preparation and hydration preservation to perform quantitative analysis of fresh locust tibiae and beetle elytra. The results obtained in this thesis have shown that it is possible with these techniques to analyse toughening in fresh and dry cuticles, by qualitative visualisation of the interaction between microstructure and crack propagation and quantitative measurement of toughness values from standardised test samples. It was shown that microstructure is responsible for the numerous extrinsic toughening mechanisms present in cuticle, of which many were previously unreported, and that hydration is responsible for improving the effectiveness and frequency of their occurrence. Furthermore, it was found that the exocuticle contributes little to toughness and that the difference in angle between the crack direction and the fibre orientations of a lamina directly affected the toughening capability of that lamina. To summarise, this thesis displays how a combined approach using state-of-the-art 3D imaging, in situ mechanical testing, sample preparation and hydration preservation techniques can provide us with new insights in how arthropod cuticle shape, microstructure, composition and mechanical properties interact. 3

DECLARATION

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

I. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the "Copyright") and she has given the University of Manchester certain rights to use such Copyright, including for administrative purposes.

II. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

III. The ownership of certain Copyright, patents, designs trademarks and other intellectual property (the "Intellectual Property") and any reproductions of copyright works in this thesis, for example graphs and tables ("Reproductions"), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

IV. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=24420), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/) and in The University’s policy on Presentation of Theses.

ACKNOWLEDGEMENTS

I would like to thank everyone who has helped and supported me during my PhD.

Firstly, I would like to thank my supervisors Russell Garwood, Philip Withers and Shelley

Rawson for their help and guidance throughout my project, their suggestions for solutions to the problems I encountered and their recommendations on how to improve my experiments, manuscripts and figures. I would also like to thank my colleagues and ex- colleagues Stuart Morse, Julia Behnsen, James Carr, Rob Bradley, Rebecca Hartwell and

Chris Egan. During our shared time at the University of Manchester they provided invaluable help, guidance and recommendations.

I am very grateful to my parents, Christopher Sykes and Sandra Sykes, and my brother,

Gary Sykes, for their continual encouragement and support. The last three and a half years would have been impossible for me without it. I would particularly like to thank my parents for their support from a young age to achieve my ambitions, from waking up in the early hours every day to transport me to the train station so I could go to Greenhead college; to supporting me in my decision to study in Southampton and supporting me throughout my time there; to the continual help, visits and support they have given me during my time at the Natural History Museum and the University of Manchester. I would also like to thank my extended family, my grandma, my uncles and aunties: Tracey, Simon and Lynne, my cousins: Evie, Alice, Karen, Suzanne, Jade, Holly and Nick, and my dearly missed grandparents: Molly, Cyril and Donald. Although I have not been able to spend as much time with them during my PhD as I would have liked, the time I did manage to have was a really appreciated respite from my PhD.

Some of my greatest support during my PhD has come from my friendships that started during my undergraduate studies. Benjamin Evans, Rebecca Summerfield, Charlotte

Brooks, Kasimir Marks, Anna Mogridge, James Mogridge, Ella Smith and Adam Lomax are great friends who I also wish I had more time with in the last three and a half years, but the time I have had has helped me immeasurably. I developed many lasting friendships during my time at the Natural History Museum that have been a great support to me, but

I would like to particularly thank Natasha Vasiliki Almeida. She has been an unwavering support and the similarities between the courses of our PhDs meant her help has been invaluable and vital at all times. I would also like to thank the friends I have made here in

Manchester and ones I already had that arrived here before me. Russell Garwood,

Charlotte Brassey, James O’Sullivan and all the members of Bill Seller’s lab have made me feel welcome here. I would particularly like to thank Thomas Puschel Rouliez, who has been a great distraction from my PhD by being a true friend and allowing me to have fun and relax once in a while.

Without a doubt the most important person during my PhD and in my life is Fernanda

Bribiesca Contreras. Since we met in our first year of PhD she has been an undying source of love, support, happiness, kindness and inspiration. She has inspired me to be a better person and scientist by example, and has made my life a happier and more enjoyable one.

Even in the toughest times, she has always known how to make me laugh away pain and make everything easier. Muchas gracias Fer por todo, te amo mucho y estoy muy emocionado de pasar mi vida contigo.

THE AUTHOR

Daniel Sykes

Research interests

My main research focus is on studying the links between biological forms and structures and their function. I apply computed tomography and advanced 3D image analyses to produce 3D volumetric data of the external form and internal structure of biological materials.

Education

Ph.D Materials

Thesis topic: Arthropod cuticle: Time-lapse 3D imaging to assess toughening and failure mechanisms

School of Materials, University of Manchester, UK.

Supervisors: Prof. Philip J. Withers, Dr. Russell J. Garwood and Dr. Shelley D. Rawson

M.Sci Marine Biology

Dissertation topic: Three-dimensional (3D) imaging of polychaete internal anatomy using micro-computed X-ray tomography

University of Southampton, UK

Supervisor: Dr. Gordon L.J. Paterson

Refereed journal publications

1. Paterson GL, Sykes D, Faulwetter S, Merk R, Ahmed F, Hawkins LE, Dinley J, Ball AD, Arvanitidis C. 2014. The pros and cons of using micro-computed tomography in gross and micro-anatomical assessments of polychaetous annelids. Memoirs of Museum Victoria 71:237–46. 2. Reumont BM von, Campbell LI, Richter S, Hering L, Sykes D, Hetmank J, Jenner RA, Bleidorn C. 2014. A Polychaete’s Powerful Punch: Venom Gland Transcriptomics of Glycera Reveals a Complex Cocktail of Toxin Homologs. Genome Biology and Evolution 6:2406–2423. DOI: 10.1093/gbe/evu190.

3. Amon DJ, Sykes D, Ahmed F, Copley JT, Kemp KM, Tyler PA, Young CM, Glover AG. 2015. Burrow forms, growth rates and feeding rates of wood-boring Xylophagaidae bivalves revealed by micro-computed tomography. Frontiers in Marine Science 2:10. DOI: 10.3389/fmars.2015.00010. 4. Andersen T, Baranov V, Goral T, Langton P, Perkovsky E, Sykes D. 2015. First record of a Chironomidae pupa in amber. Geobios 48:281–286. DOI: 10.1016/j.geobios.2015.06.004. 5. Brereton NJB, Ahmed F, Sykes D, Ray MJ, Shield I, Karp A, Murphy RJ. 2015. X-ray micro-computed tomography in willow reveals tissue patterning of reaction wood and delay in programmed cell death. BMC Plant Biology 15:83. DOI: 10.1186/s12870-015-0438-0. 6. Dunn JC, Halenar LB, Davies TG, Cristobal-Azkarate J, Reby D, Sykes D, Dengg S, Fitch WT, Knapp LA. 2015. Evolutionary Trade-Off between Vocal Tract and Testes Dimensions in Howler Monkeys. Current Biology 25:2839–2844. DOI: 10.1016/j.cub.2015.09.029. 7. Hall AC, Sherlock E, Sykes D. 2015. Does Micro-CT scanning damage DNA in museum specimens? Journal of Natural Sciences Collections 2:8. 8. Parry LA, Wilson P, Sykes D, Edgecombe GD, Vinther J. 2015. A new fireworm (Amphinomidae) from the Cretaceous of Lebanon identified from three- dimensionally preserved myoanatomy. BMC Evolutionary Biology 15:256. DOI: 10.1186/s12862-015-0541-8. 9. Soisook P, Struebig MJ, Noerfahmy S, Bernard H, Maryanto I, Chen S-F, Rossiter SJ, Kuo H-C, Deshpande K, Bates PJJ, Sykes D, Miguez RP. 2015. Description of a New Species of the Rhinolophus trifoliatus-Group (Chiroptera: Rhinolophidae) from Southeast Asia. Acta Chiropterologica 17:21–36. DOI: 10.3161/15081109ACC2015.17.1.002. 10. Adams NF, Collinson ME, Smith SY, Bamford MK, Forest F, Malakasi P, Marone F, Sykes D. 2016. X-rays and virtual taphonomy resolve the first Cissus (Vitaceae) macrofossils from Africa as early-diverging members of the genus. American Journal of Botany 103:1657–1677. DOI: 10.3732/ajb.1600177. 11. Agbesi MPK, Borsuk HS, Hunt JN, Maclaine JS, Abel RL, Sykes D, Ramsey AT, Wang Z, Cox JPL. 2016a. Motion-driven flow in an unusual piscine nasal region. Zoology 119:500–510. DOI: 10.1016/j.zool.2016.06.008. 12. Agbesi MPK, Naylor S, Perkins E, Borsuk HS, Sykes D, Maclaine JS, Wang Z, Cox JPL. 2016b. Complex flow in the nasal region of guitarfishes. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 193:52–63. DOI: 10.1016/j.cbpa.2015.12.007. 13. Britz R, Doherty-Bone TM, Kouete MT, Sykes D, Gower DJ. 2016. Monopterus luticolus, a new species of swamp eel from Cameroon (Teleostei: Synbranchidae). Ichthyological Exploration of Freshwaters 27:309–323. 14. Collinson ME, Adams NF, Manchester SR, Stull GW, Herrera F, Smith SY, Andrew MJ, Kenrick P, Sykes D. 2016. X-ray micro-computed tomography (micro-CT) of pyrite-permineralized fruits and seeds from the London Clay Formation (Ypresian)

conserved in silicone oil: a critical evaluation 1. Botany 94:697–711. DOI: 10.1139/cjb-2016-0078. 15. Gower DJ, Wade EOZ, Spawls S, Böhme W, Buechley ER, Sykes D, Colston TJ. 2016. A new large species of Bitis Gray, 1842 (Serpentes: Viperidae) from the Bale Mountains of Ethiopia. Zootaxa 4093:41–63. DOI: 10.11646/zootaxa.4093.1.3. 16. Heron DPL, Alderton DHM, Collinson ME, Grassineau N, Sykes D, Trundley AE. 2016. A eukaryote assemblage intercalated with Marinoan glacial deposits in South Australia. Journal of the Geological Society 173:560. DOI: 10.1144/jgs2015- 156. 17. Karunaratne A, Xi L, Bentley L, Sykes D, Boyde A, Esapa CT, Terrill NJ, Brown SDM, Cox RD, Thakker RV, Gupta HS. 2016. Multiscale alterations in bone matrix quality increased fragility in steroid induced osteoporosis. Bone 84:15–24. DOI: 10.1016/j.bone.2015.11.019. 18. Mitchell RL, Cuadros J, Duckett JG, Pressel S, Mavris C, Sykes D, Najorka J, Edgecombe GD, Kenrick P. 2016. Mineral weathering and soil development in the earliest land plant ecosystems. Geology 44:1007–1010. DOI: 10.1130/G38449.1. 19. Smith DB, Bernhardt G, Raine NE, Abel RL, Sykes D, Ahmed F, Pedroso I, Gill RJ. 2016. Exploring miniature insect brains using micro-CT scanning techniques. Scientific Reports 6:21768. DOI: 10.1038/srep21768. 20. Stull GW, Adams NF, Manchester SR, Sykes D, Collinson ME. 2016. Revision of Icacinaceae from the Early Eocene London Clay flora based on X-ray micro-CT. Botany 94:713–745. DOI: 10.1139/cjb-2016-0063. 21. Swart P, Wicklein M, Sykes D, Ahmed F, Krapp HG. 2016. A quantitative comparison of micro-CT preparations in Dipteran flies. Scientific Reports 6:39380. 22. Manchester SR, Collinson ME, Soriano C, Sykes D. 2017. Homologous Fruit Characters in Geographically Separated Genera of Extant and Fossil Torricelliaceae (Apiales). International Journal of Plant Sciences 178:567–579. DOI: 10.1086/692988. 23. Strekopytov S, Brownscombe W, Lapinee C, Sykes D, Spratt J, Jeffries TE, Jones CG. 2017. Arsenic and mercury in bird feathers: Identification and quantification of inorganic pesticide residues in natural history collections using multiple analytical and imaging techniques. Microchemical Journal 130:301–309. DOI: 10.1016/j.microc.2016.10.009. 24. Britz R, Sykes D, Gower DJ, Kamei RG. 2018. Monopterus rongsaw, a new species of hypogean swamp eel from the Khasi Hills in Northeast India (Teleostei: Synbranchiformes: Synbranchidae). Ichthyological Exploration of Freshwaters 1086:1–12. DOI: 10.23788/IEF-1086. 25. Gooday AJ, Sykes D, Góral T, Zubkov MV, Glover AG. 2018. Micro-CT 3D imaging reveals the internal structure of three abyssal xenophyophore species (Protista, Foraminifera) from the eastern equatorial Pacific Ocean. Scientific Reports 8:12103. DOI: 10.1038/s41598-018-30186-2. 26. Daemi SR, Lu X, Sykes D, Behnsen J, Tan C, Palacios-Padros A, Cookson J, Petrucco E, J. Withers P, L. Brett DJ, R. Shearing P. 2019. 4D visualisation of in situ nano-

compression of Li-ion cathode materials to mimic early stage calendering. Materials Horizons 6:612–617. DOI: 10.1039/C8MH01533C. 27. Martin-Silverstone E, Sykes D, Naish D. 2019. Does postcranial palaeoneurology provide insight into pterosaur behaviour and lifestyle? New data from the azhdarchoid Vectidraco and the ornithocheirids Coloborhynchus and Anhanguera. Palaeontology 62:197–210. DOI: 10.1111/pala.12390. 28. Sykes D, Hartwell R, Bradley RS, Burnett TL, Hornberger B, Garwood RJ, Withers PJ. 2019. Time-lapse three-dimensional imaging of crack propagation in beetle cuticle. Acta Biomaterialia 86:109–116. DOI: 10.1016/j.actbio.2019.01.031.

INTRODUCTION

Arthropods are a highly successful group of animals in terms of abundance, distribution and diversity, they can be found in every habitat and account for 66% of all known species on Earth (Grimaldi & Engel, 2005). Part of their success results from their characteristic external skeletal system, the exoskeleton, which is a multi-layered, hierarchically organised composite (Neville, 1975). Despite their importance to arthropod success, our understanding of the nature of their cuticular exoskeleton still contains significant gaps.

For example, in comparison to internal skeletal systems it must provide a diverse range of functions as it directly interacts with the environment. It is required to act as a physiological barrier, a structural support, a defence mechanism, a means of inter- and intra-specific attack, an adhesive surface, a sensory network and even to form wings enabling flight (Neville, 1975; Vincent, 2002; Vincent & Wegst, 2004; Cribb et al., 2007).

This diversity of function is all provided by cuticle, a biological material which is composed of chitin-protein fibres and often lacks mineralisation (Neville, 1975; Vincent & Wegst,

2004). Furthermore, this composite, despite consisting of simple biological constituents, can possess high stiffness, hardness and/or toughness (Sun, Tong & Ma, 2008; Dirks &

Taylor, 2012a; Weaver et al., 2012; Amini & Miserez, 2013). It is assumed that the hierarchical organisation of the chitin-protein fibres within the cuticle microstructure has the most significant impact in determining its mechanical properties (Neville, 1975;

Vincent & Wegst, 2004; Sachs, Fabritius & Raabe, 2008; Nikolov et al., 2010). This has important implications from a biological standpoint and for composite design. By modulating the organisation, composition and hydration of the chitin-based fibres

(Neville, 1975; Vincent & Wegst, 2004; Klocke & Schmitz, 2011), it is possible for

12 arthropods to produce a lightweight material covering their entire bodies that has different functions and properties in different structures.

This thesis will concentrate on one of the least studied mechanical properties of cuticle – toughness (Vincent & Wegst, 2004; Dirks & Taylor, 2012a). In particular, it investigates how cuticle microstructure impacts crack propagation, and the toughening mechanisms cuticle possesses to protect the vulnerable soft tissues contained within the exoskeleton.

As for vertebrates, damage to the structural integrity of the skeletal system can be fatal, and the larger the crack that propagates through the cuticle, the harder it is to repair

(Parle & Taylor, 2013; Parle, Dirks & Taylor, 2016, 2017). There is a clear evolutionary advantage to limiting crack growth as much as possible, in order to reduce the level of self-repair that is required of the organism (Parle, Dirks & Taylor, 2017). By understanding how cuticle microstructure impedes crack growth, we can deduce how different cuticles prioritise different properties. This can give us an understanding of which aspects of the life history of an arthropod species are driving the evolutionary development of their cuticle, and inform decisions in the development and production of tougher composite materials.

Fracture mechanics of arthropod cuticle

The fracture mechanics of arthropod cuticle has been studied in a handful of cuticle types: the tibiae and wings of the desert locust Schistocercia gregaria (Dirks & Taylor,

2012a,b); the claws of the American lobster Homarus americanus (Sachs, Fabritius &

Raabe, 2008), the claws of the stone crab Menippe mercenaria (Melnick, Chen &

Mecholsky, 1996), and the elytra (wing cases) of the dung beetle Copris ochus (Sun &

Tong, 2007). Despite this narrow sampling of arthropod diversity, impressive values of fracture toughness (KIC) and work of fracture (GC) have been reported. Fracture toughness

13 is the mechanical property that describes its resistance to propagation of cracks; and work of fracture (also known as strain energy release rate) is the energy required to grow

½ a crack by unit area. KIC in cuticle has been found to range between 1.0 MPa m in M. mercenaria claws (Melnick, Chen & Mecholsky, 1996) and 4.1 MPa m½ in S. gregaria tibiae

(Dirks & Taylor, 2012a), values that are comparable to mineralised, biological composites such as bone and nacre (Wegst & Ashby, 2004), and to ceramics (Gogotsi, 2003). GC in locust tibiae has been found to be as high as 5.56 kJ m-2 (Dirks & Taylor, 2012a), making cuticle one of the toughest biological composites studied so far, similar to antler and horn

(Vincent & Wegst, 2004), and even tougher than some mineralised biological materials like bone (Wegst & Ashby, 2004). However, the aspects of cuticle composition and microstructure which control its toughness are poorly understood. Both mineralised and unmineralised cuticles can have similar values of KIC (Melnick, Chen & Mecholsky, 1996;

Dirks & Taylor, 2012b). The effect of the microstructure on toughening has not been investigated at all. Currently, the only aspect that has been tested is the effect of water content in the cuticle. Dried cuticle is known to be stiffer and harder than fresh (i.e. hydration level maintained as in vivo) cuticle (Schöberl & Jäger, 2006; Klocke & Schmitz,

2011). A study that compared the fracture toughness and work of fracture of dried and fresh locust tibiae (Dirks & Taylor, 2012a), found that desiccation caused KIC to decrease

½ ½ by two times (4.12 MPa m in fresh tibiae to 2.06 MPa m in dry tibiae) and GC to decrease by eight times (5.56 kJ m-2 in fresh tibiae to 0.68 kJ m-2 in dry tibiae).

Significant gaps exist in our knowledge of fracture mechanics in cuticle. Although hydration has been shown to have a significant impact on the toughness of cuticle (Dirks

& Taylor, 2012a), it is not known how this impacts the toughening mechanisms that cuticle possesses to hinder crack growth. Studies of the fracture surfaces and crack propagation in cuticle have been restricted to 2D imaging (Chen et al., 2001; Sachs, 14

Fabritius & Raabe, 2006; Kundanati et al., 2018). As such, the interaction between the complex 3D microstructure of cuticle and damage progression is poorly understood.

Furthermore, only a few toughening mechanisms have been identified in cuticle: intralaminar crack bridging in the dried elytra of the beetles Allomyrina dichotoma (Chen et al., 2001) and Lucanus cervus (Kundanati et al., 2018), and delamination, crack deflection along fibre edges and crack bridging in the fresh claws of the lobster Homarus americanus (Sachs, Fabritius & Raabe, 2006). The mechanisms by which the remarkable values of fracture toughness reported are achieved is one such gap in the knowledge.

Also, there is little understanding of why there is such a wide variation in toughness between species of arthropods. To begin to fill these knowledge gaps, studies of how crack propagation interacts with the 3D, hierarchically organised microstructures of cuticle are needed, to compare the effect of different levels of hydration and identify the effects of interspecific differences in microstructure.

Toughening mechanisms in composite materials

In a material where a crack develops or a defect exists, the application of repeated loads or both load and further environmental damage will cause either the crack to grow or one to form and then grow from the defect (Broek, 2013). Cracks and defects produce greater stress concentrations around them, especially at the crack tip, and the larger the crack or defect the higher the stress concentration it induces (Ritchie et al., 2005; Broek, 2013).

This implies that the rate of crack propagation will increase over time, as the same load that causes a crack to grow slightly will then cause the same, elongated crack to grow faster. The stress intensity factor (KI) can be used to calculate the stress at the crack tip for a given applied load, crack length and sample size. Crack propagation occurs when the stress at the crack tip reaches a critical value therefore a critical stress intensity factor

(KIC) can be determined at which fracture will occur, this value is a material parameter that indicates the fracture toughness of the material (Broek, 2013). The energy required for a crack to propagate is dependent on a range of factors, including the crack length, whether the material is ductile or brittle and the strength of the material (Ritchie et al.,

2005). One key factor is the hierarchical organisation of the material, which can possess features to either facilitate or impede crack growth. For example, in a material with a heterogeneous particle size composition, the large particles are likely to fail at lower strains than the small particles, therefore many large voids could be formed within a material within these particles, increasing the stress concentrations within the material and reducing its toughness (Broek, 2013). When these features improve fracture toughness by impeding crack growth by increasing the energy needed to grow the crack, they are called toughening mechanisms (Ritchie et al., 2005; Liu et al., 2018).

Toughening mechanisms increase the energy required for a crack to propagate in either of a number of ways or even a combination of them. One way is to increase the surface area of the crack, this spreads the stress concentrations and therefore lowers the stress concentration at the crack tip while increasing the KIC. Another method is to increase the energy required to break chemical bonds ahead of the crack, for example materials with a large plastic deformation zone ahead of the crack tip can endure greater stress before the crack propagates into it. Toughening mechanisms are divided into two types: intrinsic mechanisms which impede crack propagation ahead of the crack tip, and extrinsic mechanisms which work from behind the crack tip (Liu et al., 2018). Laminate composites, of which arthropod cuticle is a type, are able to exploit both intrinsic and extrinsic toughening mechanisms to arrest crack growth (Liu et al., 2018).

In laminate composites, extrinsic mechanisms are the most common and provide the largest contribution towards increasing toughness (Liu et al., 2018), so will be discussed in 16 further detail here. A commonly found toughening mechanism is crack deflection, which can lead to crack bifurcation (Ritchie et al., 2005; Liu et al., 2018; Häsä & Pinho, 2019). As a propagating crack encounters a lamina or fibre with a different orientation than itself, the path of least resistance is likely to be along the edge of that feature, this causes the crack to deflect to the new lamina orientation (Häsä & Pinho, 2019). The growth of the crack in this new direction causes the stress concentrations to be redistributed away from the original direction, this results in a wider spread of stress concentrations and therefore reduces the stress at the crack tip. The effectiveness of this mechanism is determined by the difference in angle between the crack orientation and the orientation of the incident lamina or fibre, and by the relative strength of the interfacial bonds. The larger change in angle then the greater deflection of the crack that occurs, and relatively weak interfacial bonds allow the deflection to spread further before the path of least resistance becomes through the lamina or fibre itself (Liu et al., 2018; Häsä & Pinho, 2019). When cracks pass through multiple laminae or fibres in this manner, it can lead to the production of a rough-edged crack which has a greater friction coefficient than a straight crack. This increase in friction increases the load required to propagate the crack and can cause further crack deflections as the crack grows (Ballarini & Plesha, 1987).

Another important toughening mechanism in laminate composites is crack bridging (Liu et al., 2018). When cracks deflect or bifurcate around a lamina or fibre, it can be left unbroken while the crack tip spreads further through the sample, this leaves behind a bridging fibre or lamina that connects both side of the fracture surface (Barth et al., 2010;

Häsä & Pinho, 2019). This requires delamination around the bridge followed by frictional sliding of the bridge against the surrounding matrix, both of which cause significant energy dissipation reducing the crack driving force (Liu et al., 2018; Häsä & Pinho, 2019).

In addition, the presence of a bridging element sustains part of the applied load, thereby 17 reducing the stress concentration at the crack tip (Ritchie et al., 2005). It is, therefore, the concentration of bridging elements that determines the amount of applied load they can sustain, and their impact on fracture toughness. In some laminated composites, such as bone, crack bridging is thought to be the most important toughening mechanism (Ritchie et al., 2005), although in other such materials fibre or lamina pull-out was found to contribute the most energy dissipation (Häsä & Pinho, 2019). Fibre pull-out occurs when the applied stress causes interfacial bonds to break resulting in delamination. This opens up a new crack front and as the fibre is broken within the matrix, the fibre is pulled out of the matrix as the crack wake grows. This improves toughness as energy is required to both break interfacial bonds and to overcome friction when sliding the fibre out of the matrix (Ritchie et al., 2005; Häsä & Pinho, 2019), which reduces the crack driving force.

The longer the fibres (i.e. the greater the distance between the crack and fracture of the fibre itself) the greater the friction produced and more energy is required for crack propagation. The quantity of fibre pull-out events during crack propagation is dependent on the strength of the interfaces between the fibres and the matrix and between laminae.

If the interfacial bonds are relatively weak, then greater delamination occurs which will result in longer fibre lengths when pull-out occurs, therefore weak interfaces results in greater toughness (Liu et al., 2018). The majority of the extrinsic toughening mechanisms found in laminated composites are strongly affected by the strength of the interfaces between laminae or between the fibres and matrix (Ritchie et al., 2005; Barth et al.,

2010). Brittle-ductile laminates with weak interfaces are the toughest of these composites, as the weak interfaces improve the efficacy of their extrinsic toughening mechanisms (Liu et al., 2018). Unlike laminated composites, however, in arthropod cuticle very few toughening mechanisms have been identified in the microstructure (see Paper 1) and their relative contribution to its toughness is unknown.

The use of X-ray computed tomography to investigate microstructure

To investigate fracture in a complex, three-dimensional material with a hierarchical organisation from the nanometre to millimetre range, only one technique can currently non-invasively image the internal microstructure whilst performing mechanical testing – computed tomography (Patterson et al., 2016). X-ray computed tomography (CT) requires that an X-ray beam, usually a cone-beam, is passed through a sample and its X-ray attenuation is collected (i.e. an X-ray projection) via a scintillator and detector panel

(Buzug, 2008). To produce tomographic images, the sample is rotated through 180 or 360 degrees whilst collecting thousands of X-ray projections, then the resulting sinograms can be reconstructed into a series of tomograms through the sample using an algorithm such as the Feldkamp backprojection algorithm (Feldkamp, Davis & Kress, 1984). This produces grayscale slices through the sample where the grayscale values represent the X-ray attenuation (a proxy for density) of its internal and external features (Sutton, Rahman &

Garwood, 2014). This technique was initially developed for medical imaging (Buzug, 2008) but has since been developed to operate at the micrometre scale, i.e. µCT (Elliott &

Dover, 1982) and even down to the nanometre scale, i.e. nCT (Withers, 2007).

Currently, the only methods available for X-ray CT imaging of soft biological materials, which have weak X-ray attenuation so are difficult to image, is to use either phase contrast imaging (Momose et al., 1996; Holzner et al., 2010) or contrast agents

(Metscher, 2009; Gignac & Kley, 2014; Paterson et al., 2014; Bribiesca-Contreras &

Sellers, 2017; Koç et al., 2019). Contrast agents add heavy elements such as iodine to the sample, increasing its X-ray attenuation and producing tomograms with excellent contrast

(Paterson et al., 2014; Bribiesca-Contreras & Sellers, 2017). This technique, however, requires the submersion of samples in preservatives such as ethanol and the contrast

19 agents themselves, this significantly alters the mechanical properties of the samples

(Aberle, Jemmali & Dirks, 2017) and can cause tissue shrinkage (Buytaert et al., 2014;

Bribiesca-Contreras & Sellers, 2017). In contrast, phase contrast imaging requires no additional sample preparation and involves only changes to the X-ray imaging configuration. Phase contrast uses interferometry to produce two X-ray beams, one of the beams passes through the sample and the X-ray wavefront bends while the other is phase shifted (Momose et al., 1996). When the two beams are combined together, the interference fringes that are formed produce intensity modulations in the X-ray images that are caused by the phase modulations of the sample (Holzner et al., 2010). This allows the edges of low density materials to have high contrast, and it is this sensitivity to light elements that enables tomographic imaging of unstained biological materials (Momose et al., 1996). In laboratory-based X-ray CT systems phase contrast imaging can be achieved using Zernike phase contrast (Tkachuk et al., 2007; Holzner et al., 2010). Zernike phase contrast uses a phase ring to produce a phase shift in the X-ray beam that is not diffracted by the sample of interest. This undiffracted, phase shifted light interferes with the light that is diffracted by the sample, resulting in a phase contrast boost on the detector

(Neuhausler et al., 2003; Tkachuk et al., 2007; Holzner et al., 2010). This technique enhances contrast at the interfaces of materials and thus allows low density materials, such as arthropod cuticle, to be visualised in 3D. Although Zernike phase contrast produces contrast without chemical alteration of the sample, the X-ray intensity on the detector is much lower so longer scans are required to collect the same quality information as a typical (i.e. absorption contrast) CT scan. In addition, it produces a ‘halo’ effect in the tomograms where there is high X-ray attenuation at the edges of materials, which is beneficial for visualisation but also reduces the sharpness of those edges, potentially masking finer details of those surfaces.

The combination of X-ray CT and in situ mechanical testing has been used to track crack progression in a time-lapse 3D manner in synthetic and biological materials (Patterson et al., 2016). X-ray CT allows 3D visualisation of the internal structures of materials, down to the nanoscale using nano-computed tomography (Withers, 2007). This technique is non- destructive allowing images to be collected at multiple stages of an in situ mechanical test, which is performed using a novel nano-mechanical test stage (Patterson et al., 2016) where very small samples can be used. However, a limitation of this technique is that each nCT scan has a very long acquisition time, which could impact hydration in fresh cuticle samples. Nonetheless, this technique allows us to study interesting cuticular structures such as beetle elytra, which possess a unique pseudo-orthogonal arrangement of macrofibres (van de Kamp & Greven, 2010) and displays a multitude of properties including: abrasion resistance for burrowing, stiffness to resist aerodynamic forces and toughness to protect the delicate wings underneath (Sun & Tong, 2007; van de Kamp &

Greven, 2010; Yu et al., 2013; Kundanati et al., 2018). The macrofibres of the pseudo- orthogonal microstructure in beetle elytra are theoretically large enough, at 0.5-10 µm diameter (van de Kamp, Riedel & Greven, 2016), to visualise by laboratory nCT, which can achieve a spatial resolution of 127 nm (Patterson et al., 2016). Although the elemental constituents of cuticle have weak X-ray attenuation, Zernike phase contrast could be used to visualise clearly the edges between features (Mayo, Chen & Evans, 2010; Burnett et al.,

2016). This could allow the macrofibres to be visualised due to the gaps between them

(van de Kamp & Greven, 2010).

To achieve the overall aim of this thesis, time-lapse 3D nCT imaging with in situ mechanical tests was used to investigate toughness and damage progression in arthropod cuticle. The following section explains the aims of each paper within this thesis that contribute towards the overall aim. 21

Aims of Papers 1, 2, 3 & 4

Paper 1 aims to review the central issues required for an understanding of arthropod cuticle mechanics and to put the analytical papers in this thesis (i.e. 2-4) into a wider context. It compares the effects of evolutionary relationships, cuticle compositions, and functions by performing a novel meta analysis of the mechanical properties reported in the literature. It also discusses the advantages and limitations of current techniques to analyse cuticle mechanics, and the known effects of variations in hierarchical organisation and composition of cuticle on mechanical properties.

Paper 2 aims to characterise how cracks propagate through dry beetle elytra under tensile stress, in the scarab beetle Macraspis lucida, and it investigates the impact of its hierarchical cuticle structure on toughening using 3D, time-lapse nCT with an in situ crack propagation experiment.

Paper 3 aims to test petroleum jelly as a hydration preservation method for arthropod cuticle and test its effect on the mechanical properties of cuticle. It also aims to investigate potential effects of X-ray CT radiation on the mechanical properties of cuticle.

To show the likely benefits of applying hydration preservation to time-lapse nCT, in situ 3- point bend tests of fresh and dry locust tibiae are reported in order to visualise differences in damage and buckling in 3D.

Paper 4 aims to investigate the effects of desiccation on fracture toughness and toughening mechanisms in beetle elytra, in a standardised manner. A new sample preparation method is applied that combines hydration preservation of the cuticle with sample dissection using laser micromachining, to image fresh and dry elytra using time- lapse nCT with in situ crack propagation experiments. The elytra of different species of beetle in a fresh condition are also tested in order to investigate the natural variation in 22 toughening, when sample geometry is standardised and hydration level preserved as in vivo.

Alternative format

This thesis is being presented in the alternative format in accordance with the rules and regulations of the University of Manchester. Papers 1, 3 and 4 presented here have been prepared for publication in peer-reviewed journals and are ready for submission, as such they are included here as manuscripts. Paper 2 has been published as an article in a peer- reviewed journal during the course of my PhD. Therefore, each of these chapters differs in layout and referencing style based on their actual or intended journal destination.

Listed below are the details of each paper, including the intended or final journal destination and the contribution of each author to the paper.

Paper 1: Arthropod cuticle: a biological material with diverse mechanical properties

Authors: Dan Sykes, Russell J. Garwood and Philip J. Withers

Destination: PeerJ (intended)

Contribution of Authors: The design and analyses for this review paper were conceived by myself with input from Russell J. Garwood and Philip J. Withers. I carried out all the data collection from the literature, data analyses and wrote the manuscript, which includes recommendations from Russell J. Garwood and Philip J. Withers.

Paper 2: Time-lapse three-dimensional imaging of crack propagation in beetle cuticle

Authors: Dan Sykes, Rebecca Hartwell, Rob S. Bradley, Timothy L. Burnett, Benjamin

Hornberger, Russell J. Garwood and Philip J. Withers

Destination: Acta Biomaterialia (2019), 86: 109-116. doi: 10.1016/j.actbio.2019.01.031

Contribution of Authors: Philip J. Withers and Rebecca Hartwell conceived the design and methodology of the study. Rebecca Hartwell and Rob S. Bradley carried out the data collection and reconstructions. I carried out the data analyses, interpreted the results and wrote the manuscript, which includes recommendations from five anonymous reviewers and all the co-authors.

Paper 3: Preservation of mechanical properties in locust tibiae for in situ time-lapse three- dimensional imaging

Authors: Dan Sykes, Russell J. Garwood, Shelley D. Rawson and Philip J. Withers

Destination: PeerJ (intended)

Contribution of Authors: I conceived the design of the study, and developed the methodology with Shelley D. Rawson. I carried out the data collection and analyses, interpreted the results and wrote the manuscript, which includes recommendations from

Russell J. Garwood and Shelley D. Rawson.

Paper 4: Effect of hydration on crack propagation in beetle elytra using time-lapse three- dimensional imaging Authors: Dan Sykes, Russell J. Garwood and Philip J. Withers

Destination: Acta Biomaterialia (intended)

Contribution of Authors: Russell J. Garwood, Philip J. Withers and myself conceived the design of the study. I developed the methodology, carried out the data collection and analyses, interpreted the results and wrote the manuscript, which includes recommendations from Russell J. Garwood and Philip J. Withers.

References

Aberle B, Jemmali R, Dirks J-H. 2017. Effect of sample treatment on biomechanical

properties of insect cuticle. Arthropod Structure & Development 46:138–146. DOI:

10.1016/j.asd.2016.08.001.

Amini S, Miserez A. 2013. Wear and abrasion resistance selection maps of biological

materials. Acta Biomaterialia 9:7895–7907. DOI: 10.1016/j.actbio.2013.04.042.

Ballarini R, Plesha ME. 1987. The effects of crack surface friction and roughness on crack

tip stress fields. International Journal of Fracture 34:195–207. DOI:

10.1007/BF00019717.

Barth HD, Launey ME, MacDowell AA, Ager JW, Ritchie RO. 2010. On the effect of X-ray

irradiation on the deformation and fracture behavior of human cortical bone.

Bone 46:1475–1485. DOI: 10.1016/j.bone.2010.02.025.

Bribiesca-Contreras F, Sellers WI. 2017. Three-dimensional visualisation of the internal

anatomy of the sparrowhawk ( Accipiter nisus ) forelimb using contrast-enhanced

micro-computed tomography. PeerJ 5:e3039. DOI: 10.7717/peerj.3039.

Broek D. 2013. Elementary engineering fracture mechanics. The Hague: Martinus Nijhoff.

Burnett TL, Kelley R, Winiarski B, Contreras L, Daly M, Gholinia A, Burke MG, Withers PJ.

2016. Large volume serial section tomography by Xe Plasma FIB dual beam

microscopy. Ultramicroscopy 161:119–129. DOI: 10.1016/j.ultramic.2015.11.001.

Buytaert J, Goyens J, De Greef D, Aerts P, Dirckx J. 2014. Volume Shrinkage of Bone, Brain

and Muscle Tissue in Sample Preparation for Micro-CT and Light Sheet

Fluorescence Microscopy (LSFM). Microscopy and Microanalysis 20:1208–1217.

DOI: 10.1017/S1431927614001329.

Buzug TM. 2008. Computed Tomography. Berlin, Heidelberg: Springer-Verlag.

Chen JX, Iwamoto M, Ni QQ, Kurashiki K, Saito K. 2001. Laminated Structure and Its

Mechanical Properties of the Fore-Wing of Beetle. Journal of the Society of

Materials Science, Japan 50:455–460.

Cribb BW, Stewart A, Huang H, Truss R, Noller B, Rasch R, Zalucki MP. 2007. Insect

mandibles—comparative mechanical properties and links with metal

incorporation. Naturwissenschaften 95:17–23. DOI: 10.1007/s00114-007-0288-1.

Dirks J-H, Taylor D. 2012a. Fracture toughness of locust cuticle. Journal of Experimental

Biology 215:1502–1508. DOI: 10.1242/jeb.068221.

Dirks J-H, Taylor D. 2012b. Veins Improve Fracture Toughness of Insect Wings. PLOS ONE

7:e43411. DOI: 10.1371/journal.pone.0043411.

Elliott JC, Dover SD. 1982. X-ray microtomography. Journal of Microscopy 126:211–213.

DOI: 10.1111/j.1365-2818.1982.tb00376.x.

Feldkamp LA, Davis LC, Kress JW. 1984. Practical cone-beam algorithm. JOSA A 1:612–619.

Gignac PM, Kley NJ. 2014. Iodine-enhanced micro-CT imaging: Methodological

refinements for the study of the soft-tissue anatomy of post-embryonic

vertebrates. Journal of Experimental Zoology Part B: Molecular and Developmental

Evolution 322:166–176. DOI: 10.1002/jez.b.22561.

Gogotsi GA. 2003. Fracture toughness of ceramics and ceramic composites. Ceramics

International 29:777–784. DOI: 10.1016/S0272-8842(02)00230-4.

Grimaldi DA, Engel MS. 2005. Evolution of the insects. Cambridge [U.K.] ; New York:

Cambridge University Press.

Häsä R, Pinho ST. 2019. Failure mechanisms of biological crossed-lamellar microstructures

applied to synthetic high-performance fibre-reinforced composites. Journal of the

Mechanics and Physics of Solids 125:53–73. DOI: 10.1016/j.jmps.2018.12.008.

Holzner C, Feser M, Vogt S, Hornberger B, Baines SB, Jacobsen C. 2010. Zernike phase

contrast in scanning microscopy with X-rays. Nature Physics 6:883–887. DOI:

10.1038/nphys1765. van de Kamp T, Greven H. 2010. On the Architecture of Beetle Elytra. Entomologie heute

22:191–204. van de Kamp T, Riedel A, Greven H. 2016. Micromorphology of the elytral cuticle of

beetles, with an emphasis on weevils (Coleoptera: Curculionoidea). Arthropod

Structure & Development 45:14–22. DOI: 10.1016/j.asd.2015.10.002.

Klocke D, Schmitz H. 2011. Water as a major modulator of the mechanical properties of

insect cuticle. Acta Biomaterialia 7:2935–2942. DOI:

10.1016/j.actbio.2011.04.004.

Koç MM, Aslan N, Kao AP, Barber AH. 2019. Evaluation of X-ray tomography contrast

agents: A review of production, protocols, and biological applications. Microscopy

Research and Technique 0. DOI: 10.1002/jemt.23225.

Kundanati L, Signetti S, Gupta HS, Menegon M, Pugno NM. 2018. Multilayer stag beetle

elytra perform better under external loading via non-symmetric bending

properties. Journal of The Royal Society Interface 15:1–13. DOI:

10.1098/rsif.2018.0427.

Liu B, Huang L, Geng L, Yin F. 2018. Multiscale Hierarchical Structure and Laminated

Strengthening and Toughening Mechanisms. In: Osheku CA ed. Lamination -

Theory and Application. InTech,. DOI: 10.5772/intechopen.69976.

Mayo SC, Chen F, Evans R. 2010. Micron-scale 3D imaging of wood and plant

microstructure using high-resolution X-ray phase-contrast microtomography.

Journal of Structural Biology 171:182–188. DOI: 10.1016/j.jsb.2010.04.001.

Melnick CA, Chen Z, Mecholsky JJ. 1996. Hardness and toughness of exoskeleton material

in the stone crab, Menippe mercenaria. Journal of Materials Research 11:2903–

2907. DOI: 10.1557/JMR.1996.0367.

Metscher BD. 2009. MicroCT for developmental biology: A versatile tool for high-contrast

3D imaging at histological resolutions. Developmental Dynamics 238:632–640.

DOI: 10.1002/dvdy.21857.

Momose A, Takeda T, Itai Y, Hirano K. 1996. Phase–contrast X–ray computed tomography

for observing biological soft tissues. Nature Medicine 2:473–475. DOI:

10.1038/nm0496-473.

Neuhausler U, Schneider G, Ludwig W, Meyer MA, Zschech E, Hambach D. 2003. X-ray

microscopy in Zernike phase contrast mode at 4 keV photon energy with 60 nm

resolution. Journal of Physics D: Applied Physics 36:A79–A82. DOI: 10.1088/0022-

3727/36/10A/316.

Neville AC. 1975. Biology of the arthropod cuticle. Springer-Verlag.

Nikolov S, Petrov M, Lymperakis L, Friák M, Sachs C, Fabritius H-O, Raabe D, Neugebauer

J. 2010. Revealing the Design Principles of High-Performance Biological

Composites Using Ab initio and Multiscale Simulations: The Example of Lobster

Cuticle. Advanced Materials 22:519–526. DOI: 10.1002/adma.200902019.

Parle E, Dirks J-H, Taylor D. 2016. Bridging the gap: wound healing in insects restores

mechanical strength by targeted cuticle deposition. Journal of The Royal Society

Interface 13:20150984. DOI: 10.1098/rsif.2015.0984.

Parle E, Dirks J-H, Taylor D. 2017. Damage, repair and regeneration in insect cuticle: The

story so far, and possibilities for the future. Arthropod Structure & Development

46:49–55. DOI: 10.1016/j.asd.2016.11.008.

Parle E, Taylor D. 2013. The self-healing properties of insect cuticle. Journal of

Postgraduate Research, Trinity College Dublin 12:90–111.

Paterson GL, Sykes D, Faulwetter S, Merk R, Ahmed F, Hawkins LE, Dinley J, Ball AD,

Arvanitidis C. 2014. The pros and cons of using micro-computed tomography in

gross and micro-anatomical assessments of polychaetous annelids. Memoirs of

Museum Victoria 71:237–46.

Patterson BM, Cordes NL, Henderson K, Mertens JCE, Clarke AJ, Hornberger B, Merkle A,

Etchin S, Tkachuk A, Leibowitz M, Trapp D, Qiu W, Zhang B, Bale H, Lu X, Hartwell

R, Withers PJ, Bradley RS. 2016. In Situ Laboratory-Based Transmission X-Ray

Microscopy and Tomography of Material Deformation at the Nanoscale.

Experimental Mechanics. DOI: 10.1007/s11340-016-0197-3.

Ritchie RO, Kinney JH, Kruzic JJ, Nalla RK. 2005. A fracture mechanics and mechanistic

approach to the failure of cortical bone. Fatigue

ascii="&"/> Fracture of Engineering Materials and Structures 28:345–371.

DOI: 10.1111/j.1460-2695.2005.00878.x.

Sachs C, Fabritius H, Raabe D. 2006. Experimental investigation of the elastic–plastic

deformation of mineralized lobster cuticle by digital image correlation. Journal of

Structural Biology 155:409–425. DOI: 10.1016/j.jsb.2006.06.004.

Sachs C, Fabritius H, Raabe D. 2008. Influence of microstructure on deformation

anisotropy of mineralized cuticle from the lobster Homarus americanus. Journal of

Structural Biology 161:120–132. DOI: 10.1016/j.jsb.2007.09.022.

Schöberl T, Jäger IL. 2006. Wet or Dry – Hardness, Stiffness and Wear Resistance of

Biological Materials on the Micron Scale. Advanced Engineering Materials 8:1164–

1169. DOI: 10.1002/adem.200600143.

Sun J, Tong J. 2007. Fracture Toughness Properties of Three Different Biomaterials

Measured by Nanoindentation. Journal of Bionic Engineering 4:11–17. DOI:

10.1016/S1672-6529(07)60007-9.

Sun J, Tong J, Ma Y. 2008. Nanomechanical Behaviours of Cuticle of Three Kinds of Beetle.

Journal of Bionic Engineering 5:152–157. DOI: 10.1016/S1672-6529(08)60087-6.

Sutton MD, Rahman IA, Garwood RJ. 2014. Techniques for virtual palaeontology.

Hoboken, NJ ; Chichester, West Sussex: Wiley Blackwell.

Tkachuk A, Duewer F, Cui H, Feser M, Wang S, Yun W. 2007. X-ray computed tomography

in Zernike phase contrast mode at 8 keV with 50-nm resolution using Cu rotating

anode X-ray source. Zeitschrift für Kristallographie - Crystalline Materials 222. DOI:

10.1524/zkri.2007.222.11.650.

Vincent JFV. 2002. Arthropod cuticle: a natural composite shell system. Composites Part

A: Applied Science and Manufacturing 33:1311–1315.

Vincent JFV, Wegst UGK. 2004. Design and mechanical properties of insect cuticle.

Arthropod Structure & Development 33:187–199. DOI: 10.1016/j.asd.2004.05.006.

Weaver JC, Milliron GW, Miserez A, Evans-Lutterodt K, Herrera S, Gallana I, Mershon WJ,

Swanson B, Zavattieri P, DiMasi E, Kisailus D. 2012. The Stomatopod Dactyl Club: A

Formidable Damage-Tolerant Biological Hammer. Science 336:1275–1280. DOI:

10.1126/science.1218764.

Wegst UGK, Ashby MF. 2004. The mechanical efficiency of natural materials. Philosophical

Magazine 84:2167–2186. DOI: 10.1080/14786430410001680935.

Withers PJ. 2007. X-ray nanotomography. Materials today 10:26–34.

Yu M, Hermann I, Dai Z, Gitis N. 2013. Mechanical and Frictional Properties of the Elytra

of Five Species of Beetles. Journal of Bionic Engineering 10:77–83. DOI:

10.1016/S1672-6529(13)60201-2.

PAPER 1

Arthropod cuticle: a biological material with diverse mechanical properties

Paper 1

1 Arthropod cuticle: a biological material with diverse 2 mechanical properties 3 4 5 Dan Sykes1, Russell J. Garwood2,3 and Philip J. Withers1 6 7 1Henry Moseley X-ray Imaging Facility, The Royce Institute, School of Materials, The 8 University of Manchester, Manchester, M13 9PL, UK. 9 2School of Earth and Environmental Sciences, University of Manchester, UK, M13 9PL. 10 3Department of Earth Sciences, The Natural History Museum, Cromwell Road, London, UK, 11 SW7 5BD. 12 13 Corresponding author: 14 Dan Sykes1 15 Henry Moseley X-ray Imaging Facility, Photon Science Institute, The University of Manchester, 16 Manchester, M13 9PL, UK 17 Email address: [email protected] 18 19 Abstract 20 Arthropod cuticle is a natural composite present in the most speciose group of organisms on 21 Earth. It has a remarkable range of mechanical properties and functions, but it is poorly 22 understood how this diversity is achieved. Little attempt has been made to date to pull together 23 the reported data across a range of mechanical properties with the aim of establishing a better 24 understanding between structure, composition and performance. Here, we discuss: the different 25 hierarchical architectures and compositions of cuticle and their reported effects on mechanical 26 properties; and the methodologies used to investigate them including their advantages and 27 limitations. Also, we review and analyse the reported mechanical properties, taking into 28 consideration the test method, cuticular composition, taxonomy of the species tested, and 29 function of the cuticle. We aim to provide a perspective on current knowledge with biological 30 context, to identify new avenues worthy of research, to stimulate the design of synthetic 31 materials along biomimetic principles. 32 33 1. Introduction 34 Cuticle is a natural composite of chitin (a long-chain polysaccharide) and structural 35 proteins (fibrous proteins that confer stiffness or elasticity to the material). It forms the 36 exoskeleton of arthropods, and as such is present in an estimated 66% of all known species on 37 Earth (Grimaldi & Engel, 2005). Only arthropod cuticle will be discussed in this review; other 38 materials called cuticle exist in vertebrates and other invertebrates but they are built from 39 different fundamental components, likewise this review won’t apply to non-arthropod chitinous

Paper 1

40 materials. Arthropods are invertebrate animals that possess an exoskeleton, a segmented body 41 and jointed appendages. Arthropod cuticle: provides protection; gives structural support; acts as 42 a barrier for water control in homeostasis (Vincent, 2002); is used in mandibles and fangs for 43 mastication and penetration (Cribb et al., 2007); can provide an adhesive surface for climbing; 44 provides points for muscle attachment; forms wings for flight; and is used for mechano- and 45 chemoreception, to name a few of its many specialised functions. Despite the diverse range of 46 important mechanical functions that cuticle in its many guises performs, its mechanical 47 properties have received very little attention, and there is a tendency to focus on a narrow range 48 of isolated examples. This is unfortunate, as there is much we can learn from a material that has 49 so much functional variety, both in terms of our understanding of biological materials as well as 50 inspiring new designs of synthetic products. 51 This review provides an overview of the current state of knowledge of arthropod cuticle 52 from the literature, and adds biological context, where possible. Furthermore, we identify the 53 main factors affecting the mechanical properties of cuticle; the problems associated with the 54 mechanical testing of cuticle; cuticles with interesting structures, compositions and/or extreme 55 properties (compared to other cuticles); and we highlight recent findings of the past 10 years that 56 represent major advances in both our understanding of hierarchical biological materials and in 57 the methods used to study them. This review aims to help researchers identify potentially 58 productive avenues for future research for the improvement of composite design, for further 59 understanding of hierarchical biological materials, and of biomechanics in arthropods. 60 Cuticle is built from simple building blocks of carbon-based fibres which are rapidly 61 assembled by the organism. Current work suggests that the majority of the variation in the cuticle 62 exploits the hierarchical organisation of these fibres; there is a limited reliance on mineralisation 63 from calcium and metals to alter properties. This hierarchical organisation allows it to act as a 64 row of fibres (tows) in some directions and like plates in others. The resulting structure is 65 inherently anisotropic, and can be exploited based on the cuticle’s function. The role of the 66 cuticle necessitates specific mechanical properties such as stiffness for support, low friction for 67 burrowing (Dai & Yang, 2010) or high hardness for cutting (Cribb et al., 2007). However, as 68 cuticle in any anatomical location is often required to serve multiple purposes, there may be 69 compromises in the optimal structure to accommodate this. For example, the cuticle of insects 70 needs to be both lightweight to allow efficient flight (Dai & Yang, 2010) and tough to protect 71 against predators (Evans & Sanson, 2005). Further limitations are imposed by the unique form of 72 growth in cuticle. It cannot be built up in successive layers like shells, or be extended like bone, 73 so to allow growth it must be shed and a new layer rapidly built. This demonstrates that cuticle 74 can be produced with a much lower energy expenditure than other animal skeletons and can be 75 produced very quickly - in a matter of hours (Neville, 1975). This mode of growth also produces 76 a period of vulnerability after moults, where the newly produced layer requires ~12 hours to 77 stiffen (sclerotise – see section 3.4; Hepburn & Joffe, 1974). Typically, cuticle deposition then 78 continues for another three weeks after a moult (Neville, 1975). In spite of these compromises, 79 cuticle displays a remarkable range of properties. Some - such as the elastic modulus (Young’s

Paper 1

80 modulus) - have been found to vary by over seven orders of magnitude (Vincent & Wegst, 2004) 81 despite a conserved macrostructure and composition across all arthropods. 82 Arthropod species account for the majority of biodiversity on the planet (fig. 1): over one 83 million species have been established, and it is estimated that between 5 and 20 million species 84 exist. This provides a large range of cuticle types adapted to different functions and 85 environments, and great potential for new and unique solutions to mechanical problems. For 86 example, the huge diversity of individual species means that through studies to date only around 87 0.004% of arthropod species have been mechanically evaluated at all. Even acknowledging that 88 there are likely to be broad trends within specific groups, even at the ordinal level only ~12% of 89 arthropod orders have been evaluated, the majority of which (~7% of all orders) are hexapods 90 (insects and kin).

91 92 Figure 1. Infographic representing the numbers of all described species on Earth and estimations 93 of the total number of species, data sourced from Chapman (2009). 94 To conclude our introductory remarks, the diversity of cuticular organisms is so large that 95 only a small fraction of all known arthropod species have currently been investigated. As such 96 we caution that our current picture of cuticle as a material may change as further studies are 97 conducted. 98 In this review we will discuss the structure of cuticle in hierarchical order from the 99 molecular to ultrastructural level and the different compositions of cuticle found in nature, 100 including the known effects of these variations on mechanical properties. We will describe the

Paper 1

101 different mechanical tests available for investigating cuticle and the advantages and limitations 102 of those techniques. We will then examine the mechanical properties reported in the literature 103 and perform a meta analysis on them, comparing the effects of evolutionary relationships, cuticle 104 compositions, and functions. We conclude by discussing the current state of knowledge 105 regarding arthropod cuticle mechanical properties and potential future areas of investigation. 106 107 2. Survey Methodology 108 We searched using Google Scholar and journal databases to find the articles reviewed in 109 this paper. The search terms we used included: arthropod cuticle mechanical properties, cuticle 110 structure, cuticle design, arthropod exoskeleton, chitin composites and arthropod biomechanics. 111 Only peer reviewed articles were included. Articles that provide quantitative data on mechanical 112 properties in arthropods were included, with those published between 2000 and 2018 reviewed 113 and discussed in more detail. This is due to significant improvements in the quality of data 114 provided in more recent papers and in the methodologies used, in general. The results were used 115 to perform an analysis of mechanical properties for cuticles with different material conditions, 116 compositions, and functions, from a variety of different taxa. Any qualitative or quantitative 117 articles that suggested relationships between aspects of cuticle structure, composition, form and 118 function that underlie or explain differences in mechanical properties between cuticles were 119 included as well. 120 121 3. Architectural Design and Composition 122 3.1 Molecular level 123 Cuticle is a heterogeneous and anisotropic composite material. It is primarily composed 124 of chitin and structural proteins, such as resilin (Vincent & Wegst, 2004). Chitin, poly-beta-(1- 125 4)-N-acetyl-D-glucosamine), is the second most abundant natural polymer on Earth after 126 cellulose (Fabritius et al., 2009). It is found in cuticle in its α-chitin crystalline form (Meyers et 127 al., 2008a), arranged in an antiparallel configuration of chitin chains (fig. 2). There are strong 128 hydrogen bonds between the chains that provide rigidity, and chitin has very high stiffness (100- 129 200 GPa; Xu et al., 1994). It is also at this spatial level where mineralisation and hydration have 130 an impact on the structure and properties of the cuticle, which is explored later in this review.

Paper 1

131 132 Figure 2. Hierarchical structure of cuticle: a chitin polysaccharide molecule within α-chitin 133 crystalline chains, which are arranged in an antiparallel alignment. The α-chitin chains are 134 wrapped in structural proteins to form nanofibrils, clusters of which form microfibres. 135 Microfibres are arranged into planar layers (laminae) which are helicoidally stacked (twisted 136 plywood/Bouligand structure). The helicoid structure runs through the exocuticle and 137 endocuticle of the ultrastructure, which also contains an outer epicuticle and internal spaces, the 138 haemolymph space and pore canals. 139 140 3.2 Microstructure 141 In arthropod cuticle, several α-chitin crystalline chains wrapped in structural proteins 142 form nanofibrils (2-5 nm in diameter; fig. 2). A bundle of chitin-protein nanofibrils then form 143 microfibres 50-100 nm in diameter. The microfibres are aligned in parallel to form lamina, which 144 can also contain proteins and biominerals (fig. 2; Raabe, Sachs & Romano, 2005). The laminae 145 (plies) are stacked upon each other in a helicoid arrangement (also referred to as a twisted 146 plywood or Bouligand structure (fig. 2; Bouligand, 1972). This helicoid structure is a common 147 feature in biological materials - examples include the arrangement of mineralised collagen fibrils

Paper 1

148 in vertebrate bone (Giraud-Guille, 1988; Weiner & Wagner, 1998; Peterlik et al., 2006; Zhang, 149 Zhang & Gao, 2011), tendon (Zhang, Zhang & Gao, 2011) and fish scales (Giraud et al., 1978; 150 Bigi et al., 2001), aragonite lamellae in mollusc shells (Jackson, Vincent & Turner, 1988; Kamat 151 et al., 2000), and cellulose microfibrils in plant cuticle (Reis, Vian & Roland, 1994). It has been 152 shown to provide greater fracture toughness (Peterlik et al., 2006; Zhang, Zhang & Gao, 2011). 153 Throughout the Arthropoda the degree of rotation between laminae varies from 30°-90° 154 (van de Kamp, Riedel & Greven, 2016). This helicoid structure produces a composite which will 155 act like a plate in some directions and like rows of fibres in others. The result is a composite with 156 an inherent anisotropy, where the degree of anisotropy can be adjusted by changing the degree of 157 rotation between laminae. However, it has been demonstrated that some arthropods use 158 variations of this arrangement. For example, in some beetle elytra there are bundles of tightly 159 packed microfibers (and matrix) forming macrofibres or 'balken' (van de Kamp & Greven, 2010) 160 which pack laterally as blocks within the laminae. Within this arrangement, known as pseudo- 161 orthogonal, the laminae rotate between two alternative directions with occasional misalignments 162 (Kapzov, 1911), in contrast to the continuously rotating layers of microfibres found in the 163 laminae of the twisted plywood structure. Van de Kamp, Riedel & Greven (2016) reported a 164 wide diversity in the microstructure of beetle elytra, not only that some possess either a helicoid 165 or pseudo-orthogonal structure, but that within pseudo-orthogonal cuticles there are different 166 macrofibre types. Further recent findings that deviate from a helicoid structure include the 167 hunting appendages (dactyls) of mantis shrimp which have a sinusoidal helicoid structure (a 168 herringbone structure; fig. 3; Yaraghi et al., 2016), shown to enhance stress redistribution and 169 increase out-of-plane compressive stiffness and the claws of the lobster Homarus americanus 170 which have a hexagonal honeycomb structure instead of parallel fibres within each lamina (fig. 171 4; Raabe et al., 2005; Romano, Fabritius & Raabe, 2007). This is shown to reduce anisotropy in 172 the material, which is unusual in honeycomb-like structures, by increasing stiffness in the 173 transverse direction; it also has a lower weight than other mineralised cuticles (Fabritius et al., 174 2009).

Paper 1

A B

175 176 Figure 3. A comparison of A) the helicoid structure and B) the sinusoidal helicoid structure 177 found in mantis shrimp dactyls; from Yaraghi et al. (2016).

A B

178 179 Figure 4. A comparison of A) the parallel fibre arrangement in lamina and B) the honeycomb 180 fibre arrangement typical of H. americanus lobsters (modified from Raabe, Sachs & Romano 181 (2005) and Raabe et al. (2007)). 182 183 3.3 Ultrastructure 184 The cuticle comprises three different layers: the epicuticle, exocuticle and endocuticle 185 (fig. 2). The epicuticle forms a thin outermost layer which is itself made of a number of different 186 layers – exactly how many depends on the species. All arthropods possess a wax layer which 187 prevents water loss in terrestrial species, and control osmotic pressure in aquatic ones (Alarie,

Paper 1

188 Joly & Dennie, 1998). Some possess an outer cement layer which is believed to protect the wax 189 but this function is not proven (Neville, 1975). The layer beneath the wax forms the bulk of the 190 epicuticle, and provides a site for muscle attachment. 191 The exocuticle is much thicker and forms the middle layer of the cuticle; it has greater 192 strength and stiffness than the other layers (Sachs, Fabritius & Raabe, 2006a). Typically, the 193 helicoid structure runs through the exocuticle, and the microfibres are more densely distributed 194 with more tightly packed plies than in the endocuticle below. It is usually the site of 195 sclerotisation and mineralisation (both described and discussed in section 3.4 Composition). The 196 endocuticle is the innermost structural layer, and shares the helicoid structure. However, the 197 microfibres are larger and less-densely packed, and the gaps between plies are larger. This layer 198 tends to be the thickest and is usually unsclerotised. It appears to provide increased elasticity to 199 the cuticle (Sachs, Fabritius & Raabe, 2006a). However, as described in section 3.3, other 200 microstructural arrangements are possible and the exocuticle and endocuticle do not necessarily 201 have the same microstructure (van de Kamp & Greven, 2010). 202 Throughout these cuticular layers are other features that by necessity compromise the 203 structure, such as pore canals. Pore canals are extensions of epidermal cells that run through the 204 endocuticle and exocuticle to the epicuticle (Green & Neff, 1972) which are needed to transport 205 wax to the epicuticle and contain muscle attachment fibres (Neville, 1975). Muscle attachment 206 fibres run throughout the cuticle to increase the attachment surface area (Neville, 1975). Both 207 these structures twist at same rate as the laminae until they reach the epicuticle (Compére & 208 Goffinet, 1987). There is space for haemolymph (analogous function to blood in vertebrates 209 supplying nutrients to the tissue) running through the lower half of the endocuticle, which has a 210 complex, dorsoventrally flattened, three dimensional shape. The bottom layer, the epidermis is a 211 layer of cells that generate new cuticle layers when the arthropod is moulting. 212 213 3.4 Composition 214 Changes in the composition of cuticle alter its mechanical properties. Sclerotisation is 215 one such modification, which involves cross-linking (the forming of hydrogen bonds) between 216 protein fibres to increase stiffness. It is a common feature in cuticle. The highest degree of 217 sclerotisation is often found in protective and stiff cuticle, such as beetle elytra (wing cases). In 218 contrast, elastic cuticle such as that in joints or in the bodies of larvae will be unsclerotised. 219 Sclerotisation occurs rapidly and is accompanied by melanisation (darkening of the cuticle 220 caused by melanin deposition). It begins immediately upon moulting, but cuticle can take several 221 days to become fully sclerotised (Lomakin et al., 2011). The process is limited to the exocuticle, 222 which makes this layer insoluble and chemically inert. As a result it is not possible for 223 arthropods to reabsorb the exocuticle during moulting, whereas the endocuticle is always 224 reabsorbed prior to moulting. Because an entirely new layer must be synthesised at every moult 225 there is an energetic constraint on the exocuticle thickness. Some arthropods counter this by 226 consuming freshly moulted cuticle to reclaim some of the nutrients, particularly when it contains 227 metals and minerals.

Paper 1

228 Mineralisation is the incorporation of metal ions or minerals into the cuticular protein 229 matrix. These include calcium carbonate, zinc and manganese (Ansenne, Compère & Goffinet, 230 1990; Thorez et al., 1992; Schofield, Nesson & Richardson, 2002; Schofield, 2005). Whereas 231 sclerotisation tends to be limited to the exocuticle, mineralisation occurs in both the exocuticle 232 and endocuticle (although the degree of mineralisation may vary between layers). Calcified 233 tissues are formed through the large inclusions of pure minerals deposited by the pore canals 234 within the protein matrix (Roer & Dillaman, 1984; Luquet & Marin, 2004). These tissues vary in 235 both the calcium mineral concentration and the ratio of crystalline to amorphous minerals. In 236 general terms, the higher the calcium mineral concentration, the harder and stiffer the tissue 237 (Boßelmann et al., 2007). Crystalline calcium carbonate (calcite) offers a greater hardness than 238 amorphous minerals, but a mixture is often present. This is presumably because a mechanically- 239 graded structure - with a hard material on a softer base - can more effectively absorb impact and 240 deflect cracks (Currey, 2005). It has also recently been demonstrated that there is variation in the 241 types of calcium minerals found in cuticle: calcium carbonate is most common, but calcium 242 phosphate has been found in crustacean mandibles (Bentov et al., 2016). In some species, metal 243 ions are present within halides. It is not known whether these are incorporated into the cuticle in 244 the same way as calcified tissues or through secondary bonds to protein side-chains. 245 Mineralisation is rare in insects as the added weight would inhibit flight. The primary site for 246 mineralisation in the Hexapoda and also Chelicerata (arachnids and kin) are in the tips of 247 mandibles/fangs and claws, where an increased hardness improves the ability to puncture or cut 248 other tissues (Hillerton & Vincent, 1982; Schofield, 2005; Cribb et al., 2007). The only extant 249 groups with mineralisation throughout their exoskeleton are the crustaceans and millipedes 250 (Shrivastava, 1970; Albanese Carmignani & Zaccone, 1977; Ansenne, Compère & Goffinet, 251 1990; Thorez et al., 1992) which incorporate calcium carbonate within their cuticle for greatly 252 increased strength. For Crustacea, as a generally marine group, the added weight of the metal is 253 less limiting, and the requisite minerals are more readily available from the water column. 254 Swimming crustaceans have been found to have thinner, less mineralised cuticle to reduce their 255 weight (Pütz & Buchholz, 1991). Millipedes, in contrast, are believed to possess mineralised 256 tissue to withstand the stresses caused by predation and burrowing (Borrell, 2004). 257 The literature also reports that the level of hydration of the cuticle has a significant impact on 258 its properties and behaviour (Neville, 1975; Vincent & Wegst, 2004; Schöberl & Jäger, 2006; 259 Klocke & Schmitz, 2011). Softer cuticles tend to be more hydrated (Vincent & Wegst, 2004), 260 and comparisons of the same cuticle type in a wet versus dry condition found that wetting 261 resulted in a reduction in both the stiffness and hardness of the cuticle (Schöberl & Jäger, 2006; 262 Klocke & Schmitz, 2011). It appears that the effects of altering hydration may be reversible, 263 suggesting that water has a rapid and fundamental physical effect (Klocke & Schmitz, 2011). It 264 has been suggested that water interacts with the structural proteins surrounding the crystalline 265 chitin (fig. 2) allowing “free” water to be added around the fibres so that they can move more 266 freely (Klocke & Schmitz, 2011). 267

Paper 1

268 4. Mechanical Testing 269 Numerous mechanical tests have been used to assess the properties of cuticle. The suitability 270 of each for measuring certain properties is important to consider, especially given the difficulty 271 of using standard test-piece geometries in natural samples. We discuss the advantages and 272 limitations of these tests for studying cuticle, and general issues with cuticle samples in this 273 section. 274 275 4.1 Sample preparation 276 The first consideration, and often difficulty, when investigating cuticle is how to prepare 277 samples in a manner that preserves their properties as in vivo. As discussed in section 3.4, 278 changes to the hydration level has a significant effect of the mechanical properties of cuticle. 279 Currently, only one method has been employed to successfully circumvent this issue, which is to 280 test cuticle that has been freshly excised. In theory this means the cuticle will retain it’s natural 281 hydration level during testing (Enders et al., 2004; Barbakadze, 2006; Lomakin et al., 2011; 282 Dirks & Taylor, 2012a; Clark & Triblehorn, 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 283 2017). However, a significant limitation of this technique is that samples have to be prepared and 284 tested quickly, leading to an increased risk of human error and forcing the entire experiment 285 (from euthanization to completed mechanical test) to be performed on one sample at a time, a 286 much slower system than performing each process on batches of samples. A recent study by 287 Aberle, Jemmali & Dirks (2017) investigated different sample preparation techniques (e.g. 288 fixation in ethanol, storage in water and freezing) and their effect on the mechanical properties of 289 cuticle. They found that freezing maintains hydration levels without affecting mechanical 290 properties, this could allow future studies to perform tests where fast sample preparation is not 291 possible (Aberle, Jemmali & Dirks, 2017). The sample preparation required for indentation 292 usually involves drying the sample, so some studies have attempted to re-hydrate (wet) the 293 cuticle (Cribb et al., 2007, 2010), while this improvement does bring the material condition 294 closer to the in vivo state than drying, it is still a different hydration state than would be found 295 naturally. 296 Another important consideration is that cuticle comprised of two layers (exo- and endo- 297 cuticle) that have different mechanical properties. Separation of these layers is difficult and 298 involves a chemical treatment, such as NaOH, which is likely to produce changes in the 299 mechanical properties itself (Kundanati et al., 2018). Hence, most studies treat cuticle as a bulk 300 material, however, in a few cases it has been possible to physically remove the exocuticle to 301 perform tests exclusively on the endocuticle (Sachs, Fabritius & Raabe, 2008; Chen et al., 302 2008b). 303 An additional issue is how to physically prepare samples and produce them in standard 304 test-piece geometries from a natural material. The majority of studies simply use a scalpel, razor 305 blade or punch to cut their samples from the arthropod specimens or introduce notches in them 306 e.g. Hepburn & Chandler (1976); Thompson & Hepburn (1978); Sun et al. (2010); Dirks & 307 Taylor (2012a,b); Clark & Triblehorn (2014); Patterson et al. (2016); Kundanati et al. (2018).

Paper 1

308 These techniques are all likely to introduce mechanical damage into the samples which will 309 affect the properties measured. To minimise this damage, two new methods have been employed. 310 (Chen et al., 2008b) used a laser cutting machine to produce numerous dog-bone shapes of 311 consistently uniform dimensions. Similarly, other studies have used a CNC milling machine to 312 manufacture a range of 3x2 mm test-pieces in different orientations (Sachs, Fabritius & Raabe, 313 2006b, 2008), producing comparable tests of anisotropy within cuticle. Standard test-piece 314 geometries require uniform, flat shapes with a specific ratio of dimensions (e.g. 315 length:thickness). Cuticle, however, is usually curved, has an irregular shape, and has variable 316 thicknesses and sizes. This tends to limit studies to specific parts of the arthropod anatomy (e.g. 317 elytra and legs). For future studies, some potential routes to avoid this issue have become 318 available. The use of laser cutting machines and milling machines allows smaller and more 319 precise test-pieces to be produced than can be achieved with traditional techniques (Sachs, 320 Fabritius & Raabe, 2006b, 2008; Chen et al., 2008b), this in combination with high sensitivity 321 load cells can open up a much greater diversity of cuticles to be tested. Another benefit of these 322 advances is the potential to test cuticles with a comparable, standardised method for the first 323 time. 324 The anisotropic nature of cuticle is also important to consider, in particular with respect 325 to the direction of loading. Cuticle exhibits anisotropy throughout every level of its hierarchical 326 structure (fig. 5), although the anisotropy decreases as the hierarchical unit increases from the 327 crystalline chitin (4.25) to the helicoid structure (1.25; Nikolov et al., 2010). The greatest impact 328 of this issue on our understanding of cuticle is the variation in Poisson’s ratio it produces. This 329 property is an important factor when running a finite element analysis (FEA). Within the 330 literature this value commonly assumed to be 0.3 (van der Meijden, Kleinteich & Coelho, 2012; 331 Hörnschemeyer, Bond & Young, 2013; Goyens et al., 2014), but in reality has been found to 332 vary greatly (see section 5.4). 333 334 4.2 Comparison of methods 335 The six most commonly used test methods are compared in Table 1, it is noteworthy that 336 all standard tests possess some limitations when applied to arthropod cuticle in addition to those 337 identified in section 4.1.

Paper 1

Test type Specimen geometry Properties measured Advantages Limitations Tension Dog-bone or rod shape. Fracture strength  Provides unique properties  Anisotropic material Ideal dimensions ratio (gauge Maximum elongation that are necessary to produces different results length : gauge width : thickness): Poisson’s ratio calculate some properties. dependent on loading  100:10:1 Ultimate tensile strength  Can be used with digital direction. Practical dimensions (mm): Yield strength image correlation (DIC) to  Cuticle is often too thin to  12:2:1 (Sachs, Fabritius Young’s (elastic) modulus determine Poisson’s ratio test in the through- & Raabe, 2006b) and produce strain maps. thickness direction.  6.35:2.29:1.81 (Chen et  Production of uniform test- al., 2008b) pieces difficult with traditional techniques.

Compression Cylinder, tube or cube shape. Elastic modulus  A common stress for  Using a small sample Ideal dimensions ratio Fracture strength arthropod tools (as defined requires sensitive load cells (height:width): Ultimate compressive strength in section 5.1). in most cases.  2:1 for homogeneous Yield strength  Uses smaller and more compression easily produced shapes than  3:1 for buckling tension tests. Bend Rectangular or cylindrical beam Flexural elastic modulus  A common stress for  Only useful for cuticle with or a tube shape. Ultimate flexural strength arthropod beams (as beam geometries (e.g. Ideal dimensions ratio Yield strength defined in section 5.1). legs). (span:thickness):  Can use a variety of set ups  32:1 for different geometries Practical dimensions (mm): (cantilever, 3-point and 4-  10:1.2 (Dirks & Taylor, point bends). 2012a)  3- and 4-point bend tests don’t require grips reducing sample preparation time significantly. Fracture Same as tension but with a Fracture toughness  Possible to observe crack  Notching methods are lateral notch in the middle of the Strain energy release rate propagation and currently inaccurate and gauge length. toughening mechanisms cause mechanical damage.

Paper 1

with either in situ or a posteriori imaging. Indentation Shape independent as sample is Fracture toughness  Nanoindentation available  Samples need to be dried mounted in a block of resin, Hardness allowing very small and coated, changing their however, a smooth surface is Reduced modulus volumes to be tested. properties. required. Wear resistance  Can be combined with  High spatial resolution Work to adhesion atomic force microscopy increases risk that surface (AFM) to produce high roughness will significantly resolution maps of affect the results. topography, work to  Indenter tip blocks views of adhesion and local elastic propagating cracks, modulus (Griepentrog, limiting in situ imaging. Krämer & Cappella, 2013).  May be unreliable when measuring the fracture toughness of anisotropic biological materials, such as cuticle (Kruzic et al., 2009). Punch Thin, flattened disc shape. Punch strength  Small sample sizes in  Sample is in a triaxial stress Ideal dimensions: Ultimate tensile strength comparison to tension tests. state complicating specimen diameter ≥ punch size. Work to punch  Sample preparation is fast comparisons with uniaxial as whole cuticles can be tests. used (Astrop et al., 2015).  Flat disc specimen geometry can be difficult to produce from cuticle. 338 Table 1. Comparison of mechanical test methodologies applicable to cuticle. The required sample geometries, mechanical properties measurable, and 339 advantages and limitations of each method is listed. The definitions of these mechanical properties, which are discussed in this review, are provided 340 in section 5.

Paper 1

341 4.3 Computational approaches to investigating mechanical properties 342 Another means of understanding the mechanical properties of cuticle is to produce 343 theoretical computer models of the relationship between its structure and properties. A recent 344 paper (Nikolov et al., 2010) used modelling, based on ab initio multiscale simulations, to 345 investigate hierarchical stiffness in lobster cuticle (fig. 5). The authors used electron microscopy 346 to produce models of the hierarchical structure of cuticle, then added the experimental properties 347 of the α-chitin crystalline chains. This facilitated modelling of the properties at higher 348 hierarchical levels, and in different directions. A major advantage of this technique is that key 349 parameters of the model can be altered, and the effects of that change identified. Compositional 350 changes can also be investigated in this way, such as the level of mineralisation (Nikolov et al., 351 2011). This provides a theoretical framework identifying which elements of the structure have 352 the greatest effect on different mechanical properties (Nikolov et al., 2011).

353 354 Figure 5. Hierarchical stiffness modelling for the cuticle of the lobster, Homarus americanus, 355 from Nikolov et al. (2010). 356 An alternative modelling approach is to produce FEA models from three-dimensional 357 computed-tomography (CT) scan data. The 3D mesh data produced from CT scans can be used 358 as the basis of FEA models that investigate the stresses, strains and displacement of the material 359 under different conditions. However, to achieve this FEA requires the density, elastic modulus 360 and Poisson’s ratio of the material modelled - something which is complicated by the anisotropic 361 nature of cuticle. FEA models can be represented in von Mises stress or maximum principal

Paper 1

362 stress. Von Mises stress is based on distortion energy theory, which proposes that the strain 363 energy is composed of two components (volumetric and shape strain). When the strain energy 364 (von Mises stress) exceeds the yield strength for a simple tensile test, then failure occurs. 365 Maximum principal stress is based on maximum principal stress theory, where failure occurs 366 when the maximum of the three principal stresses exceeds the yield strength in simple tension. A 367 few recent papers have used FEA to model the distribution of stress under conditions similar to 368 those caused by the natural mechanical use of the cuticular component being investigated. These 369 have included studies of stresses on mandibles of beetles during feeding (Hörnschemeyer, Bond 370 & Young, 2013), the bite forces of the mandibles of stag beetles (Goyens et al., 2014), the effect 371 of scorpion claw shape on its function (van der Meijden, Kleinteich & Coelho, 2012), the effect 372 structural gradients on mechanics in a spider fang (Bar-On et al., 2014), the deformation and 373 failure mechanism in beetle elytra subject to a 3-point bend test (Kundanati et al., 2018), and the 374 dynamic distribution of stress during the impact of a mantis shrimp dactyl (Weaver et al., 2012) 375 (fig. 6). A major benefit of FEA models is that the shape of the cuticle is taken into 376 consideration, and cuticle shapes vary based on their function. Ideally mechanical tests should 377 also be performed to validate the FEA model, as FEA usually provides an approximate solution 378 to the problem, and has a number of underlying assumptions which can be a source of errors. A B

379 380 Figure 6. A) Dynamic FEA model of mantis shrimp dactyl with colour-coding corresponding to 381 different elastic properties. B) Evolution of maximum principal stress during an impact 382 according to the dynamic FEA model (from Weaver et al. (2012)). 383 384 5. Mechanical Properties 385 To compare the mechanical properties of different types of cuticle, we have collated reported 386 values from the literature for this paper. Where density wasn’t reported, one was estimated based 387 on values reported for similar types of cuticle. We present these values grouped based on 388 composition, test conditions, taxonomic group and the function of the cuticle. These groups were 389 first statistically compared using a Kruskal-Wallis test and then a Dunn’s pairwise comparison, 390 with the null hypothesis that there was no significant difference between them. Cuticle data and 391 that from a range of natural materials were plotted to summarise the distribution of values on

Paper 1

392 material property charts. Some of the comparisons are of limited robustness due to low sample 393 numbers tested in certain groupings. This is particularly a problem for fracture toughness. 394 It is also noteworthy that cuticle is anisotropic and heterogeneous at all scales (see section 395 3.3) and so the measured properties are highly dependent on the scale at which they are 396 measured: for example the anisotropy of an individual fibre differs from the tows they form, 397 which in turn differ from the laminae they form. Also, between reported values, there is great 398 diversity in the methods and cuticle types used, for example, which anatomical locations were 399 sampled and the type of mechanical test used (Table 2). Hence our property maps should be 400 interpreted with caution. While some trends are clear, there are numerous sources of error in our 401 comparison of reported values. We hope this paper provides a framework to which more data 402 and more certainty can be added in future. 403 404 5.1 Stiffness 405 The majority of studies have measured Young’s modulus, mostly using tensile tests 406 although load-displacement curves can also be inferred from indentation tests, allowing smaller 407 samples to be characterised. The Young’s modulus measured to date extend over seven orders of 408 magnitude ranging from 5 kPa in unsclerotised, extensible, intersegmental membrane (ISM) of a 409 locust (Vincent & Prentice, 1973) to 70 GPa in mineralised mantis shrimp dactyls (Weaver et al., 410 2012) as illustrated in figure 7. 411

412 413 Figure 7. Material property chart plotting Young’s modulus against density for a selection of 414 natural materials, extremes within cuticle are highlighted (figures created using data compiled 415 from the literature (Warburton, 1948; Zemlin, 1968; Hepburn & Ball, 1973; Vincent & Prentice, 416 1973; Reilly, Burstein & Frankel, 1974; Hepburn et al., 1975; Reynolds, 1975; Denny, 1976;

Paper 1

417 Thompson & Hepburn, 1978; Gibson, 1985; Ashby et al., 1985; Kitchener & Vincent, 1987; 418 Bertram & Gosline, 1987; Calvert, 1988; Jackson, Vincent & Turner, 1988; Cunniff et al., 1994; 419 Alexander, Blodig & Hsieh, 1995; Bonser & Purslow, 1995; Bonser, 1996; Melnick, Chen & 420 Mecholsky, 1996; Kasapi & Gosline, 1997; Rho, Kuhn-Spearing & Zioupos, 1998; Bonser & 421 Dawson, 1999; Currey, 1999; Zioupos, Currey & Casinos, 2000; Pérez-Rigueiro et al., 2000, 422 2001; Smith et al., 2000; Marshall et al., 2001; Currey et al., 2001; Tesch et al., 2001; 423 Lichtenegger et al., 2002, 2003; Ikoma et al., 2003; Cameron, Wess & Bonser, 2003; Kohane et 424 al., 2003; Kinney et al., 2003; Taylor, Bonser & Farrent, 2004; Enders et al., 2004; Borrell, 2004; 425 Raabe, Sachs & Romano, 2005; Eichler et al., 2006; Barbakadze, 2006; Franck et al., 2006; 426 Sachs, Fabritius & Raabe, 2006a,b; Broomell et al., 2006, 2007; Moses et al., 2006; Seki et al., 427 2006; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & Tong, 2007; Cribb et al., 2007; 428 Taylor, Hebrank & Kier, 2007; Meyers et al., 2008a; Wang, Li & Shi, 2008; Miserez et al., 2008; 429 Chen et al., 2008b,a, 2012; Müller et al., 2008; Bruet et al., 2008; Sun, Tong & Ma, 2008; Sun, 430 Tong & Zhang, 2009; Tombolato et al., 2010; Dai & Yang, 2010; Sun et al., 2010; Ha et al., 431 2011; Lomakin et al., 2011; Trim et al., 2011; Yang et al., 2011; Klocke & Schmitz, 2011; Lin et 432 al., 2011; Wang, Zhang & Fan, 2011; Marino Cugno Garrano et al., 2012; Dirks & Taylor, 433 2012a,b; Weaver et al., 2012; Politi et al., 2012; Wagner, Pittendrigh & Raman, 2012; Yu et al., 434 2013; Peisker, Michels & Gorb, 2013; Erko et al., 2013; Bennemann et al., 2014; Goyens et al., 435 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 2017; Kundanati et al., 2018) and from 436 Vincent & Wegst (2004). Note the log scale for Young’s modulus. 437 Stiffness is a key property for most cuticular structures: the material forms the 438 exoskeleton to which muscles are attached, and must hence withstand the strain of muscle 439 contractions. Stiffness is also important for structures such as mandibles which are used for 440 mastication. However, in other anatomical regions such as joints elasticity is more important. 441 Chitin and resilin, two key constituents of cuticle have been identified as strongly influencing 442 stiffness. Chitin, the principal molecular building block of cuticle, has very high stiffness (~150 443 GPa; Gupta, 2011) whilst a structural protein within cuticle called resilin has rubber-like 444 properties to provide elasticity (Weis-Fogh, 1960). Other factors influencing stiffness include, 445 for example, the presence of sclerotisation and mineralisation (fig. 8A), and the level of 446 hydration (fig. 8B). It is shown in figure 8D that different species will have reached similar 447 levels of stiffness through different mechanisms: which they use depends on factors such as the 448 evolutionary history, availability of calcium for mineralisation and functional compromises. 449 Comparing all the reported measurements of the Young’s modulus of arthropod cuticle 450 between different compositions of cuticle (unsclerotised, sclerotised and mineralised), there are 451 clearly significant differences between them all (H-statistic = 108.4, probability of statistically 452 significant difference (P) < 0.001). Comparing between individual groups there is a significant 453 difference between every group (fig. 8A). Mineralised cuticle is the stiffest and unsclerotised 454 cuticle the least stiff, in agreement with previous findings that sclerotisation and mineralisation 455 increase the stiffness of cuticle (Thompson & Hepburn, 1978; Schofield, Nesson & Richardson, 456 2002; Cribb et al., 2007). Figure 9 appears to demonstrate that mineralised cuticle has

Paper 1

457 consistently higher stiffness and a greater density than can be achieved by other cuticles, but we 458 caution that this may be somewhat biased by the majority of mineralised cuticle being sampled 459 from thick crustacean armours.

460 461 Figure 8. Box plots of Young’s modulus for different categories of cuticle type or test condition. 462 The whiskers represent the range of values, the ends of the boxes indicate the 25% and 75% 463 inter-quartile range and the middle line represents the median values. Data collected from the 464 literature (Hepburn & Ball, 1973; Vincent & Prentice, 1973; Hepburn et al., 1975; Reynolds, 465 1975; Thompson & Hepburn, 1978; Alexander, Blodig & Hsieh, 1995; Melnick, Chen & 466 Mecholsky, 1996; Smith et al., 2000; Kohane et al., 2003; Enders et al., 2004; Borrell, 2004; 467 Raabe, Sachs & Romano, 2005; Eichler et al., 2006; Barbakadze, 2006; Sachs, Fabritius & 468 Raabe, 2006a,b; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & Tong, 2007; Cribb et 469 al., 2007; Taylor, Hebrank & Kier, 2007; Wang, Li & Shi, 2008; Chen et al., 2008b; Müller et 470 al., 2008; Sun, Tong & Ma, 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 2010; Sun et al., 471 2010; Ha et al., 2011; Lomakin et al., 2011; Klocke & Schmitz, 2011; Dirks & Taylor, 2012a,b;

Paper 1

472 Weaver et al., 2012; Politi et al., 2012; Wagner, Pittendrigh & Raman, 2012; Yu et al., 2013; 473 Peisker, Michels & Gorb, 2013; Erko et al., 2013; Bennemann et al., 2014; Goyens et al., 2014; 474 Parle et al., 2016; Aberle, Jemmali & Dirks, 2017; Kundanati et al., 2018). A) Young’s modulus 475 for different compositions of cuticle. B) Young’s modulus for different conditions of cuticle 476 during testing. C) Young’s modulus for different functional groups of cuticle. D) Young’s 477 modulus for the cuticle of different taxonomic groups of arthropods.

478 479 Figure 9. Material property chart plotting Young’s modulus against density for different 480 compositions of cuticle, extremes within each group are highlighted (figures created using data 481 compiled from the literature (Hepburn & Ball, 1973; Vincent & Prentice, 1973; Hepburn et al., 482 1975; Reynolds, 1975; Thompson & Hepburn, 1978; Hillerton, Reynolds & Vincent, 1982; 483 Alexander, Blodig & Hsieh, 1995; Melnick, Chen & Mecholsky, 1996; Smith et al., 2000; 484 Kohane et al., 2003; Enders et al., 2004; Borrell, 2004; Raabe, Sachs & Romano, 2005; Eichler 485 et al., 2006; Barbakadze, 2006; Sachs, Fabritius & Raabe, 2006a,b; Sun, Tong & Zhou, 2006; 486 Schöberl & Jäger, 2006; Sun & Tong, 2007; Cribb et al., 2007; Taylor, Hebrank & Kier, 2007; 487 Wang, Li & Shi, 2008; Chen et al., 2008b; Müller et al., 2008; Sun, Tong & Ma, 2008; Sun, 488 Tong & Zhang, 2009; Dai & Yang, 2010; Sun et al., 2010; Ha et al., 2011; Lomakin et al., 2011; 489 Klocke & Schmitz, 2011; Dirks & Taylor, 2012a,b; Weaver et al., 2012; Politi et al., 2012; 490 Wagner, Pittendrigh & Raman, 2012; Yu et al., 2013; Peisker, Michels & Gorb, 2013; Erko et 491 al., 2013; Bennemann et al., 2014; Goyens et al., 2014; Parle et al., 2016; Aberle, Jemmali & 492 Dirks, 2017; Kundanati et al., 2018)). Note the log scale for Young’s modulus. 493 When comparing the test condition of cuticle (wet, dry and fresh) there is a significant 494 difference between the groups (H-statistic = 71.4, P < 0.001). We define tested wet as rewetted 495 (submerged in water for a period of time, dependent on the sample size, until fully hydrated) 496 after excision and fresh as cuticle tested immediately following excision. Comparing groups

Paper 1

497 pairwise, there is a significant difference between wet and dry cuticle and between fresh and dry 498 cuticle (fig. 8B). Dry cuticle is stiffer than fresh and wet cuticle but no significant difference 499 exists between wet and fresh cuticle. Previous findings have identified wet cuticle as the least 500 stiff, and dry cuticle as the most (Klocke & Schmitz, 2011). It is clear in comparing between the 501 different values in figure 8B that although some fresh and wet cuticles have similar values of 502 stiffness to dry, fresh and wet cuticles can achieve higher elasticity than dry. 503 By dividing the types of cuticle into basic functional groups, based on their mechanical 504 function, and comparing their stiffness, it is evident that there is a significant difference between 505 the groups (H-statistic = 118.8, P < 0.001). The functional groups (beams, joints, plates, tools 506 and wings) are defined as follows. Beams comprise cylindrical or tubular structures such as legs 507 that are used in compression (support of the organism) or bending (locomotion or movement of 508 other anatomical features). Joints are defined as the elastic connections between anatomical 509 features such as body segments and the components of the leg, which provide flexibility to the 510 organism. Plates are defined as flat or curved sheets of cuticle, which are generally used to resist 511 impact and friction (protection) or provide skeletal support, such as the wing cases and body 512 segments. Tools are defined as anatomical features used for penetration or grasping, such as 513 teeth, claw tips and fangs. We recognise that this definition could also include external sensory 514 organs such as antennae and adhesive organs such as locust feet. The mechanical function of 515 such structures vary greatly, and they have been excluded entirely from our functional groups. In 516 the future such tools could be categorised separately but at present the database is too sparse for 517 this to be useful. Wings, exclusively present in insects within arthropods, include only those used 518 in flight. They are usually protected behind wing cases, and possess unique properties, being 519 flexible to tuck away when not in use, stiff in flight, and light to reduce the energy required for 520 flapping. When the groups are compared between each other, tools are found to be significantly 521 stiffer than all the other groups (fig. 8C), but for the data analysed there is no significant 522 difference between all the other groups. We note that some general trends are apparent (fig. 10): 523 wings have a high stiffness:density ratio as is required for their function; tools reach much higher 524 densities than other groups likely due to more mineralisation within this category; and joints tend 525 to have higher elasticity per unit density than other cuticles of the same density. However, many 526 factors have affected the evolution of these cuticles and therefore it is not surprising that there is 527 a lot of overlap between the functional groups.

Paper 1

528 529 Figure 10. Material property chart plotting Young’s modulus against density for different 530 functional groups of cuticle, extremes within each group are highlighted (figures created using 531 data compiled from the literature (Hepburn & Ball, 1973; Vincent & Prentice, 1973; Hepburn et 532 al., 1975; Reynolds, 1975; Thompson & Hepburn, 1978; Alexander, Blodig & Hsieh, 1995; 533 Melnick, Chen & Mecholsky, 1996; Smith et al., 2000; Kohane et al., 2003; Enders et al., 2004; 534 Borrell, 2004; Raabe, Sachs & Romano, 2005; Eichler et al., 2006; Barbakadze, 2006; Sachs, 535 Fabritius & Raabe, 2006a,b; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & Tong, 536 2007; Cribb et al., 2007; Taylor, Hebrank & Kier, 2007; Wang, Li & Shi, 2008; Chen et al., 537 2008b; Müller et al., 2008; Sun, Tong & Ma, 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 538 2010; Sun et al., 2010; Ha et al., 2011; Lomakin et al., 2011; Klocke & Schmitz, 2011; Dirks & 539 Taylor, 2012a,b; Weaver et al., 2012; Politi et al., 2012; Wagner, Pittendrigh & Raman, 2012; 540 Yu et al., 2013; Peisker, Michels & Gorb, 2013; Erko et al., 2013; Bennemann et al., 2014; 541 Goyens et al., 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 2017; Kundanati et al., 2018)). 542 Note the log scale for Young’s modulus. 543 When comparing the cuticle across taxonomic groups, significant differences become 544 apparent (H-statistic = 45.2, P < 0.001). We exclude myriapods (millipedes, centipedes and kin) 545 from the statistical analyses due to the sparsity of data. Across groups there is a significant 546 difference between insects and crustaceans and between insects and chelicerates (fig. 8D). 547 Chelicerate and crustacean cuticle is found to be stiffer than insect cuticle but there is no 548 significant difference between chelicerates and crustaceans. This may be because only 549 chelicerate fangs have been studied which are mineralised like crustacean cuticle (although in a 550 different way (Schofield, 2005; Politi et al., 2012)). We note that insect cuticle appears limited to 551 be low density, presumably to aid flight. Although few studies have investigated non-flying 552 insects, those that have been studied occupy the highest densities within the insect group.

Paper 1

553 However, regardless of density, insect cuticle can still have a stiffness within an order of 554 magnitude of mineralised crustacean cuticle (Fig. 11).

555 556 Figure 11. Material property chart plotting Young’s modulus against density for cuticle from 557 different taxonomic groups, extremes within each group are highlighted (figures created using 558 data compiled from the literature (Hepburn & Ball, 1973; Vincent & Prentice, 1973; Hepburn et 559 al., 1975; Reynolds, 1975; Thompson & Hepburn, 1978; Alexander, Blodig & Hsieh, 1995; 560 Melnick, Chen & Mecholsky, 1996; Smith et al., 2000; Kohane et al., 2003; Enders et al., 2004; 561 Borrell, 2004; Raabe, Sachs & Romano, 2005; Eichler et al., 2006; Barbakadze, 2006; Sachs, 562 Fabritius & Raabe, 2006a,b; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & Tong, 563 2007; Cribb et al., 2007; Taylor, Hebrank & Kier, 2007; Wang, Li & Shi, 2008; Chen et al., 564 2008b; Müller et al., 2008; Sun, Tong & Ma, 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 565 2010; Sun et al., 2010; Ha et al., 2011; Lomakin et al., 2011; Klocke & Schmitz, 2011; Dirks & 566 Taylor, 2012a,b; Weaver et al., 2012; Politi et al., 2012; Wagner, Pittendrigh & Raman, 2012; 567 Yu et al., 2013; Peisker, Michels & Gorb, 2013; Erko et al., 2013; Bennemann et al., 2014; 568 Goyens et al., 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 2017; Kundanati et al., 2018)). 569 Note the log scale for Young’s modulus. 570 571 5.2 Hardness 572 The second most widely measured property is hardness. Hardness provides a measure of 573 the resistance of a material to permanent deformation under compressive force. In all the 574 available literature nanoindentation and microindentation has been used to measure hardness 575 values in Pascals. The values measured vary from 70 MPa in sclerotised locust legs (Hillerton, 576 Reynolds & Vincent, 1982) to 1.6 GPa in mineralised beetle larvae mandibles (fig. 12; Cribb et 577 al., 2010) . Puncture resistance - a measure of the maximum force required to penetrate a

Paper 1

578 material with a puncture pin - was measured in three papers (Roseland, Kramer & Hopkins, 579 1987; Czapla, Hopkins & Kramer, 1990; Clark & Triblehorn, 2014). Also, specific punch 580 strength, the force required to break the sample over the volume of the flat punch, was measured 581 in two papers (Evans & Sanson, 2005; Astrop et al., 2015). Both these properties provide 582 information about the hardness of a material. Puncture resistance was found to vary between 0.01 583 N in unsclerotised beetle elytra (Roseland, Kramer & Hopkins, 1987) to 30 N in sclerotised 584 cockroach cuticle (Clark & Triblehorn, 2014). Specific punch strength was found to vary from 585 0.12 GPa m-1 in mineralised clam shrimp carapace (Astrop et al., 2015) to 300 GPa m-1 in 586 sclerotised beetle elytra (Evans & Sanson, 2005). Hardness is particularly important in tools such 587 as mandibles where the ability to penetrate or cut prey is necessary, and in claws which are used 588 for combat, climbing, locomotion and manipulating food. It may also be important for burrowing 589 arthropods where abrasion is high. It appears that sclerotisation and mineralisation are used to 590 increase hardness, either independently or in conjunction, by increasing the stiffness of the 591 cuticle (Melnick, Chen & Mecholsky, 1996; Cribb et al., 2007, 2010).

592 593 Figure 12. Material property chart plotting hardness against density for a selection of natural 594 materials, extremes within cuticle are highlighted (figures created using data compiled from the 595 literature (Hillerton, Reynolds & Vincent, 1982; Currey, Nash & Bonfield, 1982; Melnick, Chen 596 & Mecholsky, 1996; Zioupos, Currey & Casinos, 2000; Marshall et al., 2001; Tesch et al., 2001; 597 Lichtenegger et al., 2002, 2003; Kinney et al., 2003; Enders et al., 2004; Raabe, Sachs & 598 Romano, 2005; Barbakadze, 2006; Sachs, Fabritius & Raabe, 2006a; Broomell et al., 2006, 599 2007; Moses et al., 2006; Seki et al., 2006; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; 600 Sun & Tong, 2007; Cribb et al., 2007, 2010; Müller et al., 2008; Bruet et al., 2008; Sun, Tong & 601 Ma, 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 2010; Yang et al., 2011, 2013; Klocke &

Paper 1

602 Schmitz, 2011; Lin et al., 2011; Chen et al., 2012; Weaver et al., 2012; Politi et al., 2012; Yu et 603 al., 2013; Erko et al., 2013)). 604 By comparing hardness measurements between various compositions of cuticle 605 (unsclerotised, sclerotised and mineralised), we can report a significant difference (H-statistic = 606 8.6, P = 0.014). This is unsurprising given we find a similar trend when comparing Young’s 607 modulus for these groups (fig. 8A). Pairwise comparison between these compositions 608 demonstrates a significant difference between unsclerotised and mineralised cuticle (fig. 13A), 609 but not between any others. Mineralised cuticle is harder than unsclerotised but at the cost of a 610 higher density; we note that sclerotised cuticles occupy a similar range of hardnesses to 611 mineralised cuticle whilst maintaining a lower density (fig. 14).

612 613 Figure 13. Box plots of hardness for different categories of cuticle type or test condition. The 614 whiskers represent the range of values, the ends of the boxes indicate the 25% and 75% inter- 615 quartile range and the middle line represents the median values. Data collected from the 616 literature (Hillerton, Reynolds & Vincent, 1982; Currey, Nash & Bonfield, 1982; Melnick, Chen 617 & Mecholsky, 1996; Enders et al., 2004; Raabe, Sachs & Romano, 2005; Barbakadze, 2006; 618 Sachs, Fabritius & Raabe, 2006a; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & 619 Tong, 2007; Cribb et al., 2007, 2010; Müller et al., 2008; Sun, Tong & Ma, 2008; Sun, Tong & 620 Zhang, 2009; Dai & Yang, 2010; Klocke & Schmitz, 2011; Weaver et al., 2012; Politi et al.,

Paper 1

621 2012; Yu et al., 2013; Erko et al., 2013). A) Hardness for different compositions of cuticle. B) 622 Hardness for different conditions of cuticle during testing. C) Hardness for different functional 623 groups of cuticle. D) Hardness for the cuticle of different taxonomic groups of arthropods.

624 625 Figure 14. Material property chart plotting hardness against density for different compositions of 626 cuticle, extremes within each group are highlighted (figures created using data compiled from the 627 literature (Hillerton, Reynolds & Vincent, 1982; Currey, Nash & Bonfield, 1982; Melnick, Chen 628 & Mecholsky, 1996; Enders et al., 2004; Raabe, Sachs & Romano, 2005; Barbakadze, 2006; 629 Sachs, Fabritius & Raabe, 2006a; Sun, Tong & Zhou, 2006; Schöberl & Jäger, 2006; Sun & 630 Tong, 2007; Cribb et al., 2007, 2010; Müller et al., 2008; Sun, Tong & Ma, 2008; Sun, Tong & 631 Zhang, 2009; Dai & Yang, 2010; Klocke & Schmitz, 2011; Weaver et al., 2012; Politi et al., 632 2012; Yu et al., 2013; Erko et al., 2013)). 633 Between the different test conditions of cuticle (wet, dry and fresh) there is a significant 634 difference in hardness (H-statistic = 10.5, P = 0.005). When compared pairwise there is a 635 significant difference between wet and dry cuticle and between fresh and dry cuticle (fig. 13B). 636 Dry cuticle was found to be harder than fresh and wet cuticle but there is a wide range of values 637 reported for dry cuticle. 638 No hardness measurements have been reported to date for the joints and wings functional 639 groups. However, when comparing the other groups, there is a significant difference between 640 them (H-statistic = 17.7, P < 0.001). There is a significant difference between tools and beams 641 and between tools and plates (fig. 13C). Tools were found to be harder than beams and plates as 642 would be expected given the benefits of hardness for grasping, cutting and other tool uses (fig. 643 15).

Paper 1

644 645 Figure 15. Material property chart plotting hardness against density for different functional 646 groups of cuticle, extremes within each group are highlighted (figures created using data 647 compiled from the literature (Hillerton, Reynolds & Vincent, 1982; Currey, Nash & Bonfield, 648 1982; Melnick, Chen & Mecholsky, 1996; Enders et al., 2004; Raabe, Sachs & Romano, 2005; 649 Barbakadze, 2006; Sachs, Fabritius & Raabe, 2006a; Sun, Tong & Zhou, 2006; Schöberl & 650 Jäger, 2006; Sun & Tong, 2007; Cribb et al., 2007, 2010; Müller et al., 2008; Sun, Tong & Ma, 651 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 2010; Klocke & Schmitz, 2011; Weaver et al., 652 2012; Politi et al., 2012; Yu et al., 2013; Erko et al., 2013)). 653 When comparing the cuticles of different taxonomic groups there is a significant 654 difference between the groups (H-statistic = 31.7, P < 0.001), however no data have been 655 reported in the literature for myriapodans. The significant differences lie between chelicerates 656 and crustaceans and between insects and chelicerates (fig. 13D). Chelicerate cuticle is found to 657 be harder than insect and crustacean cuticle but this may again result from sampling bias - only 658 chelicerate fangs have been measured, which have a high level of hardness in order to penetrate 659 cuticle (Politi et al., 2012). We note again that sclerotisation in insects and mineralisation in 660 crustaceans appear capable of achieving similar properties as is the case for stiffness (fig. 16).

Paper 1

661 662 Figure 16. Material property chart plotting hardness against density for cuticle from different 663 taxonomic groups, extremes within each group are highlighted (figures created using data 664 compiled from the literature (Hillerton, Reynolds & Vincent, 1982; Currey, Nash & Bonfield, 665 1982; Melnick, Chen & Mecholsky, 1996; Enders et al., 2004; Raabe, Sachs & Romano, 2005; 666 Barbakadze, 2006; Sachs, Fabritius & Raabe, 2006a; Sun, Tong & Zhou, 2006; Schöberl & 667 Jäger, 2006; Sun & Tong, 2007; Cribb et al., 2007, 2010; Müller et al., 2008; Sun, Tong & Ma, 668 2008; Sun, Tong & Zhang, 2009; Dai & Yang, 2010; Klocke & Schmitz, 2011; Weaver et al., 669 2012; Politi et al., 2012; Yu et al., 2013; Erko et al., 2013)). 670 671 5.3 Fracture toughness 672 Fracture toughness (KIc) is the energy required to grow a crack from a pre-existing defect. 673 The linear-elastic fracture toughness is determined by the critical stress intensity factor at which 674 a crack begins to grow in a material. The stress intensity factor is the level of stress near the tip 675 of a crack, and is affected by the geometry of the crack and loading conditions. The values 676 reported in the literature show a variation in fracture toughness from 0.88 MPa √m in locust 677 wing membranes (Dirks & Taylor, 2012b) to 4.12 MPa √m in locust tibia (Dirks & Taylor, 678 2012a). Other measurements of fracture toughness are plastic-elastic fracture toughness (JIc; the 679 energy needed to initiate stable crack growth in a nonlinear elastic body containing a crack) and 680 strain energy release rate (GIc; work of fracture i.e. the energy released by the crack propagation 681 for the newly created crack area). Strain energy release rate varies between 0.07 kJ/m2 in crab 682 claws (Melnick, Chen & Mecholsky, 1996) to 5.56 kJ/m2 in locust tibia (Dirks & Taylor, 2012a). 683 Toughness (amount of energy per unit volume that a material can absorb without fracturing) has 684 also been reported, it varies between 0.2 MPa in unsclerotised beetle elytra (Lomakin et al., 685 2011) to 8.3 MPa in crab legs (Chen et al., 2008b).

Paper 1

686 Fracture toughness is an expression of a material’s resistance to fracture: a cuticle with a 687 low fracture toughness would fracture easily after damage is introduced whereas a high 688 toughness requires more energy to keep growing the crack until fracture occurs. As such it is a 689 very important property, as much cuticle needs to resist fracture from – for example – predation 690 and the fatigue of locomotion. Despite this there are only four sources of values in the literature 691 (Melnick, Chen & Mecholsky, 1996; Sun & Tong, 2007; Dirks & Taylor, 2012a,b). But based on 692 this limited literature, the fracture toughness found in cuticle is relatively uniform (fig. 17). 693 However, in terms of strain energy release rate, cuticle is a remarkably tough natural composite 694 (Dirks & Taylor, 2012a). 695 There are two types of toughening mechanisms - intrinsic and extrinsic. Intrinsic 696 mechanisms act ahead of the crack tip to hinder crack initiation, primarily, and crack growth. In 697 cuticle, the only intrinsic mechanisms that have been discovered are agglomeration of pores in 698 dry lobster claw and some plastic deformation in fresh lobster claw (Sachs, Fabritius & Raabe, 699 2006b). These mechanisms reduce the stress driving force of the crack by creating a high surface 700 area for the stress to act upon. Extrinsic mechanisms act behind the crack tip to resist crack 701 growth by absorbing energy. Thus far, evidence of the following extrinsic toughening 702 mechanisms have all been found with 2D microscopy in cuticle: crack deflection along the 703 longitudinal edges of microfibres in dry and fresh lobster claw (Sachs, Fabritius & Raabe, 704 2006b); delamination between laminae in fresh lobster claw (Sachs, Fabritius & Raabe, 2006b); 705 and intra- and inter-laminar crack bridging in beetle elytra (Chen et al., 2001; Kundanati et al., 706 2018). A study by Sykes et al. (2019) used X-ray nano-computed tomography with in situ 707 mechanical testing to observe crack propagation through the microstructure of dry beetle elytra; 708 and many extrinsic toughening mechanisms were found, such as crack bifurcation, shear tearing, 709 intralaminar and interlaminar crack bridging, differential crack deflection between laminae and 710 the production of rough crack surfaces. As the majority of toughening mechanisms found in 711 cuticle are extrinsic, the crack path is a significant factor in determining the fracture toughness. 712 As such, the hierarchical organisation of cuticle is important for modifying its toughness. Despite 713 this, few studies have investigated the links between microstructure and toughness in cuticle, 714 however there are some studies of these links within cuticle biomimetics (Chen, 2018; Zaheri et 715 al., 2018; Xu et al., 2019; Jia, Yu & Wang, 2019).

Paper 1

716 717 Figure 17. Material property chart plotting fracture toughness against density for a combination 718 of natural and engineering materials, extremes within cuticle are highlighted (figures created 719 using data compiled from the literature (Ashby et al., 1985; Jackson, Vincent & Turner, 1988; 720 Melnick, Chen & Mecholsky, 1996; Zioupos, Currey & Casinos, 2000; Kamat et al., 2000; 721 Marshall et al., 2001; Currey et al., 2001; Cambridge Engineering Selector software, 2003; 722 Nalla, Kinney & Ritchie, 2003; Broomell et al., 2007; Sun & Tong, 2007; Meyers et al., 2008b; 723 Dirks & Taylor, 2012a,b; Yang et al., 2013)). Note the log scale for fracture toughness. 724 725 5.4 Other properties 726 In the literature, other properties have been measured, albeit rarely. Extensibility 727 (maximum elongation) has been measured in two papers, investigating moth wings (Reynolds, 728 1977) and cockroach thoraxes (Clark & Triblehorn, 2014). Extensibility is a measurement of 729 how far a material can be stretched before failure as a percentage of the sample length. This is a 730 particularly important property for cuticle such as in the abdomen of insects which can extend to 731 twice its length to deposit eggs, through morphology coupled with extensible arthrodial 732 membrane. Values reported varied from 3-9% in cockroach thoraxes (Clark & Triblehorn, 2014) 733 to 43% in moth wings (Reynolds, 1977). Poisson’s ratio has been measured in two papers, 734 investigating beetle wings (Ha et al., 2011) and lobster claws (Sachs, Fabritius & Raabe, 2006b). 735 Poisson’s ratio is the ratio of transverse strain to longitudinal strain under uniaxial stress 736 (Greaves et al., 2011). For example, cork has a Poisson’s ratio of almost 0 as the width of 737 uncompressed regions don’t expand when other regions are compressed. Generally isotropic 738 material will lie between 0 and 0.5 but anisotropic materials can have a Poisson’s ratio between - 739 1 and 1 depending on the direction of stress. This property hence provides a measure of 740 anisotropy, and can be compared with other properties to indicate the level of ductility (Greaves

Paper 1

741 et al., 2011). It is also valuable for finite element analyses (van der Meijden, Kleinteich & 742 Coelho, 2012; Goyens et al., 2014). Poisson’s ratio was found to vary between 0.3 and 0.7 in 743 beetle wings (Ha et al., 2011), when tested in different directions and between 0.3 and 0.4 in 744 lobster claws (Sachs, Fabritius & Raabe, 2006b), where only the longitudinal direction was 745 tested. 746 Fatigue by cyclic loading has been reported in a single paper, which examined locust 747 hind legs and wings (Dirks, Parle & Taylor, 2013). This is an important property for any 748 anatomical region that is used repeatedly. It was found that there was a large difference between 749 wings, which can withstand 100,000 cycles at 46% of the stress required for instantaneous failure 750 (ultimate tensile strength), and legs, which can withstand 76% of their ultimate tensile strength 751 over 100,000 cycles (Dirks, Parle & Taylor, 2013). Only recently have studies begun to 752 investigate whether arthropods can repair their cuticle after damage, or if they must moult to 753 grow a new exoskeleton instead. The limited evidence reported to date suggests that for minor 754 damage, self-repair mechanisms exist that can maintain the pre-damage mechanical properties 755 until a new layer of cuticle can be grown. However, major damage invokes localised growth of 756 new cuticle (Parle & Taylor, 2013; Parle, Dirks & Taylor, 2016). This ability to heal minor 757 damage makes fatigue a complex problem as damage can be repaired before affecting 758 performance, which significantly improves fatigue life but in a manner that is difficult to 759 quantify.

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured

Araneae Cupiennius saleia,b Fanga,b Mineraliseda,b Nanoindentationa, Drya,b Hardnessa,b aPoliti et al. (2012) (spiders) Sclerotiseda b Reduced modulusa,b bErko et al. (2013) Xiphosura Limulus polyphemus Leg ISM Unsclerotised Tensile tests Wet Elastic modulus Hepburn & Chandler (horseshoe Peak strain (1976)

Chelicerata crabs) Tensile strength Blattodea Blaberus craniiferc Abdomenc Mineralisedb Bending testsd Dryb Elastic modulusd aCzapla, Hopkins & (cockroaches Blaberus discoidalisd Legd Sclerotiseda,b,c,d Nanoindentationb Freshc,d Hardnessb Kramer (1990) & termites) Blatella germanicaa Mandibleb Unsclerotiseda Puncture testsa,c Frozena Peak strainc bCribb et al. (2007) b a,c c b a,c Cryptotermes primus Thorax Tensile tests Wet Puncture resistance cClark & Triblehorn Coptotermes acinaciformisb Reduced modulusb (2014) Gromphodorhina Strain energy release dParle et al. (2016) portentosab,c ratec

Hexapoda Mastotermes darwiniensisb Tensile strengthc,d Nasutitermes magnusb Periplaneta americanac,d Reticulitermes hesperusb

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured Coleoptera Allomyrina dichotoman,p,s Abdomenb,e Mineralisedm Bending testsu Dryd,f,g,h,i,j,k,l, Coupling strengthn aHepburn & Ball (1973) (beetles) Coccinella septempunctatar Adhesive organsr Sclerotiseda,c,d,e, Dynamic and m,n,q,r,s,t,u Elastic bHepburn & Chandler Copris ochush,i,k,l,o Elytrac,e,h,i,k,n,o,q,s,u g,f,g,h,i,j,k,l,m,n,o,q,s,u transient tensile Fixedt modulusa,b,d,f,g,n,o,p,q,r,s, (1976) Cybister japonicusn Headd,f,j,k Unsclerotisedb,c, testsg,q Fresha,d,e,f,g,n, t,u cRoseland, Kramer & Cybister tripunctatuss Legk e,g,p,q,r Nanoindentationd, o,p,q,r,t Flexural modulusu Hopkins (1987) Cyclommatus metallifert Mandiblem,t Unspecifiedt f,h,i,j,k,l,m,r,s,t Frozenc Flexural strengthu dEnders et al. (2004) Dorcus titanusn Thoraxe,k Punch testse Wetb,m Fracture strengthg,q,u eEvans & Sanson (2005) Geotrupes stercorariusk Wingp Puncture testsc Fracture toughnessi fBarbakadze (2006) Holotrichia sichotanak Tensile Frictional gEichler et al. (2006) Holotrichia titaniss testsa,b,n,o,p,u coefficients hSun, Tong & Zhou Lucanus cervusu Hardnessd,f,h,i,j,k,l,m,n,s (2006) Pachnoda marginatad,f,j Peak strainb,n iSun & Tong (2007)

Pachnoda sinuataa,b Poisson’s ratiop jMüller et al. (2008) Potosia brevitarsisn Puncture resistancec kSun, Tong & Ma (2008) Protaetia brevitarsiss Reduced lSun, Tong & Zhang Pseudotaenia frenchim modulush,i,j,k,l,m (2009)

Hexapoda Scarabaeus sacers Shear modulusa mCribb et al. (2010) Tenebrio molitorg,q Specific punch nDai & Yang (2010) Tribolium castaneumc,g,q strengthe oSun et al. (2010) Unspecifiede Specific strengthn pHa et al. (2011) Specific work to qLomakin et al. (2011) punche rPeisker, Michels & Strain energy release Gorb (2013) rateq sYu et al. (2013) Tensile strengtha.b.n,o tGoyens et al. (2014) Toughnessq,u uKundanati et al. (2018) Von Mises stresst Yield strengtha,o

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured Diptera Drosophila melanogastera,b Abdomena Sclerotiseda Nanoindentationa, Fresha,b Elastic modulusb aKohane et al. (2003) (flies) Heada Unsclerotiseda b Reduced modulusa bWagner, Pittendrigh & Larval Unspecifiedb Work of adhesionb Raman (2012) membranea Puparium membranea Thoraxa Wingb Hemiptera Rhodnius prolixus Larval membrane Unsclerotised Tensile tests Fresh Elastic modulus Reynolds (1975) (bugs)

Hymenoptera Apis ceranad Abdomena Mineralisedb Bending testsc Dryd Elastic modulusa,c,d aThompson & Hepburn a c a,c,d b, b,c a (bees, wasps Apis mellifera adansonii Leg Sclerotised Nanoindentation Fresh Fracture strength (1978) and ants) Atta sexdens rubipilosab Mandibleb d Weta Hardnessb,d bSchofield, Nesson & Bombus terrestris audaxc Stingerd Tensile testsa Tensile strengthc Richardson (2002) Vespula vulgarisd Von mises stressd cParle et al. (2016)

Hexapoda dDas et al. (2018) Lepidoptera Bombyx moria Abdomena,d Sclerotisedd Extension testsb Freshb,d Elastic modulusa,c aHepburn & Chandler (butterflies Manduca sextab,c Larval Unsclerotiseda,b, Punch testsd Weta,c Peak straina,b,c (1976) and moths) Unspecifiedd membranec c,d Tensile testsa,c Specific punch bReynolds (1977) Thoraxd strengthd cWolfgang & Riddiford Wingsb Specific work to (1987) punchd dEvans & Sanson (2005) Tensile strengtha Odonata Pantala falvescens Wings Unspecified Bending tests Unspecified Flexural elastic Wang, Li & Shi (2008) (dragonflies) modulus Tensile strength Torsional moment Von Mises stress

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured Orthoptera Locusta migratoria Abdomena,c,d,e,i Mineralisede,h Bending testsf,j,m,o Drye,h,i,j,n,o Compressive aVincent & Prentice (grasshoppers manilensisn Heade Sclerotisedb,e,f,i,j, Compression Fixedo strengthj (1973) & locusts) Locusta migratoria Legb,e,f,j,l,m,n,o l,m,n,o testsj Freshb,e,f,g,j,k,l Dampingo bHepburn & Joffe (1974) migratorioidesa,b,c,d,e,i Mandiblee,h Unsclerotiseda,b, Fatigue testsl ,m,o Elastic cVincent (1975) Schistocerca Thoraxe c,d,i Microhardness Frozeno modulusa,b,d,g,j,k,m,o dHepburn & Chandler gregariae,f,g,h,j,k,l,m,o Wingg,k,l Unspecifiedg,k,l testse Weta,d,h,i,n,o Fatigue limitl (1976) Nanoindentationh, Unspecified Flexural elastic eHillerton, Reynolds & i,n c modulusf,j Vincent (1982) Shear testa Flexural strengthj fKatz & Gosline (1992) Tensile Fracture toughnessj,k gSmith et al. (2000) testsb,c,d,g,k,l Friction coefficienth hSchöberl & Jäger e,h,i,n

Hardness (2006) Peak straind iKlocke & Schmitz Reduced modulush,i,n (2011) Shear stressa jDirks & Taylor (2012a) k

Hexapoda Strain energy release Dirks & Taylor (2012b) ratej,k lDirks, Parle & Taylor Stress softeningc (2013) Tensile mParle et al. (2016) strengthb,d,k,l,m nWan, Hao & Feng Von Mises stressn (2016) Wear resistanceh oAberle, Jemmali & Dirks (2017) Phasmatodea Carausius morosusa,b Adhesive organsb Sclerotiseda,c Bending testsa,c Drya,b Dampinga aDirks & Dürr (2011) (stick insects) Parapachymorpha zomproic Antennaa Unsclerotisedb Tensile testsb Fresha,c Elastic modulusb,c bBennemann et al. Legc Livea Tensile strengthc (2014) cParle et al. (2016)

Spinicaudata Cyzicus gynecia Carapace Mineralised Punch tests Fresh Specific punch Astrop et al. (2015) (clam Eulimnadia feriensis strength shrimps) Leptestheria compleximanus Specific work to punch

Crustacea 65

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured Decapoda Callinectes sapidush Abdomenc,l Mineraliseda,b,c,d Bending testsh Drya,b,d,e,f,g,i,j, Compressive aHepburn et al. (1975) (crabs & Homarus americanuse,f,g,j,k,l Carapaceb,k ,e,f,g,h,i,j,k,l Compression k modulusi bJoffe et al. (1975) lobsters) Loxorhynchun grandisi Clawd,e,f,g,j Unsclerotisedh,l testsi,j Weta,b,c,d,g,i,l Compressive cHepburn & Chandler Menippe mercenariad Jointl Microindentation strengthi (1976) Penaeus mondonb,c Lega,c,h,i e,i Elastic dMelnick, Chen & Scylla serrataa,c Nanoindentationd, modulusa,b,c,d,e,g,h,i,j,k,l Mecholsky (1996) f,k Fracture eRaabe, Sachs & Shear testsj strengthb,d,g,i,j Romano (2005) Tensile Flexural elastic fSachs, Fabritius & testsa,b,c,g,h,i,l modulush Raabe (2006a) Fracture toughnessd gSachs, Fabritius & Hardnessd,e,f,i,k Raabe (2006b)

Peak strainb,c hTaylor, Hebrank & Kier Poisson’s ratiog,j (2007) Reduced modulusf iChen et al. (2008b) Strain energy release jSachs, Fabritius &

Crustacea rated,i Raabe (2008) Tensile strengthc,d,h,i,l kErko et al. (2013) Toughnessi,l lWu et al. (2019) Von Mises straing Yield strengthg,i,j Isopoda Pentidotea resecata Leg Mineralised Bending tests Wet Elastic modulus Alexander, Blodig & Pentidotea wosnesenskii Flexural elastic Hsieh (1995) modulus Stomatopoda Gonodactylus sp.a Dactyla,b Mineraliseda,b Microhardness Dryb Elastic modulusb aCurrey, Nash & (mantid Odontodactylus scyllarusb testsa Fixeda Hardnessa,b Bonfield (1982) shrimps) Nanoindentationb Freshb Maximum principal bWeaver et al. (2012) stressb

Paper 1

Subphylum Arthropod Species investigated Anatomical Material Mechanical Material Mechanical Source(s) order locations composition test condition properties tested performed measured

Diplopoda Nyssodesmus python Trunk Mineralised Compressive C- Fresh Elastic modulus Borrell (2004) (millipedes) ring test Flexural elastic modulus Fracture strength

Myriapoda 760 Table 2. Data compiled from all the available literature showing the species of arthropod tested divided into their respective subphyla, including 761 information on the location of the cuticle tested, its composition and hydration, the mechanical tests used and the mechanical properties measured.

Paper 1

762 6. Discussion 763 Based on the reported values, cuticle appears to be able to be very tough (Dirks & Taylor, 764 2012a), hard (Cribb et al., 2010) and stiff (Weaver et al., 2012) for a material of such a low 765 molecular weight. It often achieves these remarkable properties without resorting to any kind of 766 mineralisation. Equally impressive is its ability to form optimal designs and material properties 767 for such diverse functional morphology, whether it is a wing, armour plate, fang or joint. This 768 review demonstrates the wide range of reported mechanical properties. Even similar anatomical 769 regions in closely related species can vary significantly - for example, sclerotised beetle elytra 770 have an elastic moduli between 0.5 GPa (Dai & Yang, 2010) and 14.5 GPa (Sun et al., 2010). 771 This is could represent genuine variation, but may also be impacted by inconsistent testing 772 methods. Table 2 demonstrates the breadth of mechanical testing approaches that have been 773 applied, as well as the properties measured. We can identify some key patterns in work to date. 774 The majority of published studies rely on indentation tests which, as discussed earlier, may make 775 some results such as elastic modulus and hardness inaccurate, but with large datasets they are 776 comparable, and thus can be used to show qualitative trends. Elastic moduli have been collected 777 using a mixture of indentation, bending, tension and compression tests. Sample shapes and sizes 778 remain relatively limited due to the low availability of standard sample shapes and sizes in 779 arthropod cuticles. Most samples are of bulk cuticle; few studies test different layers of cuticle to 780 separate their functions (Raabe, Sachs & Romano, 2005; Sachs, Fabritius & Raabe, 2006a; Chen 781 et al., 2008b). Therefore, most of the data reported so far are comparable but lack agreed 782 standards and methodologies. Of the different functional regions of cuticle investigated to date, 783 plates are by far the most tested, especially beetle elytra. The literature is consistent in that the 784 same properties tend to be measured within any given functional groups – for example, stiffness 785 is measured in plate regions and hardness is measured in tools. Most species have only been 786 studied in a single paper, increasing the diversity of tested cuticle, but allowing limited 787 assessment of the reproducibility of the findings. Only the lobster, Homarus gammarus, the 788 locusts, Locusta migratoria migratoides and Schistocercia gregaria, and the beetle, Copris 789 ochus, have been reported on in more than three papers. 790 Historically, the decision of what to test and how to do so was generally based on 791 practical considerations such as the availability of sample types. However, more recently papers 792 have shown greater consideration for function, form and material condition (Sachs, Fabritius & 793 Raabe, 2006b, 2008; Chen et al., 2008b; Cribb et al., 2010; Klocke & Schmitz, 2011; Dirks & 794 Taylor, 2012a,b; Weaver et al., 2012; Politi et al., 2012; Erko et al., 2013; Hörnschemeyer, Bond 795 & Young, 2013; Goyens et al., 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 2017; 796 Kundanati et al., 2018). They have found new ways to overcome the practical problems of 797 investigating cuticle properties, for example: using milling to create standard geometry test 798 samples (Sachs, Fabritius & Raabe, 2006b, 2008); performing tests on the different 799 ultrastructural layers within cuticle (Chen et al., 2008b; Kundanati et al., 2018); combining 800 mechanical tests with in situ imaging (Sachs, Fabritius & Raabe, 2006b, 2008; Dirks & Taylor, 801 2012b; Sykes et al., 2019); creating accurate three-dimensional computer models of cuticle

Paper 1

802 structures with CT (Dirks & Taylor, 2012a; Weaver et al., 2012; Politi et al., 2012; 803 Hörnschemeyer, Bond & Young, 2013; Aberle, Jemmali & Dirks, 2017; Sykes et al., 2019); 804 testing in more than one direction (i.e. longitudinal, transverse and normal; Chen et al., 2008b; 805 Klocke & Schmitz, 2011; Politi et al., 2012; Kundanati et al., 2018); and testing the same 806 material in a mixture of conditions (i.e. wet and dry or fresh and dry; Sachs, Fabritius & Raabe, 807 2006b, 2008; Klocke & Schmitz, 2011; Dirks & Taylor, 2012a; Aberle, Jemmali & Dirks, 2017). 808 This last point is important, because material condition can greatly alter properties. For example, 809 the hydration of samples has varied greatly, but cuticle is built to operate at a homeostatically 810 controlled level of hydration. In order to test the cuticle in its optimal condition the hydration 811 level should be maintained as it is in vivo. Many papers use wet or dry cuticle however only 812 marine arthropod cuticle (such as crustaceans) would be closer to a wet state in life, and no 813 cuticle would be completely dry in its natural condition. As can be seen in figures 8B and 13B 814 condition has a significant effect on mechanical properties. In the case of crustacean cuticle the 815 salinity of the tissue should be taken into consideration, as the effect of salinity on the 816 mechanical properties of cuticle hasn’t been tested. Some papers attempt to counter this problem 817 by testing the material fresh from freshly euthanized arthropods, with the aim of completing the 818 experiment before significant dehydration has occurred. However, drying occurs quickly in 819 dissected arthropod cuticle, so we caution that samples may change hydration states over the 820 course of an experiment. Klocke and Schmitz (2011) compared wet versus dry cuticle and found 821 as much as a 15x difference in the elastic modulus between the two states. Hardness has also 822 been found to decrease when cuticle is wet (Schöberl & Jäger, 2006), so this effect is significant, 823 and testing samples in their native hydration state should be an important consideration. There is 824 also an interesting idea to be investigated that arthropods may themselves be using control of the 825 hydration of their cuticle to alter the properties of particular areas of cuticle (Klocke & Schmitz, 826 2011). There has recently been a much greater use of FEA to model cuticle functional 827 morphology (van der Meijden, Kleinteich & Coelho, 2012; Weaver et al., 2012; Hörnschemeyer, 828 Bond & Young, 2013; Bar-On et al., 2014; Goyens et al., 2014; Kundanati et al., 2018), which 829 overcomes limitations imposed by the unusual geometries of cuticle, but does rely on the 830 inclusion of mechanical properties in the analysis. 831 It is also apparent from the literature that the links between the biology of arthropods and 832 their respective cuticle structure and properties are only just starting to be investigated in 833 conjunction. This likely results from the increasingly interdisciplinary nature of scientific 834 research. We hope that this will clarify the factors that dictate variations in cuticle structure and 835 composition. Further movement in this direction will benefit the field of biomimetics by 836 identifying biological tissues with unique or unusual functions, which are likely to be the most 837 interesting in terms of composition and structure. An excellent example of such a discovery is 838 the recently recognised sinusoidal structure in the very powerful mantis shrimp appendages 839 (Yaraghi et al., 2016). Future work could further benefit from investigating how the same 840 function or property has been achieved using different solutions, possibly along different 841 evolutionary pathways. A systematic approach to the species and taxa investigated would aid

Paper 1

842 such work and could also be used to investigate developmental pathways and concomitant 843 limitations to cuticle structure. 844 The links between cuticle structure, properties and functions aren’t well understood. This 845 in part results from limited consideration of the biological context and in vivo condition of the 846 cuticle when measuring biologically relevant mechanical properties. However, we note that some 847 general trends are apparent. Perhaps the clearest trend is the link between metal incorporation 848 and increased hardness and strength. An excellent example is arthropod mouthparts, where there 849 is clear evidence of metal incorporation in the tips where the greatest strain and wear occurs 850 (Hillerton & Vincent, 1982; Schofield, 2005; Cribb et al., 2007; Erko et al., 2013), presumably to 851 increase their hardness. Few studies have compared similar material in sclerotised versus 852 mineralised forms, although a study by Cribb et al. (Cribb et al., 2007) suggests that metal 853 incorporation did increase the hardness of sclerotised termite mandibles by 20%. This represents 854 a significant increase, but much less so than the reported 200 – 300% increase previously 855 reported, which did not correct for different levels of sclerotisation (Schofield, Nesson & 856 Richardson, 2002). In most cases, structural differences between cuticles - such as between joints 857 which have greater flexibility (Dirks & Taylor, 2012a) and beetle elytra which have greater 858 toughness (Lomakin et al., 2011) - are not well understood. The literature suggests that there is a 859 greater prevalence of the elastic protein, resilin, in flexible cuticle (Weis-Fogh, 1960) and 860 sclerotisation usually occurs in stiffer cuticle (Thompson & Hepburn, 1978). 861 Some aspects of cuticle structure had not been investigated at all until recently. One 862 example is, quantitative analysis on the relationship between the shape of cuticle structures and 863 their mechanical function, which has only been investigated in the past five years (van der 864 Meijden, Kleinteich & Coelho, 2012; Weaver et al., 2012; Taylor & Dirks, 2012; 865 Hörnschemeyer, Bond & Young, 2013; Bar-On et al., 2014; Goyens et al., 2014). While there are 866 studies of the links between the general design of engineering materials and their properties 867 (Weaver & Ashby, 1996), the shapes that evolve in nature are more elaborate and complex than 868 equivalent synthetic designs, and thus required FEA analysis to examine. 869 870 7. Conclusions 871 This paper has demonstrated the wide variation that occurs in cuticle and the impact of 872 some factors, such as sclerotisation, that have been investigated in the literature. It has also 873 identified interesting cuticle materials in terms of both novel microstructures and extreme 874 mechanical properties, which are worthy of further study. However, significant gaps in our 875 knowledge still exist. It is not known how the ultrastructure responds to stress, or how failure 876 occurs in the microstructure. The effect of salinity on the properties of aquatic and marine 877 arthropods hasn’t been investigated, and is rarely taken into consideration in experimental 878 design. In situ imaging is a potential future strategy that will provide a better understanding of 879 the microstructure response to stress, crack propagation and failure in cuticle (Patterson et al., 880 2016; Sykes et al., 2019). We believe that future studies will benefit greatly from using 881 multimodal techniques and modelling to analyse cuticle; composition and design are used in

Paper 1

882 conjunction so both need to be taken into consideration in any analysis. This would provide 883 scope for phylogenetically informed analysis of cuticle, between species or phylogenetic clades, 884 to determine the evolutionary pathways of design and composition, as well as the ontogeny and 885 genetic controls of these structures. The links between microstructure and hierarchy and the 886 mechanical properties of cuticle is perhaps where our understanding is most lacking, it is 887 probably the next major step required to create a more comprehensive model of the arthropod 888 cuticle. The rise in papers in the past 10 years investigating these links appears to support this 889 view and future work in this area could revolutionise our understanding of hierarchical design. 890 Ultimately, there are many factors influencing the mechanical properties of arthropod cuticle: 891 hydration, sclerotisation, mineralisation, architectural design of the microstructure, fibre 892 orientation and density, thickness of exocuticle relative to endocuticle, shape of the cuticle, and 893 intrinsic components (such as pore canals). These combine to create a complicated and difficult 894 to understand material, but as a result, also one that has the potential to provide many insights 895 into novel composite design. 896 897 Acknowledgements 898 PJW acknowledges the support of the European Research Council (ERC) under grant [H2020- 899 ERC-ADG project number 695638] CORREL-CT (Multiscale Correlative Research). 900 901 References 902 Aberle B, Jemmali R, Dirks J-H. 2017. Effect of sample treatment on biomechanical properties 903 of insect cuticle. Arthropod Structure & Development 46:138–146. DOI: 904 10.1016/j.asd.2016.08.001. 905 Alarie Y, Joly H, Dennie D. 1998. Cuticular hydrocarbon analysis of the aquatic beetle Agabus 906 anthracinus Mannerheim (Coleoptera: Dytiscidae). The Canadian Entomologist 130:615– 907 629. 908 Albanese Carmignani MP, Zaccone G. 1977. Morphochemical study of the cuticle in the 909 millepede Pachyiulus flavipes--C. Koch (Diplopoda, Myriapoda). Cellular and 910 Molecular Biology, Including Cyto-Enzymology 22:163–168. 911 Alexander DE, Blodig J, Hsieh S-Y. 1995. Relationship between Function and Mechanical 912 Properties of the Pleopods of Isopod Crustaceans. Invertebrate Biology 114:169–179. 913 DOI: 10.2307/3226889. 914 Ansenne A, Compère P, Goffinet G. 1990. Ultrastructural organization and chemical 915 composition of the mineralized cuticle of Glomeris marginata (Myriapoda, Diplopoda). 916 Brill Academic Publishers. 917 Ashby MF, Easterling KE, Harrysson R, Maiti SK. 1985. The Fracture and Toughness of Woods. 918 Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering 919 Sciences 398:261–280. DOI: 10.1098/rspa.1985.0034. 920 Astrop TI, Sahni V, Blackledge TA, Stark AY. 2015. Mechanical properties of the chitin- 921 calcium-phosphate “clam shrimp” carapace (Branchiopoda: Spinicaudata): implications

Paper 1

922 for taphonomy and fossilization. Journal of Crustacean Biology 35:123–131. DOI: 923 10.1163/1937240X-00002332. 924 Barbakadze N. 2006. Local mechanical properties of the head articulation cuticle in the beetle 925 Pachnoda marginata (Coleoptera, Scarabaeidae). Journal of Experimental Biology 926 209:722–730. DOI: 10.1242/jeb.02065. 927 Bar-On B, Barth FG, Fratzl P, Politi Y. 2014. Multiscale structural gradients enhance the 928 biomechanical functionality of the spider fang. Nature Communications 5:3894. DOI: 929 10.1038/ncomms4894. 930 Bennemann M, Backhaus S, Scholz I, Park D, Mayer J, Baumgartner W. 2014. Determination of 931 the Young’s modulus of the epicuticle of the smooth adhesive organs of Carausius 932 morosus using tensile testing. The Journal of Experimental Biology 217:3677–3687. 933 DOI: 10.1242/jeb.105114. 934 Bentov S, Aflalo ED, Tynyakov J, Glazer L, Sagi A. 2016. Calcium phosphate mineralization is 935 widely applied in crustacean mandibles. Scientific Reports 6. DOI: 10.1038/srep22118. 936 Bertram JE, Gosline JM. 1987. Functional design of horse hoof keratin: the modulation of 937 mechanical properties through hydration effects. Journal of Experimental Biology 938 130:121–136. 939 Bigi A, Burghammer M, Falconi R, Koch MHJ, Panzavolta S, Riekel C. 2001. Twisted Plywood 940 Pattern of Collagen Fibrils in Teleost Scales: An X-ray Diffraction Investigation. Journal 941 of Structural Biology 136:137–143. DOI: 10.1006/jsbi.2001.4426. 942 Bonser RHC. 1996. The mechanical properties of feather keratin. Journal of Zoology 239:477– 943 484. DOI: 10.1111/j.1469-7998.1996.tb05937.x. 944 Bonser RHC, Dawson C. 1999. The structural mechanical properties of down feathers and 945 biomimicking natural insulation materials. Journal of Materials Science Letters 18:1769– 946 1770. DOI: 10.1023/A:1006631328233. 947 Bonser R, Purslow P. 1995. The Young’s modulus of feather keratin. Journal of Experimental 948 Biology 198:1029–1033. 949 Borrell BJ. 2004. Mechanical properties of calcified exoskeleton from the neotropical millipede, 950 Nyssodesmus python. Journal of Insect Physiology 50:1121–1126. DOI: 951 10.1016/j.jinsphys.2004.09.012. 952 Boßelmann F, Romano P, Fabritius H, Raabe D, Epple M. 2007. The composition of the 953 exoskeleton of two crustacea: The American lobster Homarus americanus and the edible 954 crab Cancer pagurus. Thermochimica Acta 463:65–68. DOI: 10.1016/j.tca.2007.07.018. 955 Bouligand Y. 1972. Twisted fibrous arrangements in biological materials and cholesteric 956 mesophases. Tissue and Cell 4:189–217. DOI: 10.1016/S0040-8166(72)80042-9. 957 Broomell CC, Khan RK, Moses DN, Miserez A, Pontin MG, Stucky GD, Zok FW, Waite JH. 958 2007. Mineral minimization in nature’s alternative teeth. Journal of The Royal Society 959 Interface 4:19–31. DOI: 10.1098/rsif.2006.0153. 960 Broomell CC, Mattoni MA, Zok FW, Waite JH. 2006. Critical role of zinc in hardening of 961 Nereis jaws. Journal of Experimental Biology 209:3219–3225. DOI: 10.1242/jeb.02373.

Paper 1

962 Bruet BJF, Song J, Boyce MC, Ortiz C. 2008. Materials design principles of ancient fish armour. 963 Nature Materials 7:748–756. DOI: 10.1038/nmat2231. 964 Calvert PD. 1988. Encyclopedia Materials Science and Engineering. Biological 965 Macromolecules. Pergamon Press, Oxford. 966 Cambridge Engineering Selector software. 2003. Rustat House, 62 Clifton Rd, Cambridge, CB1 967 7EG.: Granta Design Limited. 968 Cameron GJ, Wess TJ, Bonser RHC. 2003. Young’s modulus varies with differential orientation 969 of keratin in feathers. Journal of Structural Biology 143:118–123. DOI: 10.1016/S1047- 970 8477(03)00142-4. 971 Chapman AD. 2009. Numbers of living species in Australia and the world. Toowoomba, 972 Australia: Australian Biodiversity Information Services. 973 Chen JX. 2018. Biomimetic research on beetle forewings in twenty years: Internal structure, 974 model and integrated honeycomb plates. Scientia Sinica (Technologica) 48:701–718. 975 Chen JX, Iwamoto M, Ni QQ, Kurashiki K, Saito K. 2001. Laminated Structure and Its 976 Mechanical Properties of the Fore-Wing of Beetle. Journal of the Society of Materials 977 Science, Japan 50:455–460. 978 Chen P, Lin A, Lin Y, Seki Y, Stokes A, Peyras J, Olevsky E, Meyers M, Mckittrick J. 2008a. 979 Structure and mechanical properties of selected biological materials. Journal of the 980 Mechanical Behavior of Biomedical Materials 1:208–226. DOI: 981 10.1016/j.jmbbm.2008.02.003. 982 Chen P-Y, Lin AY-M, McKittrick J, Meyers MA. 2008b. Structure and mechanical properties of 983 crab exoskeletons. Acta Biomaterialia 4:587–596. DOI: 10.1016/j.actbio.2007.12.010. 984 Chen P-Y, Schirer J, Simpson A, Nay R, Lin Y-S, Yang W, Lopez MI, Li J, Olevsky EA, 985 Meyers MA. 2012. Predation versus protection: Fish teeth and scales evaluated by 986 nanoindentation. Journal of Materials Research 27:100–112. DOI: 987 10.1557/jmr.2011.332. 988 Clark AJ, Triblehorn JD. 2014. Mechanical properties of the cuticles of three cockroach species 989 that differ in their wind-evoked escape behavior. PeerJ 2:e501. DOI: 10.7717/peerj.501. 990 Compére P, Goffinet G. 1987. Ultrastructural shape and three-dimensional organization of the 991 intracuticular canal systems in the mineralized cuticle of the green crab Carcinus maenas. 992 Tissue and Cell 19:839–857. DOI: 10.1016/0040-8166(87)90024-3. 993 Cribb BW, Lin C-L, Rintoul L, Rasch R, Hasenpusch J, Huang H. 2010. Hardness in arthropod 994 exoskeletons in the absence of transition metals. Acta Biomaterialia 6:3152–3156. DOI: 995 10.1016/j.actbio.2010.02.009. 996 Cribb BW, Stewart A, Huang H, Truss R, Noller B, Rasch R, Zalucki MP. 2007. Insect 997 mandibles—comparative mechanical properties and links with metal incorporation. 998 Naturwissenschaften 95:17–23. DOI: 10.1007/s00114-007-0288-1. 999 Cunniff PM, Fossey SA, Auerbach MA, Song JW, Kaplan DL, Adams WW, Eby RK, Mahoney 1000 D, Vezie DL. 1994. Mechanical and thermal properties of dragline silk from the spider

Paper 1

1001 Nephila clavipes. Polymers for Advanced Technologies 5:401–410. DOI: 1002 10.1002/pat.1994.220050801. 1003 Currey JD. 1999. The design of mineralised hard tissues for their mechanical functions. Journal 1004 of Experimental Biology 202:3285–3294. 1005 Currey JD. 2005. Hierarchies in Biomineral Structures. Science 309:253. DOI: 1006 10.1126/science.1113954. 1007 Currey JD, Nash A, Bonfield W. 1982. Calcified cuticle in the stomatopod smashing limb. 1008 Journal of Materials Science 17:1939–1944. DOI: 10.1007/BF00540410. 1009 Currey JD, Zioupos P, Peter D, Casinos A. 2001. Mechanical properties of nacre and highly 1010 mineralized bone. Proceedings of the Royal Society of London B: Biological Sciences 1011 268:107–111. DOI: 10.1098/rspb.2000.1337. 1012 Czapla TH, Hopkins TL, Kramer KJ. 1990. Cuticular strength and pigmentation of five strains of 1013 adult Blattella germanica (L.) during sclerotization: Correlations with catecholamines, β- 1014 alanine and food deprivation. Journal of Insect Physiology 36:647–654. DOI: 1015 10.1016/0022-1910(90)90069-R. 1016 Dai Z, Yang Z. 2010. Macro-/Micro-Structures of Elytra, Mechanical Properties of the 1017 Biomaterial and the Coupling Strength Between Elytra in Beetles. Journal of Bionic 1018 Engineering 7:6–12. DOI: 10.1016/S1672-6529(09)60187-6. 1019 Das R, Yadav RN, Sihota P, Uniyal P, Kumar N, Bhushan B. 2018. Biomechanical Evaluation of 1020 Wasp and Honeybee Stingers. Scientific Reports 8:14945. DOI: 10.1038/s41598-018- 1021 33386-y. 1022 Denny M. 1976. The physical properties of spider’s silk and their role in the design of orb-webs. 1023 Journal of Experimental Biology 65:483–506. 1024 Dirks J-H, Dürr V. 2011. Biomechanics of the stick insect antenna: Damping properties and 1025 structural correlates of the cuticle. Journal of the Mechanical Behavior of Biomedical 1026 Materials 4:2031–2042. DOI: 10.1016/j.jmbbm.2011.07.002. 1027 Dirks J-H, Parle E, Taylor D. 2013. Fatigue of insect cuticle. Journal of Experimental Biology 1028 216:1924–1927. DOI: 10.1242/jeb.083824. 1029 Dirks J-H, Taylor D. 2012a. Fracture toughness of locust cuticle. Journal of Experimental 1030 Biology 215:1502–1508. DOI: 10.1242/jeb.068221. 1031 Dirks J-H, Taylor D. 2012b. Veins Improve Fracture Toughness of Insect Wings. PLOS ONE 1032 7:e43411. DOI: 10.1371/journal.pone.0043411. 1033 Eichler C, Lomakin J, Arakane Y, Kramer KJ, Kanost MR, Gehrke SH, Beeman RW. 2006. 1034 Insect Cuticle as a Biomimetic Material. In: Proceedings of American Institute of 1035 Chemical Engineers Annual Meeting. 4309–4313. 1036 Enders S, Barbakadse N, Gorb S n., Arzt E. 2004. Exploring Biological Surfaces by 1037 Nanoindentation. Journal of Materials Research 19:880–887. DOI: 1038 10.1557/jmr.2004.19.3.880. 1039 Erko M, Hartmann MA, Zlotnikov I, Valverde Serrano C, Fratzl P, Politi Y. 2013. Structural and 1040 mechanical properties of the arthropod cuticle: Comparison between the fang of the

Paper 1

1041 spider Cupiennius salei and the carapace of American lobster Homarus americanus. 1042 Journal of Structural Biology 183:172–179. DOI: 10.1016/j.jsb.2013.06.001. 1043 Evans AR, Sanson GD. 2005. Biomechanical properties of insects in relation to insectivory: 1044 cuticle thickness as an indicator of insect “hardness” and “intractability.” Australian 1045 Journal of Zoology 53:9. DOI: 10.1071/ZO04018. 1046 Fabritius H-O, Sachs C, Triguero PR, Raabe D. 2009. Influence of Structural Principles on the 1047 Mechanics of a Biological Fiber-Based Composite Material with Hierarchical 1048 Organization: The Exoskeleton of the Lobster Homarus americanus. Advanced Materials 1049 21:391–400. DOI: 10.1002/adma.200801219. 1050 Franck A, Cocquyt G, Simoens P, De Belie N. 2006. Biomechanical Properties of Bovine Claw 1051 Horn. Biosystems Engineering 93:459–467. DOI: 10.1016/j.biosystemseng.2006.01.007. 1052 Gibson LJ. 1985. The mechanical behaviour of cancellous bone. Journal of Biomechanics 1053 18:317–328. DOI: 10.1016/0021-9290(85)90287-8. 1054 Giraud MM, Castanet J, Meunier FJ, Bouligand Y. 1978. The fibrous structure of coelacanth 1055 scales: A twisted ‘Plywood.’ Tissue and Cell 10:671–686. DOI: 10.1016/0040- 1056 8166(78)90054-X. 1057 Giraud-Guille MM. 1988. Twisted plywood architecture of collagen fibrils in human compact 1058 bone osteons. Calcified Tissue International 42:167–180. DOI: 10.1007/BF02556330. 1059 Goyens J, Soons J, Aerts P, Dirckx J. 2014. Finite-element modelling reveals force modulation 1060 of jaw adductors in stag beetles. Journal of The Royal Society Interface 11:20140908. 1061 DOI: 10.1098/rsif.2014.0908. 1062 Greaves GN, Greer AL, Lakes RS, Rouxel T. 2011. Poisson’s ratio and modern materials. 1063 Nature Materials 10:823–837. DOI: 10.1038/nmat3134. 1064 Green JP, Neff MR. 1972. A survey of the fine structure of the integument of the fiddler crab. 1065 Tissue and Cell 4:137–171. DOI: 10.1016/S0040-8166(72)80013-2. 1066 Griepentrog M, Krämer G, Cappella B. 2013. Comparison of nanoindentation and AFM methods 1067 for the determination of mechanical properties of polymers. Polymer Testing 32:455–460. 1068 DOI: 10.1016/j.polymertesting.2013.01.011. 1069 Grimaldi DA, Engel MS. 2005. Evolution of the insects. Cambridge [U.K.] ; New York: 1070 Cambridge University Press. 1071 Gupta NS (ed.). 2011. Chitin: formation and diagenesis. Dordrecht ; New York: Springer. 1072 Ha NS, Jin TL, Goo NS, Park HC. 2011. Anisotropy and non-homogeneity of an Allomyrina 1073 Dichotoma beetle hind wing membrane. Bioinspiration & Biomimetics 6:046003. DOI: 1074 10.1088/1748-3182/6/4/046003. 1075 Han C-S, Sanei SHR, Alisafaei F. 2016. On the origin of indentation size effects and depth 1076 dependent mechanical properties of elastic polymers. Journal of Polymer Engineering 36. 1077 DOI: 10.1515/polyeng-2015-0030. 1078 Hepburn HR, Ball A. 1973. On the structure and mechanical properties of beetle shells. Journal 1079 of Materials Science 8:618–623. DOI: 10.1007/BF00561216.

Paper 1

1080 Hepburn HR, Chandler HD. 1976. Material properties of arthropod cuticles: the arthrodial 1081 membranes. Journal of Comparative Physiology B: Biochemical, Systemic, and 1082 Environmental Physiology 109:177–198. 1083 Hepburn HR, Joffe I. 1974. Locust solid cuticle—a time sequence of mechanical properties. 1084 Journal of Insect Physiology 20:497–506. 1085 Hepburn HR, Joffe I, Green N, Nelson KJ. 1975. Mechanical properties of a crab shell. 1086 Comparative Biochemistry and Physiology Part A: Physiology 50:551IN13–554. 1087 Hillerton JE, Reynolds SE, Vincent JFV. 1982. On the indentation hardness of insect cuticle. 1088 Journal of Experimental Biology 96:45–52. 1089 Hillerton JE, Vincent JFV. 1982. The Specific Location of Zinc in Insect Mandibles. Journal of 1090 Experimental Biology 101:333–336. 1091 Hörnschemeyer T, Bond J, Young PG. 2013. Analysis of the Functional Morphology of 1092 Mouthparts of the Beetle Priacma serrata, and a Discussion of Possible Food Sources. 1093 Journal of Insect Science 13. DOI: 10.1673/031.013.12601. 1094 Ikoma T, Kobayashi H, Tanaka J, Walsh D, Mann S. 2003. Microstructure, mechanical, and 1095 biomimetic properties of fish scales from Pagrus major. Journal of Structural Biology 1096 142:327–333. DOI: 10.1016/S1047-8477(03)00053-4. 1097 Jackson AP, Vincent JFV, Turner RM. 1988. The Mechanical Design of Nacre. Proceedings of 1098 the Royal Society of London. Series B. Biological Sciences 234:415. DOI: 1099 10.1098/rspb.1988.0056. 1100 Jia Z, Yu Y, Wang L. 2019. Learning from nature: Use material architecture to break the 1101 performance tradeoffs. Materials & Design 168:107650. DOI: 1102 10.1016/j.matdes.2019.107650. 1103 Joffe I, Hepburn HR, Nelson KJ, Green N. 1975. Mechanical properties of a crustacean 1104 exoskeleton. Comparative Biochemistry and Physiology Part A: Physiology 50:545–549. 1105 DOI: 10.1016/0300-9629(75)90312-6. 1106 Kamat S, Su X, Ballarini R, Heuer AH. 2000. Structural basis for the fracture toughness of the 1107 shell of the conch Strombus gigas. Nature 405:1036–1040. DOI: 10.1038/35016535. 1108 van de Kamp T, Greven H. 2010. On the Architecture of Beetle Elytra. Entomologie heute 1109 22:191–204. 1110 van de Kamp T, Riedel A, Greven H. 2016. Micromorphology of the elytral cuticle of beetles, 1111 with an emphasis on weevils (Coleoptera: Curculionoidea). Arthropod Structure & 1112 Development 45:14–22. DOI: 10.1016/j.asd.2015.10.002. 1113 Kapzov S. 1911. Untersuchungen über den feineren Bau der Cuticula bei Insekten. Zeitschrift für 1114 wissenschaftliche Zoologie 98:297–337. 1115 Kasapi MA, Gosline JM. 1997. Design complexity and fracture control in the equine hoof wall. 1116 Journal of Experimental Biology 200:1639–1659. 1117 Katz SL, Gosline JM. 1992. Ontogenetic Scaling and Mechanical Behaviour of the Tibiae of the 1118 African Desert Locust (Schistocerca Gregaria). Journal of Experimental Biology 1119 168:125–150.

Paper 1

1120 Kinney JH, Habelitz S, Marshall SJ, Marshall GW. 2003. The Importance of Intrafibrillar 1121 Mineralization of Collagen on the Mechanical Properties of Dentin. Journal of Dental 1122 Research 82:957–961. DOI: 10.1177/154405910308201204. 1123 Kitchener A, Vincent JFV. 1987. Composite theory and the effect of water on the stiffness of 1124 horn keratin. Journal of Materials Science 22:1385–1389. DOI: 10.1007/BF01233138. 1125 Klocke D, Schmitz H. 2011. Water as a major modulator of the mechanical properties of insect 1126 cuticle. Acta Biomaterialia 7:2935–2942. DOI: 10.1016/j.actbio.2011.04.004. 1127 Kohane M, Daugela A, Kutomi H, Charlson L, Wyrobek A, Wyrobek J. 2003. Nanoscale in vivo 1128 evaluation of the stiffness of Drosophila melanogaster integument during development. 1129 Journal of Biomedical Materials Research Part A 66A:633–642. DOI: 1130 10.1002/jbm.a.10028. 1131 Kruzic JJ, Kim DK, Koester KJ, Ritchie RO. 2009. Indentation techniques for evaluating the 1132 fracture toughness of biomaterials and hard tissues. Journal of the Mechanical Behavior 1133 of Biomedical Materials 2:384–395. DOI: 10.1016/j.jmbbm.2008.10.008. 1134 Kundanati L, Signetti S, Gupta HS, Menegon M, Pugno NM. 2018. Multilayer stag beetle elytra 1135 perform better under external loading via non-symmetric bending properties. Journal of 1136 The Royal Society Interface 15:1–13. DOI: 10.1098/rsif.2018.0427. 1137 Lichtenegger HC, Schöberl T, Bartl MH, Waite H, Stucky GD. 2002. High Abrasion Resistance 1138 with Sparse Mineralization: Copper Biomineral in Worm Jaws. Science 298:389–392. 1139 DOI: 10.1126/science.1075433. 1140 Lichtenegger HC, Schöberl T, Ruokolainen JT, Cross JO, Heald SM, Birkedal H, Waite JH, 1141 Stucky GD. 2003. Zinc and mechanical prowess in the jaws of Nereis, a marine worm. 1142 Proceedings of the National Academy of Sciences 100:9144–9149. DOI: 1143 10.1073/pnas.1632658100. 1144 Lin YS, Wei CT, Olevsky EA, Meyers MA. 2011. Mechanical properties and the laminate 1145 structure of Arapaima gigas scales. Journal of the Mechanical Behavior of Biomedical 1146 Materials 4:1145–1156. DOI: 10.1016/j.jmbbm.2011.03.024. 1147 Lomakin J, Huber PA, Eichler C, Arakane Y, Kramer KJ, Beeman RW, Kanost MR, Gehrke SH. 1148 2011. Mechanical Properties of the Beetle Elytron, a Biological Composite Material. 1149 Biomacromolecules 12:321–335. DOI: 10.1021/bm1009156. 1150 Luquet G, Marin F. 2004. Biomineralisations in crustaceans: storage strategies. Comptes Rendus 1151 Palevol 3:515–534. DOI: 10.1016/j.crpv.2004.07.015. 1152 Marino Cugno Garrano A, La Rosa G, Zhang D, Niu L-N, Tay FR, Majd H, Arola D. 2012. On 1153 the mechanical behavior of scales from Cyprinus carpio. Journal of the Mechanical 1154 Behavior of Biomedical Materials 7:17–29. DOI: 10.1016/j.jmbbm.2011.07.017. 1155 Marshall GW, Balooch M, Gallagher RR, Gansky SA, Marshall SJ. 2001. Mechanical properties 1156 of the dentinoenamel junction: AFM studies of nanohardness, elastic modulus, and 1157 fracture. Journal of Biomedical Materials Research 54:87–95. DOI: 10.1002/1097- 1158 4636(200101)54:1<87::AID-JBM10>3.0.CO;2-Z.

Paper 1

1159 van der Meijden A, Kleinteich T, Coelho P. 2012. Packing a pinch: functional implications of 1160 chela shapes in scorpions using finite element analysis. Journal of Anatomy 220:423– 1161 434. DOI: 10.1111/j.1469-7580.2012.01485.x. 1162 Melnick CA, Chen Z, Mecholsky JJ. 1996. Hardness and toughness of exoskeleton material in 1163 the stone crab, Menippe mercenaria. Journal of Materials Research 11:2903–2907. DOI: 1164 10.1557/JMR.1996.0367. 1165 Meyers MA, Chen P-Y, Lin AY-M, Seki Y. 2008a. Biological materials: Structure and 1166 mechanical properties. Progress in Materials Science 53:1–206. DOI: 1167 10.1016/j.pmatsci.2007.05.002. 1168 Meyers MA, Lin AY-M, Chen P-Y, Muyco J. 2008b. Mechanical strength of abalone nacre: 1169 Role of the soft organic layer. Journal of the Mechanical Behavior of Biomedical 1170 Materials 1:76–85. DOI: 10.1016/j.jmbbm.2007.03.001. 1171 Miserez A, Schneberk T, Sun C, Zok FW, Waite JH. 2008. The Transition from Stiff to 1172 Compliant Materials in Squid Beaks. Science 319:1816–1819. DOI: 1173 10.1126/science.1154117. 1174 Moses DN, Mattoni MA, Slack NL, Waite JH, Zok FW. 2006. Role of melanin in mechanical 1175 properties of Glycera jaws. Acta Biomaterialia 2:521–530. DOI: 1176 10.1016/j.actbio.2006.05.002. 1177 Müller M, Olek M, Giersig M, Schmitz H. 2008. Micromechanical properties of consecutive 1178 layers in specialized insect cuticle: the gula of Pachnoda marginata (Coleoptera, 1179 Scarabaeidae) and the infrared sensilla of Melanophila acuminata (Coleoptera, 1180 Buprestidae). Journal of Experimental Biology 211:2576–2583. DOI: 1181 10.1242/jeb.020164. 1182 Nalla RK, Kinney JH, Ritchie RO. 2003. Mechanistic fracture criteria for the failure of human 1183 cortical bone. Nature Materials 2:164–168. DOI: 10.1038/nmat832. 1184 Neville AC. 1975. Biology of the arthropod cuticle. Springer-Verlag. 1185 Nikolov S, Fabritius H, Petrov M, Friák M, Lymperakis L, Sachs C, Raabe D, Neugebauer J. 1186 2011. Robustness and optimal use of design principles of arthropod exoskeletons studied 1187 by ab initio-based multiscale simulations. Journal of the Mechanical Behavior of 1188 Biomedical Materials 4:129–145. DOI: 10.1016/j.jmbbm.2010.09.015. 1189 Nikolov S, Petrov M, Lymperakis L, Friák M, Sachs C, Fabritius H-O, Raabe D, Neugebauer J. 1190 2010. Revealing the Design Principles of High-Performance Biological Composites 1191 Using Ab initio and Multiscale Simulations: The Example of Lobster Cuticle. Advanced 1192 Materials 22:519–526. DOI: 10.1002/adma.200902019. 1193 Parle E, Dirks J-H, Taylor D. 2016. Bridging the gap: wound healing in insects restores 1194 mechanical strength by targeted cuticle deposition. Journal of The Royal Society 1195 Interface 13:20150984. DOI: 10.1098/rsif.2015.0984. 1196 Parle E, Herbaj S, Sheils F, Larmon H, Taylor D. 2016. Buckling failures in insect exoskeletons. 1197 Bioinspiration & Biomimetics 11:016003. DOI: 10.1088/1748-3190/11/1/016003.

Paper 1

1198 Parle E, Taylor D. 2013. The self-healing properties of insect cuticle. Journal of Postgraduate 1199 Research, Trinity College Dublin 12:90–111. 1200 Patterson BM, Cordes NL, Henderson K, Mertens JCE, Clarke AJ, Hornberger B, Merkle A, 1201 Etchin S, Tkachuk A, Leibowitz M, Trapp D, Qiu W, Zhang B, Bale H, Lu X, Hartwell 1202 R, Withers PJ, Bradley RS. 2016. In Situ Laboratory-Based Transmission X-Ray 1203 Microscopy and Tomography of Material Deformation at the Nanoscale. Experimental 1204 Mechanics. DOI: 10.1007/s11340-016-0197-3. 1205 Peisker H, Michels J, Gorb SN. 2013. Evidence for a material gradient in the adhesive tarsal 1206 setae of the ladybird beetle Coccinella septempunctata. Nature Communications 4:1661. 1207 DOI: 10.1038/ncomms2576. 1208 Pérez-Rigueiro J, Elices M, Llorca J, Viney C. 2001. Tensile properties of Argiope trifasciata 1209 drag line silk obtained from the spider’s web. Journal of Applied Polymer Science 1210 82:2245–2251. DOI: 10.1002/app.2072. 1211 Pérez-Rigueiro J, Viney C, Llorca J, Elices M. 2000. Mechanical properties of single-brin 1212 silkworm silk. Journal of Applied Polymer Science 75:1270–1277. DOI: 1213 10.1002/(SICI)1097-4628(20000307)75:10<1270::AID-APP8>3.0.CO;2-C. 1214 Peterlik H, Roschger P, Klaushofer K, Fratzl P. 2006. From brittle to ductile fracture of bone. 1215 Nature Materials 5:52–55. DOI: 10.1038/nmat1545. 1216 Politi Y, Priewasser M, Pippel E, Zaslansky P, Hartmann J, Siegel S, Li C, Barth FG, Fratzl P. 1217 2012. A Spider’s Fang: How to Design an Injection Needle Using Chitin-Based 1218 Composite Material. Advanced Functional Materials 22:2519–2528. DOI: 1219 10.1002/adfm.201200063. 1220 Pütz K, Buchholz F. 1991. Comparative ultrastructure of the cuticle of some pelagic, 1221 nektobenthic and benthic malacostracan crustaceans. Marine Biology 110:49–58. 1222 Raabe D, Al-Sawalmih A, Yi SB, Fabritius H. 2007. Preferred crystallographic texture of α- 1223 chitin as a microscopic and macroscopic design principle of the exoskeleton of the lobster 1224 Homarus americanus. Acta Biomaterialia 3:882–895. DOI: 10.1016/j.actbio.2007.04.006. 1225 Raabe D, Romano P, Sachs C, Al-Sawalmih A, Brokmeier H-G, Yi S-B, Servos G, Hartwig HG. 1226 2005. Discovery of a honeycomb structure in the twisted plywood patterns of fibrous 1227 biological nanocomposite tissue. Journal of Crystal Growth 283:1–7. DOI: 1228 10.1016/j.jcrysgro.2005.05.077. 1229 Raabe D, Sachs C, Romano P. 2005. The crustacean exoskeleton as an example of a structurally 1230 and mechanically graded biological nanocomposite material. Acta Materialia 53:4281– 1231 4292. DOI: 10.1016/j.actamat.2005.05.027. 1232 Reilly DT, Burstein AH, Frankel VH. 1974. The elastic modulus for bone. Journal of 1233 Biomechanics 7:271–275. DOI: 10.1016/0021-9290(74)90018-9. 1234 Reis D, Vian B, Roland J-C. 1994. Cellulose-glucuronoxylans and plant cell wallstructure. 1235 Micron 25:171–187. DOI: 10.1016/0968-4328(94)90041-8. 1236 Reynolds SE. 1975. The mechanical properties of the abdominal cuticle of Rhodnius larvae. 1237 Journal of Experimental Biology 62:69–80.

Paper 1

1238 Reynolds S E. 1977. Control of cuticle extensibility in the wings of adult Manduca at the time of 1239 eclosion: effects of eclosion hormone and bursicon. Journal of Experimental Biology 1240 70:27–39. 1241 Rho J-Y, Kuhn-Spearing L, Zioupos P. 1998. Mechanical properties and the hierarchical 1242 structure of bone. Medical Engineering & Physics 20:92–102. DOI: 10.1016/S1350- 1243 4533(98)00007-1. 1244 Roer R, Dillaman R. 1984. The structure and calcification of the crustacean cuticle. American 1245 Zoologist 24:893–909. 1246 Romano P, Fabritius H, Raabe D. 2007. The exoskeleton of the lobster Homarus americanus as 1247 an example of a smart anisotropic biological material. Acta Biomaterialia 3:301–309. 1248 DOI: 10.1016/j.actbio.2006.10.003. 1249 Roseland CR, Kramer KJ, Hopkins TL. 1987. Cuticular strength and pigmentation of rust-red 1250 and black strains of Tribolium castaneum: Correlation with catecholamine and β-alanine 1251 content. Insect Biochemistry 17:21–28. DOI: 10.1016/0020-1790(87)90139-9. 1252 Sachs C, Fabritius H, Raabe D. 2006a. Hardness and elastic properties of dehydrated cuticle 1253 from the lobster Homarus americanus obtained by nanoindentation. Journal of Materials 1254 Research 21:1987–1995. DOI: 10.1557/jmr.2006.0241. 1255 Sachs C, Fabritius H, Raabe D. 2006b. Experimental investigation of the elastic–plastic 1256 deformation of mineralized lobster cuticle by digital image correlation. Journal of 1257 Structural Biology 155:409–425. DOI: 10.1016/j.jsb.2006.06.004. 1258 Sachs C, Fabritius H, Raabe D. 2008. Influence of microstructure on deformation anisotropy of 1259 mineralized cuticle from the lobster Homarus americanus. Journal of Structural Biology 1260 161:120–132. DOI: 10.1016/j.jsb.2007.09.022. 1261 Schöberl T, Jäger IL. 2006. Wet or Dry – Hardness, Stiffness and Wear Resistance of Biological 1262 Materials on the Micron Scale. Advanced Engineering Materials 8:1164–1169. DOI: 1263 10.1002/adem.200600143. 1264 Schofield RM. 2005. Metal-halogen biomaterials. American Entomologist 51:45–47. 1265 Schofield RM, Nesson MH, Richardson KA. 2002. Tooth hardness increases with zinc-content in 1266 mandibles of young adult leaf-cutter ants. Naturwissenschaften 89:579–583. 1267 Seki Y, Kad B, Benson D, Meyers MA. 2006. The toucan beak: Structure and mechanical 1268 response. Materials Science and Engineering: C 26:1412–1420. DOI: 1269 10.1016/j.msec.2005.08.025. 1270 Shrivastava SC. 1970. Cuticular components of common Indian arachnids and myriapods. 1271 Experientia 26:1028–1029. DOI: 10.1007/BF02114179. 1272 Smith CW, Herbert R, Wootton RJ, Evans KE. 2000. The hind wing of the desert locust 1273 (Schistocerca gregaria Forskal). II. Mechanical properties and functioning of the 1274 membrane. Journal of Experimental Biology 203:2933–2943. 1275 Sun J, Tong J. 2007. Fracture Toughness Properties of Three Different Biomaterials Measured 1276 by Nanoindentation. Journal of Bionic Engineering 4:11–17. DOI: 10.1016/S1672- 1277 6529(07)60007-9.

Paper 1

1278 Sun J, Tong J, Chen D, Lin J, Liu X, Wang Y. 2010. Micro-tensile Testing of the Lightweight 1279 Laminated Structures of Beetle Elytra Cuticle. Advances in Natural Science 3:225–234. 1280 DOI: 10.3968/j.ans.1715787020100302.027. 1281 Sun J, Tong J, Ma Y. 2008. Nanomechanical Behaviours of Cuticle of Three Kinds of Beetle. 1282 Journal of Bionic Engineering 5:152–157. DOI: 10.1016/S1672-6529(08)60087-6. 1283 Sun J, Tong J, Zhang Z. 2009. Nanomechanical properties and the hierarchical structure of elytra 1284 cuticle of dung beetle (Copris ochus Motschulsky). In: 2009 International Conference on 1285 Mechatronics and Automation. 4277–4282. DOI: 10.1109/ICMA.2009.5246542. 1286 Sun J y, Tong J, Zhou J. 2006. Application of nano-indenter for investigation of the properties of 1287 the elytra cuticle of the dung beetle (Copris ochus Motschulsky). IEE Proceedings - 1288 Nanobiotechnology 153:129–133. DOI: 10.1049/ip-nbt:20050050. 1289 Sykes D, Hartwell R, Bradley RS, Burnett TL, Hornberger B, Garwood RJ, Withers PJ. 2019. 1290 Time-lapse three-dimensional imaging of crack propagation in beetle cuticle. Acta 1291 Biomaterialia 86:109–116. DOI: 10.1016/j.actbio.2019.01.031. 1292 Taylor AM, Bonser RHC, Farrent JW. 2004. The influence of hydration on the tensile and 1293 compressive properties of avian keratinous tissues. Journal of Materials Science 39:939– 1294 942. DOI: 10.1023/B:JMSC.0000012925.92504.08. 1295 Taylor D, Dirks J-H. 2012. Shape optimization in exoskeletons and endoskeletons: a 1296 biomechanics analysis. Journal of The Royal Society Interface 9:3480–3489. DOI: 1297 10.1098/rsif.2012.0567. 1298 Taylor JRA, Hebrank J, Kier WM. 2007. Mechanical properties of the rigid and hydrostatic 1299 skeletons of molting blue crabs, Callinectes sapidus Rathbun. Journal of Experimental 1300 Biology 210:4272–4278. DOI: 10.1242/jeb.007054. 1301 Tesch W, Eidelman N, Roschger P, Goldenberg F, Klaushofer K, Fratzl P. 2001. Graded 1302 Microstructure and Mechanical Properties of Human Crown Dentin. Calcified Tissue 1303 International 69:147–157. DOI: 10.1007/s00223-001-2012-z. 1304 Thompson PR, Hepburn HR. 1978. Changes in chemical and mechanical properties of honeybee 1305 (Apis mellifera adansonii L.) cuticle during development. Journal of Comparative 1306 Physiology ? B 126:257–262. DOI: 10.1007/BF00688935. 1307 Thorez A, Compere P, Goffinet G, IRSIA B. 1992. Ultrastructure and mineral composition of the 1308 tergite cuticle of the iulid millipede Ophyiulus pilosus. Berichte des 1309 Naturwissenschaftlich-medizinischen Vereins Innsbruck 10:63–70. 1310 Tombolato L, Novitskaya EE, Chen P-Y, Sheppard FA, McKittrick J. 2010. Microstructure, 1311 elastic properties and deformation mechanisms of horn keratin. Acta Biomaterialia 1312 6:319–330. DOI: 10.1016/j.actbio.2009.06.033. 1313 Trim MW, Horstemeyer MF, Rhee H, El Kadiri H, Williams LN, Liao J, Walters KB, 1314 McKittrick J, Park S-J. 2011. The effects of water and microstructure on the mechanical 1315 properties of bighorn sheep (Ovis canadensis) horn keratin. Acta Biomaterialia 7:1228– 1316 1240. DOI: 10.1016/j.actbio.2010.11.024.

Paper 1

1317 Vincent JFV. 1975. Locust Oviposition: Stress Softening of the Extensible Intersegmental 1318 Membranes. Proceedings of the Royal Society of London B: Biological Sciences 1319 188:189–201. DOI: 10.1098/rspb.1975.0012. 1320 Vincent JFV. 2002. Arthropod cuticle: a natural composite shell system. Composites Part A: 1321 Applied Science and Manufacturing 33:1311–1315. 1322 Vincent JFV, Prentice JH. 1973. Rheological properties of the extensible intersegmental 1323 membrane of the adult female locust. Journal of Materials Science 8:624–630. DOI: 1324 10.1007/BF00561217. 1325 Vincent JFV, Wegst UGK. 2004. Design and mechanical properties of insect cuticle. Arthropod 1326 Structure & Development 33:187–199. DOI: 10.1016/j.asd.2004.05.006. 1327 Wagner R, Pittendrigh BR, Raman A. 2012. Local elasticity and adhesion of nanostructures on 1328 Drosophila melanogaster wing membrane studied using atomic force microscopy. 1329 Applied Surface Science 259:225–230. DOI: 10.1016/j.apsusc.2012.07.023. 1330 Wan C, Hao Z, Feng X. 2016. Structures, properties, and energy-storage mechanisms of the 1331 semi-lunar process cuticles in locusts. Scientific Reports 6. DOI: 10.1038/srep35219. 1332 Wang X-S, Li Y, Shi Y-F. 2008. Effects of sandwich microstructures on mechanical behaviors of 1333 dragonfly wing vein. Composites Science and Technology 68:186–192. DOI: 1334 10.1016/j.compscitech.2007.05.023. 1335 Wang L, Zhang H, Fan Y. 2011. Comparative study of the mechanical properties, micro- 1336 structure, and composition of the cranial and beak bones of the great spotted woodpecker 1337 and the lark bird. Science China Life Sciences 54:1036–1041. DOI: 10.1007/s11427-011- 1338 4242-2. 1339 Warburton FL. 1948. Determination Of The Elastic Properties Of Horn Keratin. Journal of the 1340 Textile Institute Proceedings 39:P297–P308. DOI: 10.1080/19447014808663158. 1341 Weaver PM, Ashby MF. 1996. The Optimal Selection of Material and Section-shape. Journal of 1342 Engineering Design 7:129–150. DOI: 10.1080/09544829608907932. 1343 Weaver JC, Milliron GW, Miserez A, Evans-Lutterodt K, Herrera S, Gallana I, Mershon WJ, 1344 Swanson B, Zavattieri P, DiMasi E, Kisailus D. 2012. The Stomatopod Dactyl Club: A 1345 Formidable Damage-Tolerant Biological Hammer. Science 336:1275–1280. DOI: 1346 10.1126/science.1218764. 1347 Weiner S, Wagner HD. 1998. The Material Bone: Structure-Mechanical Function Relations. 1348 Annual Review of Materials Science 28:271–298. DOI: 10.1146/annurev.matsci.28.1.271. 1349 Weis-Fogh T. 1960. A rubber-like protein in insect cuticle. Journal of Experimental Biology 1350 37:889–907. 1351 Wolfgang WJ, Riddiford LM. 1987. Cuticular Mechanics During Larval Development Of The 1352 Tobacco Hornworm, Manduca Sexta. Journal of Experimental Biology 128:19–33. 1353 Wu J, Qin Z, Qu L, Zhang H, Deng F, Guo M. 2019. Natural hydrogel in American lobster: a 1354 soft armor with high toughness and strength. Acta Biomaterialia. DOI: 1355 10.1016/j.actbio.2019.01.067.

Paper 1

1356 Xu W, Mulhern PJ, Blackford BL, Jericho MH, Templeton I, Dahn DC, Moy V, Chernoff DA. 1357 1994. A new atomic force microscopy technique for the measurement of the elastic 1358 properties of biological materials. Scanning Microscopy 8:499–506. 1359 Xu M, Pan L, Chen J, Zhang X, Yu X. 2019. The flexural properties of end-trabecular beetle 1360 elytron plates and their flexural failure mechanism. Journal of Materials Science. DOI: 1361 10.1007/s10853-019-03488-7. 1362 Yang W, Gludovatz B, Zimmermann EA, Bale HA, Ritchie RO, Meyers MA. 2013. Structure 1363 and fracture resistance of alligator gar (Atractosteus spatula) armored fish scales. Acta 1364 Biomaterialia 9:5876–5889. DOI: 10.1016/j.actbio.2012.12.026. 1365 Yang W, Kashani N, Li X-W, Zhang G-P, Meyers MA. 2011. Structural characterization and 1366 mechanical behavior of a bivalve shell (Saxidomus purpuratus). Materials Science and 1367 Engineering: C 31:724–729. DOI: 10.1016/j.msec.2010.10.003. 1368 Yaraghi NA, Guarín-Zapata N, Grunenfelder LK, Hintsala E, Bhowmick S, Hiller JM, Betts M, 1369 Principe EL, Jung J-Y, Sheppard L, Wuhrer R, McKittrick J, Zavattieri PD, Kisailus D. 1370 2016. A Sinusoidally Architected Helicoidal Biocomposite. Advanced Materials:1–10. 1371 DOI: 10.1002/adma.201600786. 1372 Yu M, Hermann I, Dai Z, Gitis N. 2013. Mechanical and Frictional Properties of the Elytra of 1373 Five Species of Beetles. Journal of Bionic Engineering 10:77–83. DOI: 10.1016/S1672- 1374 6529(13)60201-2. 1375 Zaheri A, Fenner JS, Russell BP, Restrepo D, Daly M, Wang D, Hayashi C, Meyers MA, 1376 Zavattieri PD, Espinosa HD. 2018. Revealing the Mechanics of Helicoidal Composites 1377 through Additive Manufacturing and Beetle Developmental Stage Analysis. Advanced 1378 Functional Materials 28:1803073. DOI: 10.1002/adfm.201803073. 1379 Zemlin JC. 1968. A study of the mechanical behavior of spider silks. Collaborative Research Inc 1380 Waltham MA. 1381 Zhang Z, Zhang Y-W, Gao H. 2011. On optimal hierarchy of load-bearing biological materials. 1382 Proceedings of the Royal Society of London B: Biological Sciences 278:519–525. DOI: 1383 10.1098/rspb.2010.1093. 1384 Zioupos P, Currey JD, Casinos A. 2000. Exploring the Effects of Hypermineralisation in Bone 1385 Tissue by Using an Extreme Biological Example. Connective Tissue Research 41:229– 1386 248. DOI: 10.3109/03008200009005292. 1387 1388

PAPER 2

Time-lapse three-dimensional imaging of crack propagation in beetle cuticle

Acta Biomaterialia 86 (2019) 109–116

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier.com/locate/actabiomat

Full length article Time-lapse three-dimensional imaging of crack propagation in beetle cuticle ⇑ Dan Sykes a, , Rebecca Hartwell a,1, Rob S. Bradley a,2, Timothy L. Burnett a, Benjamin Hornberger b,3, Russell J. Garwood c,d, Philip J. Withers a a Henry Moseley X-ray Imaging Facility, The Royce Institute, School of Materials, The University of Manchester, Manchester M13 9PL, UK b Carl Zeiss X-ray Microscopy Inc., Pleasanton, CA, USA c School of Earth and Environmental Science, The University of Manchester, Manchester M13 9PL, UK d Earth Sciences Department, Natural History Museum, London SW7 5BD, UK article info abstract

Article history: Arthropod cuticle has extraordinary properties. It is very stiff and tough whilst being lightweight, yet it is Received 6 September 2018 made of rather ordinary constituents. This desirable combination of properties results from a hierarchical Received in revised form 9 January 2019 structure, but we currently have a poor understanding of how this impedes damage propagation. Here we Accepted 14 January 2019 use non-destructive, time-lapse in situ tensile testing within an X-ray nanotomography (nCT) system to Available online 17 January 2019 visualise crack progression through dry beetle elytron (wing case) cuticle in 3D. We find that its hierarchical pseudo-orthogonal laminated microstructure exploits many extrinsic toughening mechanisms, Keywords: including crack deflection, fibre and laminate pull-out and crack bridging. We highlight lessons to be Arthropod cuticle learned in the design of engineering structures from the toughening methods employed. Time-lapse imaging X-ray tomography Biological composites Statement of Significance

We present the first comprehensive study of the damage and toughening mechanisms within arthropod cuticle in a 3D time-lapse manner, using X-ray nanotomography during crack growth. This technique allows lamina to be isolated despite being convex, which limits 2D analysis of microstructure. We report toughening mechanisms previously unobserved in unmineralised cuticle such as crack deflection, fibre and laminate pull-out and crack bridging; and provide insights into the effects of hierarchical microstructure on crack propagation. Ultimately the benefits of the hierarchical microstructure found here can not only be used to improve biomimetic design, but also helps us to understand the remarkable success of arthropods on Earth. Ó 2019 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction impact of moulting) and ensuring low density, to facilitate locomotion such as flight [1,3]. An extreme case of the balance between Arthropod exoskeletons typically comprise a lightweight, hier- protection and weight reduction are beetle elytra, which are archically organised chitin-protein composite (i.e. cuticle) [1] that unmineralised chitinous wing cases [4,5]. Indeed, they show one can exhibit a diverse range of properties [2,3]. Many biological of the highest weight to stiffness ratios of any arthropod cuticle materials employ mineralisation to increase strength and stiffness. (Fig. 1), which makes elytra a potentially fruitful biomaterial for However, most terrestrial arthropod cuticles are not mineralised to study. allow rapid, energy-efficient production (thus minimising the Cuticle exhibits an extraordinary combination of high stiffness (Fig. 1) and fracture toughness (the highest reported is 4.12 MPa m½ [7]) using ordinary biological constituents [3]. Like ⇑ Corresponding author. many protective biological materials [8–12], the cuticle is made E-mail address: [email protected] (D. Sykes). 1 Current address: Glass and Facade Technology Research Group, Department of up of fibres organised in laminae with orientations that rotate Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, UK. between successive layers. Within arthropods the following 2 Current address: Geotek Ltd, Daventry NN11 8PB, UK. arrangements have been found: helicoid [1,4,13,14], where each 3 Current address: Lyncean Technologies, Inc., 47633 Westinghouse Drive, Fre- unidirectional lamina of microfibres is rotated a fixed angle mont, CA 94539, USA. https://doi.org/10.1016/j.actbio.2019.01.031 1742-7061/Ó 2019 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. 110 D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116

Manchester; the elytra of this species were used as they are large, ﬂattened and have a pseudo-orthogonal structure. The elytra were removed with tweezers and a strip of cuticle (0.39 ± 0.02 Â 6.4 ± 0.32 mm) was dissected using a razor blade. A scalpel was used to create a 0.076 ± 0.004 mm width notch (Fig. 4a). The notched cuticle sample was then attached to aluminium tabs using epoxy glue and left to dry for two hours. The aluminium tabs holding the sample, were then attached to the nano-mechanical test stage [29] with epoxy glue (Fig. 4b).

2.2. Time-lapse nano-computed tomography (nCT) with in situ crack propagation experiment

An initial nCT scan was performed using a Zeiss Xradia 810 Ultra X-ray microscope with a chromium source. It was operated at 5.4 keV and Zernike phase contrast was achieved using a phase ring. The sample was centred and a total of 608 equi-angularly Fig. 1. Ashby plot of Young’s modulus against density for arthropod cuticle and a spaced X-ray projections were taken over a 156° rotation; a full range of engineering materials [6]. Arthropod cuticle and beetle elytra data were 180° rotation is not possible using the nano-mechanical test stage collected from the literature (see Supplementary Table 1). as it attenuates all the X-rays at some projection angles. This intro- duces some minor cupping artefacts to the data [30]. An exposure time of 15 s/projection was used giving a total acquisition time of between successive laminae; pseudo-orthogonal or ‘balken’ [15], 2.5 h. An in situ nano-mechanical tensile test was performed which is described in Fig. 2; and honeycomb and sinusoidal heli- between scans: the sample was loaded until the maximum dis- coids, which have only been found in the lobster Homarus ameri- placement of 0.5 mm, at a strain rate of 0.01 mm/s. The gauge canus [16–18] and the mantis shrimp Odontodactylus scyllarus length of the sample was 2.87 ± 0.14 mm, the sample width was [19], respectively. 0.39 ± 0.02 mm, and thickness was 0.169 ± 0.008 mm. Five load Unfortunately, our current knowledge of cuticle crack propaga- stages were scanned and volumetric data collected before failure tion is relatively poor. Two types of toughening mechanism can occurred (Fig. 3). The sample was scanned at each loading stage prevent crack growth: intrinsic ones, where energy is absorbed using 457 X-ray projections each with an exposure time of 12 s, ahead of the crack tip (e.g. plasticity in metals); and extrinsic ones, giving a total acquisition time of 1.5 h per load stage. Due to the where crack growth is inhibited by shielding mechanisms absorb- extended time for acquiring time-lapse tomographic sequences ing energy and/or reducing the local stresses and strains at the tip only one sample was studied by CT; we did however test a number (e.g. crack deflection in ceramics). To date fracture has been stud- of samples off-line and these showed broadly similar behaviour. ied in two mineralised cuticles, the stone crab Menippe mercenaria [20] and H. americanus [21]. In the former, which has a helicoid 2.3. nCT image reconstruction cuticle [22], a correlation was found between lower fracture toughness and an increased pore density [20]. For dried samples of the All volumetric data were collected with a spatial resolution of latter, brittle fracture occurred due to agglomeration of pores, 127 nm and reconstructed using filtered back projection, via the but some extrinsic toughening mechanisms were evident, i.e. crack TomoTools interface to the ASTRA reconstruction toolbox [31]. deflection along the longitudinal edges of fibres [21]. For fresh TomoTools is a Matlab based interface and is available from samples, fracture occurred by stepwise crack propagation with www.github/rsbradley/tomotools. Ring artefacts were reduced by extrinsic toughening mechanisms (delamination, crack deflection applying a median filter approach to the sinograms [32], and the and crack bridging) implied alongside intrinsic mechanisms (plas- sinograms were padded to reduce the cupping artefacts arising tic deformation) [21]. Fracture surfaces have been examined in the from the region-of-interest scanning [30]. elytra of two beetles, Allomyrina dichotoma [23] and Lucanus cervus The grey levels of the reconstructed volumetric data were [24]: in both cases evidence of intralaminar crack bridging was matched to that of the initial scan, taken with no load, by finding found. Studies of cuticle structure have led to the development of the linear transform of the data which led to the best agreement biomimetics with improved anti-peeling properties [23,25–28], (in terms of least squared difference) between the grey level cumu- likewise studying crack propagation and structure simultaneously lative density functions. This was carried out using TomoTools and could lead to biomimetics with improved toughness. the data were exported as 3D TXM files. Our aim is to characterise how cracks propagate through beetle elytra in the scarab beetle Macraspis lucida and to investigate the 2.4. Image analysis impact of its hierarchical cuticle structure, both to better understand biological design, and to inform the engineering of synthetic The scan data (TXM files) were imported into Avizo 9.0.0, where structures. In this paper a dry sample is studied which is less tough the laminae were visible in XZ plane views and manually seg- than fresh elytra but is more stable under prolonged X-ray exam- mented (Fig. 4d) every 100 slices, and then an interpolation of ination and shows many of the same micromechanisms of tough- the segmentations was performed to isolate all the plies individu- ening if with less energy absorption. ally through the whole scan volume. This was performed for each load stage. The segmented plies were exported as 3D TIFF files for 2. Materials and methods import into Drishti [33], where 3D visualisations of the individual plies can be made to show crack propagation through the individ- 2.1. Sample preparation ual plies. The angles of the fibre orientations in each lamina were measured in Avizo. All laminae were measured and the angle A dried (less than1% water content) Macraspis lucida beetle reported in relation to the load direction. Macrofibre and lamina specimen was donated by the Manchester Museum, University of sizes were also measured in Avizo. D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116 111

Fig. 2. Diagram of the hierarchical, pseudo-orthogonal structure of beetle elytra cuticle. The a-chitin molecules form long anti-parallel arrangements of a-chitin crystalline chains. These are incorporated into bundles of chitin fibres that are bound in structural proteins to form nanofibrils. The nanofibres in turn are bundled into microfibres 100 s of nanometres in diameter, which are also bundled into macrofibres (‘balken’) [15] tens of micrometres wide. These form unidirectional laminae which have a pseudo- orthogonal arrangement, whereby laminae rotate between alternate directions. The pseudo-orthogonal structure represented here provides a reference for the orientations in Figs. 4–6. In beetles, these angles vary between 30 and 90° with angles 60° being the most common [4] and runs through the exocuticle and endocuticle layers of the cuticle [1].

stage was analysed using Levene’s test [34]. This analysis was performed using the SciPy package [35] in Python 3.7.

3. Results

3.1. Computed nanotomography of cuticle structure

As a non-destructive technique, computed nanotomography (nCT) allows a sequence of 3D scans to be acquired in a time- lapse manner to image damage accumulation and deformation over time and load [29]. This provides a unique perspective on the interactions between crack growth and microstructure, allowing the toughening mechanisms to be visualised during testing. For low density materials like chitin-protein fibres, X-ray attenuation is weak giving rise to low contrast. Therefore, we used Zernike phase contrast imaging [36] that generates much stronger contrast Fig. 3. Stress-strain curve for in situ tensile test of dry elytra sample. Load stages in weakly absorbing specimens [37], and highlights interfaces such (LS) 1–5 are marked to indicate where nCT scans were collected. Load-displacement m data were collected at each load stage over a total of 53 h, with a pause for each load as those between fibres and cracks. A 65 m region of interest, local stage (to allow for the displacement to settle) which took 90 min per nCT scan. to the notch tip of the beetle elytron sample (Fig. 4c), was imaged using nCT: sufficient to capture 10 laminae of the cuticle structure (Fig. 4d). The macrofibre laminae are convex as a result of the shape of the elytron, making them difficult to visualise in ortho- 2.5. Statistical analysis slices and thus the 3D datasets required segmentation (e.g. Fig. 4d). Each lamina in the cuticle is one macrofibre thick; from All measurements were taken five times and the mean and 95% the nCT scan, the average lamina thickness is 7.37 ± 0.15 mm, com- confidence intervals calculated. The fibre orientations of the lami- prising macrofibres 7.37 ± 0.16 mm wide. In other words, the nae were measured at LS 0 and 5, then the variance for each load macrofibres have a square cross-section and form a ‘balken’ or 112 D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116

Fig. 4. a, Photo of dissected M. lucida elytron (dorsal view). b, Dissected strip of cuticle removed from elytron, notched to create a single edge notched tension (SENT) specimen, mounted in nano-mechanical testing rig. c, Optical microscope image of notched cuticle, the notch tip corresponds to Z = 0 in the nCT images. d, 3D nCT image of endocuticle at notch tip segmented into laminae labelled L1-L10, L1 being the dorsal-most lamina and L10 the most ventral. pseudo-orthogonal arrangement [15]. The orientation of laminae are required for the fibres to fracture. However, in some instances, follows an alternating pattern of +/- 30-40° with respect to loading cracks across macrofibres are deflected internally causing a zig-zag axis (+/- 80° rotation between laminae), as can be seen in Table 1. shear tear to occur (Fig. 5 – LS 4, lamina 4 and LS 5, lamina 5). This There is an increase in the rotation between more ventral laminae is likely due to the combination of lamina orientation, and the (8–10), which have orientations of À63°,+4° and À106° respec- direction of principal stress or the arrangement and properties of tively, this represents a À99°, +67° & À110° rotation between these nanofibrils and microfibres within the macrofibres. In either case laminae (Table 1). it greatly increases the fracture load required to break an individual macrofibre. 3.2. 3D time-lapse imaging of crack propagation in dry cuticle As the crack grows, evidence of delaminations within the cuticle appear, for example an instance of crack bridging between The stress-strain curve (Fig. 3) shows a linear elastic response to macrofibres is shown in lamina 5 (Fig. 5 - LS 5). Also, fibre pull- strain until crack initiation occurs at load stage (LS) 4. In the dry out can be seen (Fig. 5 – LS 4, lamina 5 and LS 5, lamina 4) implying state, very little energy is dissipated before catastrophic failure that the macrofibres and matrix have debonded to allow the crack occurs by brittle fracture at LS 5. Pre-existing cracks in the cuticle to grow along the macrofibres. This increases the crack area per- show no growth until LS 4. Based on our data, the modulus of pendicular to the principal stress so that the macrofibres must be toughness (UT, the area under the stress-strain curve prior to crack fractured and pulled out of the matrix. If the macrofibre is not bro- initiation) is 3.0 MPa. Given that the current samples are in the dry ken at the same location then it will bridge the gap (Fig. 5 –LS4, condition, our aim was not to quantify the stiffness and toughness, lamina 5), and requires more work to be pulled out of the matrix. but rather to observe the damage sequence. Using a dry sample has Fibre pull-out is a significant toughening mechanism for compos- the advantage that the condition of the sample (and therefore its ites as the elastic strain energy has to pull the fibres against fric- properties) won’t change during the experiment. tion, or shear the matrix parallel to the microfibres, while A time lapse sequence for the segmented central laminae (4, 5 propagating the crack through the matrix [39]. Another toughening and 6) showing the damage sequence during tensile straining for mechanism being exploited by the elytron appears to be the pro- the three central laminae is shown in Fig. 5, with some of the duction of a rough-edged crack. At LS 5 this is particularly evident toughening mechanisms highlighted. This clearly shows that in lamina 5 (Fig. 5), where this roughness increases the frictional cracking initially occurs along pre-existing channels between the sliding force, hence shielding the crack and slowing its growth macrofibres and along minor damage likely introduced while han- [40]. Between LS 0 and 5 the orientations of the laminae have dling the sample (Fig. 5 – LS 1 and 4, lamina 5). These lines of not significantly changed at the macroscale; rotations between weakness open slightly (240–280 nm) during the initial loading À3.4° À 3.1° occurred. However, the variance of measured angles stages 1–3. These pre-existing channels are a protein matrix that within each lamina did significantly increase in LS 5 provides an interface between the chitin-protein macrofibres with (p = 3.2 Â 10-6; Levene’s test), an increase in the standard deviation covalent bonds and/or H-bonding [38]. Within a lamina the crack from 0.92° in LS 0 to 3.92° in LS 5. This implies that delaminations grows until it encounters the chitin-protein macrofibres, where it between laminae and between macrofibres are taking place in the either travels along the interface between the fibres (Fig. 5 –LS damage zone; these rotations can also absorb energy. 4), as the path of least resistance, or connects between interfaces While Fig. 5 predominantly shows the damage and energy by perpendicular fracturing of the macrofibres (Fig. 5 – LS 4, lamina absorption mechanisms within laminae, the perpendicular cross 6). This causes the crack to be deflected away from the load direc- sections in Fig. 6 demonstrate how the damage propagates tion, reducing the stress intensity and increasing the surface area between laminae (with the tip of the notch representing 0 mm). It of the crack, thus increasing the energy required for it to propagate. confirms that the crack does not have the same shape in adjacent The pseudo-orthogonal macrofibre orientations within the cuti- laminae (Fig. 5 – LS 4 and Fig. 6 –LS5,34mm) giving rise to cle laminae cause the crack to develop in alternating directions in pull-out between laminae. Some delaminations can be observed adjacent laminae. This means that energy is absorbed by sliding between laminae at the first loading stage (Fig. 6 –LS1,10mm) and pull-out between the laminae (though this sliding energy is which appear to be caused by the creation of the notch, whilst small in the current dry case). Perpendicular macrofibre fracture some of these grow with increasing load (Fig. 6 – LS5, at 10 mm) also slows crack growth when the macrofibres are not aligned with either because of higher strain in that area or due to the pre- the load direction. This is because the axial macrofibre stress is existing damage. Deeper into the cuticle (further along the crack) smaller than the applied stress, so additional local elastic stresses delaminations are not evident (Fig. 6 – 19 and 34 mm), however, D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116 113

Table 1 Lamina orientations with respect to load direction (y axis in Fig. 4), laminae numbered as shown in Fig. 4d.

Lamina 1 2 3 4 5 6 7 8 9 10 Orientation +33° À44° +36° À36° +39° À39° +36° À63° +4° À106°

Fig. 5. 3D images (thick sections in the XY plane) of the individually segmented laminae (4–6) showing crack propagation within laminae during loading.

since fibre pull-out is occurring in these areas, some delamination that they occur due to the nature of the pseudo-orthogonal and interlaminar sliding must be occurring also. arrangement of macrofibres. As every alternate lamina has cracks There is evidence of further crack deflection in the YZ plane, by travelling in the same direction, in general, it appears that shear bifurcation of the crack (Fig. 6 –LS4,19mm) and shear tears along tears cross the intermediate lamina to join these cracks. In both macrofibres (Fig. 6 –LS4,34mm). Bifurcations occur when the the XY and YZ plane the crack can be seen to follow pre-existing crack propagates both across and along the macrofibres simultane- defects within the cuticle (Fig. 5 – LS1, lamina 5 and Fig. 6 – LS1, ously, however this is a relatively rare occurrence. The more likely 19 mm). Pre-existing defects are the result of naturally occurring outcome is that the crack is blunted by a macrofibre, forcing the damage during the beetle’s life and unavoidable damage due to crack around it and allowing the macrofibre to act as a bridge sample preparation. In both cases, our results show the ability of between laminae (Fig. 6 –LS5,at34mm). Both bifurcation and beetle elytra to withstand damage. crack blunting act to increase the surface area of the crack and thus increase the energy required for crack growth, and both result in 4. Discussion crack bridging which acts as an additional and important extrinsic toughening mechanism. Shear tears along macrofibres (i.e. Time-lapse nCT has provided a unique perspective: the interac- between laminae) (Fig. 6 –LS4,at34mm), increase the energy tion between the hierarchical structure of elytra and crack propa- required to propagate the crack across the macrofibre. It appears gation, at a resolution sufficient to see interaction at the 114 D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116

Fig. 6. Virtual (YZ) CT cross sections for different distances along the X-axis (i.e. away from the notch tip), showing crack propagation along the crack length through whole scanned region from LS 1–5 with the 10 laminae segmented using different colours (L1-10). It is noteworthy that L9 has macrofibres aligned with the loading axis and L10 has macrofibres aligned perpendicular to the loading axis. nanoscale between individual macrofibres and crack growth. This intrinsic toughening. However, dried samples are reported to have has allowed us: to observe and identify the toughening mecha- a lower fracture toughness than ‘fresh’ samples [7] and so more nisms; and to visualise the intralaminar and interlaminar damage extensive toughening may be expected in the natural state. It is mechanisms, within beetle elytra for the first time. The ability to interesting to note that we have found more evidence of active segment the curved layers of macrofibres was necessary to achieve toughening mechanisms than were present in both dry and wet these results. This study considers only the fracture of dry beetle lobster claw [21], however, a 3D time-lapse study of crack propa- elytra – recently we have developed methods to preserve the cuti- gation in a mineralised cuticle such as this would likely reveal fur- cle in a ‘fresh’ state (i.e. hydration level maintained as in vivo) over ther toughening mechanisms. It will be valuable for future studies a period long enough to capture a sequence of nano-tomographs to investigate how crack propagation is affected by hydration, as (around 24–48 h using a lab source) and work is ongoing to pro- fresh beetle elytron samples are likely to have a progressive and duce results to compare with dry cuticle. Nevertheless the current more energy-absorbing tensile failure response than dry elytra results indicate that the variety of interfaces present in cuticle [21]. Due to the time constraints associated with carrying out a across a range of scales can inhibit and frustrate crack growth. time-lapse nanoCT experiment, the results presented here are from It is clear from our results that cuticle employs many toughen- a single sample. Nevertheless we have tested other samples offline ing mechanisms that have not previously been reported, such as and achieved very similar results. We believe the results are repre- crack bifurcation, shear tearing, interlaminar crack bridging, differ- sentative of the wide range of toughening mechanisms present in ential crack deflection between laminae and rough crack surfaces. dry beetle elytra. In summary, the majority of toughening observed Although cuticle utilises both intrinsic and extrinsic toughening in this sample is extrinsic, as for ceramics. A unique aspect of mechanisms to slow crack growth (i.e. ahead of and behind the toughening in cuticle, however, is that it is applied across a crack tip), dry cuticle at least appears to have little capacity for hierarchical structure. This simultaneously allows toughening D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116 115 mechanisms to occur on a number of laminae, and therefore late that one potential benefit of a pseudo-orthogonal structure increase their effectiveness. For example, here, crack deflection over an orthogonal arrangement is that the differently orientated occurs in all laminae, but the deflections occur in different direc- lamina (L8-10) could shield deeper layers from surface damage. tions depending upon the orientation of the macrofibres within We found that many extrinsic toughening mechanisms are used each. This greatly increases the volume of the damage zone by cuticle, which are intimately linked to its structural organisa- thereby increasing the energy dissipated, and slows crack growth tion, however a largely brittle fracture still occurred. It is likely that - more so than if the fibres had the same oblique orientation with this is due to dehydration causing the interfaces between fibres at respect to the crack direction. The hierarchical arrangement of every scale of the hierarchical structure to become weaker. Further fibres also allows delaminations and pull-out to happen along study of the effect that different microstructures, compositions and many different interfaces within the material. They can occur both conditions of cuticle have upon its toughening mechanisms could within laminae (between macrofibres and surrounding matrix) or provide many lessons in biomimetic design of new materials. between laminae; this increases the energy dissipated and spreads Another important reason for understanding why arthropods use the damage away from the principal stress direction. We also different cuticle microstructures and how these are advantageous, report unexpected variations in the pseudo-orthogonal laminate is that it could aid our understanding of why arthropods are the lay-up structure. In contrast to the regular ± 80° misorientation most successful group of animals on Earth. between successive lamina observed for laminates 1–8, a quite different misorientation was observed for lamina 9 and 10 (those Acknowledgements nearest to the interior). We speculate that these laminae may act to contain damage within different zones of the cuticle by prevent- We would like to thank Richard Preziosi and the Manchester ing cracks spreading from the outer surface to deeper lamina Museum for guidance and donation of beetles for this study. We within the cuticle. The crack path and toughening mechanisms are very grateful for the helpful comments of the anonymous found visually, imply a brittle-ductile laminate structure, however reviewers. Also we would like to acknowledge the assistance of the stress-strain curve does not reflect this and appears to imply a the Manchester X-ray Imaging Facility, which was funded in part brittle structure. Ex situ loading and unloading tests of elytra found by the EPSRC (grants EP/F007906/1, EP/F001452/1 and some evidence of stress relaxation and plastic deformation in their EP/I02249X/1). stress-strain curves, however it may be that in dried samples the interfaces are too weak to absorb much energy, resulting in a brittle fracture. These interfaces need to be reasonably weak to provide Author contributions the best combination of properties in the composite material, however it would be expected that in a hydrated sample the interfaces P.J.W. and R.H. conceived the study. R.H. and R.S.B. planned and would be more elastic [7,41] therefore increasing both the energy performed the nCT experiment. R.S.B. reconstructed the nCT data. they absorb and that is required for delaminations to occur. This D.S. analysed the data, interpreted the results and wrote the manu- would result in a more ductile fracture and a higher fracture tough- script. All the authors discussed the results and commented on the ness in the cuticle [3,21]. manuscript.

5. Conclusions Appendix A. Supplementary data

This study found that dry M. lucida elytra has a high modulus of Supplementary data to this article can be found online at toughness relative to most other arthropod cuticles. This implies https://doi.org/10.1016/j.actbio.2019.01.031. that these beetle elytra are adapted to tolerate high impact loads [3], and they also have high punch strength [42]. These properties References are advantageous for protecting the delicate wings underneath the elytra from both accidental damage and predators. However, this is [1] A.C. Neville, Biology of the arthropod cuticle, Springer-Verlag, 1975. for a dried elytron, cuticle has been found to be tougher when [2] M.F. Ashby, L.J. Gibson, U. Wegst, R. Olive, The Mechanical Properties of Natural Materials. I. Material Property Charts, Proc. R. Soc. Math. Phys. Eng. Sci. 450 more hydrated [3]. For meaningful comparisons of fracture proper- (1995) 123–140, https://doi.org/10.1098/rspa.1995.0075. ties in relation to cuticle structure, composition and condition, [3] J.F.V. Vincent, U.G.K. Wegst, Design and mechanical properties of insect cuticle, more cuticles need to be measured, ideally in a fresh state. Never- Arthropod Struct. Dev. 33 (2004) 187–199, https://doi.org/10.1016/j. asd.2004.05.006. theless, this study highlights the effectiveness of a hierarchical [4] T. van de Kamp, A. Riedel, H. Greven, Micromorphology of the elytral cuticle of laminated structure for producing tough materials even when beetles, with an emphasis on weevils (Coleoptera: Curculionoidea), Arthropod the biomaterial itself is produced from simple constituents. These Struct. Dev. 45 (2016) 14–22, https://doi.org/10.1016/j.asd.2015.10.002. [5] Z. Dai, Z. Yang, Macro-/Micro-Structures of Elytra, Mechanical Properties of the hierarchical structures are used in many protective biomaterials Biomaterial and the Coupling Strength Between Elytra in Beetles, J. Bionic Eng. like fish scales [8,9], mollusc shells [10,11], and plant cuticle 7 (2010) 6–12, https://doi.org/10.1016/S1672-6529(09)60187-6. [12]. This implies a general benefit for toughness in using hierar- [6] M. Ashby, D. Cebon, Materials selection in mechanical design, J. Phys. IV Colloq. 3 (1993) 1–9, https://doi.org/10.1051/jp4:1993701. chical organisation, and in the case of arthropod cuticle, we have [7] J.-H. Dirks, D. Taylor, Fracture toughness of locust cuticle, J. Exp. Biol. 215 found many complimentary toughening mechanisms across this (2012) 1502–1508, https://doi.org/10.1242/jeb.068221. structure. It seems that arthropods benefit from an ability to pro- [8] A. Bigi, M. Burghammer, R. Falconi, M.H.J. Koch, S. Panzavolta, C. Riekel, duce tough cuticle for defence quickly and at low energetic cost, Twisted Plywood Pattern of Collagen Fibrils in Teleost Scales: An X-ray Diffraction Investigation, J. Struct. Biol. 136 (2001) 137–143, https://doi.org/ as it can be assembled at room temperature in less than 24 h [1]. 10.1006/jsbi.2001.4426. Cuticle has one of the highest reported fracture toughnesses of [9] M.M. Giraud, J. Castanet, F.J. Meunier, Y. Bouligand, The fibrous structure of any biomaterial [7] and yet until now the full range of crack shield- coelacanth scales: A twisted ‘Plywood’, Tissue Cell. 10 (1978) 671–686, https:// doi.org/10.1016/0040-8166(78)90054-X. ing mechanisms and their interactions during crack propagation [10] S. Kamat, X. Su, R. Ballarini, A.H. Heuer, Structural basis for the fracture had not been captured. Our study has found that misorientations toughness of the shell of the conch Strombus gigas, Nature 405 (2000) 1036– between laminae of unidirectional fibres has a large impact on 1040, https://doi.org/10.1038/35016535. [11] A.P. Jackson, J.F.V. Vincent, R.M. Turner, The Mechanical Design of Nacre, Proc. the direction of crack growth and could be a useful design to apply R. Soc. Lond. B Biol. Sci. 234 (1988) 415, https://doi.org/10.1098/ to the production of tougher carbon-fibre composites. We specu- rspb.1988.0056. 116 D. Sykes et al. / Acta Biomaterialia 86 (2019) 109–116

[12] D. Reis, B. Vian, J.-C. Roland, Cellulose-glucuronoxylans and plant cell [26] J.X. Chen, Biomimetic research on beetle forewings in twenty years: Internal wallstructure, Micron. 25 (1994) 171–187, https://doi.org/10.1016/0968- structure, model and integrated honeycomb plates, Sci. Sin. Technol. 48 (2018) 4328(94)90041-8. 701–718. [13] Y. Bouligand, Twisted fibrous arrangements in biological materials and [27] B. Chen, X. Peng, J. Fan, Fiber-reinforced characteristics of chafer cuticle and cholesteric mesophases, Tissue Cell. 4 (1972) 189–217, https://doi.org/ research on structure of branched fiber, Chin. J. Mater. Res. 17 (2003) 630–636. 10.1016/S0040-8166(72)80042-9. [28] J.X. Chen, Q.Q. Ni, J. Xie, The Lightweight Composite Structure and Mechanical [14] H.-O. Fabritius, A. Ziegler, M. Friák, S. Nikolov, J. Huber, B.H.M. Seidl, S. Properties of the Beetle Forewing, in: N. Hu (Ed.), Compos. Their Prop., InTech, Ruangchai, F.I. Alagboso, S. Karsten, J. Lu, A.M. Janus, M. Petrov, Li-Fang Zhu, P. 2012. doi:10.5772/48807. Hemzalová, S. Hild, D. Raabe, J. Neugebauer, Functional adaptation of [29] B.M. Patterson, N.L. Cordes, K. Henderson, J.C.E. Mertens, A.J. Clarke, B. crustacean exoskeletal elements through structural and compositional Hornberger, A. Merkle, S. Etchin, A. Tkachuk, M. Leibowitz, D. Trapp, W. Qiu, B. diversity: a combined experimental and theoretical study, Bioinspir. Biomim. Zhang, H. Bale, X. Lu, R. Hartwell, P.J. Withers, R.S. Bradley, In Situ Laboratory- 11 (2016) 055006. doi:10.1088/1748-3190/11/5/055006. Based Transmission X-Ray Microscopy and Tomography of Material [15] T. van de Kamp, H. Greven, On the Architecture of Beetle Elytra, Entomol. Deformation at the Nanoscale, Exp. Mech. (2016), https://doi.org/10.1007/ Heute. 22 (2010) 191–204. s11340-016-0197-3. [16] D. Raabe, P. Romano, C. Sachs, A. Al-Sawalmih, H.-G. Brokmeier, S.-B. Yi, G. [30] A. Kyrieleis, V. Titarenko, M. Ibison, T. Connolley, P.J. Withers, Region-of- Servos, H.G. Hartwig, Discovery of a honeycomb structure in the twisted interest tomography using filtered backprojection: assessing the practical plywood patterns of fibrous biological nanocomposite tissue, J. Cryst. Growth. limits, J. Microsc. 241 (2011) 69–82, https://doi.org/10.1111/j.1365- 283 (2005) 1–7, https://doi.org/10.1016/j.jcrysgro.2005.05.077. 2818.2010.03408.x. [17] P. Romano, H. Fabritius, D. Raabe, The exoskeleton of the lobster Homarus [31] W. van Aarle, W.J. Palenstijn, J. De Beenhouwer, T. Altantzis, S. Bals, K.J. americanus as an example of a smart anisotropic biological material, Acta Batenburg, J. Sijbers, The ASTRA Toolbox: A platform for advanced algorithm Biomater. 3 (2007) 301–309, https://doi.org/10.1016/j.actbio.2006.10.003. development in electron tomography, Ultramicroscopy 157 (2015) 35–47, [18] H.-O. Fabritius, C. Sachs, P.R. Triguero, D. Raabe, Influence of Structural https://doi.org/10.1016/j.ultramic.2015.05.002. Principles on the Mechanics of a Biological Fiber-Based Composite Material [32] R.A. Ketcham, New algorithms for ring artifact removal, in: U. Bonse (Ed.), Proc. with Hierarchical Organization: The Exoskeleton of the Lobster Homarus SPIE, 2006: pp. 1–7. doi:10.1117/12.680939 americanus, Adv. Mater. 21 (2009) 391–400, https://doi.org/10.1002/ [33] A. Limaye, Drishti: a volume exploration and presentation tool, in: S.R. Stock adma.200801219. (Ed.), 2012: p. 85060X. doi:10.1117/12.935640. [19] N.A. Yaraghi, N. Guarín-Zapata, L.K. Grunenfelder, E. Hintsala, S. Bhowmick, J. [34] H. Levene, Robust tests for equality of variances, in: Contrib. Probab. Stat. M. Hiller, M. Betts, E.L. Principe, J.-Y. Jung, L. Sheppard, R. Wuhrer, J. McKittrick, Essays Honor Harold Hotell., Stanford University Press, 1960: pp. 278–292. P.D. Zavattieri, D. Kisailus, A Sinusoidally Architected Helicoidal Biocomposite, [35] E. Jones, T. Oliphant, P. Peterson, SciPy: open source scientific tools for, Python Adv. Mater. (2016) 1–10, https://doi.org/10.1002/adma.201600786. (2001). [20] C.A. Melnick, Z. Chen, J.J. Mecholsky, Hardness and toughness of exoskeleton [36] P.J. Withers, X-ray nanotomography, Mater. Today. 10 (2007) 26–34. material in the stone crab, Menippe mercenaria, J. Mater. Res. 11 (1996) 2903– [37] R.S. Bradley, P.J. Withers, Correlative multiscale tomography of biological 2907, https://doi.org/10.1557/JMR.1996.0367. materials, MRS Bull. 41 (2016) 549–556, https://doi.org/10.1557/ [21] C. Sachs, H. Fabritius, D. Raabe, Experimental investigation of the elastic– mrs.2016.137. plastic deformation of mineralized lobster cuticle by digital image correlation, [38] J.F.V. Vincent, Arthropod cuticle: a natural composite shell system, Compos. J. Struct. Biol. 155 (2006) 409–425, https://doi.org/10.1016/j.jsb.2006.06.004. Part Appl. Sci. Manuf. 33 (2002) 1311–1315. [22] G.F. Warner, The biology of crabs, Elek, 1977. [39] A. Kelly, Strong solids, 2nd edn. Clarendon, Oxford, 1973. [23] J.X. Chen, M. Iwamoto, Q.Q. Ni, K. Kurashiki, K. Saito, Laminated Structure and [40] R. Ballarini, M.E. Plesha, The effects of crack surface friction and roughness on Its Mechanical Properties of the Fore-Wing of Beetle, J. Soc. Mater. Sci. Jpn. 50 crack tip stress fields, Int. J. Fract. 34 (1987) 195–207, https://doi.org/10.1007/ (2001) 455–460. BF00019717. [24] L. Kundanati, S. Signetti, H.S. Gupta, M. Menegon, N.M. Pugno, Multilayer stag [41] D. Klocke, H. Schmitz, Water as a major modulator of the mechanical beetle elytra perform better under external loading via non-symmetric properties of insect cuticle, Acta Biomater. 7 (2011) 2935–2942, https://doi. bending properties, J. R. Soc. Interface. 15 (2018) 1–13, https://doi.org/ org/10.1016/j.actbio.2011.04.004. 10.1098/rsif.2018.0427. [42] A.R. Evans, G.D. Sanson, Biomechanical properties of insects in relation to [25] J.X. Chen, Q. Zu, G. Wu, J. Xie, W. Tuo, Review of beetle forewing structures and insectivory: cuticle thickness as an indicator of insect ‘‘hardness” and their biomimetic applications in China: (II) On the three-dimensional ‘‘intractability”, Aust. J. Zool. 53 (2005) 9, https://doi.org/10.1071/ZO04018. structure, modeling and imitation, Mater. Sci. Eng. C 55 (2015) 620–633, https://doi.org/10.1016/j.msec.2015.04.045.

PAPER 3

Preservation of mechanical properties in locust tibiae for in situ time-lapse three-dimensional imaging

Paper 3

Preservation of mechanical properties in locust tibiae for in situ time- lapse three-dimensional imaging

Dan Sykesa, Russell J. Garwoodb,c, Shelley D. Rawsona and Philip J. Withersa a Henry Moseley X-ray Imaging Facility, The Royce Institute, School of Natural Sciences, The University of Manchester, Manchester M13 9PL, UK b School of Earth and Environmental Science, The University of Manchester, Manchester M13 9PL, UK c Earth Sciences Department, Natural History Museum, London SW7 5BD, UK Abstract

The development of mechanical tests that run over a long time scale is at odds with the necessity of preserving arthropod cuticle hydration and therefore its mechanical properties. Time-lapse X-ray CT with in situ mechanical testing enables three-dimensional imaging of damage progression over time. However, preserving arthropod cuticle hydration is necessary to maintain its mechanical properties, and the effects of X-ray irradiation on cuticle mechanical properties is unknown.. Here, we develop a method to preserve hydration by coating samples with white petrolatum (petroleum jelly) and test the impact of this method, and exposure to X-ray radiation, on the mechanical properties of locust tibiae in 3-point bending. We then apply this methodology to a time-lapse in situ 3-point bend with X-ray µCT imaging of both a petroleum jelly coated and a desiccated tibia to investigate the effect of desiccation on their failure mechanisms. Petroleum jelly was an effective coating to preserve cuticle hydration and mechanical properties. In addition, X-ray irradiation did not significantly affect the Young’s modulus and flexural strength of cuticle. In situ time-lapse imaging revealed that the failure of a tibia coated with petroleum jelly occurred predominantly by buckling in the compressive stress region with limited crack propagation in the tensile stress region, where there was evidence of extrinsic toughening mechanisms. In contrast, desiccated cuticle failed due to catastrophic crack propagation throughout the tibia with only limited buckling under compressive stress. Combining petroleum jelly coating with time-lapse X-ray CT paves the way for further studies of the mechanical behaviours of biological materials, sensitive to hydration levels, in a three-dimensional, time-lapse manner, which could advance our understanding of the biomechanics of arthropod cuticles. Introduction

Arthropod cuticle is a hierarchically-organised biological material which possesses diverse mechanical properties, whilst maintaining a low density. Therefore, there is a lot of interest in biomimicry of cuticle to develop advanced composites with the impressive combination of stiffness, hardness and toughness that cuticles possess. In the past 15 years, this interest has coincided with a rapid advance in our ability to investigate this unique material, with the use of more advanced modelling (Nikolov et al., 2010; Weaver et al., 2012; Hörnschemeyer, Bond & Young, 2013; Goyens et al., 2014; Kundanati et al., 2018), three-dimensional non-invasive imaging (Dirks & Taylor, 2012a; van der Meijden, Kleinteich & Coelho, 2012; Goyens et al., 2014; Aberle, Jemmali & Dirks, 2017; Das et al., 2018; Sykes et al., 2019) and more sensitive mechanical testing equipment (Dirks & Taylor, 2012a,b; Dirks, Parle & Taylor, 2013; Bennemann et al., 2014; Aberle, Jemmali & Dirks, 2017; Das et al., 2018). Our understanding of cuticle has also been vastly improved by studies demonstrating that dehydration significantly increases its Young’s modulus, strength and hardness and decreases its

Paper 3 fracture toughness (Schöberl & Jäger, 2006; Klocke & Schmitz, 2011; Dirks & Taylor, 2012a). Therefore, dehydration needs to be taken into account for any mechanical test. Dehydration starts immediately upon dissection and progresses rapidly in cuticle; it can take as little as three hours for a dissected locust tibia to become completely dry (Dirks & Taylor, 2012a). For most mechanical tests, the solution to this problem has been the test cuticle samples immediately after euthanasia and excision (fresh), as most mechanical tests can be completed before major changes in hydration levels affect mechanical properties. However, some experiments, such as fatigue tests, have long durations during which the cuticle will desiccate and its properties change. Currently, the only available solution is to immerse the sample in water for the duration of the experiment (Dirks, Parle & Taylor, 2013; Aberle, Jemmali & Dirks, 2017). Unfortunately this could increase hydration levels above the fresh state and could potentially alter the mechanical properties, as was found in some cases of long-term submersion of locust tibia in water (Aberle, Jemmali & Dirks, 2017). In addition, water immersion is not possible for all long-duration mechanical tests. For example, a newly developed technique, time-lapse three-dimensional X-ray nano-computed tomography (nCT) with in situ mechanical testing would be challenging to conduct in water (Patterson et al., 2016; Sykes et al., 2019).

Time-lapse CT involves collecting X-ray projections, over a 360° rotation, to produce three- dimensional tomographic images of a sample, then collecting repeat scans of the sample at different stages of an in situ mechanical test. It can achieve sufficient resolution to image macrofibres within the cuticle microstructure and observe their response to mechanical testing (Sykes et al., 2019). A single time-lapse nCT experiment can run for more than 12 hours (Sykes et al., 2019), much more time than is required to desiccate the cuticle; however, keeping samples submerged in water would reduce X-ray contrast (Metscher, 2009), potentially alter mechanical properties (Klocke & Schmitz, 2011; Aberle, Jemmali & Dirks, 2017) and increase the risk of sample movement during rotation. As such, there is a need to develop methods to preserve the mechanical properties of cuticle for longer periods of time to allow these long-duration experiments to be performed. An additional challenge with time- lapse nCT is the unknown effect of exposure to X-ray radiation on cuticle. Laboratory source µCT and nCT produce a relatively low dose of radiation in comparison to synchrotron sources, but X-ray irradiation has been shown to affect the mechanical properties of other biological materials, such as human cortical bone (Barth et al., 2010, 2011), so the potential effects must be considered.

Locust tibiae cuticle is of particular interest due to its impressive energy storage capabilities. Locusts use a catapult jump mechanism to reach heights of up to 250 mm and distances of up to 1 m, with take-off velocities of around 3 m/s. This extraordinary jump capability requires the storage of elastic energy within the tibiae (Bennet-Clark, 1975; Bayley, Sutton & Burrows, 2012) and the semi-lunar processes of the femorotibial joint (Bennet-Clark, 1975; Wan, Hao & Feng, 2016). Jumping exposes tibiae to high stress, yet their safety factor is reported to be very low, only 1.7 (Parle, Larmon & Taylor, 2016). Failure has been found to occur by buckling under compressive stress for a number of different insect tibia, including locusts (Parle et al., 2016). However, the differences in mechanical behaviour and progression of buckling damage between hydrated (to in vivo levels) and desiccated locust tibae have not been studied in a time-lapse manner.

To investigate methods to preserve the mechanical properties of cuticle, we performed 3-point bend tests on adult locust (Schistocercia gregaria) tibiae after being treated with different sample preparation techniques. White petrolatum also known as petroleum jelly (PJ) is a commonly used moisture barrier, which we tested as a method to coat samples and preserve their hydration. Also, freezing samples as a storage method prior to testing has recently been found to have no significant

Paper 3 impact on the mechanical properties of cuticle (Aberle, Jemmali & Dirks, 2017). Therefore, we tested this hydration preservation method on frozen samples, to check for any interacting effects.

The aims of this study are: to test petroleum jelly as a hydration preservation method for arthropod cuticle and test its effect on the mechanical properties of cuticle; to test the effect of X-ray CT radiation on the mechanical properties of cuticle; and to apply the method to locust tibiae to visualise damage and buckling in 3D in both a dry and fresh condition. Materials and methods

Hydration preservation

Desert locusts (S. gregaria) were bought as mature adults and immediately euthanized. To determine the effectiveness of petroleum jelly to preserve cuticle hydration levels, whole adult locusts were split into two groups of 15: a control group were left to air dry at room temperature over the course of the experiment; and the other group were coated in a thin layer of petroleum jelly (Vaseline), then left exposed to air also. Samples were weighed immediately on a KERN PCB (Germany) digital balance (0.001g level of accuracy), every hour for the subsequent six hours, and then every 24 hours for 144 hours to observe changes in weight over the length of a maximum experiment run-time.

Effect of sample preparation on mechanical properties

For all sample preparations, the hind tibiae (jumping legs) were dissected from the locusts by cutting with a razor blade 1 mm below the femorotibial joint and 1 mm above the tibiotarsal joint, leaving a hollow tube. For each individual S. gregaria, the left tibia was used for a different sample preparation than the right, to allow pairwise statistical testing which reduces the impact of large natural variation on the results. The different sample preparations used in this study are described in table 1. To examine if freezing, petroleum jelly coating or exposure to X-ray CT radiation affected the mechanical properties of cuticle, five different paired tests were performed: Fresh vs Frozen, Fresh vs Fresh+PJ, Frozen vs Frozen+PJ, Fresh+PJ vs Fresh+Irr, and Frozen+PJ vs Frozen+Irr. µCT scans were used instead of nCT (which were used in (Sykes et al., 2019)) as µCT transmits a significantly larger dose of X-ray radiation to an entire tibia.

Paper 3

Sample Fresh Frozen Fresh+PJ Frozen+PJ Fresh+Irr Frozen+Irr preparation (PJ and X-ray irradiation) (PJ and X-ray irradiation) Method  Dissected  Fresh tibia  Fresh tibia coated  Frozen tibia  Fresh+PJ tibia  Frozen+PJ tibia tibia tested frozen upon in petroleum jelly coated in exposed to 2 hours of exposed to 2 hours of immediately dissection at  Stored at room petroleum jelly X-ray µCT radiation X-ray µCT radiation -18°C temperature for after thawing  Stored at - 24 hours  Stored at room 18°C for 24 temperature for a hours before further 24 hours thawing Table 1. Sample preparation methods used for all the paired groups, with descriptions of each. Samples were exposed to X-ray µCT radiation in a Zeiss Xradia 520 Versa X-ray microscope (Carl Zeiss X-ray Microscopy, Pleasanton, California, United States) with an X-ray beam of 140 kV and 5 W with no X-ray filter. The µCT scans were composed of 2,000 X-ray projections, each was collected through a 0.39x objective lens with a one second X-ray exposure.

Paper 3

All tibiae were tested on an Instron 3344 rig (Instron, Norwood, Massachusetts, United States), with a 10 N load cell, in a 3-point bend set up (fig. 1a). The samples were loaded in the medial-lateral direction, to avoid interference from the ventral spines of the tibia. All 3-point bend tests were performed at a strain rate of 1 mm/min over a 10 mm span. To calculate mechanical properties from the 3-point bend tests, an additional µCT scan was performed on a static tibia to calculate the average moment of inertia (fig. 1b). Locust tibiae were glued upright to a pin with superglue then mounted within a Zeiss Xradia 520 Versa X-ray microscope. Scans were operated at 60 kV and 5 W with a 6.3 µm voxel size, over a 360° rotation 1,600 X-ray projections were collected with a 5 second exposure each. The reconstructed scans were segmented by thresholding in Avizo 9.0. The segmented legs were imported into FIJI where the minimum moment of inertia (Imin) i.e. in the test direction (medial-lateral) was calculated for 607 slices, which were located in the middle region of the tibia and where no ventral spines were present, using the BoneJ plugin (Doube et al., 2010). Young’s modulus was calculated using the force F, the deflection d¸ the span (the spacing between the supports) S, and the moment of inertia I, thus (Aberle, Jemmali & Dirks, 2017): 퐹 ( ) . 푆3 퐸 = 푑 48퐼 Flexural strength (maximum stress) was also calculated using the mean radius (r), thus: 퐹푆푟 휎 = 8퐼

Figure 1. A) 3-point bend test configuration, with dissected locust tibia. Deflection (d) and span (S) are indicated. B) cross-section from µCT scan of a tibia with the Imax and Imin indicated.

In situ µCT scans of locust tibiae buckling behaviour

A desiccated and a fresh, petroleum jelly coated S. gregaria tibia had a 3-point bend test applied in situ with a series of time-lapse three-dimensional µCT scans collected at different stages. A customised Perspex rig was used to apply a 3-point bend to the tibiae in situ within a Zeiss Xradia 520 Versa X-ray microscope (Carl Zeiss X-ray Microscopy, Pleasanton, California, United States). Scans were operated at 80 kV and 7 W with a 2.25 µm voxel size and 4x objective lens, and no filter was used. 1,600 X-ray projections were collected over a 360° rotation with a 10 second exposure

Paper 3 each. The reconstructed scan data were reconstructed with Zeiss Scout and Scan software and then imported into FIJI where a non-local means filter was applied to reduce noise. The data were then segmented using the 3D region growing tool in Avizo 9.0.0 (ThermoFisher Scientific, Waltham, Massachusetts, United States). Images of 2D cross-sections in different orthogonal planes were produced using FIJI and Drishti 2.6.5 (Limaye, 2012) was used to generate 3D visualisations.

Statistical analysis

The preservation of sample hydration by PJ coating was analysed with least squares linear regression models using a custom script in Python 3.7, that utilises the SciPy (Jones, Oliphant & Peterson, 2001), NumPy (Oliphant, 2006) and Statsmodels (Seabold & Perktold, 2010) modules. The treatments (control and PJ coated) were considered to be significantly different from a slope of 0 (no change in mass) when P < 0.05. Another custom Python script was used to statistically compare Young’s modulus and flexural strength between paired groups of tibia. Shapiro-Wilk normality tests were performed on all data. Where data were normally distributed a paired t-test was performed, otherwise a Wilcoxon paired test was performed. Groups were considered to be significantly different when P < 0.05. All the plots were generated using the Matplotlib module (Hunter, 2007) in Python. All values are reported with 95% confidence intervals. Results

Hydration preservation

Whole locusts left to air dry were observed to lose mass at a mean rate of 0.0028 ± 0.0010 g/hour, equivalent to a 0.192 ± 0.003 % loss per hour. In contrast, locusts coated in petroleum jelly were observed to lose mass at only 0.0008 ± 0.0001 g/hour. When the linear regression models of these two conditions are compared to a slope of 0, equivalent to no reduction in mass via desiccation, the dried samples exhibited a significant change in mass (P < 0.001). However, no significant rate of loss in mass was found for PJ coated samples (P = 0.127). The reduction in mass that did occur in PJ coated samples over 144 hours (mean 9.2 ± 1.7 %) started after 24 hours and coincided with the onset of blackening of the internal soft tissues, indicating that decomposition had begun.

Effect of sample preparation on mechanical properties

4 From the µCT scan of a static tibia, an average Imin of 0.020 ± 0.0008 mm and Imax 0.033 ± 0.0007 4 mm were calculated. These values are similar to those found in previous studies: Imax ranging 4 4 between 0.027 mm (Aberle, Jemmali & Dirks, 2017) and 0.041 mm (Dirks & Taylor, 2012a), and Imin values ranging between 0.015 mm4 (Aberle, Jemmali & Dirks, 2017) and 0.016 mm4 (Dirks & Taylor,

2012a). The Imin was used to calculate the mean Young’s modulus and flexural strength for each treatment (fig. 2). Figure 2 shows a high degree of natural variation in the stiffness (fig. 2a) and flexural strength (fig. 2b) of all the tibiae but no statistically significant difference was found between any of the paired groups.

Paper 3

Figure 2. Bar charts comparing the A) mean Young’s modulus and B) mean flexural strength for each paired group. Error bars indicate 95% confidence intervals.

To further compare the variation in results of different sample preparations, the stress-strain curves for each tibia were averaged within its treatment group, and then plotted alongside its paired group (fig. 3). From figure 3 it can be seen that the stress-strain responses of each paired group were highly similar but that the 95% confidence intervals (representing the natural variation) are very wide. Figure 2 and figure 3a highlight that the least similarity occurs between the fresh and frozen groups, whilst X-ray irradiation appears to have the most similar response (least effect) between treatment groups (fig. 3d and 3e). However, any differences observable are not statistically significant (fig. 2).

100

Paper 3

Figure 3. Average stress-strain curves for each treatment plotted alongside its corresponding paired group: A) Fresh vs Frozen, B) Fresh vs Fresh+PJ, C) Frozen vs Frozen+PJ, D) Fresh+PJ vs Fresh+Irr and E) Frozen+PJ vs Frozen+Irr. 95% confidence intervals for every average stress are represented as dashed lines.

101

Paper 3

In situ µCT 3-point bend tests

The impact of desiccation on mechanical properties of locust tibiae is known (Dirks & Taylor, 2012a) and therefore desiccated tibia were not included in the above comparisons, however the effect of desiccation on damage progression in tibia is unknown. The damage progression from an in situ 3- point bend test of a fresh, petroleum jelly coated tibia is shown in figure 4. At 1.6 mm displacement, it can be seen that all the damage occurring is from buckling under the pointer (fig. 4e), which is assumed to be the region of peak compressive stress (fig. 4c). At 2.3 mm displacement, there is a significant increase in buckling damage (fig. 4e and f) and a lateral crack starts to form in the opposite side (fig. 4g), which is assumed to be the region of peak tensile stress (fig. 4c), spreading towards the centre (fig. 4d). The crack penetrates only 45 µm of the ~100 µm thickness of the tibia wall (fig. 4g). At 4.0 mm displacement, the highest level of buckling in this experiment can be seen (fig. 4b and e), producing an almost completely flattened cross-section (fig. 4f). Also, the crack now crosses the entire thickness of the tibia, but only 347 µm of the 1140 µm width (30.4%) in that region (fig. 4d). The crack propagates in a zig-zag manner (fig. c and d) which suggests extrinsic toughening mechanisms are affecting crack propagation. Furthermore, two more cracks have begun to propagate in the tensile stress region parallel to the first crack (fig. 4d).

102

Paper 3

Figure 4. Time-lapse in situ µCT images of a fresh, petroleum jelly coated locust tibia at different levels of displacement. A) 3D render of the lateral view prior to the in situ 3-point bend test, dashed lines indicate the cross-sectional series E), F) and G). B) 3D render of the lateral view with 4 mm displacement showing the region of compressive stress, dashed lines indicate the cross-sectional series E), F) and G). C) 3D render of the dorsal view at 4 mm displacement with the different regions of stress and 3D zig-zag crack indicated. D) 3D render of the lateral view with 4 mm displacement showing the region of tensile stress. E) Longitudinal cross-sections showing progression of buckling damage in the compressive stress region and crack formation in the tensile stress region. F) Lateral cross-sections showing buckling in the area of greatest deformation. G) Longitudinal cross-sections showing crack propagation in the tensile stress region. 103

Paper 3

The damage sustained by the dried tibia involves a small amount of buckling in the compressive stress region, compared to the fresh tibia, also it produces transverse, longitudinal and lateral cracks in this region of the cuticle (fig. 5e, f and g). In the tensile stress region, many cracks spread both transversely and laterally (fig. 5c). Between these two regions two large longitudinal cracks grow along the length of the tibia (fig. 5d). The desiccated tibia suffered catastrophic failure at 3.2 mm displacement due to longitudinal, transverse and lateral crack propagation spanning across the entire width and height of the tibia, as well as a significant portion of its length (fig. 5a and b). The cracks are relatively straight compared to the ones that form in the fresh tibia (fig. 4x and 5a and b), indicating that few extrinsic toughening mechanisms are present.

104

Paper 3

Figure 5. µCT images of dry locust tibia, at 3.2 mm displacement, the point of catastrophic failure. A) 3D render of the lateral view with cross-sections C) and D) indicated as dashed lines. B) 3D render of the dorsal view with cross-sections E), F) and G) indicated as dashed lines. C) Cross-section in longitudinal plane showing minimal buckling and transverse and lateral cracks. D) Lateral cross- section with longitudinal cracks forming halfway between tensile and compressive stress regions marked. E and F) Transverse longitudinal cross-sections highlighting the presence of transverse and lateral cracks in the compressive stress region. G) 3D cross- section in the transverse longitudinal plane showing minimal buckling and the presence of transverse, longitudinal and lateral cracks in the compressive stress region. 105

Paper 3

Discussion

Efficacy of petroleum jelly coating method

For the duration of the experiment, 144 hours i.e. the longest duration that an in situ nCT experiment has been run on arthropod cuticle (Sykes et al., 2019), petroleum jelly coating reduced the rate of water loss in arthropod cuticle to a level that is not significantly different from a rate of zero water loss. However, after 144 hours the mass of PJ coated locusts had fallen by an average of 9.2 ± 1.7 %, which is sufficient for a change in mechanical properties to occur (Dirks & Taylor, 2012a). We propose two possible mechanisms for this reduction despite a non-volatile moisture barrier coating the surface of the locusts. Firstly, locusts have a complex topography that is difficult to ensure it is completely covered in PJ, therefore some areas may be exposed and drying. This could explain the wide variation in total percent mass loss after 144 hours between samples, 3.8 – 16.0% and 21.8 – 39.6% in the PJ coated and air dried samples respectively. Secondly, whole locusts possess a relatively large amount of soft tissues, which are vulnerable to rapid decomposition, and visible decomposition of the internal tissues was evident in some samples. However for mechanical testing, exoskeletal material is used exclusively where material degradation is much slower than soft tissues.

The Young’s modulus and flexural strength of fresh tibia (2.92 GPa and 81.3 MPa respectively) are similar to previously reported values of 3.05 GPa and 72.1 MPa respectively (Dirks & Taylor, 2012a). There is large natural variation in these properties amongst all fresh tibia tested here. These properties for PJ-coated tibiae had no significant difference with either fresh or frozen tibiae, providing further evidence of the efficacy of PJ as a hydration control method for mechanical testing of cuticle. The geometry of tibiae are much simpler than a whole locust, it is therefore likely that PJ provided a more effective coating than for whole locusts, as shown by the very similar stress-strain responses of PJ-coated groups to their paired groups (fig. 3b and c). For comparison, dry locust tibae have been found to have a significantly higher Young’s modulus (6.2 GPa) and flexural strength (217.4 MPa) than reported values for fresh cuticle and the values reported here (Dirks & Taylor, 2012a).

Effect of freezing and X-ray irradiation on mechanical properties

We found that the mechanical properties of frozen samples do not significantly change from the fresh state, as previously found by Aberle, Jemmali & Dirks (2017). Furthermore, we found no interaction between freezing and other treatments, showing that this sample preservation technique can be combined with the PJ-coating method developed here without sacrificing the natural behaviour of these cuticles.

We also found no significant change in mechanical properties resulting from X-ray µCT irradiation. This is likely to be due to the fact that the X-ray energy or dose is too low to cause significant damage to the chitin-protein fibres of cuticle. However, different sources can produce X-rays of higher energies and doses (Saam et al., 2013), so the effect of other sources of X-ray radiation (such as synchrotrons) is still unknown and requires further investigation. However, the X-ray energy used and dose received in µCT are much greater than a nCT lab source, which was used in previous studies (Sykes et al., 2019), can produce. Therefore, we suggest that X-ray nCT radiation also would not affect mechanical properties of cuticle for time-lapse nCT tests.

Failure mechanisms of desiccated and fresh locust tibia

106

Paper 3

The methodology developed here was applied to investigate the failure mechanisms of tibia under 3 point bending in both the fresh and dry conditions, to compare and highlight the differences in failure mechanisms observed using in situ µCT imaging. Failure mechanisms differed significantly between the two conditions of cuticle. In the fresh cuticle, failure occurred primarily from significant buckling with small lateral zig-zag cracks forming in the tensile stress region. Failure by buckling has been found previously in fresh tibia but no evidence of crack formation has been found before (Parle et al., 2016). The crack itself spread through the entire wall thickness of the tibia only at the maximum displacement, and was unable to spread across the entire width or height of the tibia. The zig-zag shape of the crack path indicates the presence of extensive extrinsic toughening, which could explain the high fracture toughness value of 4.12 MPa m½ for fresh locust tibia found by Dirks & Taylor (2012a). In contrast, the desiccated tibia failed with minimal buckling and catastrophic crack propagation in all orthogonal directions, also, there was 25% less displacement before failure occurred. Cracks also developed under both compressive and tensile stress and at the mid-point between the two regions. Also in contrast to fresh tibia, the crack paths were very straight, with only minor deviations present in some cracks (fig. 5). This highlights that the capacity of cuticle for extrinsic toughening is significantly reduced when dried, and as a result becomes a more brittle material. However, these results are from a single sample for fresh cuticle and one for dry cuticle, therefore further studies with larger sample sizes for each condition of cuticle are needed to support these findings.

We propose that failure by buckling would be advantageous in vivo compared to crack propagation as buckling could potentially be repaired (Parle & Taylor, 2013; Parle, Dirks & Taylor, 2016, 2017) or at least the damaged limbs could be used with significantly reduced ability. Whereas in the dried case, cracks could grow easily over time as the lack of toughening would hinder repair, this would either eventually or immediately (dependant on the level of damage) result in the loss of the appendage entirely. This potential to reduce damage and employ extrinsic toughening mechanisms could explain why locusts keep their tibia hydrated rather than using drier tibia which would have more advantageous biomechanical properties(Schöberl & Jäger, 2006; Klocke & Schmitz, 2011; Dirks & Taylor, 2012a). This suggests a potential evolutionary advantage to having less efficient and even potentially slightly damaged but usable jumping legs rather than risk losing their tibiae entirely.

Time-lapse CT has the potential to vastly improve our understanding of the interaction between microstructure and crack propagation (Sykes et al., 2019), variation in mechanical responses to stress as shown here, and a multitude of other potential uses including: investigating the biomechanical benefits of the unusual shapes produced in nature and three-dimensional imaging that can be directly compared with FEA models (van der Meijden, Kleinteich & Coelho, 2012; Goyens et al., 2014). Moreover, the addition of the petroleum jelly coating means that the mechanical properties and responses studied are the same as would be found when testing fresh cuticle. Conclusions

This paper provides an effective methodology to perform in situ mechanical tests in combination with three-dimensional time-lapse computed tomography, whilst preserving the mechanical properties and structural response to stress of arthropod cuticle, by utilising petroleum jelly coating. We also found that PJ coating preserves hydration for the duration of long time scale mechanical tests such as time-lapse nCT, up to 144 hours. Further, we showed that X-ray µCT irradiation had no significant effect on mechanical properties. However, more work should be done to explore the effect a wider range of X-ray doses on cuticle. In situ 3-point bend tests of fresh and dry tibiae showed that changes in water content causes significant changes in the mechanical response to

107

Paper 3 stress, with desiccation making cuticle more brittle. The fresh tibia failed by buckling with some minor lateral crack propagation; in contrast the dry tibia failed by catastrophic and widespread crack propagation and minor buckling. The combination of these techniques opens a new path to investigate the mechanical behaviours of biological materials, sensitive to hydration levels, in a three-dimensional, time-lapse manner, which could aid our understanding of the biomechanical function of the many varied shapes and types of arthropod cuticle. Acknowledgements

We would like to thank Stuart Morse (Mechanical Characterisation lab, School of Materials, University of Manchester) for his advice in the experimental design and guidance in running the 3- point bend experiments. Also, we would also like to thank Mason Dean and Ronald Seidel (Max Planck Institute of Colloids & Interfaces, Potsdam, Germany) for their suggestions of potential hydration preservation techniques. References

Aberle B, Jemmali R, Dirks J-H. 2017. Effect of sample treatment on biomechanical properties of

insect cuticle. Arthropod Structure & Development 46:138–146. DOI:

10.1016/j.asd.2016.08.001.

Barth HD, Launey ME, MacDowell AA, Ager JW, Ritchie RO. 2010. On the effect of X-ray irradiation

on the deformation and fracture behavior of human cortical bone. Bone 46:1475–1485. DOI:

10.1016/j.bone.2010.02.025.

Barth HD, Zimmermann EA, Schaible E, Tang SY, Alliston T, Ritchie RO. 2011. Characterization of the

effects of x-ray irradiation on the hierarchical structure and mechanical properties of human

cortical bone. Biomaterials 32:8892–8904. DOI: 10.1016/j.biomaterials.2011.08.013.

Bayley TG, Sutton GP, Burrows M. 2012. A buckling region in locust hindlegs contains resilin and

absorbs energy when jumping or kicking goes wrong. Journal of Experimental Biology

215:1151–1161. DOI: 10.1242/jeb.068080.

Bennemann M, Backhaus S, Scholz I, Park D, Mayer J, Baumgartner W. 2014. Determination of the

Young’s modulus of the epicuticle of the smooth adhesive organs of Carausius morosus

using tensile testing. The Journal of Experimental Biology 217:3677–3687. DOI:

10.1242/jeb.105114.

108

Paper 3

Bennet-Clark HC. 1975. The energetics of the jump of the locust Schistocerca gregaria. Journal of

Experimental Biology 63:53–83.

Das R, Yadav RN, Sihota P, Uniyal P, Kumar N, Bhushan B. 2018. Biomechanical Evaluation of Wasp

and Honeybee Stingers. Scientific Reports 8:14945. DOI: 10.1038/s41598-018-33386-y.

Dirks J-H, Parle E, Taylor D. 2013. Fatigue of insect cuticle. Journal of Experimental Biology 216:1924–

1927. DOI: 10.1242/jeb.083824.

Dirks J-H, Taylor D. 2012a. Fracture toughness of locust cuticle. Journal of Experimental Biology

215:1502–1508. DOI: 10.1242/jeb.068221.

Dirks J-H, Taylor D. 2012b. Veins Improve Fracture Toughness of Insect Wings. PLOS ONE 7:e43411.

DOI: 10.1371/journal.pone.0043411.

Doube M, Kłosowski MM, Arganda-Carreras I, Cordelières FP, Dougherty RP, Jackson JS, Schmid B,

Hutchinson JR, Shefelbine SJ. 2010. BoneJ: Free and extensible bone image analysis in

ImageJ. Bone 47:1076–1079. DOI: 10.1016/j.bone.2010.08.023.

Goyens J, Soons J, Aerts P, Dirckx J. 2014. Finite-element modelling reveals force modulation of jaw

adductors in stag beetles. Journal of The Royal Society Interface 11:20140908. DOI:

10.1098/rsif.2014.0908.

Hörnschemeyer T, Bond J, Young PG. 2013. Analysis of the Functional Morphology of Mouthparts of

the Beetle Priacma serrata, and a Discussion of Possible Food Sources. Journal of Insect

Science 13. DOI: 10.1673/031.013.12601.

Hunter JD. 2007. Matplotlib: A 2D Graphics Environment. Computing in Science & Engineering 9:90–

95. DOI: 10.1109/MCSE.2007.55.

Jones E, Oliphant T, Peterson P. 2001. SciPy: open source scientific tools for Python.

Klocke D, Schmitz H. 2011. Water as a major modulator of the mechanical properties of insect

cuticle. Acta Biomaterialia 7:2935–2942. DOI: 10.1016/j.actbio.2011.04.004.

109

Paper 3

Kundanati L, Signetti S, Gupta HS, Menegon M, Pugno NM. 2018. Multilayer stag beetle elytra

perform better under external loading via non-symmetric bending properties. Journal of The

Royal Society Interface 15:1–13. DOI: 10.1098/rsif.2018.0427.

Limaye A. 2012. Drishti: a volume exploration and presentation tool. In: Stock SR ed. 85060X. DOI:

10.1117/12.935640. van der Meijden A, Kleinteich T, Coelho P. 2012. Packing a pinch: functional implications of chela

shapes in scorpions using finite element analysis. Journal of Anatomy 220:423–434. DOI:

10.1111/j.1469-7580.2012.01485.x.

Metscher BD. 2009. MicroCT for developmental biology: A versatile tool for high-contrast 3D imaging

at histological resolutions. Developmental Dynamics 238:632–640. DOI:

10.1002/dvdy.21857.

Nikolov S, Petrov M, Lymperakis L, Friák M, Sachs C, Fabritius H-O, Raabe D, Neugebauer J. 2010.

Revealing the Design Principles of High-Performance Biological Composites Using Ab initio

and Multiscale Simulations: The Example of Lobster Cuticle. Advanced Materials 22:519–

526. DOI: 10.1002/adma.200902019.

Oliphant TE. 2006. A guide to NumPy. Trelgol Publishing USA.

Parle E, Dirks J-H, Taylor D. 2016. Bridging the gap: wound healing in insects restores mechanical

strength by targeted cuticle deposition. Journal of The Royal Society Interface 13:20150984.

DOI: 10.1098/rsif.2015.0984.

Parle E, Dirks J-H, Taylor D. 2017. Damage, repair and regeneration in insect cuticle: The story so far,

and possibilities for the future. Arthropod Structure & Development 46:49–55. DOI:

10.1016/j.asd.2016.11.008.

Parle E, Herbaj S, Sheils F, Larmon H, Taylor D. 2016. Buckling failures in insect exoskeletons.

Bioinspiration & Biomimetics 11:016003. DOI: 10.1088/1748-3190/11/1/016003.

Parle E, Larmon H, Taylor D. 2016. Biomechanical Factors in the Adaptations of Insect Tibia Cuticle.

PLOS ONE 11:e0159262. DOI: 10.1371/journal.pone.0159262.

110

Paper 3

Parle E, Taylor D. 2013. The self-healing properties of insect cuticle. Journal of Postgraduate

Research, Trinity College Dublin 12:90–111.

Patterson BM, Cordes NL, Henderson K, Williams JJ, Stannard T, Singh SS, Ovejero AR, Xiao X,

Robinson M, Chawla N. 2016. In situ X-ray synchrotron tomographic imaging during the

compression of hyper-elastic polymeric materials. Journal of Materials Science 51:171–187.

DOI: 10.1007/s10853-015-9355-8.

Saam T, Herzen J, Hetterich H, Fill S, Willner M, Stockmar M, Achterhold K, Zanette I, Weitkamp T,

Schüller U, Auweter S, Adam-Neumair S, Nikolaou K, Reiser MF, Pfeiffer F, Bamberg F. 2013.

Translation of Atherosclerotic Plaque Phase-Contrast CT Imaging from Synchrotron Radiation

to a Conventional Lab-Based X-Ray Source. PLOS ONE 8:e73513. DOI:

10.1371/journal.pone.0073513.

Schöberl T, Jäger IL. 2006. Wet or Dry – Hardness, Stiffness and Wear Resistance of Biological

Materials on the Micron Scale. Advanced Engineering Materials 8:1164–1169. DOI:

10.1002/adem.200600143.

Seabold S, Perktold J. 2010. Statsmodels: Econometric and Statistical Modeling with Python. :6.

Sykes D, Hartwell R, Bradley RS, Burnett TL, Hornberger B, Garwood RJ, Withers PJ. 2019. Time-lapse

three-dimensional imaging of crack propagation in beetle cuticle. Acta Biomaterialia 86:109–

116. DOI: 10.1016/j.actbio.2019.01.031.

Wan C, Hao Z, Feng X. 2016. Structures, properties, and energy-storage mechanisms of the semi-

lunar process cuticles in locusts. Scientific Reports 6. DOI: 10.1038/srep35219.

Weaver JC, Milliron GW, Miserez A, Evans-Lutterodt K, Herrera S, Gallana I, Mershon WJ, Swanson B,

Zavattieri P, DiMasi E, Kisailus D. 2012. The Stomatopod Dactyl Club: A Formidable Damage-

Tolerant Biological Hammer. Science 336:1275–1280. DOI: 10.1126/science.1218764.

111

PAPER 4

Effect of hydration on crack propagation in beetle elytra using time-lapse three-dimensional imaging

112

Paper 4

Effect of hydration on crack propagation in beetle elytra using time-

lapse three-dimensional imaging

Dan Sykes1*, Russell J. Garwood2,3 and Philip J. Withers1

1Henry Moseley X-ray Imaging Facility, The Royce Institute, School of Natural Sciences, The University of Manchester, Manchester, M13 9PL, UK.

2School of Earth and Environmental Science, The University of Manchester, Manchester, M13 9PL, UK.

3Earth Sciences Department, Natural History Museum, London, SW7 5BD, UK

*Corresponding author

Abstract

Arthropod cuticle has extraordinary properties, it is stiff, hard, tough and lightweight, whilst being made of chitin-based biological constituents and often lacking mineralisation. The hierarchical structure, composition and hydration of cuticle produce this combination of properties, but we still have a poor understanding of the links between these aspects of cuticle and damage progression.

Here we use non-destructive, time-lapse X-ray nanotomography (nCT) with novel sample preparation and hydration preservation techniques to study crack propagation in situ, through fresh and dry beetle elytra (wing cases) cuticle in 3D. Also, we compare fresh elytra of two different species of beetle (Xylotrupes pubescens and Eudicella aurata). We find that fresh elytra possess extensive toughening mechanisms throughout the endocuticle laminae. The same toughening mechanisms are present but are less extensive and effective in dry elytra, resulting in a lower fracture toughness. In all tested elytra we show that fibre orientation affects the type and efficacy of toughening mechanisms within the laminae. We highlight insights into the design of biomimetics including the impact of alternating fibre orientations and modulating hydration levels.

113

Paper 4

1. Introduction

Arthropod cuticle exhibits an extraordinary diversity of mechanical properties, including high stiffness [1], hardness [2] and fracture toughness (the highest reported value is 4.12 MPa m½) [3] whilst using ordinary biological constituents [4], i.e. hierarchically organised chitin-protein fibres.

However, most cuticles require a balance of different properties for different functions, which is often achieved by specialisation of the outer exocuticle and inner endocuticle layers to provide different properties to the bulk cuticle material [5,6]. An extreme case of the trade-offs in achieving beneficial properties is beetle elytra (unmineralised wing cases), which require a low weight to facilitate flight and other forms of locomotion whilst protecting the beetles from a diverse range of hazards including impact, abrasion and fracture [7,8]. Beetle elytra also have an unusual hierarchical organisation of their microstructure compared to other cuticles, called pseudo-orthogonal or

‘balken’ [9]. This hierarchical organisation is as follows: the exocuticle is formed of laminae of a single layer of microfibres (~100 nm diameter) with orientations that rotate between successive laminae, and the endocuticle is formed of laminae of a single layer of macrofibres (~10 µm width) with orientations that rotate between successive laminae [6,9,10].

The links between cuticle microstructure and its mechanical properties are not well understood as most studies have been limited to 2D imaging of cuticle surfaces, which reveals little of the complex

3D nature of the internal microstructure. This also restricts time-lapse studies to strain mapping using digital image correlation [11–14]. However, a recent study employed three-dimensional time- lapse nano-computed tomography (nCT) to study the interaction between a pseudo-orthogonal microstructure and crack propagation in the elytra of the scarab beetle Macraspis lucida [15]. By studying cuticle in this manner, it was possible to determine that beetle elytra employ extrinsic toughening mechanisms almost exclusively. Furthermore, many of the toughening mechanisms found had not been previously described in arthropod cuticle. These included crack bifurcation, shear tearing, interlaminar crack bridging, and differential crack deflection between laminae, which

114

Paper 4 all produce a rough crack surface [15]. This is in addition to intralaminar crack bridging [15], which has been previously reported from fracture surface images of the elytra of Allomyrina dichotoma

[16] and Lucanus cervus [17]. This ability to visualise damage progression across the microstructure in 3D could lead to biomimetics designed to resist damage progression and therefore have improved toughness.

However, there are limitations in the study of M. lucida by time-lapse nCT [15]. Firstly, a dried elytron was tested, and dried samples have a lower fracture toughness than fresh ones (i.e. the same hydration level as in vivo) [3]. Therefore it may be expected that fresh elytra have more extensive toughening and a progressive, more energy-absorbing tensile failure response [14]. This is due to a decrease in stiffness and hardness with increasing water content [5,18], a trend that is particularly evident in the endocuticle [5], where the lamina studied were located [15].

Unfortunately, the acquisition times that were required for this study (24-48 hours) [15] are much longer than the time necessary for cuticle to desiccate completely, preventing the testing of fresh cuticle in this manner. However, a method to preserve the hydration of cuticle for long durations (at least one week) during mechanical testing has recently been developed (Paper 3). By coating cuticle samples in petroleum jelly immediately after euthanasia and dissection, it was possible to preserve the in vivo hydration levels of the cuticle, and therefore investigate the natural mechanical behaviour of the cuticle (Paper 3). Secondly, in the study of M. lucida [15], the samples were dissected using a razor blade and notched with a scalpel making the production of standardised test samples difficult to achieve, and introducing mechanical damage into the sample that could affect crack propagation. However, the recent development of laser-based micromachining systems allows the very small sample sizes that are necessary for mechanical tests of cuticle to be produced, in a standardised and precise way [19].

Here, we investigate the effects of desiccation on fracture toughness and toughening mechanisms in beetle elytra, by applying a new sample preparation method that combines hydration preservation

115

Paper 4 of the cuticle (Paper 3) with sample dissection using laser micromachining [19], to image fresh and dry elytra using time-lapse nCT with in situ crack propagation experiments. We also perform this test on the elytra of different species of beetle in a fresh condition to investigate the natural variation in toughening when sample geometry is standardised and hydration level preserved.

2. Materials and methods

2.1. Sample preparation

Two species of scarab beetles, the flower chafer Eudicella aurata (Westwood, 1841) and the rhinoceros beetle Xylotrupes pubescens (Waterhouse, 1841), were purchased as live mature adults and immediately euthanised. Both these species belong to the same family (Scarabaeidae), as well as M. lucida, in which the interaction between crack propagation and microstructure was studied using the same methodology [15]. Scarab beetles were chosen for having a pseudo-orthogonal arrangement of macrofibres [7,9,15], which can be visualised by nCT [15], and large, flattened elytra that ease sample preparation and mechanical testing. In addition, X. pubescens (Fig. 1a) and E. aurata (Fig. 1b) are members of the subfamilies Dynastinae (rhinoceros beetles) and Cetoniinae

(flower chafers), respectively. The elytra of Allomyrina dichotoma, a member of the rhinoceros beetles, have a relatively low Young’s modulus, and the elytra of Protaetia brevitarsis, a member of the flower chafers, have a relatively high Young’s modulus compared to other scarab beetles [20]. It is therefore interesting to compare these groups given their similar microstructure but apparent prioritisation of different mechanical properties.

To produce SENT samples from the elytra, the euthanised whole beetles were glued to a SEM stub

(Fig. 1c) and mounted in a Gatan MicroPREP laser system (Gatan Inc., Pleasanton, CA, United States) to cut uniform samples (Fig. 1c and d). This technique has the advantage of causing minimal damage to the sample and rapidly producing a geometry with a consistent notch size, which would be impossible using previous methods, e.g. using a scalpel. The speed of this technique (

116

Paper 4 per sample) means that the samples don’t have time to dry prior to mechanical testing. Fine tweezers were used to remove the SENT samples from the beetle elytra. Samples to be tested dry were left at room temperature to desiccate for 24 hours. In contrast, samples to be tested in a fresh condition (i.e. the same hydration level as in vivo) were immediately attached to the nano- mechanical test stage [21] with epoxy glue, as described in Sykes et al. (2019) [15]. To preserve the cuticle hydration and mechanical properties, the fresh (F) samples were immediately coated with petroleum jelly following the method of Paper 3, while dry (D) samples were attached to the nano- mechanical test stage without further treatment.

Figure 1. A) Photo of whole Xylotrupes pubescens (male) beetle, scale bar represents 5 mm; B) photo of whole Eudicella aurata (male) beetle, scale bar represents 5 mm; C) photo of whole E. aurata

(female) mounted on a SEM stub with a SENT sample laser cut from both elytra, scale bar represents

5 mm and the white dashed rectangle highlights the location of D) a magnified image of the laser cut

SENT sample with a schematic of the final sample for mechanical testing overlaid in white, scale bar represents 1 mm.

117

Paper 4

2.2. Time-lapse nano-computed tomography (nCT) with in situ crack propagation experiment

The same methodology as Sykes et al. (2019) [15] was used, and is summarised here. nCT scans were performed using a Zeiss Xradia 810 Ultra X-ray microscope (Carl Zeiss X-ray Microscopy Inc.,

Pleasanton, CA, United States) with a chromium source, operated at 5.4 keV with Zernike phase contrast. For each scan (load stage), a total of 721 equiangular X-ray projections were collected over a 170° rotation; a full 180° rotation is not possible as the nano-mechanical test stage attenuates all the X-rays at some projection angles, which causes some minor cupping artefacts in the data [22]. An exposure time of 60 seconds per projection was used giving a total acquisition time of ~12 hours per scan. An in situ nano-mechanical tensile test was performed between scans with the sample loaded at a strain rate of 0.01 mm/s, scans were collected when energy dissipation events were observed until failure occurred (Fig. 2). After failure, two nCT scans were obtained to capture the damage through the entire thickness of the cuticle. Sample dimensions were measured from X-ray projections: X. pubescens (D) – gauge length 0.90 mm, sample width 0.17 mm, thickness 0.07 mm and notch length 0.04 mm; X. pubescens (F) – gauge length 0.90 mm, sample width 0.16 mm, thickness 0.07 mm and notch length 0.07 mm; and E. aurata (F) – gauge length 0.98 mm, sample width 0.20 mm, thickness 0.06 mm and notch length 0.05 mm. Due to the very long acquisition times for each time-lapse tomographic sequence (2-3 days continuous nCT image collection) only one sample per condition and species were analysed; we did however test a number of samples offline, which displayed similar behaviour.

2.3. nCT image reconstruction and analysis

The volumetric data, collected from each nCT scan were reconstructed with a spatial resolution of

127 nm, using a filtered back projection algorithm in the Zeiss XMReconstructor software (Carl Zeiss

X-ray Microscopy Inc., Pleasanton, CA, United States), and exported as 8-bit 3D TIFF images. The reconstructed nCT data were then imported into the Dragonfly software (version 4.0; Object

118

Paper 4

Research Systems Inc., Montreal, Canada) where the laminae within the elytron sample were visible in the XY plane views. The laminae were manually segmented every 100 slices then an interpolation of the segmentations was performed to isolate all the individual lamina in each scan. The segmented laminae were exported as 3D TIFF images for import into the Drishti software [23], to produce 3D visualisations of crack propagation through each lamina for every load stage. The thickness and fibre orientation of each lamina were measured using Dragonfly.

2.4. Fracture toughness and estimated work of fracture

The load-displacement data collected for each sample during the in situ crack propagation experiment was converted into stress-strain data using the sample dimensions. Stress-strain curves were produced using a custom script in Python 3.7 that utilises the Matplotlib module [24]. From the stress-strain data and nCT scans of crack growth we calculated the fracture toughness (KIC) of each sample, using the formula:

퐾퐼퐶 = 푌휎√휋훼 where Y is a calibration factor that depends on the geometry of the sample, here a value of 1.12 was used [25], σ is the applied stress and α is the crack length, which was determined from 2D radiographs of the notch prior to crack propagation (e.g. Appendix A, Supplementary figure 6). The work of fracture (GC; strain energy release rate) was estimated using the formula:

퐾2 퐺 = 퐼퐶 퐶 퐸 where E is the Young’s modulus. As the Young’s modulus is not known for X. pubescens and E. aurata, we used the values for the closely related species (same subfamily) A. dichotoma (4.22 GPa) and P. brevitarsis (8.29 GPa), respectively [20], to estimate GC.

119

Paper 4

3. Results

3.1. Fracture toughness and estimated work of fracture

The stress-strain curves (Fig. 2) show a highly varied response to tensile stress between both cuticle condition and species. The dry X. pubescens elytron exhibited a linear elastic response to strain until load stage (LS) 1, where crack initiation occurs resulting in significant energy dissipation. This is followed by another linear response to strain until 8.4 MPa stress, where some energy dissipation occurs until catastrophic failure at LS2. The fresh X. pubescens elytron exhibited a similar stress- strain profile to the dry elytron with significant energy dissipation also occurring at 0.011 strain, although at a higher level of stress. Similarly, there is a second linear response to strain but a ~1.5x higher level of stress is reached. Furthermore, there is a greater degree of energy dissipation over a greater amount of strain until catastrophic failure occurs at LS2. In contrast, the fresh E. aurata elytron reaches far lower levels of stress (4.5 – five times less) before crack initiation occurs than X. pubescens in both conditions. However, after crack initiation occurs at LS1, the elytron sample undergoes three more steps of increasing stress followed by crack growth and energy dissipation until LS2, at which point the level of strain is greater than was required for catastrophic failure in X. pubescens elytra. This pattern repeats another two times until the sample’s peak tensile stress is reached and catastrophic failure occurs at LS3.

120

Paper 4

Figure 2. Stress-strain curves for the in situ tensile tests of the three tested samples, with the load stages (LS) 1-3 marked to indicate where nCT scans were collected. The nCT scans at load stage 0 were collected before applying any load therefore LS0 is not indicated here. (F) indicates a fresh sample and (D) indicates a dry sample.

From the stress-strain curves and nCT images of crack propagation, the fracture toughness was calculated and a work of fracture was estimated based on the Young’s modulus of closely related species (Table 1). The fresh X. pubescens elytra has the highest KIC, ~1.25x and ~3.2x greater than the dried X. pubescens elytra and fresh E. aurata elytra, respectively. The estimated GC was highest in the fresh X. pubescens elytra also, ~1.33x and ~5x greater than the dried X. pubescens elytra and fresh E. aurata elytra, respectively. The ultimate tensile strength was higher in fresh X. pubescens elytra (9.95 MPa) than in dry X. pubescens elytra (8.03 MPa), while fresh E. aurata elytra has an ultimate tensile strength of 2.03 MPa.

121

Paper 4

X. pubescens (F) X. pubescens (D) E. aurata (F) ½ Fracture toughness - KIC (MPa m ) 0.25 0.20 0.06 -2 Estimated work of fracture - GC (kJ m ) 0.008 0.006 0.0004

Table 1. The fracture toughness (KIC) and estimated work of fracture (GC) for the beetle elytra. Work of fracture estimated for X. pubescens and E. aurata using reported values of Young’s modulus for A. dichotoma and P. brevitarsis, respectively [20].

3.2. Computed nanotomography of cuticle structure

The thickness and fibre orientation of each lamina were measured from the nCT images of each sample and are reported in Table 2. The exocuticle, the outermost, dorsal lamina, is comprised of a helicoid arrangement of microfibre laminae that are too small (~100 nm) to resolve via nCT imaging, so we consider it as a single layer for the purposes of our study. The laminae thicknesses are consistent between the dry and fresh X. pubescens elytra, with the exocuticle accounting for 21.9 and 23.8 µm of the total thickness, respectively. The endocuticle is comprised of 10 macrofibre laminae (L1-10) in a pseudo-orthogonal arrangement, which are large enough to visualise and segment individually. In both samples, L1-9 have thicknesses between 4.0 – 5.2 µm and contain macrofibres with a square cross-section. L10, the innermost lamina, is thicker occupying 7.7 and 6.8

µm of the total thickness, respectively. E. aurata elytra have the same helicoid arrangement of microfibres in the exocuticle and pseudo-orthogonal arrangement of macrofibres in the endocuticle as X. pubescens. However, the elytron itself is ~10 µm thinner than in X. pubescens, which results from a reduction in the thickness of most laminae, but the relative proportion of the laminae is still similar. For example, the exocuticle occupies 30.9% of the total thickness, which is comparable to X. pubescens (30.8 – 33.5%). L1-9 have thicknesses between 3.0 – 4.8 µm and L10 has a thickness of 7.4

µm, which is similar in size to L10 in X. pubescens but a higher proportion of the total cuticle thickness.

122

Paper 4

The fibre orientations in the laminae of the dry X. pubescens elytron increase incrementally by an average of 84 ± 8° clockwise from the outer (L1) to inner (L10) lamina, which is consistent with the fresh X. pubescens elytron that exhibits a 81 ± 9° mean clockwise incremental increase in fibre orientation (Table 2). In contrast, in the E. aurata elytron, the fibre orientation increases incrementally by an average of 99 ± 11° in the clockwise direction.

123

Paper 4 Xylotrupes pubescens (F) Xylotrupes pubescens (D) Eudicella aurata (F) Fibre Fibre Fibre Fibre Fibre Fibre Thickness orientation by Thickness orientation by Thickness orientation by Lamina orientation orientation orientation (µm) clockwise (µm) clockwise (µm) clockwise w.r.t. load (°) w.r.t. load (°) w.r.t. load (°) increment (°) increment (°) increment (°) Exocuticle 23.8 ± 1.2 21.9 ± 1.4 18.7 ± 0.4 L1 4.7 ± 0.1 21 21 4.8 ± 0.4 140 140 3.1 ± 0.4 62 62 L2 4.8 ± 0.5 100 100 5.2 ± 0.6 52 232 3.4 ± 0.4 163 163 L3 4.6 ± 0.3 168 168 4.5 ± 0.2 120 300 4.5 ± 0.2 68 248 L4 4.0 ± 0.03 54 234 4.3 ± 0.2 14 374 4.8 ± 0.3 172 352 L5 4.7 ± 0.2 132 312 4.7 ± 0.5 94 454 4.4 ± 0.1 102 462 L6 4.8 ± 0.3 31 391 4.8 ± 0.2 177 537 4.1 ± 0.1 13 553 L7 4.7 ± 0.3 121 481 4.7 ± 0.3 84 624 3.7 ± 0.2 123 663 L8 4.2 ± 0.4 27 567 4.6 ± 0.5 175 715 3.0 ± 0.2 22 742 L9 4.0 ± 0.3 121 661 4.0 ± 0.1 84 804 3.5 ± 0.3 135 855 L10 6.8 ± 0.4 32 752 7.7 ± 0.7 177 897 7.4 ± 0.2 50 949

Total 71.0 71.1 60.6 Table 2. The mean thickness and the fibre orientations of each lamina within each sample, the total thickness of the elytron is also reported for each sample. L1-10 indicates lamina 1-10 (1 being the outermost lamina and 10 the innermost) of the endocuticle. Thickness is reported as the mean ± standard deviation. Fibre orientations are provided both with respect to the load direction (the vertical direction in Figs. 3, 4 and 5) and by increment in the clockwise direction, i.e. the total angle assuming a continuous clockwise rotation between laminae.

124

Paper 4

Figure 3. 3D nCT images of individually segmented laminae of the fresh E. aurata elytron sample showing crack propagation within laminae 4-10 during loading. Scale bar represents 10µm. 125

Paper 4

Figure 4. 3D nCT images of individually segmented laminae of the fresh X. pubescens elytron sample showing crack propagation within laminae Exocuticle – L7 during loading. Scale bar represents 10µm.

126

Paper 4

Figure 5. 3D nCT images of individually segmented laminae of the dry X. pubescens elytron sample showing crack propagation within laminae Exocuticle – L7 during loading. Scale bar represents 10µm.

127

Paper 4

Figure 6. 3D nCT images of individually segmented laminae of the fresh E. aurata elytron sample showing crack propagation within laminae Exocuticle – L3 at the point of failure. Scale bar represents 10µm.

Figure 7. 3D nCT images of individually segmented laminae of the fresh X. pubescens elytron sample showing crack propagation within laminae L8-10 at the point of failure. Scale bar represents 10µm.

Figure 8. 3D nCT images of individually segmented laminae of the dry X. pubescens elytron sample showing crack propagation within laminae L8-10 at the point of failure. Scale bar represents 10µm.

128

Paper 4

3.3. 3D time-lapse imaging of crack propagation in beetle elytra

Time-lapse sequences of 3D images showing damage progression during tensile straining are shown for fresh E. aurata (Figs. 3 and 6), fresh X. pubescens (Figs. 4 and 7) and dry X. pubescens (Figs. 5 and

8) elytra, with features of interest and toughening mechanisms highlighted. In E. aurata (F), crack initiation occurs in L6 at LS1, whilst in the other laminae the notch widens in comparison to LS0, but no other damage is visible (Fig. 3 – LS1). At LS2 the crack propagates through the width of the scanned region of the elytron sample in all laminae (Fig. 3 – LS2), and some toughening mechanisms can be observed: crack deflection along the macrofibre interfaces (Fig. 3 – LS2, L4-7, 9-10) as the path of least resistance; bifurcation of the crack (Fig. 3 – LS2, L7) with the crack split between a perpendicular fracture of the macrofibre and crack growth along the fibre interface; and shearing across macrofibres (Fig. 3 – LS2, L8-10). These are all extrinsic toughening mechanisms affecting crack growth at this stage. In addition, perpendicular fracture of macrofibres that are not aligned with the load direction can be observed (Fig. 3 – LS2, L7-10), which also act to slow crack growth.

Catastrophic failure occurred at LS3 and more extrinsic toughening mechanisms are observable.

Fibre pull-out is an important toughening mechanism in composites that is evident in a number of laminae (Fig. 3 – LS3, L4, 6-7; Fig. 6 – LS3, L3), it is also indicative of delamination and can result in bridging fibres, both of which act as further toughening mechanisms. Also, shearing across macrofibres occurs in laminae 1-4 (Fig. 3 – LS3, L4; Fig. 6 – LS3, L1-3), increasing the energy required to fracture the macrofibre. The alternating fibre orientations of the laminae results in some deviation of crack direction between laminae (Fig. 3 – LS3, L4-5), although the general direction of crack propagation remains consistent throughout the cuticle. Almost all the laminae produce a zig- zag, rough-edged crack (Fig. 3 – LS3, 4, 6-10; Fig. 6 – LS3, L1-3), this roughness acts to increase the frictional sliding force that shields the crack and slows its growth [26]. However, the exocuticle and

L5 produce a straight crack (Fig. 3 – LS3, L5; Fig. 3 – LS3, Exocuticle) where no clear toughening mechanisms are observable. From 2D X-ray projections of the sample taken prior to the in situ tensile test, it can be seen that the straight crack through the exocuticle formed prior to loading. In

129

Paper 4 the case of L5, the fibre orientation is closely aligned to the notch direction, allowing the crack to propagate freely along the macrofibre interface without deflection.

In the elytron of X. pubescens (F), crack initiation occurs at LS0 (Fig. 4 – LS0, L4) and there is a straight crack through the exocuticle with no evidence of any toughening mechanisms (Fig. 4 – LS0,

Exocuticle). By LS1 the crack has propagated through the width of the scanned region of the elytron and many toughening mechanisms are apparent (Fig. 4 – LS1). Crack deflection along the macrofibre interfaces is abundant within every lamina and produces large deflections (Fig. 4 – LS1, L1-7), this also produces significant amounts of perpendicular fracture of the macrofibres that are unaligned with load direction, in all such laminae (Fig. 4 – LS1, L1-2, 4-7). These mechanisms in combination with shearing of the macrofibres (Fig. 4 – LS1, L1-5) and fibre pull-out (Fig. 4 – LS1, L2-5, 7) produce a zig-zag, rough-edged crack (Fig. 4 – LS1, L1-7) which significantly slows crack growth. In L1, crack bridging by a macrofibre can be seen (Fig. 4 – LS1, L1) which provides further evidence that the macrofibres observed to be pulled out initially act as crack bridges, as this macrofibre is subsequently pulled out (Fig. 4 – LS2, L1). Catastrophic failure occurs at LS2 and the toughening mechanisms found in L1-7 are present in L8-10 as well. Crack deflection along the macrofibre interfaces (Fig. 7 – LS2, L8-10) is clearly visible but crack deflection along the microfibre interfaces within macrofibres can also be observed (Fig. 7 – LS2, L9-10). As in the other laminae, shearing across macrofibres (Fig. 7 – LS2, L9-10), perpendicular macrofibre fracture (Fig. 7 – LS2, L8-10) and fibre pull-out (Fig. 7 – LS2, L9) are present and produce a zig-zag, rough-edged crack (Fig. 7 – LS2, L8-

10). The changes in fibre orientation between laminae cause the crack to grow in alternating directions (Fig. 4 – LS2, L1-6), these changes in crack direction are more pronounced than in the E. aurata (F) elytron.

In comparison to the fresh samples, the elytron of X. pubescens (D) exhibited much fewer extrinsic toughening mechanisms. Between LS0 and LS1, the notch width increases and crack initiation occurs

(Fig. 5 – LS1, L6); unlike E. aurata (F) and X. pubescens (F) the exocuticle is undamaged at this stage

130

Paper 4

(Fig. 5 – LS1, Exocuticle). By LS2, catastrophic failure has occurred and in the exocuticle, L1 and L9 there are no clear toughening mechanisms observable (Fig. 5 – LS2, Exocuticle, L1; Fig. 8 – LS2, L9).

The crack path deviates little from the initial direction of the notch through all the laminae, with the exception of L2, 3, 8 and 10 (Fig. 5 – LS2). Also, at this stage the crack propagates across the exocuticle (Fig. 5 –LS2, Exocuticle) with a very similar morphology to the crack that is produced in the exocuticle of E. aurata (F) and X. pubescens (F). In most laminae, some extrinsic toughening mechanisms can be seen, such as: shearing across macrofibres (Fig. 5 – LS2, L2), crack deflection along the macrofibre interfaces (Fig. 5 – LS2, L2-3, 5-7; Fig. 8 – LS2, L8, 10), and perpendicular macrofibre fracture (Fig. 5 – LS2, L2). The most abundant toughening mechanism appears to be fibre pull-out which occurs in both macrofibres (Fig. 5 – LS2, L2-3, 5; Fig. 8 – LS2, L8, 10) and microfibres

(Fig. 5 – LS2, L3-4). Due to the limited effects of crack deflection in this sample the majority of laminae do not produce cracks of alternating directions with the exception of L2 and L3 (Fig. 5 –LS2,

L2-3).

4. Discussion

From the nCT data of all the elytra (Figs. 3-8) we are able to identify numerous extrinsic toughening mechanisms acting to slow crack growth behind the crack tip, such as: crack deflection along macrofibre interfaces; crack bifurcation; shearing across macrofibres; perpendicular fracture of macrofibres unaligned with the load direction; crack bridging by macrofibres; macrofibre pull-out; the production of a zig-zag, rough-edged crack; and alternating crack paths between adjacent laminae. We found two new toughening mechanisms that have not been found previously in cuticle

[15]: crack deflection within macrofibres and microfibre pull-out, although these were found in both the dry and fresh elytra. All the above evidence indicates that both dry and fresh cuticle possess the same extrinsic toughening mechanisms but that the frequency of their occurrence and scale of their effects are markedly different (Figs. 4 and 5). Our data also show that the interaction between the direction of crack growth (90° w.r.t. load) and the fibre orientation of the laminae has a significant

131

Paper 4 impact on which toughening mechanisms are present. Laminae with fibre orientations between 80-

110° (w.r.t. load) demonstrate little ability to deflect or slow crack growth, regardless of hydration level. Despite their seeming inability to resist crack growth, crack initiation was not observed in these laminae for any sample. Fibre pull-out, crack bridging and macrofibre shearing were most commonly associated with laminae where the fibre orientation was <70° or >110° w.r.t. load; in the fresh samples this contributed to producing large crack deflections but had little effect on the crack path in the dry sample. In general, crack paths were rougher, had more deflections and were more likely to run in alternating directions between adjacent laminae in the fresh samples (Figs. 3 and 4) than in the dry, where the crack paths were much straighter and similar between adjacent laminae

(Fig. 5). An interesting exception to this pattern is found in the exocuticle. In the fresh elytra of X. pubescens and E. aurata there was crack growth through the entire exocuticle at LS0, where only a small load from mounting the sample on the nano-mechanical test stage is applied, and the crack path is straight with almost no deviations at all. In the dry elytra of X. pubescens, in contrast, there is no crack growth in the exocuticle until catastrophic failure occurs, but the crack path is similarly straight. Despite this seeming disadvantage to the fresh elytra, as it could be assumed that having such a large defect would significantly weaken the sample prior to loading, the X. pubescens (F) elytra demonstrated greater resistance to crack growth and fracture toughness (Fig. 2 and Table 1).

This suggests that the exocuticle has little impact on toughness, at least from a crack growing in the lateral direction. It is known that the exocuticle and endocuticle contribute different properties to the bulk cuticle material, e.g. relatively high stiffness in the exocuticle and relatively high elasticity in the endocuticle [5,6]. From our results we can observe that the endocuticle is responsible for preventing crack growth, therefore we propose that the exocuticle and endocuticle may also provide different damage resistance properties. As exocuticle is exposed predominantly to impact and abrasion damage it may be specialised for that role; whereas endocuticle is not exposed to those types of damage so may be specialised for preventing cracks from spreading to the epidermis (the

132

Paper 4 cell layer responsible for cuticle production [6] and repair [27]), where damage could hinder the beetle’s ability to self-repair.

We found that fresh X. pubescens elytra had the highest fracture toughness and estimated work of fracture of the samples we tested (0.25 MPa m½ and 0.008 kJ m-2, respectively), which is much lower than values reported for fresh locust tibia (4.12 MPa m½ and 5.56 kJ m-2) [3]. This could be a result of different microstructure or hydration levels between these features with different forces acting upon them and functional capabilities. Desiccation is known to significantly increase the stiffness of cuticle [5,18] and the reported values we use to estimate GC are from dry elytra [20], therefore the value we report could be an underestimation. The dehydration of the elytra resulted in a reduction

½ -2 in the KIC and GC to 0.20 MPa m and 0.006 kJ m , respectively. The fracture toughness and estimated work of fracture are lower than dry locust tibia, 2.06 MPa m½ and 0.68 kJ m-2 respectively

½ [3]. The values for KIC and estimated GC for fresh E. aurata elytra found in this study (0.06 MPa m and 0.0004 kJ m-2) are much lower than in X. pubescens elytra, regardless of hydration level.

Although the estimated GC could be significantly underestimated as the stiffness for a dried elytra of a closely related species was used [20]; these values are lower than the previous lowest reported values of locust wing membrane (1.04 MPa m½ and 0.58 kJ m-2) [28]. E. aurata elytra exhibited a progressive and more energy-absorbing tensile failure response, as would be expected for fresh cuticle [5,15], than fresh X. pubescens elytra despite E. aurata having the lowest toughness. Ultimate tensile strength was found to be higher in fresh X. pubescens elytra than when dry, which is opposite to the expected response [5]. However, this may be due to natural variation as ex situ experiments found that although generally dry elytra were stiffer and stronger there is significant variation resulting in an overlap (Appendix A, Supplementary figure 2 and 3). This could be due a low level of hydration in elytra resulting in minor variation between dry and fresh elyta, or the sampling location could affect the response as there are many anatomical features that are not evenly distributed but may weaken the bulk cuticle, such as pore canals or the haemolymph space (Paper 1).

133

Paper 4

Our study has shown that the dehydration of beetle elytra reduces its toughness and even though the same toughening mechanisms are present in both fresh and dry X. pubescens elytra, the relative effect of those mechanisms and frequency of their occurrence is much higher in the fresh elytra. In the fresh elytra: every laminae of the endocuticle exhibits toughening mechanisms; the crack deflections are greater and more frequent than in dry elytra; and the roughness of the crack is greater than in dry elytra. Cuticle when dried has been shown to become stiffer and brittle [5,18] and this is supported by our ex situ tests. Although our in situ tests showed a stiffer response in fresh elytra, despite the fact that macrofibres in dry elytra had the more brittle response and tend to fracture in the direction of the crack path rather than being deflected around them as in fresh elytra.

Surprisingly, the largest difference in mechanical properties was not between fresh and dry elytra, but between fresh X. pubescens elytra and fresh E. aurata elytra. The effect of desiccation is known and significant, however our knowledge of differences between species, where the microstructure of the cuticle is the same, is very limited. This suggests that interspecific differences could be far greater than previously considered. Also, from our results it can be seen that the same toughening mechanisms are present at a similar frequency in both species. Additionally, both species do not mineralise their cuticle and their proportions of exocuticle and endocuticle in the elytra are similar.

From our data, the only structural difference between the elytra is the mean incremental increase in fibre orientation is ~15° greater in the laminae of E. aurata. The effectiveness of the toughening mechanisms in E. aurata is noticeably reduced, with smaller crack deflections and more similar crack paths between adjacent lamina than in fresh X. pubescens elytra. Based on these observations and the high stiffness reported for the elytra of the closely related P. brevitarsis [20], we propose that in this species other mechanical properties are being prioritised over toughness in comparison to X. pubescens. However, it is unclear where the trade-offs lie and whether changes in the microstructure produce these effects or if there is a more fundamental compositional difference.

Analysis of the proportions of chitin in the chitin-protein fibres, the degree of sclerotisation and the

134

Paper 4 in vivo hydration level of both cuticles would need to be compared to determine if compositional differences are responsible.

This study is the first 3D, quantitative characterisation of crack propagation in a hydration-sensitive biological material in its in vivo hydration state; where the hydration state was preserved for the entirety of a three day time-lapse nCT with in situ crack propagation experiment. However, the speed of the acquisition could be increased by using synchrotron X-ray tomography to acquire the X- ray projections at a similar resolution. However, the effects of synchrotron X-ray radiation on the mechanical properties of arthropod cuticle are unknown, unlike the laboratory X-ray nCT radiation used here (Paper 3). If synchrotron X-ray radiation is found to have no effect on cuticle mechanical behaviour it would also allow tomographic scans of multiple regions of interest at each load stage to be acquired, which would provide a more complete picture of the interaction between the microstructure and crack propagation. Future studies could use this methodology to investigate the mechanical behaviour of cuticle of a variety of forms and compositions under different mechanical stresses, all whilst preserving hydration and therefore mechanical properties. Another interesting, potential application of this technique would be to provide quantitative data of mechanical behaviour and the 3D microstructure of cuticle across different arthropod species, in order to improve our understanding of the evolutionary developmental biology of arthropods and to inform biomimetic design.

5. Conclusions

The combination of time-lapse nCT with in situ mechanical testing [15], cuticle hydration preservation (Paper 3) and sample preparation by laser micromachining has allowed us to study the interaction between microstructure of cuticle and crack propagation over time in both fresh and dry elytra for the first time. In addition, the use of laser micromachining to produce cuticle samples with a standard geometry allows us to provide a standardised, quantitative assessment of the fracture mechanics of cuticle. Desiccation of the X. pubescens elytra caused a reduction in fracture toughness

135

Paper 4 and in the extent and efficacy of the extrinsic toughening mechanisms within cuticle; particularly as the fresh elytra has the highest fracture toughness reported for any arthropod cuticle so far, that is comparable to natural, mineralised composites. The extent of extrinsic toughening mechanisms within the fresh X. pubescens elytra and the fresh E. aurata elytra was similar with all laminae exhibiting toughening, however the efficacy of these mechanisms in the X. pubescens elytra is much greater. Despite these differences in response to stress, it is apparent that the angle between the crack path and the fibre orientation of each lamina has a significant impact on how tough that lamina is and therefore the bulk toughness of the entire cuticle. This is an important lesson for biomimetic design in order to understand how different fibre arrangements can resist damage depending on the nature and direction of crack growth.

References

[1] J.C. Weaver, G.W. Milliron, A. Miserez, K. Evans-Lutterodt, S. Herrera, I. Gallana, W.J. Mershon,

B. Swanson, P. Zavattieri, E. DiMasi, D. Kisailus, The Stomatopod Dactyl Club: A Formidable

Damage-Tolerant Biological Hammer, Science. 336 (2012) 1275–1280.

doi:10.1126/science.1218764.

[2] J. Sun, J. Tong, Y. Ma, Nanomechanical Behaviours of Cuticle of Three Kinds of Beetle, Journal of

Bionic Engineering. 5 (2008) 152–157. doi:10.1016/S1672-6529(08)60087-6.

[3] J.-H. Dirks, D. Taylor, Fracture toughness of locust cuticle, Journal of Experimental Biology. 215

(2012) 1502–1508. doi:10.1242/jeb.068221.

[4] J.F.V. Vincent, U.G.K. Wegst, Design and mechanical properties of insect cuticle, Arthropod

Structure & Development. 33 (2004) 187–199. doi:10.1016/j.asd.2004.05.006.

[5] D. Klocke, H. Schmitz, Water as a major modulator of the mechanical properties of insect cuticle,

Acta Biomaterialia. 7 (2011) 2935–2942. doi:10.1016/j.actbio.2011.04.004.

[6] A.C. Neville, Biology of the arthropod cuticle, Springer-Verlag, 1975.

136

Paper 4

[7] T. van de Kamp, A. Riedel, H. Greven, Micromorphology of the elytral cuticle of beetles, with an

emphasis on weevils (Coleoptera: Curculionoidea), Arthropod Structure & Development. 45

(2016) 14–22. doi:10.1016/j.asd.2015.10.002.

[8] Z. Dai, Z. Yang, Macro-/Micro-Structures of Elytra, Mechanical Properties of the Biomaterial and

the Coupling Strength Between Elytra in Beetles, Journal of Bionic Engineering. 7 (2010) 6–12.

doi:10.1016/S1672-6529(09)60187-6.

[9] T. van de Kamp, H. Greven, On the Architecture of Beetle Elytra, Entomologie Heute. 22 (2010)

191–204.

[10] D. Raabe, P. Romano, C. Sachs, H. Fabritius, A. Al-Sawalmih, S.-B. Yi, G. Servos, H.G. Hartwig,

Microstructure and crystallographic texture of the chitin–protein network in the biological

composite material of the exoskeleton of the lobster Homarus americanus, Materials Science

and Engineering: A. 421 (2006) 143–153. doi:10.1016/j.msea.2005.09.115.

[11] J. Goyens, J. Soons, P. Aerts, J. Dirckx, Finite-element modelling reveals force modulation of jaw

adductors in stag beetles, Journal of The Royal Society Interface. 11 (2014) 20140908.

doi:10.1098/rsif.2014.0908.

[12] N.S. Ha, T.L. Jin, N.S. Goo, H.C. Park, Anisotropy and non-homogeneity of an Allomyrina

Dichotoma beetle hind wing membrane, Bioinspir. Biomim. 6 (2011) 046003. doi:10.1088/1748-

3182/6/4/046003.

[13] C. Sachs, H. Fabritius, D. Raabe, Influence of microstructure on deformation anisotropy of

mineralized cuticle from the lobster Homarus americanus, Journal of Structural Biology. 161

(2008) 120–132. doi:10.1016/j.jsb.2007.09.022.

[14] C. Sachs, H. Fabritius, D. Raabe, Experimental investigation of the elastic–plastic deformation of

mineralized lobster cuticle by digital image correlation, Journal of Structural Biology. 155 (2006)

409–425. doi:10.1016/j.jsb.2006.06.004.

137

Paper 4

[15] D. Sykes, R. Hartwell, R.S. Bradley, T.L. Burnett, B. Hornberger, R.J. Garwood, P.J. Withers, Time-

lapse three-dimensional imaging of crack propagation in beetle cuticle, Acta Biomaterialia. 86

(2019) 109–116. doi:10.1016/j.actbio.2019.01.031.

[16] J.X. Chen, M. Iwamoto, Q.Q. Ni, K. Kurashiki, K. Saito, Laminated Structure and Its Mechanical

Properties of the Fore-Wing of Beetle, J. Soc. Mat. Sci., Japan. 50 (2001) 455–460.

[17] L. Kundanati, S. Signetti, H.S. Gupta, M. Menegon, N.M. Pugno, Multilayer stag beetle elytra

perform better under external loading via non-symmetric bending properties, Journal of The

Royal Society Interface. 15 (2018) 1–13. doi:10.1098/rsif.2018.0427.

[18] T. Schöberl, I.L. Jäger, Wet or Dry – Hardness, Stiffness and Wear Resistance of Biological

Materials on the Micron Scale, Adv. Eng. Mater. 8 (2006) 1164–1169.

doi:10.1002/adem.200600143.

[19] T. Höche, M. Krause, M. Ebert, U. Wagner, M. Gebhardt, A Novel Laser Tool for High-Volume

Sample Preparation, Laser Technik Journal. 12 (2015) 42–44. doi:10.1002/latj.201500002.

[20] M. Yu, I. Hermann, Z. Dai, N. Gitis, Mechanical and Frictional Properties of the Elytra of Five

Species of Beetles, Journal of Bionic Engineering. 10 (2013) 77–83. doi:10.1016/S1672-

6529(13)60201-2.

[21] B.M. Patterson, N.L. Cordes, K. Henderson, J.C.E. Mertens, A.J. Clarke, B. Hornberger, A. Merkle,

S. Etchin, A. Tkachuk, M. Leibowitz, D. Trapp, W. Qiu, B. Zhang, H. Bale, X. Lu, R. Hartwell, P.J.

Withers, R.S. Bradley, In Situ Laboratory-Based Transmission X-Ray Microscopy and Tomography

of Material Deformation at the Nanoscale, Experimental Mechanics. (2016).

doi:10.1007/s11340-016-0197-3.

[22] A. Kyrieleis, V. Titarenko, M. Ibison, T. Connolley, P.J. Withers, Region-of-interest tomography

using filtered backprojection: assessing the practical limits, Journal of Microscopy. 241 (2011)

69–82. doi:10.1111/j.1365-2818.2010.03408.x.

138

Paper 4

[23] A. Limaye, Drishti: a volume exploration and presentation tool, in: S.R. Stock (Ed.), 2012: p.

85060X. doi:10.1117/12.935640.

[24] J.D. Hunter, Matplotlib: A 2D Graphics Environment, Computing in Science & Engineering. 9

(2007) 90–95. doi:10.1109/MCSE.2007.55.

[25] M. Liu, Y. Gan, D.A.H. Hanaor, B. Liu, C. Chen, An improved semi-analytical solution for stress at

round-tip notches, Engineering Fracture Mechanics. 149 (2015) 134–143.

doi:10.1016/j.engfracmech.2015.10.004.

[26] R. Ballarini, M.E. Plesha, The effects of crack surface friction and roughness on crack tip stress

fields, International Journal of Fracture. 34 (1987) 195–207. doi:10.1007/BF00019717.

[27] E. Parle, J.-H. Dirks, D. Taylor, Damage, repair and regeneration in insect cuticle: The story so far,

and possibilities for the future, Arthropod Structure & Development. 46 (2017) 49–55.

doi:10.1016/j.asd.2016.11.008.

[28] J.-H. Dirks, D. Taylor, Veins Improve Fracture Toughness of Insect Wings, PLOS ONE. 7 (2012)

e43411. doi:10.1371/journal.pone.0043411.

139

DISCUSSION

140

DISCUSSION

The fundamental issue addressed in this thesis can be summarised as being analysis of the interaction between arthropod cuticle microstructure and crack propagation. The novel methodologies of time-lapse computed tomography (CT) with in situ mechanical testing, laser micromachining and a novel hydration preservation technique, were developed and combined to address the effects of hydration and microstructure on toughness in cuticle.

The main aims of each paper can be summarised as follows:

 Paper 1: To review the central issues required for an understanding of arthropod

cuticle mechanics and to put the analytical papers in this thesis (i.e. 2-4) into a

wider context. It aims to compare the effects of evolutionary relationships, cuticle

compositions, and functions by performing a meta analysis of the mechanical

properties reported in the literature. It also discusses the advantages and

limitations of current techniques to analyse cuticle mechanics and the known

effects of variations in hierarchical organisation and composition of cuticle on

mechanical properties.

 Paper 2: To characterise how cracks propagate through dry beetle elytra (wing

cases) under tensile stress and to investigate the impact of its hierarchical cuticle

structure on toughening using 3D, time-lapse nanoCT (nCT) with an in situ crack

propagation experiment.

 Paper 3: To test petroleum jelly as a hydration preservation method for arthropod

cuticle and test its effect on the mechanical properties of cuticle. It also aims to

investigate potential effects of X-ray CT radiation on the mechanical properties of

141

cuticle. To show the likely benefits of applying hydration preservation to time-

lapse CT by performing in situ 3-point bend tests of fresh and dry locust tibiae to

visualise differences in damage and buckling, in 3D.

 Paper 4: To investigate the effects of desiccation on fracture toughness and

toughening mechanisms in beetle elytra, in a standardised manner, by applying a

new sample preparation method that combines hydration preservation of the

cuticle with sample dissection using laser micromachining, to image fresh and dry

elytra using time-lapse nCT with in situ crack propagation experiments. It also aims

to test the elytra of different species of beetle in a fresh condition to investigate

the natural variation in toughening when sample geometry is standardised and

hydration level preserved as in vivo.

The results of each analytical paper (i.e. 2-4) can be summarised as follows:

 Paper 2: The results obtained in this paper were the first three-dimensional, time-

lapse images of crack propagation in arthropod cuticle. They showed that the

pseudo-orthogonal arrangement of macrofibres in the microstructure is

responsible for increasing the toughness of the cuticle of beetle elytra. This study

identified a much greater range of extrinsic toughening mechanisms present in

cuticle than previously reported, including: shear tearing across macrofibres, crack

deflection along the interfaces of macrofibres, intra- and inter-laminar crack

bridging, fibre pull-out, crack bifurcation, delamination and crack deflection in

alternating directions between adjacent laminae (due to alternating macrofibre

orientations), all of which produced a zig-zag rough-edged crack. A dried elytron

was tested and despite evidence of extensive, extrinsic toughening a brittle

fracture occurred. This paper has shown that time-lapse nCT with in situ

142

mechanical testing is able to provide unique 3D information about the interaction

between microstructure and damage progression.

 Paper 3: Petroleum jelly coating is an effective method to preserve the hydration

and, therefore, the mechanical properties of cuticle as in vivo. Additionally, that

laboratory source X-ray CT radiation has no effect on cuticle’s mechanical

properties. These results contribute to the development of a protocol whereby

mechanical testing of fresh cuticle can be carried out in situ, in combination with

time-lapse CT without affecting mechanical properties. The results of testing this

protocol, by performing in situ 3-point bends with time-lapse microCT (µCT) on

fresh and dry tibiae, showed significant differences in damage progression. In the

dry tibia failure was predominantly caused by catastrophic crack propagation in all

orthogonal directions, with minor buckling damage. In contrast, in the fresh tibia

failure was predominantly caused by major buckling damage, although some crack

propagation due to tensile stress was also found, which has not been reported

before.

 Paper 4: The results obtained in this paper were the first three-dimensional, time-

lapse images of crack propagation in a fresh cuticle. They are also the first

mechanical tests using laser micromachining to produce cuticle samples with

standardised geometries, and without causing desiccation or mechanical damage.

Previous findings that water loss causes a significant reduction in fracture

toughness and work of fracture was confirmed, by comparing fresh and dry elytra

of the same species, Xylotrupes pubescens. In addition, the results showed that

hydration has a significant impact on the extrinsic toughening mechanisms within

cuticle. Although the mechanisms present in dry and fresh elytra are the same, in

fresh elytra the toughening mechanisms are more extensive and have a greater

143

effect on crack propagation. In X. pubescens, these mechanisms contribute to a

1.25x increase in fracture toughness over dried elytra. However, the difference in

toughness between fresh X. pubescens and Eudicella aurata elytra was the

greatest, despite possessing the same microstructural organisation and the

presence of the same toughening mechanisms. Interspecific differences therefore

may be more significant than previously realised. The reduced effectiveness of the

toughening mechanisms of E. aurata may be caused by differences in composition

or in vivo hydration level, in order to prioritise different mechanical properties

such as stiffness. This paper also found evidence of a relationship between

macrofibre orientation and the type of toughening mechanisms present, and of a

significant difference in the toughness of exocuticle and endocuticle, with

specialisation of the endocuticle for impeding crack growth.

Each paper presented here is a discrete body of work, either already peer reviewed, or written to be submitted for peer review in the near future. Therefore, every paper contains its own discussion of how it fits into the broader context of arthropod cuticle mechanics and composite design. Nonetheless, there are two key themes, which run through all the analytical papers (i.e. 2-4) and will be discussed in depth. First, the development of a methodology for in situ mechanical testing of hydration-sensitive biological materials. Second, the causes and variations of toughness in arthropod cuticle.

In addition, some future research directions are also discussed.

144

Development of a methodology for in situ mechanical testing of hydration- sensitive biological materials

Initially, characterisation studies of the mechanical properties of arthropod cuticle would often use dried cuticles, as it was easier to source and prepare dried material. Also, it was not known that hydration had a significant impact on its behaviour until recently

(Schöberl & Jäger, 2006; Klocke & Schmitz, 2011). This discovery resulted in a re- evaluation of the properties of arthropod cuticle, and a realisation that the same material in the same organism could have different properties purely by modulation of the material’s water content (Lomakin et al., 2011; Klocke & Schmitz, 2011; Dirks & Taylor,

2012a,b; Weaver et al., 2012; Wagner, Pittendrigh & Raman, 2012; Peisker, Michels &

Gorb, 2013; Goyens et al., 2014; Parle et al., 2016; Aberle, Jemmali & Dirks, 2017).

Mechanical testing of fresh cuticle was achieved simply by performing experimental work on cuticle dissected immediately after euthanasia. However, this technique doesn’t solve the problem of characterisation with in situ mechanical testing, which can take hours or even weeks to complete (Belkas et al., 2005; Barth et al., 2010). Particularly, as cuticle can experience a significant change in hydration and, therefore, mechanical properties in less than 20 minutes (Klocke & Schmitz, 2011).

Our proposed solution of coating samples in petroleum jelly (PJ) after euthanasia and dissection (Paper 3), is shown to preserve hydration and mechanical properties for at least one week. The longest time-lapse CT experiments that have been performed, to my knowledge, are the ones lasting 2-3 days carried out in Papers 2 & 4 of this thesis. Paper 4 demonstrates that the differences in mechanical behaviour and properties between fresh and dry cuticle can be observed with PJ coating (e.g. the ductile failure observed in the fresh elytra of E. aurata), providing further corroboration of the efficacy of this method.

145

Furthermore, enabling time-lapse nCT of crack propagation in fresh cuticles provides a unique analysis of the impact of 3D microstructure in a hydration-sensitive biological material, that is not possible with any other technique. However, the use of X-ray CT to image a biological material raises the concern of potential X-ray radiation damage causing a change in mechanical properties, as has been observed in other biological materials such as bone (Barth et al., 2010, 2011). Paper 3 provides preliminary evidence that X-ray radiation from a laboratory CT source has no effect on the mechanical properties of cuticle, and ex situ testing of beetle elytra produced similar stress-strain responses to in situ tests (Papers 2 & 4). Therefore, it appears that time-lapse CT with a laboratory source does not cause significant radiation damage, however, further studies to investigate this in depth and with different sources of X-ray radiation are still required.

Toughness in arthropod cuticle

The toughness of cuticle has only been investigated in a few types of cuticle (Melnick,

Chen & Mecholsky, 1996; Sun & Tong, 2007; Sachs, Fabritius & Raabe, 2008; Dirks &

Taylor, 2012a,b), but high toughness values were reported, especially for unmineralised locust tibiae (Dirks & Taylor, 2012a). In addition, few toughening mechanisms have been identified from cuticle fracture surfaces (Chen et al., 2001; Sachs, Fabritius & Raabe,

2006; Kundanati et al., 2018), which leaves a significant knowledge gap between the high reported values for toughness and the scarcity of evidence of what provided those properties. Although a reduction in hydration level has been found to significantly reduce the fracture toughness and work of fracture of locust tibiae (Dirks & Taylor, 2012a).

The use of three-dimensional, time-lapse CT with in situ crack propagation experiments in this thesis makes significant progress in filling this knowledge gap, although there is still much more work to be done before a complete understanding can be reached. By

146 visualising crack propagation in 3D by nCT, in both fresh and dry beetle elytra, a wide range of extrinsic toughening mechanisms could be identified, as described in the summary of results of Papers 2 and 4. Crack deflection leading to the production of a zig- zag rough-edged crack was also evident in 3-point bending of fresh locust tibiae (Paper 3).

Although lower resolution microCT imaging was used and given the high fracture toughness of locust tibiae (Dirks & Taylor, 2012a), it is likely that at the nanoscale many more toughening mechanisms would be identifiable. All the toughening mechanisms identified in cuticle act behind the crack tip to slow crack growth, they work by significantly increasing the surface area of the crack or the energy required for it to grow.

Many of the mechanisms can work in tandem to produce multiplied effects, for example: crack deflection along the interface of a fibre can lead to crack bridging by that fibre, as the crack grows there is delamination around the fibre that can eventually lead to fibre pull-out, which absorbs significant amounts of energy and results in the production of a rough-edged crack that requires more energy to grow (Lawrence, 1972; Ballarini & Plesha,

1987; Kim, Baillie & Mai, 1992).

Paper 4 reports a significant reduction in fracture toughness for cuticle in the elytra of X. pubescens when dried, as previously reported (Dirks & Taylor, 2012a). However, it was found that dry elytra possess the same toughening mechanisms as fresh elytra, even between different species (Papers 2 & 4). This suggests that the microstructural organisation is responsible for the presence of the extrinsic toughening mechanisms of cuticle, in particular, the fibre orientations of the laminae with respect to the direction of crack growth. Where the fibre orientations are misaligned with the crack direction by 20° or more is where fibre pull-out, large crack deflections and shearing of fibres are most common (Paper 4). In addition, the exocuticle was found to have little impact on the bulk toughness of the cuticle with no toughening mechanisms found within this layer, in both 147 dry and fresh elytra. The key difference between fresh and dry cuticle appears to be the frequency and effectiveness of the toughening mechanisms in each lamina (Paper 4).

Fresh elytra have multiple types of toughening mechanism and a higher frequency of them in each lamina, in comparison to dry elytra. In addition, the toughening mechanisms are more effective in fresh elytra. For example, crack deflections in fresh elytra occur more frequently and cause greater deviations in the crack path, significantly increasing the surface area of the crack and therefore increasing toughness. It is likely that this increased toughness is due to increased hydrogen bonding between water and the protein matrix surrounding the chitin fibres, which reduces stiffness and increases the ability of the matrix to absorb energy (Klocke & Schmitz, 2011).

Future work

There are a number of weaknesses in these analyses of fracture mechanics in arthropod cuticle that should be addressed in future work. A major limitation of using time-lapse nCT with a 127 nm spatial resolution to investigate microstructure, is its inability to resolve the finer microfibres that form the entirety of the cuticular microstructure in locust tibiae (Paper 3), the exocuticle of beetle elytra (Paper 4) and the constituents of the macrofibres in the pseudo-orthogonal arrangement found in beetle elytra (Paper 1).

As the microfibres have a diameter of ~100 nm (Neville, 1975; Sachs, Fabritius & Raabe,

2008), one solution is to use higher resolution nCT, but this comes with a trade-off in increased acquisition times and a smaller field of view. Therefore many scans at multiple scales, loadings and locations would need to be collected to fully characterise crack propagation in a single sample, which would result in a total acquisition time for each time-lapse sequence of longer than a week. Another solution would be to perform serial section tomography by Focused Ion Beam-Scanning Electron Microscopy (FIB-SEM) to

148 produce very high resolution 3D models of the microfibre orientations and arrangement of untested samples (Burnett et al., 2016). These 3D models could then be compared to fracture surfaces from the in situ mechanical tests in order to obtain a complete picture of the interaction between crack propagation and microstructure.

The sample geometries of cuticle prepared for crack propagation experiments were significantly improved by the work of this thesis, using laser micromachining to rapidly produce standardised, undamaged samples (Paper 4). However, in nature, crack propagation is unlikely to spread laterally from a defect through all the layers until after significant damage has already occurred. This is due to the fact that the exocuticle forms a complete barrier between the internal laminae and the environment (Neville, 1975), therefore cracks would need to form in the exocuticle before spreading into the endocuticle. Although the results presented here show that exocuticle is an ineffective barrier to crack propagation, the direction of crack growth would progress in the normal direction through the laminae. As fibre orientation has been shown here to have a significant effect on crack propagation (Papers 2 & 4), the toughness of the cuticle may be anisotropic as has been found for stiffness and hardness (Romano, Fabritius & Raabe,

2007; Sachs, Fabritius & Raabe, 2008; Lin, Dorfmann & Trimmer, 2009; Ha et al., 2011).

However, to study damage progression in a natural context would require the formation of a notch in the exocuticle. To do this, a sample would need to be machined from the cuticle, extracted, rotated 90°, re-mounted and a notched sample machined out.

Preparing such a sample from a material that has a pre-determined thickness between 5 –

250 µm, whilst preserving the hydration throughout the preparation process, presents a significant challenge. However, if a method can be developed to do this, it could reveal interesting new discoveries about anisotropy in fracture toughness due to hierarchical organisation of fibres. 149

An interesting and unexpected finding of Paper 4, was that the interspecific difference in toughness between X. pubescens and E. aurata elytra was greater than the differences caused by desiccation. The causes of this difference are not obvious as both species have a similar life history (burrowing species that mostly eat fruit), they are relatively closely related (same family), and they both possess the same microstructural organisation

(helicoid exocuticle and pseudo-orthogonal endocuticle). If E. aurata elytra prioritises other mechanical properties for its survival, there could be compositional differences in the cuticle, such as an increase in the relative proportion of chitin in its chitin-protein fibres or increased sclerotisation, both of which would increase the stiffness and potentially reduce fracture toughness (Vincent & Wegst, 2004). To identify which aspect, if any, is causing this difference, a number of beetle elytra sharing similar life history, phylogenetic relationship and microstructure should be analysed to find clear differences in composition prior to testing, to identify their impact.

Paper 4 is the first report of differences in toughness between the exocuticle and endocuticle, which is also the case for stiffness and hardness (Neville, 1975; Vincent &

Wegst, 2004; Sun, Tong & Zhang, 2009; Dirks & Dürr, 2011; Fabritius et al., 2016). It proposes that there is specialisation of the exocuticle and endocuticle to resist different types of damage. Exocuticle, as the outer layer, is exposed to abrasion and impact damage while the endocuticle is exposed to crack damage propagating from the exocuticle. Therefore, a full characterisation of the hardness by indentation tests of both layers and stiffness by tensile tests of both layers (if possible to separate them by peeling or laser micromachining) is required, to determine how these layers resist different forms of damage, and how they form a bulk material with combined properties.

Regardless of these limitations to the analyses of toughness in cuticle in these papers, we have gained valuable knowledge from the studies presented here and developed a 150 methodology that has great potential for future investigations of arthropod cuticle. From a biological standpoint, it would be very interesting to perform time-lapse CT with in situ mechanical testing of a specific cuticle type across a diverse range of species. A 3D characterisation of the elytra microstructure, the toughness properties, a compositional analysis of the cuticle and phylogenetic analysis could be combined to investigate a number of research questions of value to biologists and materials scientists. For example, has convergent evolution of different microstructures resulted in similar mechanical properties of elytra? Or is elytra microstructure highly conserved between species, but composition altered to respond to different selection pressures? This could also identify species that possess extreme values of toughness, leading to the determination of which microstructural organisations and compositions are the optimal for providing toughness at the micro- and nano-scale. It would also be valuable to investigate mineralised cuticles with this methodology, as values for fracture toughness have been found to be similar to dried, unmineralised cuticles (Melnick, Chen & Mecholsky, 1996; Dirks & Taylor, 2012a).

However, it is not known whether this is because the tested mineralised cuticles are intrinsically less tough than fresh, unmineralised cuticles, have a microstructure that prioritises other mechanical properties, or have some other compositional differences that have a greater contribution to the toughness of cuticle (e.g. salinity or hydration level). Another potentially fruitful direction for this technique is to apply other in situ mechanical tests to cuticle. Combining in situ tensile, compression and indentation tests with digital volume correlation would allow 3D strain mappings to be created, to further our understanding of how the fibres in this composite respond to those stresses, as has already been carried out on other biological materials like dentin (Patterson et al., 2016), and in 2D on cuticle using digital image correlation (Sachs, Fabritius & Raabe, 2008). Also, these in situ tests combined with 3D imaging allows direct comparison and validation of

151 the 3D deformations predicted by finite element analysis of cuticle, which is increasingly used to analyse the effect of shape on cuticle properties (van der Meijden, Kleinteich &

Coelho, 2012; Weaver et al., 2012; Goyens et al., 2014).

Conclusions

Our understanding of the links between shape, microstructure, composition and mechanical properties in arthropod cuticle still contains many gaps. However, the recent findings of extraordinary properties in such a lightweight biological material (Dirks &

Taylor, 2012a; Weaver et al., 2012) has intensified research interest, which looks likely to grow significantly over the coming years, especially as more unique microstructures and compositions are discovered (Raabe et al., 2005; Weaver et al., 2012; van de Kamp, Riedel

& Greven, 2016). It is now clear that no single aspect of cuticle structure, form or composition can explain its incredible diversity of mechanical properties, and that arthropods exploit all these aspects to aid their survival. In order to contribute to filling one of the largest gaps in our understanding of cuticle, the present thesis showed how microstructure and hydration affect toughness. To achieve this, three-dimensional time- lapse computed tomography was combined with in situ crack propagation experiments to visualise the interaction between 3D microstructure and damage progression; the development of novel sample preparation techniques to produce standardised samples; and the development of a method to preserve hydration levels to study fresh cuticle.

The results obtained in this thesis have shown that it is possible with these techniques to analyse toughening in both fresh and dry cuticles, by qualitative visualisation of the interaction between microstructure and crack propagation and quantitative measurement of toughness values from standardised test samples. All the analytical papers (Papers 2, 3 & 4) associated microstructure and hydration with fracture mechanics

152 in cuticle, and formed part of the development of a novel methodology to investigate mechanical behaviour in a hydration-sensitive biological material. It was shown that microstructure is responsible for the numerous extrinsic toughening mechanisms present in cuticle, of which many were previously unreported, and that hydration is responsible for improving the effectiveness and frequency of their occurrence (Papers 2 & 4).

Furthermore, it was found that the exocuticle contributes little to toughness; and that the difference in angle between the crack direction and the fibre orientations of a lamina directly affected the toughening capability of that lamina (Paper 4). To summarise, the different papers of this thesis display how a combined approach using state-of-the-art 3D imaging, in situ mechanical testing, sample preparation and hydration preservation techniques can provide us with new insights into the interaction between shape, microstructure, composition and mechanical properties in arthropod cuticle.

References

Aberle B, Jemmali R, Dirks J-H. 2017. Effect of sample treatment on biomechanical

properties of insect cuticle. Arthropod Structure & Development 46:138–146. DOI:

10.1016/j.asd.2016.08.001.

Ballarini R, Plesha ME. 1987. The effects of crack surface friction and roughness on crack

tip stress fields. International Journal of Fracture 34:195–207. DOI:

10.1007/BF00019717.

Barth HD, Launey ME, MacDowell AA, Ager JW, Ritchie RO. 2010. On the effect of X-ray

irradiation on the deformation and fracture behavior of human cortical bone.

Bone 46:1475–1485. DOI: 10.1016/j.bone.2010.02.025.

Barth HD, Zimmermann EA, Schaible E, Tang SY, Alliston T, Ritchie RO. 2011.

Characterization of the effects of x-ray irradiation on the hierarchical structure

153

and mechanical properties of human cortical bone. Biomaterials 32:8892–8904.

DOI: 10.1016/j.biomaterials.2011.08.013.

Belkas JS, Munro CA, Shoichet MS, Johnston M, Midha R. 2005. Long-term in vivo

biomechanical properties and biocompatibility of poly(2-hydroxyethyl

methacrylate-co-methyl methacrylate) nerve conduits. Biomaterials 26:1741–

1749. DOI: 10.1016/j.biomaterials.2004.05.031.

Burnett TL, Kelley R, Winiarski B, Contreras L, Daly M, Gholinia A, Burke MG, Withers PJ.

2016. Large volume serial section tomography by Xe Plasma FIB dual beam

microscopy. Ultramicroscopy 161:119–129. DOI: 10.1016/j.ultramic.2015.11.001.

Chen JX, Iwamoto M, Ni QQ, Kurashiki K, Saito K. 2001. Laminated Structure and Its

Mechanical Properties of the Fore-Wing of Beetle. Journal of the Society of

Materials Science, Japan 50:455–460.

Dirks J-H, Dürr V. 2011. Biomechanics of the stick insect antenna: Damping properties and

structural correlates of the cuticle. Journal of the Mechanical Behavior of

Biomedical Materials 4:2031–2042. DOI: 10.1016/j.jmbbm.2011.07.002.

Dirks J-H, Taylor D. 2012a. Fracture toughness of locust cuticle. Journal of Experimental

Biology 215:1502–1508. DOI: 10.1242/jeb.068221.

Dirks J-H, Taylor D. 2012b. Veins Improve Fracture Toughness of Insect Wings. PLOS ONE

7:e43411. DOI: 10.1371/journal.pone.0043411.

Fabritius H-O, Ziegler A, Friák M, Nikolov S, Huber J, Seidl BHM, Ruangchai S, Alagboso FI,

Karsten S, Lu J, Janus AM, Petrov M, Li-Fang Zhu, Hemzalová P, Hild S, Raabe D,

Neugebauer J. 2016. Functional adaptation of crustacean exoskeletal elements

through structural and compositional diversity: a combined experimental and

theoretical study. Bioinspiration & Biomimetics 11:055006. DOI: 10.1088/1748-

3190/11/5/055006.

154

Goyens J, Soons J, Aerts P, Dirckx J. 2014. Finite-element modelling reveals force

modulation of jaw adductors in stag beetles. Journal of The Royal Society Interface

11:20140908. DOI: 10.1098/rsif.2014.0908.

Ha NS, Jin TL, Goo NS, Park HC. 2011. Anisotropy and non-homogeneity of an Allomyrina

Dichotoma beetle hind wing membrane. Bioinspiration & Biomimetics 6:046003.

DOI: 10.1088/1748-3182/6/4/046003. van de Kamp T, Riedel A, Greven H. 2016. Micromorphology of the elytral cuticle of

beetles, with an emphasis on weevils (Coleoptera: Curculionoidea). Arthropod

Structure & Development 45:14–22. DOI: 10.1016/j.asd.2015.10.002.

Kim J-K, Baillie C, Mai Y-W. 1992. Interfacial debonding and fibre pull-out stresses. Journal

of Materials Science 27:3143–3154. DOI: 10.1007/BF01116004.

Klocke D, Schmitz H. 2011. Water as a major modulator of the mechanical properties of

insect cuticle. Acta Biomaterialia 7:2935–2942. DOI:

10.1016/j.actbio.2011.04.004.

Kundanati L, Signetti S, Gupta HS, Menegon M, Pugno NM. 2018. Multilayer stag beetle

elytra perform better under external loading via non-symmetric bending

properties. Journal of The Royal Society Interface 15:1–13. DOI:

10.1098/rsif.2018.0427.

Lawrence P. 1972. Some theoretical considerations of fibre pull-out from an elastic

matrix. Journal of Materials Science 7:1–6. DOI: 10.1007/BF00549541.

Lin H-T, Dorfmann AL, Trimmer BA. 2009. Soft-cuticle biomechanics: A constitutive model

of anisotropy for caterpillar integument. Journal of Theoretical Biology 256:447–

457. DOI: 10.1016/j.jtbi.2008.10.018.

155

Lomakin J, Huber PA, Eichler C, Arakane Y, Kramer KJ, Beeman RW, Kanost MR, Gehrke

SH. 2011. Mechanical Properties of the Beetle Elytron, a Biological Composite

Material. Biomacromolecules 12:321–335. DOI: 10.1021/bm1009156. van der Meijden A, Kleinteich T, Coelho P. 2012. Packing a pinch: functional implications

of chela shapes in scorpions using finite element analysis. Journal of Anatomy

220:423–434. DOI: 10.1111/j.1469-7580.2012.01485.x.

Melnick CA, Chen Z, Mecholsky JJ. 1996. Hardness and toughness of exoskeleton material

in the stone crab, Menippe mercenaria. Journal of Materials Research 11:2903–

2907. DOI: 10.1557/JMR.1996.0367.

Neville AC. 1975. Biology of the arthropod cuticle. Springer-Verlag.

Parle E, Herbaj S, Sheils F, Larmon H, Taylor D. 2016. Buckling failures in insect

exoskeletons. Bioinspiration & Biomimetics 11:016003. DOI: 10.1088/1748-

3190/11/1/016003.

Patterson BM, Cordes NL, Henderson K, Mertens JCE, Clarke AJ, Hornberger B, Merkle A,

Etchin S, Tkachuk A, Leibowitz M, Trapp D, Qiu W, Zhang B, Bale H, Lu X, Hartwell

R, Withers PJ, Bradley RS. 2016. In Situ Laboratory-Based Transmission X-Ray

Microscopy and Tomography of Material Deformation at the Nanoscale.

Experimental Mechanics. DOI: 10.1007/s11340-016-0197-3.

Peisker H, Michels J, Gorb SN. 2013. Evidence for a material gradient in the adhesive

tarsal setae of the ladybird beetle Coccinella septempunctata. Nature

Communications 4:1661. DOI: 10.1038/ncomms2576.

Raabe D, Romano P, Sachs C, Al-Sawalmih A, Brokmeier H-G, Yi S-B, Servos G, Hartwig HG.

2005. Discovery of a honeycomb structure in the twisted plywood patterns of

fibrous biological nanocomposite tissue. Journal of Crystal Growth 283:1–7. DOI:

10.1016/j.jcrysgro.2005.05.077.

156

Romano P, Fabritius H, Raabe D. 2007. The exoskeleton of the lobster Homarus

americanus as an example of a smart anisotropic biological material. Acta

Biomaterialia 3:301–309. DOI: 10.1016/j.actbio.2006.10.003.