The word bionics was coined by Jack Steele of the US Air Force in 1960 at a meeting at Wright-Patterson Air Force Base in Dayton, Ohio. He defined it as the science of systems which have some function copied from nature, or which represent characteristics of natural systems or their analogues. At another meeting at Dayton in 1963, Schmitt said

Later, Schmitt used the word biomimetics in the title of a paper (Schmitt 1969); the word made its first public appearance in Webster's Dictionary in 1974, accompanied by the following definition:

Biomimetics (which we here mean to be synonymous with ‘biomimesis’, ‘biomimicry’, ‘bionics’, ‘biognosis’, ‘biologically inspired design’ and similar words and phrases implying copying or adaptation or derivation from biology) is thus a relatively young study embracing the practical use of mechanisms and functions of biological science in engineering, design, chemistry, electronics, and so on. However, people have looked to nature for inspiration for more than 3000 years (since the Chinese first tried to make an artificial silk). Historically, we have:

The following list is by no means complete; the Internet is a large and unreliable source of further examples. We have chosen examples which are contentious, interesting or iconic.

Approximately 50 years ago in Russia, a particularly successful problem-solving system began to be developed. It was named TRIZ, the acronym of Teorija Reshenija Izobretatel'skih Zadach (loosely translated as ‘Theory of Inventive Problem Solving’).TRIZ is well known for its successful transfer of various inventions and solutions from one field of engineering to another. Since the main thrust of biomimetics is also to transfer functions, mechanisms and principles from one field to another, TRIZ seems the ideal starting point (Bogatyrev 2000; Vincent & Mann 2002). We also use TRIZ as a functional summary and definition of engineering methodology, a novel use of the system. We know of no other strategy or system which is so powerful and so general. Since TRIZ is not very well known to Western science and technology, a short description is necessary, outlining its normal use by problem solvers.

TRIZ is a collection of tools and techniques, developed by Genrich Altshuller and Rafik Shapiro (Altshuller 1999) that ensures accurate definition of a problem at a functional level and then provides strong indicators towards successful and often highly innovative solutions. At the definition stage, a number of techniques are used to ensure that the problem is placed properly within its context (simply changing the context may solve the problem…) and the available resources listed. In the most popular (though probably not the best) technique for solution, the problem is then characterized by a pair of opposing or conflicting characteristics (typically ‘what do I want’ and ‘what is stopping me getting it’, but Hegel's thesis and antithesis will do as well, suggesting that it is a form of dialectic process), which can be compared with pairs of characteristics derived from other, solved, problems derived from the examination and analysis of more than three million significant patents.

We present a case study based on a relatively simple natural fibrous composite material—the outer covering or cuticle of arthropods. This layer of material, produced by a single layer of epithelial cells, is called upon to provide a large number of functions, such as shape, structure, hinges, barrier, filter and similar functions (Neville 1975). Some of these functions are intrinsically and profoundly conflicting, although obviously since they coexist some form of compromise must have been evolved, so that the cuticle can be multifunctional. A list of these functions and the associated characteristics of cuticle was generated partly by reference to literature (such as Neville 1975) and partly from experience with insects and insect cuticle over the years.

Consider, then, the function of the cuticle in providing a stiff support or exoskeleton for the insect, attachment for muscles, mechanical protection and control of shape (Vincent 2005). A uniformly stiff skeleton does not permit movement, so hinged areas are needed. In the insect, this has been achieved by making the cuticle softer along the hinge line. This appears to be the same as TRIZ inventive principle (IP) 3: Control of local quality, which is characterized by the following statements: use gradients instead of uniformity (change an object's structure, or its environment, from homo- to heterogeneous); compartmentalize (make each part of an object more adapted to its own purpose); introduce multifunctionality (make each part of an object fulfil a different function like a pencil with an eraser; a hammer with a nail-puller; or a Swiss army knife). Translated into cuticular structure, the hinge areas have different amounts and orientation of chitin (the fibrous component), and the matrix proteins are chemically different from the stiff areas and so more hydrated and softer; the geometry of the hinge can be linear (for an intersegmental membrane) or circular (for a hair socket). Second, stiffness requires extensive cross-linking of the matrix protein, which militates against the use of the cuticle as a labile, resorbable chemical energy store (important for insects which feed only intermittently, such as Rhodnius prolixus, a blood-sucking bug). The resolution of this conflict is achieved by processes described by IP 2: Extraction—extract, isolate or remove an interfering or necessary part or property from an object. Its cuticular translation is to have a minimum of two layers of cuticle, the inner one being only partially stabilized and available for resorption. Since this layer is more likely to take loads in tension, its ability to resist compression is less important. Third, an external skeleton is a barrier to transmission of sensory information about the external environment, a function provided by sensory hairs and holes (the functional basis of the campaniform sensillum and slit sense organ). Note that translucent cuticle, needed over photoreceptors (IP 3), can still be cross-linked and stiff. Resolution here is achieved by the morphology of IP 31 Porous materials, make an object porous; use the pores to introduce a useful substance or function. Last, the animal will gain advantage if it recycles as much of the old cuticle as possible when synthesizing the new one at the moult, which stiffness will compromise since it requires extensive cross-linking. Larval and nymphal cuticles tend to be less cross-linked than adult cuticles, probably for this reason. Resolution is achieved in soft-bodied larvae by prestressing the material in tension to allow the structure to take compressive forces (i.e. a hydrostatic skeleton), which provides protection before the challenge and is described in IP 9 Prior counteraction.

In order to compare these biological resolutions of a design conflict with those which technology would use, it is necessary to convert the functions identified in the cuticle into the conflict topics that TRIZ recognizes. For instance, the functions change stiffness, protection, soft cuticle and stiff skeleton are all reduced to conflict number 11, which is defined as stress or pressure (compression, tension or bending). Similarly, keep poison out, self cleaning, surface properties and waterproof all become conflict number 30, which is external harm affects the object. The conflicting functions are similarly classified into the standard TRIZ features, which now allow the conflicts to be treated in the standard TRIZ system (Vincent & Mann 2002) and a direct comparison to be made between technical and biological solutions to the same problem.

One outcome of this study is that biology and technology solve problems in design in rather different ways. Apart from similarities in spectral filtering which allow the cuticle to let visible light through to photoreceptors, yet resist damaging UV radiation (a resolution which we would resolve in the same way that the insect does), most of the functions of cuticle are provided by detailed control of properties over a very short distance at a chemical and morphological level, summarized in IP 3. The TRIZ matrix derived from technology reveals that we tend to use a rather blunter, more global approach. This is illustrated by the fact that IP 35 is the commonest resolution in the TRIZ matrix, which involves changing a parameter, such as temperature.

TRIZ was conceived in and derived from the environment of things artificial, non-living, technical and engineering. But biomimetics operates across the border between living and non-living systems. And since the reason for looking to nature for solutions is to enhance technical functions, it is necessarily true that TRIZ does not contain many of these functions, and probably does not have the means of deriving them. Despite the fact that TRIZ is the most promising system for biomimetics, we still have a mismatch. This is conflated by a number of factors that are currently not normally observed in a technical system. For instance, the more closely an artificial system is modelled on a living prototype, which is typically complex and hierarchical, the more frequently we have emergent effects, which are unpredictable, therefore mostly unexpected and often harmful. Furthermore, one of the basic features of living systems is the appearance of autonomy or independence of action, with a degree of unexpectedness directly related to the complexity of the living system. This gives living systems great adaptability and versatility, but at the expense of the predictability of the system's behaviour by an external observer. In general, we do not accept unpredictability in technical systems; indeed, we avoid it. But we need to consider this even in our current technology, since nearly every technical system is actually a combination of a technical system in the narrow sense, and a living (usually human) system which is the operator of this technical system. This immediately suggests a broader and more general definition of the term technical system—a biological system, part of the functions of which is delegated to a device that is mostly artificial and/or non-living. This definition includes agriculture. This consideration is commonly omitted; technical systems are often considered in isolation, neglecting any broader context despite the fact that engineering is really a subset of human behaviour. At best this can lead to reduced effectiveness, at worst it produces technological catastrophes and/or social tension and unrest.

Another TRIZ concept of which we make much use is the System Operator or ‘9 Windows’. It is commonly depicted as nine squares arranged 3×3, horizontally representing time (‘before’, ‘now’ and ‘after’) and vertically representing size or hierarchy. This allows us to regularize levels of hierarchy above and below almost any object being studied that can be in a different condition (growth, death, attainment of a goal and so on) before and after the time at which they are currently being considered. In biology, the hierarchical levels are organelle, cell, tissue, organ, organism, population, ecosystem; the hierarchical level of the object under scrutiny is always referred to as the System. The super-System represents the assemblage of which the System is apart (for a cell it is the tissue; for an organism it is the population) and the sub-System represents one of the components of the System (for a cell it is an organelle; for an organism it is an organ).

We can now compare the types of solution to particular pairs of conflicts which are arrived at in technology via classical TRIZ, and in biology. Although the problems commonly are very similar, the inventive principles that nature and technologies use to solve problems can be very different (tables 1 and ​and22).

In fact, the similarity between the TRIZ and BioTRIZ matrices is only 0.12, where identity is represented by 1. Only the principles of spatial composition are significantly similar (0.73) in biology and technology. The differences are in large part to do with the pervasive presence of hierarchy in biological structures and systems. But they are also to do with the degree of detail it is possible to incorporate into a structure which, like an organism, is self-assembled and even designed by the forces of molecular interaction (Vincent 1999, 2005).

Hierarchy is exceedingly important in the solution of problems (see above); this is not obvious because our current technologies are either not significantly hierarchical, or ignore any hierarchical structure. However, a strictly scalar approach is difficult since many basic biological functions occur in organisms over a very wide range of sizes. Thus, the basic processes of cellular metabolism are more or less invariant from protista to large mammals; complexity and added functionality are achieved by adding levels of hierarchy which can be quantified quasistatically by considering the number of cell types in an organism (Bonner 1965), or in a more dynamic way by considering the provision of infrastructure (West & Brown 2005). The hierarchical approach is also difficult to use, since in engineering (taken here to represent the entire spectrum of effects which people impose on the world in an attempt to make it more habitable) hierarchy is not as well developed as in biology. The definition of cause and effect needs clarification: in physical terms, ambient pressure (cause) dictates the temperature of the boiling point of pure water (effect); in biological terms, an effect is a problem to be solved by the organism and the cause is the method of solution. Changing the boiling temperature invokes the operation field energy as does changing pressure. These are at the substance level of hierarchy. At the molecular level of hierarchy we would say, ‘the speed at which molecules move in a liquid (operation field is time) depends on the pressure (operation field is energy) and temperature (energy) which we apply to them.’ Appendix 3 of the electronic supplementary material illustrates our classification of effects.

We can now comment quantitatively on the differences between biology and technology (figures 2 and ​and3)3) At size levels of up to 1m, where most technology is sited, the most important variable for the solution of a problem is manipulation of energy usage (up to 60% of the time), closely followed by use of material (figure 2). Thus, faced with an engineering problem, our tendency is to achieve a solution by changing the amount or type of the material or changing (usually increasing) the energy requirement. But in biology the most important variables for the solution of problems at these scales are information and space (figure 3). This can be illustrated by comparing the functionality of biological and man-made polymers, proteins and polysaccharides. People have produced over 300 polymers, but none of them is as versatile or responsive as these two biological polymers. For example, at the primary level, proteins are remarkably similar in the energy required for their synthesis since the peptide bond is the pervasive motif. However, there is a wide range varying from inert fibrous (such as collagen or silk) to responsive fibrous (such as muscle) and from inert globular (such as skeletal proteins in insects) to responsive globular (such as enzymes). The difference between these proteins is less to do with the energy required for their synthesis than the complement and order of the amino acids, which is a derivation of information stored in the DNA of the nucleus or elsewhere. Space is also relevant, since the shape of the protein is an essential part of its function. The same story pertains when considering biological hard tissues. Calcium, and less commonly silicon, derivatives are commonest, and carbonates and phosphates are predominant. This limited range of chemicals (with the occasional addition of iron, zinc or manganese) suffices for nearly all biological hard materials. In insect cuticle, the main variety of function is achieved by making complex composites with anything up to 10 constituents and a range of properties (like Young's modulus), thus covering several orders of magnitude with a single material (Vincent & Wegst 2004).

This comparison between the few materials of biology and the many materials of technology has been made commonly, but never explained functionally. We can now do this. It appears that biological systems have developed relatively few synthetic processes at low size at which the contribution of energy is significant; but the main variety of function is achieved by manipulations of shape and combinations of materials at larger sizes achieved by high levels of hierarchy, where energy is not an issue. This is a very subtle biomimetic lesson. Instead of developing new materials each time we want new functionality, we should be adapting and combining the materials we already have. Obviously, we are doing this to an extent, but it is unclear whether we recognize this as a significant route rather than a route of convenience.

The data from technology and biology present a continuum of variables and contradictions at different levels of complexity—from a cellular organelle to an ecosystem, from a single transistor in a microcontroller to a fleet of aircraft. The biological data have to be structured into a framework that is compatible with technology to operate with this large amount of very varied information. Initially, we designed auxiliary conflict matrices for biological structures and environments, and for causes and limits of actions. These allow us to break natural data into engineering-like chunks of information and cover the primary TRIZ components of ‘function’, ‘effect’ and ‘conflict’. The matrix we have developed takes account of: