The new methodology proposed by Ramsey et al. (2007) has implications for studies in areas beyond the ones they highlight. The majority of their examples, and those in the cited literature, are concerned with foraging methods, or coping with novel or changeable environments. It is striking that one well-studied branch of behaviour is underrepresented. Courtship displays, usually exhibited by males, provide a basis for female choice and offer some of the most elaborate behaviours seen or heard in animals (Anderson Reference Anderson1994). Whilst many displays are morphologically fixed (long tails or bright colours) and likely strongly influenced early in life by genetic factors, others are behaviourally flexible, and candidates for innovation. Innovation may occur in the display form (e.g., song structure, movements, use of objects) and in female preferences (mate choice copying, alternative exploitation of a pre-existing bias). Two examples of sexual display in which innovations are possible are bird song (Marler & Slabbekoorn Reference Marler and Slabberkoorn2004) and bowerbird's bowers (Frith & Frith Reference Frith and Frith2004).

Why are innovations in sexual display of interest? First, the variety of elaborate sexual displays still demands an explanation. Innovation, coupled with learning (ten Cate & Rowe Reference ten and Rowe2007), provides an additional mechanism to traditional genetic mechanisms of inheritance to explain the rapid appearance and divergence of displays. If this leads to mating isolation of certain populations, then it may also inform the understanding of speciation.

Second, sexual display provides an unusual situation where selection favours innovation per se. Females may prefer innovative males, regardless of the exact form of display exhibited, favouring males that contrast with peers (Burley & Symanski Reference Burley and Symanski1998; Elias et al. Reference Elias, Hebets and Hoy2006), perhaps because unusual males are less likely to be relatives, or because innovative males possess a general cognitive ability or can bear a costly brain (Miller Reference Miller2000). Males exhibiting exaggerated displays, such as large song repertoires, have high mating success (Kroodsma Reference Kroodsma, Marler and Slabberkoorn2004) and one way to extend the repertoire is by generating novel syllables or phrases (Johnson Reference Johnson2006). These become available to other males to copy and incorporate (Slaabekoorn et al. 2003).

Third, the apparent decoupling of reward for innovation (mating or reproductive success) from the display itself provides a useful system to study how individuals decide whether to retain innovations. Sexual displays differ from the exemplar behaviours given by Ramsey et al. (2007). In most cases, such behaviours produce an immediate reward (e.g., food becomes available). A close link between behaviour and reward will likely reinforce the innovative behaviour. Additionally, the reward can be perceived by observers, offering a stimulus for copying (Caldwell & Whiten Reference Caldwell and Whiten2003). Conversely, rewards for producing an innovative display may be long delayed and concealed from observers. Females can continue to sample males and only subsequently return to a preferred novel male (Uy et al. Reference Uy, Patricelli and Borgia2001). Males may be unsure of their reward in terms of reproductive success, especially if they provide no parental care. Feedback following innovation demands further investigation.

What evidence is there that displays could have originated as innovations? There exist similar types of evidence as for non-display behaviours, such as novel foraging techniques, but currently in smaller quantities. First, individuals may spontaneously produce song syllables (Johnson Reference Johnson2006), or incorporate unusual objects into their display (Frith & Frith Reference Frith and Frith2004), only some of which they retain in subsequent displays. Second, displays may not be genetically fixed, but are refined as the individual ages (Gil et al. Reference Gil, Cobb and Slater2001; Vellenga Reference Vellenga1986). If innovative individuals are copied, then their innovations spread in the population. Finally, populations differ in the displays that males give, independent of environmental or genetic factors, and match patterns of male association (Madden et al. Reference Madden, Lowe, Fuller, Dasmahapatra and Coe2004; Marler & Tamura Reference Marler and Tamura1964; Uy & Borgia Reference Uy and Borgia2000).

Innovations in male display may be initiated in response to innovative female choice. If females copy the mate choices of individuals (e.g., Swaddle et al. Reference Swaddle, Cathey, Correll and Hodkinson2005), and make “errors” in learning – for example, learning to target an incorrect set of cues, or by copying the “wrong” individual – then a new pattern of preferences becomes available for males to exploit using innovative displays (ten Cate & Rowe Reference ten and Rowe2007). As long as the choice has neutral fitness consequences and the erroneous female is copied by others, runaway processes could rapidly favor the expression of an innovative display coupled with an innovative preference (Laland Reference Laland1994; but see Kirkpatrick & Dugatkin Reference Kirkpatrick and Dugatkin1994). Alternatively, female choice can facilitate innovative displays if her mating decision is based on a pre-existing physiological, neurological, or psychological bias (Endler & Basolo Reference Endler and Basolo1998). If the bias is generalised (e.g., colour), males potentially have numerous ways of stimulating it. For example, a bias for specific colours derived in a foraging context, is stimulated by male morphological colour patches that match that preference (Rodd et al. Reference Rodd, Hughes, Grether and Baril2002). Innovation can drive the incorporation of novel colour-matched objects into a male's display (Madden & Tanner 2003).

Why has little attention been paid to innovation in sexual display? First, innovation in display may be rare, with strong selection against novel displays that are uninformative or noisy. Second, maybe displays are not really innovations. All sexual signals ultimately serve the single purpose of attracting a mate, and so at one level, the behaviour has not changed and no new behaviour has been added to the individual's repertoire. However, by comparison with traditional innovations, there are many ways to crack a nut and these alternative solutions are classed as innovations. In the case of displays, various novel objects could stimulate the same sensory biases exhibited by females, providing scope for innovation in display form, but not its ultimate function. Third, some variation in display form between populations has been explained by environmental (Handford Reference Handford1981, Madden & Balmford Reference Madden and Balmford2004) or genetic (Baker et al. Reference Baker, Thompson, Sherman, Cunningham and Tomback1982) factors. Correlations with differences in these factors may be accepted as explanations for the causes of difference, but it is also possible that random innovative change in display could prompt certain males to favour a particular environment in which their display is most conspicuous, or to mate assortatively with certain females, leading to genetic similarities making current genetic differences meaningless when considering origins of innovations. Finally, innovations in display may be very subtle – a novel syllable interjected in a song of many different syllables, or the incorporation of one novel object on a bower already hosting a hundred different object types, providing a daunting task to scientists. The new methodology of Ramsey et al. (2007) addresses this and offers a means of rigorously assessing innovations in sexual display, and so permits expansion of the study of innovations into this novel and potentially fruitful field.