Minimal Signal Length

Counter selection on call length comes from two primary sources. First, energy invested in the production of the call may be a trade-off with energy required for current or future mating, for example the replenishment of costly accessory gland products such as the spermatophylax of many ensiferan Orthoptera (Gwynne, 1984; Thornhill and Gwynne, 1986; Simmons et al., 1999). Second, the risks incurred through predation from bats (Belwood, 1990) and parasitoid flies (Cade, 1979; Heller, 1992; Allen, 1995; Rotenberry et al., 1996) may offer selection to minimise the exposure of the advertising signal. In both cases there will be strong selection on the calling sex, which is invariably the male. But at least theoretically, while female preference may be for higher than mean values of the species' signal, such selection invariably holds the male call beneath these values (Kotiaho et al., 1998).

Evidence of a trade-off among tettigoniids between metabolic costs of sound production and mating effort is invariably dependent on both postemergence development, usually referred to as the refractory period, and food availability and quality (Simmons and Gwynne, 1991). For most insects there is a time between emergence and calling where spermatogenesis and the development of accessory gland material represent a premium over calling (Simmons, 2001). During this time, the propensity to call and the call length are usually shorter than when males are fully mature. In extreme cases, such as Kawanaphila nartee (Zaprochilinae), where calling and searching roles are reversed through a lack of nutrients within the habitat, females may compete for males and during this time males may provide the most minimal of signals (Gwynne and Bailey, 1999).

Are metabolic costs involved in changing signal components such as length, type or loudness? While Prestwich et al. (1989) suggest that the metabolic costs of calling are trivial for insects, particularly compared with frogs where energy conversion to acoustic power is less efficient, there are a number of examples where food-limited insects call less and with signals of lower amplitude. For example food-deprived bushcrickets, such as Requena verticalis, call less following mating, with the production of a spermatophore, and are influenced by the availability of food (Bailey et al., 1993). Ephippiger ephippiger males on a lowered diet produce fewer chirps per day and at significantly reduced amplitude compared with males on a normal diet (Ritchie et al., 1998). Similarly, Gryllus lineaticeps on high diet tend to sing more than their siblings on lower diets (Wagner and Hoback, 1999). One assumption is that insects can replenish their food resources with ease, and while this may be so for species in nutrient-rich environments, other species are specialist feeders and may be restricted in diet by volume and quality {Requena verticalis; Simmons et al., 1992).

The second selection pressure is predation. Here it is assumed that the call length will decrease with rising incidence of predation. The most thoroughly researched predation event is that from parasitoid flies belonging to the Tachinidae. Commonly, ormine flies locate the calling insect and either lay eggs or larvae on or close to the host (Cade, 1975, 1979; Allen, 1998; Zuk et al., 1998; Gray and Cade, 1999; Allen and Hunt, 2000). Indirect evidence can be inferred from the current calling activity of closely related species. For example, Heller (1992) suggested that the difference between searching and calling roles of two Mediterranean bushcrickets is a response to predation history, in part through parasitoid flies. More direct evidence comes from changes in call structure and female preference function in the Australian cricket, Teleogryllus oceanicus. In regions where ormine flies are abundant, not only is calling time altered, but male call length is decreased (Zuk et al., 1993a; Rotenberry et al., 1996).

Evidence for a reduction in call length is more indirect for predation by bats, and here most evidence relates to tropical bushcrickets. The large picture in regard to ensiferan Orthoptera suggests that there are many more short or intermittent calling species in tropical rain forests inhabited by foliage gleaning bats compared to more temperate or open habitats (Rentz, 1975; Belwood, 1990; Morris et al., 1994). Foliage gleaning bats are more attracted to mist-nets loaded with calling bushcrickets than nets without bushcrickets (Belwood and Morris, 1987). The effects of call length were tested by attracting foliage gleaning bats to different species producing different calls rates. Bats were more attracted to a species producing a call rate of 60/min compared with one calling infrequently at less that one call per minute (Belwood and Morris, 1987). Gleaning bats also prefer the long calls of crickets (Bailey and Haythornthwaite, 1998).

Belwood (1990) made the observation that many nocturnal species of tropical bushcricket call within a very narrow time window, with some calling during the early hours of the morning, possibly to escape the flight times of bats. By comparison, species that are continuous callers are more frequent in open grass areas where bat activity is far less (e.g. Neoconocephalus (Greenfield and Roizen, 1993)). Interestingly, the continuously calling forest-dwelling pseudophylline, Ischnomela pulchripennis, calls from refuges such as bromeliads where bats are unable to intrude (Belwood, 1990). Perhaps the most compelling evidence for call structure change is the switch by tropical forest bushcricket species from acoustic to vibratory signalling strategies. Many species combine airborne high frequency signals with substrate tremulation (Belwood, 1990; Morris et al., 1994) and, in the pseudophylline species Docidocerus gigliotosi, the call is extremely brief, consisting of a diplosyllable while the tremulation signal consists of a long series of syllables (Belwood and Morris, 1987; Belwood, 1990).

There exists a minimum signal length for information transfer. Once the signal reaches minimal values the call may no longer achieve the signaller's two primary goals of identity and location. (See Beecher (1989), Bradbury and Vehrencamp (1998), and Gerhardt and Huber (2002) for a broader discussion of the selective forces on acoustic signals that provide information for recognition.) Where signals fall into this category, the signaller's relative fitness will presumably increase with higher chances of mutuality and decrease where there is deceit, or where alternative tactics are employed. Further, the signaller's fitness will increase with its ability to eavesdrop or use its signal in an aggressive or spiteful manner (Wiley, 1993). Critical, however, in any consideration on signal effectiveness is the decision by the receiver to respond, which is often considered as its internal state. But such a decision may be influenced by more complex factors such as exposure to predation during searching or phonotaxis and also the ability to locate the sender. Gerhardt and Huber (2002) summarise the situation: "if individuals cannot be distinguished statistically by their signals, then in the long run there will be no benefit in investing time and energy in discriminating between signals, the signal differences (must) reliably reflect on average differences in state". Further, the response and recognition of signals is compounded by memory of previous events, including signals (Greenfield et al., 1989). While this is frequently considered for vertebrate signallers, there has been little experimental evidence that this may be important for insects.

The notion of acoustic memory in insects is a discussion beyond the scope of this chapter and such events have received surprisingly little attention, yet with intermittent callers or those producing extremely short signals, spatial memory would seem intuitive. There is simply not sufficient signal to locate one's mate in a complex habitat. Interestingly, there have been two recent papers that identify behaviour patterns of acoustically searching females of nocturnal species that are linked with visual patterns associated with the caller (Helversen and Wendler, 2000; Bailey et al., 2003). The evidence so far strongly favours the idea that visual cues are an integral part of the phonotaxis process for nocturnal insects.

One way to view a theoretically minimum signal length is to consider the ability of an insect to resolve repeated patterns. For example, lower limits of pattern resolution have been described for the grasshopper, Chortippus bigutulus, which produces distinctly structured sounds from each leg, but the combined stridulation of both legs results in a near continuous noise interrupted by gaps (Helversen, 1984). The behavioural and neural story of each sex's response to conspecific sounds is far from clear, and while the shape of each pulse is important, females fail to respond to songs if gaps of longer than 2 to 4 msec are inserted. These are presumably close to a recognition threshold (Ronacher and Stumpner, 1988; Stumpner et al., 1991). Evidence at a neural level, however, shows that neurons such as AN12 produce bursts with song pause durations of 20 to 25 msec and are as influenced by syllable duration. For many copiphorine bushcrickets, the calls consist of an almost continuous series of syllables produced by the wing's closure, and in some species the inter-syllable interval can be as low as 5 msec. For example, in extremely fast callers such as Neoconocephalus robustus, the syllable repetition rate is close to 150/sec, producing an interval between syllables of 7 msec (Josephson and Halverson, 1971). In the genus Hexacentrus, repetition rates reach over 200/sec (Heller, 1986), suggesting an interval resolution lower than 5 msec.

An initial survey of calls less than 0.5 sec of bushcrickets fails to provide any convincing pattern in regard to minimum syllable of pulse interval, largely because there is often extreme variation within one taxon, and some species produce near pure tone calls, which may provide as much information to the receiver as the interval between the syllables. The expectation was that there would be a lower call interval that matched the theoretically lower limit of "gap" resolution. However, two basic short signal strategies did emerge from this survey. First, males may produce a single syllable of rising amplitude formed by a single wing closure (Figure 8.4a) and, while the action of the plectrum on the file creates a series of internal tooth impacts, these impacts are not probably resolvable by the insect nervous system as they are below the theoretical limit provided by other studies. In this case the signal ends with a sharp and final loud pulse of sound; the receiver is warned of the oncoming signal by the rising phase and there is then an opportunity for recognition with final louder pulse. The signal may be repeated several seconds later and the interval between repetitions may be random; essentially the element is a syllable with a defined species' shape. The key feature of the signal is that the receiver is set up for the explosive end of the syllable.

By contrast, other species may produce distinct syllables in pairs or triplets (Figure 8.4b). In these examples, the initial signal warns the receiver that the caller is on-air and the interval between this and the proceeding signal has the capacity to provide a species' distinct signal; the latency of the diplosyllable contains temporal information. Tropical species that are presumed to be under selection from predation by foliage gleaning bats are more likely to have such extreme calling strategies. Morris et al. (1994) illustrate a number of cases that fit this style of calling. For example the pseudophylline Myopophyllum speciosum produces a doublet sound of extremely high carrier

Was this article helpful?

0 0

Post a comment