Ivory Towers: Music to a grasshopper's ears

SERENADING females is a common courtship ritual in many species, but how long is it advisable to go on singing? That important question is answered in 'Song duration and female response behaviour in the grasshopper Omocestus viridulus' by Thorleifur Eiriksson (Animal Behaviour, 1994, 47, 707-12). The conclusion is that, for grasshoppers at least, it pays not to go on singing too long.

The grasshoppers, Omocestus viridulus, used in this experiment, have a characteristic calling song of between 10 and 25 seconds. They are closely related to Omocestus rufipes, which has a very similar singing system, but markedly shorter song. If females use the song to differentiate between the two species, one would expect thae lady viridulus to go for the longest arias, while Mrs rufipes likes them as short as possible.

For the experiment, they enlisted the help of the computer from the Linguistics Department of Stockholm University to manipulate song length. By identifying a single pulse (one complete movement of the leg rubbed against the wing), electronically isolating and repeating it as often as required, songs of varying length were created, to be played to the experimental females.

Each of 20 females was placed in a grass-covered box in a sound-attenuated room then, after a period of acclimatisation, played a variety of songs, of between 3.75 and 60 seconds' duration. There was a 90-minute rest period between each song and the next. Times were noted when the female responded by walking or jumping towards the sound, or by rubbing her own legs in a responding song of her own.

The experiments showed that while all females had responded by the time a 60- second burst was finished, a 3.75-second call was enough for 60 per cent of them on the first trial and 87 per cent on the second. The female's song is characteristically shorter and softer than the male's, despite the fact that 'singing is probably very costly for the male as very little of the energy used to produce sound is transferred to acoustical energy'.

Also, if the male goes on singing too long, he may miss the female's response, as she will not wait for him to stop before she sings herself. Competing, quiet males may easily encounter the female before the original singer has noticed her.

The results showed no difference between long songs and repeated short songs, thus suggesting that species identification plays little part in the procedure. The author concludes: 'A male seems to evoke most response by singing as much as possible. However, to maximise the efficiency of his singing, he must also time his songs to decrease the overlap with the female's response singing, and increase the probability of hearing her response.'