The cricket in hand is smaller than a prune pit.
Robert Wyttenbach and his colleagues dab the cricket's back with a spot of glue, then attach him to a thin wire in their cricket flight chamber.
A fan is turned on. In the dim light the male cricket feels the wind and flies, as crickets do at night. A computer attached to speakers simulates the sound of other crickets, and most importantly, the ultrasound of bats.
The experiment makes tiny invertebrate history. Dr. Wyttenbach and his colleagues in neurobiology and behavior research at Cornell University in Ithaca, N.Y., discovered recently how crickets make life-or-death "decisions" in flight.
The insects distinguish the ultrahigh-frequency sounds of echolocating bats from cricket chirps and thus avoid becoming a bat meal.
"When the cricket hears a sound," says Wyttenbach, a postdoctoral associate in neurobiology, "it faces a choice of going toward it, away, or ignoring it. This requires the need to classify that sensory input according to certain criteria. What frequency is the sound? High-pitched or low-pitched? What pattern is it? What species?"
At night, crickets hear a rapid-fire cacophony of sounds. On their forelegs are "ears." Prior research at Cornell by Ronald Hoy in 1978 established that crickets detect the bats' ultrasound, which is beyond the range of human hearing. But the ongoing research discovered how a cricket reaches the cutoff line between "attractive" and "repulsive" sounds.
The cricket decides this by "categorical perception," a sophisticated capability used by humans and other animals. Invertebrates, as suggested by the experiments, use it too.
In humans, categorical perception comes into play, for instance, when an infant hears the sound of "pa" repeated, and becomes habituated to it. The infant responds less and less until it hears a new sound "ba," and becomes attentive.
Wyttenbach discovered that crickets categorically label sounds at and above the range of 16 kilohertz (kHz) as perilous. Below that, all is well. But out of many night sounds, the bat zings a ba-like sound at 16 kHz. The insect takes evasive action.
"It's a stretch to say that the cricket is smarter than the bat," says Wyttenbach. "You don't have to be smart to outwit a prey detection system." He also says the cricket is not always successful.
"Smart is a human concept," says Don Lewis, an entomologist at Iowa State University in Ames. "Insects are not doing things in a logical way in the same fashion that we would think," he says. "Over hundreds and hundreds of generations they have developed to be what they are."
Wyttenbach believes the cricket makes a decision, but with a limitation. "The problem with saying it makes a decision," he says, "is that this carries all sorts of baggage about consciousness. You can be agnostic about whether crickets are conscious or not, and still say they decide. I say the cricket classifies the input and makes a decision, but without the cognitive baggage we associate with that term."
Categorical perception was initially thought to be unique to human speech, but experiments proved it was present in visual stimuli in humans regarding colors. Experiments with chinchillas and monkeys indicated they also use categorical perception.
Wyttenbach used crickets common to Polynesia, Australia, and Hawaii. They fly more readily than local crickets. "You can look at this in the broad historical view that a lot of insects evolved nocturnal behavior to escape daytime predators like birds," he says, "but the bats are a counter because they evolved to detect prey in the dark to circumvent that. Much goes on at night that we aren't aware of."
Male crickets make a sound - called stridulating - between 4 to 5 kHz, just above the highest note on a piano. They produce the sound by rubbing a rough forewing across the opposite wing with some 200 teeth. Small crickets stridulate as much as 10,000 cycles a second.
Crickets use several songs or signals at night, including a courtship signal and an aggressive signal. "There is even a sort of victory song that a male will sing after winning an encounter," says Wyttenbach. "It's probably a signal that he has this territory."
Mr. Lewis says the worth of this knowledge comes up often.
"The view that everything has to have a relationship to humans," he says, "is fairly self-centered and self-serving of us. Many things are happening out there with no impact on us. But that doesn't make them any less valid as components of an ecosystem."
Wyttenbach is now turning his attention to two other questions: how an animal can tell where a sound is coming from in the presence of echoes, and how frogs and crickets decide when to call. "There are several lifetimes of research left to be done on the cricket," he says.