[Ed. note: A pioneer in the study of animal behavior and the founder of Rockefeller University’s noted centerfor ethology, Donald Griffin was one of the first scientists to challenge the dogma that animals are mindless automatons, controlled solely by instinct and reflex. “The flexibility and appropriateness of animal behavior suggest both that complex processes occur in their brains, and that these events may have much in common with our own conscious mental experiences,” he wrote in a 1976 book, The Question of Animal Awareness (Rockefeller University Press).
Griffin’s own research has ranged from probing the mysteries of bird migration to analyzing the underwater hearing capabilities of fish. But he is best known for his discovery of echolocation, a phenomenon he named back in 1944. Now Professor Emeritus at Rockefeller University and a visiting lecturer at Princeton University, Griffin looks back on the experiments that led to the discovery]
It is now well known that bats use sonar, but the refinement and acuity of this unfamiliar mode of perception, and the fits and starts by which biologists learned about it, are not generally appreciated. As an undergraduate student of biology at Harvard in the 1930s, I banded hundreds of bats to study their migrations, and it was most impressive to watch them flying through pitch-dark caves without touching stalactites or walls. At the time, however, no one suspected that they relied on sonar.
In 1793, Lazzaro Spallanzani had shown that blinded bats fly quite normally, and the accepted explanation was that they feel obstacles through tactile sense organs on the skin and wing membranes. This seemed plausible, for I had often observed them flying quite close to an obstacle before turning to avoid it, and I could feel the air currents from their wingbeats as they flew within a few inches of my face. But I wondered if that was the full story, maybe sound played a role.
Bats make lots of high-pitched noises, such as the scratchy chittering they emit when they are disturbed. And although the hats I observed seemed to be completely silent when flying through dark caves, perhaps their sounds were above the range of human hearing. So I was intrigued when I learned that a Harvard physics professor had an apparatus that would detect such sounds. I almost let the matter drop, however. Without some prodding from interested friends, I might easily have made one more trip to some cool mossy cave in Vermont instead of knocking on his laboratory door with a cage full of bats.
George Washington Pierce was a jolly and slightly chubby fellow who reminded me vaguely of a Santa Claus without beard or costume. He welcomed me cordially and immediately introduced me to his “sonic receiver,” which he was using to study the sounds of insects. This consisted of a parabolic horn with a Rochelle salt microphone at its focus, vacuum tube amplifiers, and a heterodyne receiver tunable from about 10 to 150 kHz. The apparatus reduced the pitch of high-frequency sounds so that they could be heard using a loudspeaker.
As the hats crawled about in their wire mesh cage, the loudspeaker popped and rattled impressively. And when I held one in front of the parabolic horn, the apparatus continued to click and rattle. But to our keen disappointment, the loudspeaker fell silent when the hats flew around Pierce’s laboratory Equally disappointing, we saw no evidence that the inaudible sounds we had discovered played any role in obstacle avoidance. I was tempted to turn to more hat banding, or to study harder for my final exams.
Yet these inaudible sounds were too intriguing to ignore. One of my fellow students, Robert Galambos, was especially interested because his thesis research was on the cochlear microphonic potentials that had been recently discovered by Hallowell Davis at the Harvard Medical School. Pierce was graciously willing to lend us the only apparatus in existence that could generate controlled sounds above the frequency range of human hearing. Unfortunately, we did not realize that the ordinary-looking power outlets in the medical school supplied DC and not AC power. Therefore the first experiment yielded a puff of smoke and a serious diplomatic problem. Bob was sure that he could replace the power transformer, but I still had to break the news to Professor Pierce. Fortunately, he was understanding, and with the repaired apparatus, Bob discovered that bat cochleae generate cochlear microphonics up to about 100 kHz. We also found that if we held Pierce's microphone directly in front of a flying bat, high-frequency sounds were easily and consistently detectable. Pierce and I had not realized that his apparatus could detect bats only when they flew toward the microphone, because of the directionality of both the parabolic horn and the bats emitted beams. In addition, we found that blocking either hearing or sound emission caused bats to bump into obstacles. In 1940 these were surprising results. One distinguished physiologist was so shocked by our presentation at a scientific meeting that he seized Bob by the shoulders and shook him while expostulating, “You can’t really mean that!”
After the war I analyzed the sounds used by bats more accurately than had been possible with Pierce’s original apparatus. I also found that bats could detect tiny objects, such as wires one-fifth of millimeter in diameter. Mean while, the notion of sensing objects by means of reflected signals became more widely accepted as so nar and radar escaped into public awareness from behind the veil of wartime military secrecy.
Furthermore, three psychologists—Supa, Cotzin, and Dallenbach—showed in 1944 that many blind people detect obstacles by making sounds and hearing reflections or changes in these sounds. This led me to recognize that there is an important general process, which I suggested be called echolocation, by which animals locate objects that they cannot see or touch by emitting signals and analyzing the returning echoes.
At the time, echolocation seemed to be merely a collision warning system. But I wondered whether bats could tell anything more than “There’s something out there.” After trying unsuccessfully to devise laboratory experiments to measure how well bats could distinguish between different objects, I turned to simple observations of wild bats hunting insects. I wondered vaguely whether echolocation played a role in hunting, even though the prevailing view was that bats either saw their prey, or listened for the sounds of wingbeats and located the insects by passive hearing.
When I set up my apparatus beside a small pond where bats were actively hunting, it was a dramatic surprise to see on the oscilloscope screen how many ultrasonic sounds filled the air. When bats fly up to obstacles in the laboratory they roughly double the repetition rate of their orientation sounds. But when bats are pursuing flying insects the repetition rate increases tenfold or more, so that the bats’ audible translation changes from a steady “put-put-put” to a staccato buzz of pulses at rates up to 200 per second.
My observations strongly suggested that bat sonar was being used to guide the rapid and complex maneuvers during insect pursuit and capture. But my attempts to induce bats to catch insects in the laboratory were a total failure. I would release nets full ofinsects in a room along with one or more bats, but despite everything I tried, no bat paid the slightest attention. On hearing of my dilemma, Eric Tetens-Ni, a student of mosquito behavior, invited me to bring bats to his laboratory flight chamber. I flew to Florida, taking with me a few freshly caught bats in a specially modified briefcase. There, we found that dense swarms of mosquitoes elicited active insect pursuit within a few minutes. Then in later experiments, Frederic Webster and I were able to show that bats could catch fruit flies just as rapidly in the dark or in the presence of loud audio frequency noise. But in the presence of ultrasonic noise, they stopped hunting immediately.
The fruit flies landed on the walls and had to be brushed into the air at frequent intervals to provide good hunting for the bats. To avoid this chore, I thought up a brilliant scheme. I arranged a vertical air jet using from a shallow conical funnel of sheet metal. When dozens of freshly frozen fruit flies were dumped into the jet, they were blown up four or five feet. They then fell back into the funnel only to be blown up again to form a continuous “Drosophila fountain.” I was sure that the bats would find this excellent hunting since I knew that they ate freshly killed insects. But they fooled me once again by totally ignoring the dead fruit flies. At the time, this was just one more unexplained frustration.
Eventually, Webster was successful in inducing bats to catch mealworms tossed up to them from the floor, and we began to study whether bats can discriminate between targets of roughly similar size. In a series of tosses, we randomly interspersed inedible targets with the mealworms. At first the bats would pursue and catch anything, just as wild bats chase pebbles thrown up in front of them. Webster was especially proud of his flash photograph showing a bat with its wings-wrapped around a tennis ball. But after a few days of this game, the bats learned that only some of the targets were edible, and we were able to begin quantitative experiments.
Mealworms and flying insects vary greatly in the area they present to impinging sound waves and hence in the strength of the 20 to 30 echoes per second that return to an approaching bat. I was beginning to wonder whether the bats might have some ability to detect differences among flying insects, and we therefore extended the experiments in include two sizes of disks, one smaller and the other larger than mealworms. Both disks and mealworms turn and tumble when tossed into the air, so that the successive echoes from each type of target can vary by a hundredfold or more. Yet the most proficient bats caught the great majority of the mealworms but turned away from almost all the disks.
It thus gradually became clear that bat echolocation is much more than a collision warning system. Recent elaborations of these experiments have demonstrated that bats can even distinguish between different kinds of flying insects. This may explain the failure of my “Drosophila fountain”; since the dead fruit flies had their wings folded, their echoes differed from those of real live insects.
When I first began to study fly. ing bats, I had no idea how capable these little animals would prove to be. At each stage in these investigations, I asked rather limited questions, and was somewhat reluctant even to contemplate what proved later to be the logical next steps. Animals do not perform miracles, but some of their capabilities would have seemed magical bad anyone ventured to suggest them 50 years ago.