WALKER VAN RIPER OF DENVER, Colo., is surely one of the most active amateur zoologists in the U. S. In the 14 years since his retirement from the investment-banking business he has concerned himself with, among other things, the way in which rattlesnakes strike (they do not bite; they stab) and many aspects of the behavior of hummingbirds. Accounts of his work with these animals, which has featured some remarkable high-speed photography, have appeared in SCIENTIFIC AMERICAN. Recently Van Riper has turned his attention to the question of how the hummingbird finds its way to the nectar on which it feeds; specifically, does it do so by means of the sense of smell? He writes:

"The avian sense of smell has received some notice from ornithologists, the classic experiment being one by John James Audubon on the turkey vulture. Audubon set out a stuffed deer; the vultures found it, pecked at its glass eyes and tore its skin, apparently ignoring the fact that it did not smell like food. When Audubon painted a dead deer on a large canvas and spread the canvas on the ground, the birds again sought it out. Then he hid a dead and putrid animal in a ravine over which the vultures regularly flew. They were not attracted to the bait. He concluded that these birds found their food by sight alone.

"The experiment was of considerable importance to Audubon; he reported it in Edinburgh in his maiden speech before a scientific audience. When he returned to the U. S., he found that his observations were questioned. He repeated the experiment before a group of educators and scientists, who thereupon certified that vultures used the sense of sight, and not that of smell, to find their food.

"A century later Victor Coles, a graduate student at Cornell University, repeated the experiment, with essentially the same result. In 1929 Frank M. Chapman, then the leading U. S. ornithologist, repeated one of Coles's experiments on Barro Colorado Island in the Canal Zone. Chapman hid foul-smelling food in various places; when this was found by vultures, he concluded that they must have responded to the odor. It has been noted, however, that Chapman's experiment is open to question because he did not exclude such visual clues as blowflies, beetles and similar insects that are known to find their food by smell. These could have led the keen-sighted vultures to the bait.

"A few experiments have been made on the sense of smell in other birds. Various odors have been tried as conditioning stimuli for pigeons, ducks and parakeets, mostly with negative results. In one series of experiments it was found that pigeons could not even distinguish water from a strong ammonia solution, except by dipping their beaks into the liquid.

"Similarly, in the course of extensive observations in Arizona on the behavior of the black-chinned hummingbird, the naturalist Frank Bene hid food bottles carrying an orange-scented honey solution in a flower bed. In no observed case did the birds find the food by means of its scent.

"A considerable amount of work has been done on the anatomy of the olfactory organs of birds by E. H. Craigie, R. M. Strong, C. L. Turner and others. Their findings indicate that most birds appear to have the apparatus necessary for smelling, and that it is more fully developed in some species than in others.

"The experiments so far mentioned throw some light on the question of whether or not the bird makes use of the sense of smell. But failure to observe the use of this sense under a given set of experimental conditions does not justify the conclusion that the sense is not used. Nor does the mere possession of olfactory equipment prove that an animal can or does make use of it. Considerations such as these mark important distinctions between the modern interpretation of animal behavior and older views. It must also be borne in mind that the behavior of birds differs from species to species. Pronounced variations occur even within strains of the same species. Chapman's vultures, for example, were forest-ranging birds to which the use of a sense of smell might well be more advantageous (and hence a more likely adaptation) than to the strain that inhabits open country.

"As modern students of animal behavior have made clear, what an animal can do does not necessarily throw light on what it does do. In this connection N. Tinbergen of the University of Oxford points out the following feature of instinctive behavior: An animal does not react to all the stimuli that its sense organs can perceive, but only to a small part of them. This applies to instinctive behavior, but not necessarily to learned behavior. The instinctive behavior for a given situation is thought of as a fixed, unlearned pattern of responses, the same for all individuals of the species, which is triggered or 'released' by certain definite stimuli received from the environment or induced by internal physiological changes. It is important to bear in mind that the triggering stimuli are definite, generally few in number and limited to only a small portion of the perceptions received by the animal.

"The carnivorous water beetle Dytiscus, for example, has highly developed compound eyes and can be trained to respond to visual stimuli. It does not react to the mere sight of moving prey such as a tadpole. When moving prey enclosed in a glass tube is displayed to Dytiscus, no pattern of behavior is released. Experiments show that the beetle's feeding response is never set off by the sense of sight but only by tactile and chemical stimuli. A dilute meat extract dripped into the water will induce the beetle to attack every solid object it touches.

"Similarly it has been found that a hen coming to the rescue of a chick in distress is reacting to the sound of the distress call, and not to the chick's movements. In one experiment the chick was tied behind a screen so that it could not be seen by the hen; the mother came to the rescue when she heard the chick's cry. But when the chick was subsequently put under a glass bell so that the hen could see it struggling but could not hear it, she remained entirely indifferent to its distress.

"The interesting phenomenon known as 'imprinting' is another example of the working of instinctive behavior and its releasers. The Austrian naturalist Konrad Z. Lorenz took half a clutch of eggs away from a duck and hatched them in an incubator. Subsequently he learned that if he made certain definite sounds and movements within a sharply limited period after the incubator ducklings hatched, they would adopt him as the mother and from then on follow him and react to him as their siblings did toward the real mother [see "'Imprinting' in Animals," by Eckhard H. Hess; SCIENTIFIC AMERICAN, March, 1958].

"When an egg falls out of the nest of a herring gull, it triggers in the brooding gull a series of maneuvers that results in the egg being moved back into the nest. Not so, however, with pigeons. I have observed them ignore eggs in plain view near the nest.

"For some years I have maintained a hummingbird feeding station in an area of rough pine and scrub oak on Cherokee Ranch, 20 miles south of Denver. The remarkable experiments of the Austrian zoologist Karl von Frisch on the senses and behavior of honeybees suggested a test of the sense of smell in hummingbirds. This sense has been shown to be useful to the bees, which, like hummers, feed on the nectar of flowers. Probably no other wild bird is so amenable to experiments of this sort as the hummingbird. Their indifference to the presence of man, together with the fact that a considerable number of them can be kept in a small area once they have learned to use artificial feeders, make them ideal subjects.

"The food dispenser I am now using is shown in the accompanying photograph. It is made of a bottle inverted over a plastic container. Attached to the lid of the container is a wire perch. The birds feed through four 1/8-inch holes in the lid. This arrangement excludes honeybees, wasps and other insects, as well as other birds. The outgo of the sugar-water solution is easily measured and recorded. For this experiment I installed a new set of feeders, taking pains to make them exactly alike in every detail. An inverted half-ounce vial stuffed with cotton was attached to the side of each bottle to dispense various scents.

"On May 12 the first birds were seen, all males. Seven feeders were distributed over an area about half a mile long and 100 yards wide. Several drops of a strong flower scent, rose-geranium, were added to the scent vial of each feeder. This scent, renewed from time to time, was always associated with the feeders containing syrup. By May 25 the females had arrived; the consumption of syrup indicated that some 35 birds were using the feeders.

"On May 28 I removed all the feeders. Then I hung, on the lower branches of a large pine, six feeders, three made with syrup and scent and three with plain water and no scent. Observation gave no indication that the birds had learned to associate syrup and scent. The seven regular feeders were now replaced in their original locations for four days. Then, on June 2, I put up a feeder filled with plain water and bearing no scent 10 feet away from each of the regular feeders. The positions of these pairs were interchanged every other day.

"At this point a special testing rack was installed. It was made from parts of a television antenna and consists of four five-foot arms fastened at right angles to the top of a one-foot tube. The tube slips over and is supported by the top of a six-foot 3/8-inch pipe set in the ground as shown in the accompanying photograph [above]. A feeder was attached to the end of each arm. The feeder positions could be changed by pulling a rope tied to one arm. This rack and its associated feeders was substituted for all other feeders in the area. Two syrup-filled, scented feeders were hung on opposite arms of the rack and two water-filled, nonscented feeders on the other arms. Observations were made every other day, on the average, usually for a two-hour period.

"For each bird that visited the rack three things were recorded: (1) sex, (2) type of bottle visited first and (3) behavior with respect to the nonfood bottles. Birds that proceeded directly to food bottles were allowed to feed, but not to satisfy their hunger fully. After an estimated half-feeding, the bird was disturbed by revolving the rack, and a nonfood bottle was located where the food bottle had been. Most birds returned to the rack immediately. Their behavior could thus be recorded a second time.

"Because all the feeders were identical, the apparatus gave the birds no visual clue to the dispensers that contained syrup. Positional clues were eliminated by frequent movements of the rack. Bees and wasps do not frequent the area because it provides no natural food for them; hence the experiment was not hampered by insect clues. The only differences between food and nonfood bottles were those detectable either by smell or by taste.

"The first question I had sought to answer by these experiments was: Did the birds learn to go directly to food bottles by scent alone? The answer was provided during the training period from May 12 to June 8, and it was 'no.' Visits to food and nonfood bottles were about equal in number.

"It became apparent with the first observations that a bird coming in search of food would normally light on the first feeder in its line of flight. When it chanced to select a feeder bearing syrup and scent, it inserted its bill in the dispenser and fed. What took place when it lit on a feeder containing no food? If it immediately left the feeder without tasting, or hovered near the feeder without perching and flew to one of the syrup and scent-carrying feeders and fed, this behavior could be interpreted as being directed by the sense of smell. If, on the other hand, the bird perched on a nonfood bottle, inserted its bill in the feeding hole, tasted and then flew away, it would appear reasonable to conclude that taste alone, and not smell, had been the determining factor.

"Throughout the experiment, syrup was always fed from the same dispensers and always carried the rose-geranium flower scent. The scent was quite strong and was renewed from time to time. A variety of scents and nonfood substances was offered in the remaining dispensers. The general scheme was to pair each syrup dispenser with a nonfood dispenser nearby, the positions being interchanged periodically. The several test combinations were observed for varying periods according to the following schedule.

"May 12 to May 28: Syrup bottles with the scent attached were available alone. No nonfood bottles.

"May 28 to June 4: Plain water was dispensed in nonfood bottles with no scent.

"June 8: Two tablets of saccharin were added to each nonfood bottle (a previous experiment having shown this solution to be unacceptable to the birds). Observations were made on June 13, June 16, June 18 and June 20.

"June 23: An eye-dropper of iodoform in alcohol was added to nonfood scent vials and the reaction of the birds observed on June 23, June 26 and June 28. On June 28 iodoform was also added to the water.

"June 30: Kerosene was placed in the feeding cups of nonfood bottles. Observations were made June 30 and July 4.

"July 4: Household ammonia was placed in nonfood bottles and scent vials. Observations were made on July 4, July 6, July 8, July 15 and July 20.

"At this point the question of whether the birds make use of the sense of smell in distinguishing between food and nonfood dispensers was clearly answered. The behavior of birds alighting on dispensers that carried water, saccharin, iodoform, kerosene or ammonia was characteristic. The bird tasted what was in the cup, promptly pulled away with a little shudder and flew in search of a syrup dispenser.

"The uniformity with which the birds reacted to the situation indicated plainly that they were responding to an instinctive or inborn behavior pattern and that any stimuli being received by the sense of smell were excluded. In my opinion this conclusion is clear-cut and sound. However, neither this response, nor the other examples of instinctive behavior cited, mean that all animals showing instinctive behavior patterns are mere automatons-that their lives are governed entirely by inborn mechanisms. Learning, insight and intelligence also influence an animal's total behavior.

"Having reached the conclusion that my birds were not making use of a sense of smell, I next tackled a more difficult and fundamental question: Do hummingbirds have any sense of smell? I was experimenting with wild birds that were feeding on nectar and insects, their natural foods, in addition to my offerings. I did not think that there was much chance of inducing them to change their ways. But I decided to try to demonstrate the presence or absence of the olfactory sense by using another repellent odor in fixed locations for a longer period.

"On July 27 a solution of butyric acid and water (to the human nose the most repellent of the substances I used) was put in all nonfood bottles. Six of these bottles were paired with regular food bottles and left in place for nearly a month, the idea being that the birds might learn to avoid the bad odor in a longer training period. Observations were made on July 27, August 4 and August 11, with negative results. At this time I became busy with a photographic project that required the removal of all food bottles except one for attracting birds to be photographed.

"The butyric acid bottles were left in place, however, until August 20. It was interesting to observe that, although these dispensers had been in place for nearly a month, birds looking for food would light on them and invariably take the tell-tale taste.

"The results are summarized in the accompanying table. The two columns of the table headed Taste under Nonfood Bottles and Food Bottles list the visits by males and females that behaved in the normal manner, that is, lighting on the bottle and tasting. The columns under No Taste list visits of birds that hovered near the nonfood bottles and flew away without lighting or tasting-a pattern of behavior that might possibly indicate the use of the sense of smell. As shown by the table, these instances of apparently significant behavior were more frequent in the earlier tests than later. With continued observation an alternate explanation of 'No-Taste' behavior became obvious: 'spooking' by a guarding male. As soon as all the bottles except those on the testing rack were taken down, a male (not always the same one) would take charge of the territory and chase away or try to chase away all other birds that came to feed. The intensity of this behavior varied greatly. Sometimes a guarding male would leave the job, giving all visitors a chance to feed. Then a new guardian would take over. With close guarding, a bird coming in was easily spooked or frightened by the dominant bird and often made only one pass at a bottle. Initially this behavior seemed to indicate a rejection of the food and I accordingly noted them as No Taste.

"The table shows that more passes were recorded at the nonfood than at the food bottles. This is explained by the fact that I manipulated the rack for this result. Passes at the nonfood bottles were the significant ones. Many more observations were made of females than of males, because for a considerable portion of the period the hens were feeding their young. Their food requirement was therefore two to three times that of the other birds.

"Unfortunately the butyric acid experiment was not correctly designed to show the presence or absence of a sense of smell. Some indication of the complexity of this question and of the difficulty of devising an experimental method for answering it is given in a paper by Wolfgang J. Michelsen of Harvard University in Science for September 11, 1959. He used a rather complicated apparatus for studying 'olfactory discrimination' in pigeons and was able to teach the birds to use the sense of smell to obtain food. The method gives a sound and convincing result which proves that pigeons possess a sense of smell.

"In the case of hummers it seems clear that the existence of a sense of smell might be determined by the use of captive birds. With the wild birds I cannot devise a way to do it. Proving the negative-that the wild birds do not possess a sense of smell-seems to me quite impossible.

"As mentioned earlier, my experiment was suggested by the pioneering investigation of the honeybee's senses. This work shows that bees have and make use of a sense of smell. It was observed that a scout bee finds a new source of nectar, returns to the hive and by a definite pattern of behavior communicates the distance and direction of the discovery to its fellow workers. Because the scout bears the scent of the flowers, it is also able to disclose the kind of flower to be sought by the workers. Hummingbirds, being completely nonsocial, cannot use the sense of smell in this way. Moreover, the use of the sense of smell, in my opinion, would not significantly increase the efficiency of the hummingbird's normal food-seeking behavior-a possible explanation of why they have not developed this faculty."

Suppliers and Organizations

The Society for Amateur Scientists
5600 Post Road, #114-341
East Greenwich, RI 02818
Phone: 1-401-823-7800

Internet: http://www.sas.org/