ANIMALS learn largely by trial and error, but an experimenter can make it easy for an animal to acquire a desired pattern of behavior by modifying its environment and subjecting it to the appropriate kinds of experience. The pattern of behavior thus acquired may disclose significant facts about the animal that would otherwise remain obscure. An experiment that illustrates this powerful technique for exploring animal capacities was made last year by James S. Moran, a high school student of Bridgeton, Mo.

Moran set out to investigate visual perception in pigeons and in particular to compare their color vision with that of humans. He succeeded so well that his project was numbered among the top winners at the 1970 International Science Fair. Moran describes his experiment as follows:

"The eyes of mammals contain photosensitive cells of two kinds: rods, which are extremely sensitive to light but insensitive to differences in color, and cones, which are less sensitive to light but instrumental in the perception of color. Investigators have learned by dissection that the eyes of nocturnal animals, including mice and dogs, are deficient in cones [see "Visual Cells," by Richard W. Young, page 80]. Presumably these animals see the world largely in monochrome. In contrast, the eyes of numerous fishes, reptiles and birds have many cones. The retina of the land tortoise has no rods. These animals should perceive color at least as vividly as people do.

"I should have enjoyed experimenting with tortoises to check this assumption, but they are difficult to train, and so I settled for pigeons. A friend who makes a hobby of racing pigeons gave me two males and two females, together with basic instruction m the care of pigeons. I took the birds home in small cages made of wire screen on a wood frame with a plywood floor. Larger quarters were then built with similar materials. The permanent cage consisted of three interconnecting compartments. Two of the compartments are 10 feet long, two feet wide and two feet high. They were stacked one on top of the other and connected at one end to the third compartment, a four-foot cube. Openings about 12 inches square near the floor enabled the birds to move at will among the three compartments. Open boxes at floor level extended from the outside into the cube for providing feed, grit and water. A layer of sand on the floor absorbed droppings and kept the cage reasonably decent between cleanings. No experiments were made until the birds became accustomed to their new surroundings and began to breed.

"The contrived environment to which the birds were ultimately subjected was provided by a version of an apparatus developed by B. F. Skinner of Harvard University. The apparatus essentially consists of an 18-inch cubical box. In it a pigeon is exposed to a small window that admits light of a selected color to the compartment [see illustration at left]. When the bird pecks the window, an electrical mechanism drops a measured quantity of food into the box, provided that certain switches have been placed in the correct position. The bird can thus learn to reward itself for pecking at light of a predetermined color. Conversely, it can also learn that no reward will follow the pecking at light of other colors. Skinner developed the apparatus primarily for measuring the rewarding effect of a stimulus by counting the number of times an animal would perform an act leading to a reward compared with the number of times it would perform an unrewarded act.

"The electrical circuit of my box includes an inexpensive digital counter that is actuated when the bird pecks the window, if a switch is in the proper position. The apparatus can be altered in many ways for making other experiments. For example, one interesting variation that illustrates the versatility of the apparatus consists of a box equipped with a lever at one end that causes food to be deposited at the opposite end. Two cats were put in the box, one at a time, until each had learned by trial and error to operate the lever and so obtain food. After the cats had mastered the operation both were put in the box. Thereafter one cat spent the entire time patiently working the lever until the freeloader at the other end had gorged itself!

"The Skinner box I made had been modified several times. The version I shall describe requires relatively few parts and is adequate for investigating the color perception of pigeons. The top and bottom and three of the sides of the 18-inch cube are quarter-inch plywood. I install one side of transparent plastic when I want to demonstrate the experiment, but otherwise I use a white opaque material and shine a bright light into the box. Another side of the box has a circular window an inch in diameter located seven inches above the floor. A square opening of the same size is made below the window, three inches above the floor. A rectangular tube that slants downward enters the box through the lower opening. The tube serves as a chute for admitting food to the box from an automatic dispenser. The upper opening is the window that admits colored light to the box. When the bird pecks the window, food drops through the chute at its feet, if the experimenter has preset the appropriate switch.

''The window assembly [see illustration at right] functions as an electrical switch. Its contacts are normally closed. The assembly includes a strip of quarter inch pine about two inches wide and four inches long hinged at one end to a similar strip of transparent plastic; together they close like a book. A hole near the bottom end of the pine strip matches the window in the box. The strip is mounted to the box so that the holes coincide. The bird can momentarily knock the plastic away from the window by pecking through the opening, thereby separating the heads of a pair of brass machine screws that serve as electrical contacts. One of the screws passes through the wood just above the window; the other goes through the plastic. The force of a weak compression spring restores the circuit. The spring is held in place by a third machine screw that is fastened to the wood just below the window and passes through an over size hole in the plastic. To diffuse the light that enters the box I cemented to the outer surface of the plastic half of a table-tennis ball positioned to cover the opening. Light scattered by the translucent ball can be seen easily from most positions inside the box.

"The light source is a 35-millimeter slide projector. Its beam is focused on the table-tennis ball. Colors can be provided by inserting Wratten filters in the slide holder of the projector. My experiments were made with red, with a dominant wavelength of 6,055 angstroms; yellow-green, at 5,013 angstroms, and blue, at 4,711 angstroms. When the filters are flooded by white light of constant intensity, yellow-green appears brightest and red appears brighter than

blue. I equalized the apparent intensity by inserting neutral-density filters between the projection lens and the Wratten filters. The neutral-density filters are supported at right angles to the beam by saw kerfs made in a block of wood two inches square mounted in front of and just below the projection lens.

"The food dispenser is attached to the side of the box that contains the window. Pigeon food consists typically of oats, cracked corn or another grain. The dispenser operates on the same principle as a penny candy machine. Food from a reservoir drops into a hole in a slab of plastic that can be slid back and forth [see illustration at left]. The bottom of the hole is closed by a fixed slab o f plywood that supports the movable piece, thereby trapping food in the hole. When the sliding member is moved to the outer extreme of its travel, the food drops through a matching hole in the fixed slab and into a chute that carries it to the box.

"A solenoid operates the slide to deliver food. After the delivery a helical spring returns the sliding member to its former position. Solenoids that operate on house current are inexpensive at military-surplus stores. The solenoid is energized by a relay that forms part of the food-dispenser assembly.

"The remaining parts of the apparatus were assembled in a separate control box eight inches long, six inches high and four inches deep. This box contains a 24-volt direct-current power supply; a relay with two normally closed contacts that are electrically independent; a digital counter of the ratchet type capable of registering a minimum of 500 counts per minute; three single-pole, single-throw toggle switches; a push button, and a fuse. The relays are designed for operation on direct current at 28 volts. The digital counter operates on 24 volts at 60 cycles [see illustration at right]. The control cabinet is connected; to the Skinner box by a six-wire cable.

"The operation is fairly simple. When the circuit is energized by closing the on-off switch, the armature of the relay in the control cabinet pulls up. In this position the energizing circuit is broken to both the digital counter and the relay that operates the food dispenser.

"Assume that the automatic feed switch and the counting switch are turned on. A bird that pecks the window will then break the circuit that energizes the relay in the control box. The armature of this relay will return to its normal position, and its contacts will close the circuit of both the food dispenser and the digital counter.

"Food is dispensed at once, but a time-delay relay prevents the dispenser from operating oftener than once every three seconds. Every peck is counted, even though all pecks are not necessarily rewarded. The delayed action prevents the birds from gorging themselves too quickly after they have become proficient in operating the apparatus.

"Either the food dispenser or the digital counter can be inactivated by opening the appropriate switch. The manual food switch is a push button of the kind with normally closed contacts. It is connected in series with the window switch and is normally used for testing the apparatus.

"The circuitry I have described can be modified. I happened to have access to direct-current relays with this contact configuration and to a digital counter that operates on alternating current at this voltage, and so I made a power supply to match these parts. Other components can be substituted by modifying the power supply. My present apparatus contains two windows, so that a bird can select either of two colors.

"Three pairs of birds were selected for the experiments. Each bird was weighed at the same hour every day for two weeks while being fed normally. The results (ignoring excessively low or -high weights) were averaged and recorded as the normal weight of each individual. During the next two weeks all the birds were put on a restricted diet to reduce their weight to 80 percent of normal. Thus prepared, my birds were ready to learn how to operate the Skinner box.

"I made up a list of rules to govern the experiments. Birds were tested only if their weight exceeded 75 percent of normal. A bird that fell below 75 percent of its normal weight was given extra rations until it regained minimum weight. All the birds were weighed immediately before and after tests. Considerable effort was made to maintain a daily schedule of testing at a predetermined hour. On the few occasions when the schedule was inadvertently interrupted, the performance at the was on the day that tests were resumed but was not included in the final results.

"During all tests the birds were subjected to 'white noise,' the hissing signal generated by an amplifier when the button of a carbon microphone or a similar noise generator is connected to the input terminals. The electrical noise was reproduced as sound by a loudspeaker placed near the Skinner box. My aim was to drown out audible distractions.

"The birds were rewarded with food only when they pecked at red. Their exposure to each color was limited to one minute. The color filters were put in the projector in random order, as determined by a book of random numbers. If the last color in a series of tests happened to be other than red, and if the bird had failed to peck at the last color, it was exposed to red for an additional minute to determine whether the failure to peck was in fact due to the non-red color or merely to the fact that the bird had satisfied its hunger.

"For the first series of tests the birds were exposed to red and blue because these colors are at opposite ends of the spectrum. I assumed that if the birds could not discriminate between red and blue, there would be no point in continuing the experiment with more closely related colors. A good way to train a naive bird is to fasten a bit of food to a piece of tape and attach it to the window. With the red light on, push the manual feed switch to reward the bird every time it hits the switch. Continue to reinforce only as the bird pecks closer to the red light, and stop using the manual switch once the bird has activated the window. A pigeon can be trained in this way in as little as 15 minutes.

"In the first tests a response (peck) to any color other than red yielded no food. Each color was presented for one minute. At the end of this interval the light was turned off and the number of responses indicated by the digital counter was recorded. The next color, as determined by random selection, was similarly presented to the bird for one minute, and so on for a series of 20 one-minute periods. The number of responses to each color was tabulated for the session and carried forward cumulatively on succeeding days. Each bird was subjected to only one 20-minute test session a day.

"A marked trend in favor of response to red became evident in the tabulated results of all birds within five days and increased thereafter. When the results are plotted, the graph that represents response to red rises steeply, whereas the one representing blue rises somewhat less at first and then levels off. After 15 test sessions the average response of all the birds during each test session, as adjusted for 10 periods of exposure to red and 10 to blue, was 194 responses to red and six to blue [see illustration at right]. This result, it seems to me, can be explained only by the assumption that pigeons easily discriminate between these colors, which differ in their dominant wavelength by 1,344 angstroms.

"A similar series of experiments was conducted in which the birds were exposed to red and yellow-green. The dominant wavelengths differ by only 509 angstroms and there is an overlap of more than 400 angstroms. The average results for all the birds after 15 test sessions, as adjusted for 10 exposures to red and 10 to green, were 144 responses to red and one response to green [see illustration at left]. The likelihood that these results are the product of random response is , or one chance in a million.

"Currently I am continuing the experiment with a series of filters of better quality that transmit a narrower band of color and that progressively approach red. The birds should find it increasingly difficult to distinguish between hues, and at some point discrimination may become impossible. The result should provide an approximate measure of the limit of the bird's color perception, information acquired indirectly by the ability of the pigeon to learn."

LAST month this department described a simple oscillator in the form of a hairpin loop of wire that vibrates like a child's swing when the wire is heated by an electric current. The origin of the oscillator and the explanation of how it works is now provided by A. D. Moore, professor emeritus of electrical engineering of the University of Michigan. Moore writes as follows:

"The oscillator was discovered accidentally in 1928 by P. C. Clarke of Lansdale, Pa. Clarke spent a lot of time trying to find out what made the gadget go, and within a year or so people at various institutions went to work on it. I came across the oscillator during a visit to the University of Illinois and could scarcely wait until I got home to make one of my own.

"A number of experimenters had reached the conclusion that still air is necessary for the oscillator's operation. When you put the wire in a vessel and pump out the air, you get no action. On the other hand, the wire will not vibrate in a wind. You can stop the action by fanning the wire with a piece of cardboard. I finally hit on the idea that perhaps the motion is generated by alternating differences in temperature between the windward and leeward sides of the moving wire. When the wire is moving in one direction, the windward side cools. As a result of the cooling that side contracts. The leeward side also cools, but not as much, nor does it contract as much. The net difference in contraction causes the wire to bend in the direction of the motion and thus generates a force that increases the amplitude of the swing.

"It is well known that when a cooling fluid moves past a cylinder, heat transfer is greatest on the windward side and least on the leeward side. According to my theory, the dimensions of the wire must be such that the hot and cold sides have sufficient time to reverse when the loop reaches the limit of its excursion and starts back. The temperature cannot reverse instantly, but it must do so soon enough for the expansion-contraction effect to operate over a large part of the swing.

"It is always good fun to cook up theories and even more fun to contrive experiments for testing them. To check my theory I constructed a loop of adjustable length with No. 24 Nichrome wire. At a length of approximately 10 inches and a current of 2.75 amperes the amplitude of vibration was about an inch. At greater or lesser current the amplitude decreased. Maintaining the optimum current, I gradually shortened the loop. The natural frequency of vibration increased and the amplitude decreased. At a length of six inches the wire came to rest. The frequency of vibration was too high for effective reversals of temperature. I then added mass to the bottom of the loop to reduce the frequency. The short, loaded loop promptly resumed oscillation. This result agreed with my theory.

"On the other hand, different theories have been advanced. For example, Roger Hayward, who illustrates 'The Amateur Scientist,' recently suggested that perhaps the overall expansion and contraction of the wire causes the vibrations, much the same effect that children exploit when they pump a swing. According to this theory, the full length of the loop cools and contracts during each vibration. The contraction raises the mass toward the point of suspension, just as the child raises his mass when he pumps a swing. In such a situation the driving force arises from the conservation of angular momentum.

"To further check what happens I made another oscillator of the same type but of a different design. I called the new oscillator a 'one-legged' loop. I was certain that it would work, and it did.

"The apparatus consisted of a bras tube of thin-wall stock about a yard long with an outer diameter of 1/8 inch. The upper end of the tube was rigid anchored to a support. Through the bore of the tube I ran a length of No. 30 Nichrome wire and insulated it from the brass with short lengths of glass tubing that were strung on the wire like beads. The bottom of the wire was attached to the bottom of the tube to complete the circuit.

"When current was applied, the tub behaved just as I thought it would: the free end rotated in a circle about three inches in diameter! It may be possible for pumping action to make a contribution to the motion of oscillators of the hairpin-loop type, although I doubt it. On the other hand, pumping action could not possibly drive the 'one-legged oscillator.' Its motion can be maintained only by a bending action that continuously circles the tube."

Bibliography

TRICHROMATIC VISION IN THE PIGEON AS ILLUSTRATED BY THE SPECTRAL HUE DISCRIMINATION CURVE. W. F. Hamilton and T. B. Coleman in The Journal of Comparative Psychology, Vol. 15, No. 1, pages 183-191; February, 1933.

EXPERIMENTS IN OPERANT BEHAVIOR. Ellen P. Reese. Appleton-Century-Crofts, 1964.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.