Installation and Use of Cell: A selenium cell in the dark permi{s so large a current to pass that the resulting galvanometric deflection is liable to go off the scale. In order to remedy this defect •the potential fall across the galvanometric terminals is compensated for by the introduction of a dry cell, DC, connected to the terminals of the 10,000-ohm. resistance box, Rj. To one of the plugs a wire is soldered and by means of this connection is established with a similar resistance box, Rg. which is finally connected with the galvanometer, as indicated by the drawing. This second resistance box, R2, is shunted around the first for two reasons: (1) the neces- sary, small potential drop across the galvanometric terminals may be attained; (2) a high resistance is shunted around the terminals of the galvanometer, hence, but little of the current from the circuit containing the selenium cell is diverted. In the drawing, all plugs in both Ri and R2 are removed with the exception of those indi- cated. While these particular resistances apply only to a particular selenium cell, it is readily possible to apply this method to any selenium cell, for by changing suitably the relative resistances in Ri and R2 it is always possible to bring about the desired condition, that the galvanometer shall show no deflection when the selenium cell is in darkness. In using the instrument as installed above we first allow the current to run through the cell for approximately one hour. By the end of that time the slight heating effect so pro- duced has reached a constant. We next test the cell under the influence of a standard carbon lamp. In the original calibration we established the fact that the cell gave a deflection of, say 60 mm., when exposed for 12.5 seconds to a standard 8-c. p. carbon lamp at 1 meter's distance. If, when we come to use the cell for equating energies, we find that the standard lamp gives a slightly different reading (due to temperature changes, etc.) we alter the sensibility of the galvanometer by shunting another 10,000 resistance box (not sho^\^l) around its terminals. When we are sure that our system as a whole is in the same condition as it was during the moment of calibration we are ready to equate our stimuli. Monochromatic light of known wave-length is allowed to fall upon the " stimulus patch " S for a period of 12.5 seconds. By altering the intensity of 76 APPARATUS AND METHODS the light through the use of the smoked wedges (Fig. 3) we finally get the desired reading upon the galvanometer. We then repeat the procedure for the green light. In order easily to expose the cell for the desired length of time (12.5 seconds) the author has devised an automatic electric clock which starts the instant the light is turned on and breaks the circuit at the end of this period. By convenient switches the clock may be made to break the circuit through the standard lamp or to actuate a magnetic shutter which works over the face of the slit in the collimator, Si (Fig. 3), (thus excluding the monochromatic light except for the 12.5 seconds interval ).
Device for securing a purified spectrum. — The spec- trometer system, as above described (p. 71), is arranged to test the sensitivity to differences in wave-length. An addi- tional device is needed when we come to test the limits of spectral sensitivity and to do threshold work generally. In such work great purity of the beams (freedom from white light) is demanded. It is well known that the spectrum obtained with any single spectrometer contains white light. The intensity of this light varies in different spectrometers and in the different regions of the spectrum of any given spectrometer. Ordinarily this does not matter, since where two beams are employed at high intensities, the intensity of the white light in each may be independently varied by the method shown in the auxiliary drawing in Fig. 3. In test- ing limits of sensitivity in the infra-red and ultra-violet it is obvious that the presence of white light in the beam would lead to results which could not be interpreted. White light may be excluded by the introduction of a second spectrometer. The apparatus then takes the form shown in Fig. 5.
8 SEC ^ -^ STimiVS PATCH . TOANlMAli Fig. 5. Ground Plan Double Spectrometer THE PURIFIED SPECTRUM 77 In this sketch the arrangement of parts up to Sg is the same as in Fig. 3. After issuing from S2 the light (illustration shows that green is admitted) Gi takes the following course: It passes iirst through the Brodhun sector, B sec. (which by a series of prisms rotates the light around a stationary sector, the angular opening of which can be adjusted so as to admit all gradations in the intensity of the light), then through a (1-inch aperture, 8-inch focus) single achromat which serves as a collimator, thence through a direct vision refracting prism (1-inch face), through a (1-inch aperture, 8-inch focus) single achromat which serves as an objective. As shown in the diagram, the white light in G is dispersed into a faint continuous spectrum along S3. The green appears as a very intense band. The slit opening at S3 is made just wide enough to admit this band. It passes to the total reflection prism, TR, and is reflected down upon the plaster surface below (shown also in Fig. 1). Examination with a pocket spectrometer of the beams so purified shows that the white light has been completely excluded.' A spectrum so obtained is known as the " purified spectrum of Helmholtz."
Apparatus for testing response to white light, form, and size. — Fig. 6 shows the apparatus required to test differential sensitivity to white light, to form, and to size. This apparatus consists of three chief parts: (1) a light box (Fig. 6), A; (2) an experiment box suited to the animal to be tested (Fig. 1), and between the two (3) a stimulus adapter, by means of which two illuminated areas are simultaneously exposed to view. This appears as the end (front) of the light box.
The wooden light box. A, is divided into two compartments, C and D, by the partition, B. The sides, ends, and partition of the box are y^-inch planed and seasoned lumber. The inside dimensions are as follows: Length, 3 meters; width (between sides), 52 cm.; depth, 30 cm. The bottom is 1%-inch planed stock, dowelled and glued. It is made 2 feet longer than the box, as a provision for the support of the experiment box. To the middle portion of the box are hinged two lids, E and F, of y^-\ne\\ lumber. The edges of the box and lids are grooved and rabbeted. When the lids are closed, the two compartments, C and D, are light-tight with respect to one another. Two cast-iron carriages, G and H, carry incandescent lamps, which serve as sources of photic stimuli. Each carriage rides on a pair of steel tracks, IJ and KL, placed on the floor of its compartment. To the floor of each compartment is attached a St^rrett steel tape, M and N, from wiiich the position of the source of light may be read directly in millimeters. In order that daylight, instead of artificial light, may be used when it seems desirable, a hole 12.7 cm. in diameter is cut in the end of each compartment. These holes are fitted with Aubert diaphragms, as shown at 0 and P.
5 I.e., only a single bright band appears in the field.
78 APPARATUS AND METHODS Fig. 6. Perspective of Light or " Brightness " Apparatus A, light box; C, D, compartments of A; B, partition between C and D; E, F, lids of A; G, H, metal carriages carrying tungsten lamps; IJ and KL, tracks for G and H; M, N, Starrett steel milli- meter tapes; O, P, apertures covered by Aubert diaphragms; R, Bauscli and Lomb cooling cell in light box; d, d', metal straps; y, aluminum plate sliding between d and d'; T, tracks for y; V, stop for y; z, steel plate bolted to wooden end of light box; h, screws attaching y to z; s, s, standard brass stimulus plates; p, brass frame about aperture in y; r, hard rubber ring screwed to p.
FURTHER APPARATUS IN VISION 79 The completed apparatus carries a system of ball-bearing pulleys, cords, and levers, not shown in Fig. 6, by means of which the ex- perimenter may shift the lamp carriages without approaching the light box or experiment box. Against the experiment box end of each compartment of the light box is placed a water cell, R, to serve as an adiathermal screen. The cooling cells used by Bausch and Lomb in their projection lanterns prove satisfactory. They are circular metal cells of at least 11 cm. diameter with a water space of 4.8 x 12.5 cm. The stimulus adapter is a device by means of which the experimenter is able to regulate the size, form, and position of the visual stimuli. A steel casting, % inch thick, carefully planed, and firmly bolted to the light box, is shown at z (Fig. 6). This sheet of steel contains two circular apertures, 10 cm. in diameter, through which the light passes from the light box to the experiment box. To the metal plate, z, tne stimulus adapter is attached by four bolts, one of which is labeled as h. The essential parts of the stimulus adapter are: ( 1 ) A metal frame composed of the straps d and d' and the two vertical straps (not shown). The inner edges of the horizontal straps are rabbeted and into them is fitted (2) the aluminum plate y, which slides smoothly on a pair of tracks (which do not appear in the figure), which are screwed to d and d'. y moves on six rollers. It contains three windows, each 12 cm. square, located 25 cm. apart (center to center). These windows receive the standard stimulus plates. (3) Back of each window is attached by four small screws a square frame of 1-16-inch sheet brass, p, 9.3 cm. on its inner edges and 13.3 cm. on its outer edges. A flange is thus created which serves to hold in each window (4) a brass plate, s, 12 x 12 cm. and 1-16 inch thick. This plate contains an accurately cut opening. It is held firmly in position by two screws through diagonally opposite corners of the metal frame, p. At any time by loosening the screws, the experimenter may readily remove a plate and replace it by another with an opening of diflFerent size or form. The set of standard plates is fully described in connec- tion with the table on p. 80. Between the frame, p, and the standard plate, s, is inserted a piece of opal-flash glass, 1. This serves as a diffusing surface and constitutes the stimulus area. It is held against the brass plate, s, by two screws. (5) To the metal frame, p, is screw^ed a ring of hard rubber, r, i/g cm. thick, 2 cm. wide, with a circular aperture 10 cm. in diameter. To this rubber ring is glued a ring of piano felt about 1/2 cm. thick. These hard rubber and felt rings serve to fill in the space between the metal front, z, of the light box, and the aluminum plate, y, which carries the stimulus plates. (6) By means of the bolts at h, and corresponding positions in the three other corners, the stimulus adapter may be brought into the necessary proximity to z to prevent light from passing from one stimulus opening to the other, between z and y. Each of the four bolts (h) at the corners of the stimulus adapter carries, between z and y, a coiled spring which serves to press y away from z. By putting the proper amount of pressure on the four springs the experimenter can so adjust the surfaces of the rings of piano felt to the planed front of the steel plate, z, that the light cannot pass between the two, while at the same time the surfaces may be moved over one another freely whenever it is necessary to move y, (7) At 80 APPARATUS AND METHODS either end of the frame of the stimulus adapter a stop, V, is attached so that y shall not run beyond the track, T. The most important part of the " brightness " apparatus is the set of accurately made brass stimulus plates which is briefly described below. The set, as used by us at present, consists of 26 plates. These plates fall into three groups: ( 1 ) Plates to test white light vision — three plates with a 6-cm. circular opening and three with a 5-cm. circular opening. This provides a plate for each of the three windows of the stimulus adapter with a diameter of either 5 or 6 cm. (2) Plates to test size vision — this group consists (including those of group 1) of plates with circular openings, which, between 6 cm. and 5.5 cm., differ by 1 mm. in diameter, and between 5.5 cm. and 3 cm. by 5 mm. (3) Plates to test form vision — there are in this group four plates whose openings diff'er in form, while being equal in area. They are (a) the 6-cm. circle; (b) a hexagon 3.299 cm. on the side; (c) a square 5.317 cm. on the side; (d) and an equilateral triangle 8.081 cm. on the side. The area of each opening is 28.2743 sq. cm. In addition to the above there are provided three forms which are inscribed in the aperture of the 6-cm. circle. They are (a) a hexagon 3.000 cm. on the side; (b) a square 4.243 cm. on the side; and (c) an equilateral triangle 5.196 cm. on the side.
Description of Standard Stimulus Plates Each Plate is a 12 cm. square of 1-16 inch Acid Blackened Brass, Containing an Accurately Cut and Centered Opening Description of Openings Use Number To test re- Diameter of plates sponse to Form or side Area needed white light Circle cm.
sq.
cm.
3 (( cm.
sq.
cm.
3 The 6 and 5 cm. circles (as i ibove ) and Circle cm.
sq.
cm.
(( cm.
sq.
cm.
C( cm.
sq.
cm.
(e cm.
sq.
cm.
To size ii cm.
sq.
cm.
(I cm.
sq.
cm.
a cm.
sq.
cm.
it cm.
sq.
em.
C( cm.
sq.
cm.
Circle cm.
(as above ) Hexagon cm.
(side) (( 1 To form Square Equil.
cm.
(( ii 2 triangle cm.
(C (C 2 Openings inscribed in 6.000 cm . circle Hexagon cm.
1 Square cm.
2 Equil.
triangle cm.
Total.
2 APPARATUS IN AUDITION 81 The apparatus as thus described is probably better adapted to test white light vision than to test form and size vision. Criticisms have been urged against it on ac- count of the fact that when a 5-cm. circle is contrasted with, e.g., a 6-cm. circle, there appears a difference not only with respect to size but also with respect to the amount of flux. H. M. Johnson is at present using a modification of the Ive's grating which changes the size without altering the total flux. In regard to its adaptability for testing form, Hunter has likewise urged that " pure form " has not been offered to the animal: that what is offered is really a difference in pattern. We cannot here enter into the details of these criticisms (but see p. 366).
Apparatus for producing auditory stimulation. — The auditory work in this country has been very backward largely because satisfactory apparatus have been very ex- pensive and hard to obtain. One of the serious handicaps has been the lack of a constant air supply for the Stern variators, pipes, and other wdnd instruments. It is highly desirable to make the finer tests on auditory sensitivity of animals with good tuning forks. At the same time it is essential to have control tones which can be substituted for the forks. The variators serve this purpose very well. The need for a constant air supply has been felt not only by the behaviorist but also by the human psychologist as well. The standard tuning forks are very serviceable and very accurate but they are almost useless in behavior work, un- less they are actuated by the method Helmholtz devised. We shall describe first the air supply system which has been in use in the Hopkins laboratory for some years, and then the Helmholtz system of tandem-driven forks.
Appabatus for Obtaining Constant Air Supply Figs. 7 to 10, inclusive, show the complete installation necessary for constant air supply. Figs, 7 and 8 show the interior construction of the Dureco 0 positive pressure blower, which is the essential part of the system.
82 APPARATUS AND METHODS The drum of the blower is eccentrically mounted in the case of the machine and the vanes are so arranged as to slide in and out of this drum. These vanes are supported at the ends by shoes or rollers, depending upon the size of the machine, which run in a raceway near the outside circumference of the casing heads. The purpose of these shoes or rollers is to hold the vanes in position so that they cannot come in contact with the outside casing and thus cause a great amount of friction; they also hold the vanes in a steady position so that they can slide in and out of the drum without jerk or vibration. As the drum revolves the shoes or rollers in the raceway keep the vanes in approximate contact with the outside easing, but at all times there is a clearance. On account of the position of these va^es as the drum revolves the air in the casing is displaced and Fig. 7. End View Showing Fig. 8. Showing Shoes {a, other two not lettered) Running in Raceway Near the Outside Cir- cumference of the Cas- ing Head Internal Construction of Dureco Positive Pressure Blower B shows vane in " out " posi- tion. The other two vanes are shown forced back into the drum.
the vanes force the air out of the discharge opening, at the same time sucking in the air at the intake opening. By reason of the special construction described above the machines are practically noiseless in their operation, have very little friction and require only a small amount of power.*' Fig. 9 shows the method of installation. The pulley, P,, of the blower is belted to Pa, the pulley on the shaft of an electric motor. The (114-inch) outlet pipe, OP, of heavy fireman's hose runs into the bottom of a galvanized iron tank, 2 feet in diameter and 4 feet in height. The top of this tank is supplied with a spring pop valve, R, which regulates the pressure when no air is being used. The iron ^ We are indebted to the National-Standard Company, Niles, Michigan, for the above cuts and description.
APPARATUS IN AUDITION 83 supply pipe, Sp. P (lyg-inch), leads directly down (or up) to the room in which the air is to be used. A cord, C, is attached to the valve, R, and runs over pulleys to the room where the air is to be used. This cord is attached to a screw control at H (Fig. 10). By T?
fl Motor D Tank Sep Stand V Fig. 9. System of Installing Blower, Motor, and Air Tank means of this control the air, w^hen in use, can be held at any pres- sure. After the blower has run for a time the whole system warms up and the pressure rises somewhat. When the rise in temperature has reached a maximum the valve is adjusted to give the needed pressure. It will run indefinitely with a variation in pressure too small to be detected by a water manometer. Fig. 10 shows the method of using the air system with a Stern variator.
84 APPARATUS AND METHODS Fig. 10. Method of Using Air Supply As in the previous figure, Sp. P and C represent, respectively, the supply pipe and the cord controlling the pop valve. The opening of the heaAy gate valve, L, admits the air. The valve is opened up until the gauge, G, shows about 3 pounds pressure. (The higher the pressure within certain limits the less the variation in the manometer.) By means of the small cock, vv, the air which supplies the variator, V, can be accurately regulated. The pressure required to produce a clear tone in the variator shows in the water (or coal oil) manometer, M. The manometer shows extremely small changes in pressure. A pressure change which cannot be detected by a change in the pitch of the variator will show readily on the manometer. In attempting to test changes in pitch or intensity one must be con- stantly on one's guard against changing the position of the head, standing waves, etc. Adaptive changes in the sense organ are also likely to produce apparent changes in pitch and in intensity. By making the proper double connections at w it can readily be seen that the apparatus lends itself to the uses suggested on p. 90 as well as to carrying out tests with two variators upon auditory sensitivity in animals.
APPARATUS IN AUDITION 85 O > M 05 O 1^ H S3 O W W O o l-l EH O o 86 APPARATUS AND METHODS The Helmholtz system of tandem-driven forks. — Fig. 11 shows Johnson's modification of Hehiiholtz's method of actuating tuning forks. The c-64 d. v. primary fork is mounted in a room 100 feet distant from the room in which the experiments upon the animals are being conducted. This fork runs continuously without attention from the experimenter during the w^iole course of the tests. Upon the table (Fig. 11) are placed the two stimulus forks S and S'. The system is connected as follows; The primary fork, PE, is actuated by a current passing from the positive pole of a 2-volt 6-ampere storage cell, ST, through the rheostat, R, to the binding post, a; thence through the fork, through the platinum contact, p', with the mercury cup, c, through the magnet, M, to the binding post, b; thence to the negative pole of the storage cell. When the current is made at the platinum- mercury contact the prongs of the fork are attracted by M, thus withdrawing p from c and breaking the circuit, which releases the prongs of the fork and permits the circuit to be made again at p and c. The primary fork is thus driven at its natural frequency. The secondary forks are driven by a current passing from the positive pole of a 4-volt 6-ampere storage cell, ST', to the binding post, a; thence through the fork and platinum contact, p', with the mercury cup, c', to the central pole of the double key, K; thence through an extreme pole of K through the rheostat, R, the magnet, M', of S to the negative pole of the storage cell. For S' the current passes through the corresponding connections from the other extreme pole of K to the negative pole of S', etc. By means of the double throw switch, K, the experimenter can actuate either fork at will and for any given length of time. Since there are no electrical contacts on the two stimulus forks the tone produced is almost absolutely pure.'^ As will be seen from the drawing, each fork has placed over the ends of its tines a Konig adjustable resonator which must be accur- stely tuned. By varying the height of the resonator a very great variation in the range of intensity is afforded. The only drawback to this system comes from the fact that a primary fork will actuate only forks the vibration frequencies of which stand in simple ratio to that of the primary. By choosing a primary of low frequency, 64 d. v., a fairly large number of secondary forks in the middle region of the scale will be at our disposal (256, 320, 384, 448, 512, 576 d. v., etc.) for testing differential sensitivity, response to clangs, etc.
^ If the forks are placed directly upon a solid surface one can hear faintly the undertone. If the fork is suspended and its resonator properly tuned" the undertone cannot be heard heyond a distance of two or three feet. In cases where great sensitivity in the deep tone is suspected the stimulus forks themselves may be placed some distance away.
APPARATUS IN AUDITION 87 Animal control box for work on audition. — A special form of control box is desirable for use in connection with the above method. It is shown in Fig. 12.
B F' O A, home-box; B, introductory alley; C, opening into alleys D and D'; E. E', alleys entering food compartments F and F'; G. G', punishment grills; T, table containing stimulus-forks. V, W, X, X', Y, and Y' are doors, automatically swinging in the direction indi- cated. The experimenter's place is in an adjacent room by window O in front of door V of the home-box.. (After Johnson, Animal Behav. Monographs, Ser. No. 8, p. 32.)
A is the home-box, 4 by 4 feet, in which the animal is placed be- tween tests. Entrance is made from without by the door, V; B is an introductory alley, 6 feet long and 2 feet wide, leading to the alleys, D and D'. These alleys are each 10 feet long and open to alleys E and E', which are shut off by the two doors, X and X'. These doors are made to open from the animal, and are closed automatically by a small coiled spring. Each is provided with an iron lift latch, which is heavy enough to catch when the door is closed. A string is fastened in a small hole drilled in the end of the lever of each of these latches, and run through an eye-screw^ in the door above, then through a pulley attached to the side of alley 88 APPARATUS AND METHODS E or E', as the case may be, then to the operator's place, so that by pulling the string the door may be unlatched and pulled open without the operator having to leave his place. When the string is released the door closes and latches itself. Alleys E afid E' open into two food compartments, F and F'. The covering of these boxes is pro- vided with two doors, located near the end of alleys E and E', through which food is dropped. Y and Y' are two doors opening from food compartments, F and F', into the home-box, A. These doors are not provided with latches, since they close behind the animal, flush with the jambs, and cannot be opened by an animal which has not free use of its hands. They are provided with coiled springs, like those on doors X and X'. The animals would usually open these doors from the food compartments, merely pushing their way into the home-box, but it is well to provide a means of opening them with strings, as are doors X and X', for the sake of timid animals. Door W is opened by means of a spring. A heavy gut cord is fastened to an eye-screw near the top of door W, and run through a small hole near the top of the outside framework of home-box A, to the operator's station, where it is hooked until door W is to be released. G and G' are two punishments grills — strips of brass about 3 feet long, secured to a white pine board 3 by 2 feet. Alternate strips are connected with the respective poles of the secondary coil of an inductorium, leaving the other end free. When the current is switched in, the animal's foot must rest on two or more of these strips, which are only 1 cm. wide and 1 cm. apart, thus completing the circuit and causing the animal to receive a shock. The inductorium should be placed outside the room where the animal is being worked, and far enough away so that the sparking noise will not disturb the experiment. The current may be shifted through G or G' by a double-throw switch at the operator's station. The framework of this cage is constructed of yellow pine 1 inch by 3 inches; the top and sides are covered with woven steel wire, having a mesh about 1 cm, square. Food is kept in both the food compartments, F and F'. The animal is given the problem of