FACTORS INFLUENCING THE ORIENTATION OF MIGRATING ANADROMOUS FISHES

BY GERALD B. COLLINS

FISHERY BULLETIN 73

UNITED STATES DEPARTMENT OF THE INTERIOR, Oscar L. Chapman, Secretary

FISH AND WILDLIFE SERVICE, Albert M. Day, Director ABSTRACT

The influence of certain physical and chemical characteristics of water upon the orientation of one type of anadromous fish was examined by presenting the migrating fish with a choice between two channels with different water characteristics. The orientative influence of the water properties was measured by the number of fish selecting each channel. The reactions of more than 8,000 fish of the genus Pomolobus-alewife, P. pseudoharengus (Wilson), and glut herring, P. aestivaUs (Mitchill)-were tested as the fish migrated upstream thl'Ough the Herring River at Bournedale, Mass., toward their spawning area. Presented with a choice of waters having different temperatures, 77 percent of the fl.sh entered the channel with the warmer water when the temperature difference continuously exceeded 0.5° C. The response of the fish to temperature differences near the threshold diffe1·ence decreased as the temperature level of the water increased. Presented with a choice of waters having different amounts of free CO,, 72 percent of the fish entered the channel with the water having the lower CO, content when the free CO, difference exceeded 0.3 p. p. m. The sex of the fish appeared to have no influence on its response to differences in CO, or temperature. Experiments indicated that velocity and turbulence can influence orientaticm. The relative or.ientative influence of CO, and temperature, when the two factors were in opposition, was shown to depend on their relative differences of magnitude. Behavior of the fish during the experiments indicated that the orientation was accom­ plished by a method of "trial" involving both movement of the fish and a comparison of intensities of stimulations which were successive in time. UNITED STATES DEPARTMENT OF THE INTERIOR, Oscar L. Chapman, Secretary FISH AND WILDLIFE SERVICE, Albert M. Day, Director

FACTORSINFLUENCI NG THE ORIENTATION OF MIGRATING ANADROMOUS FISHES

BY GERALD B. COLLINS

FISHERY BULLETIN 73 From Fishery Bulletin of the Fish and Wildlife Service VOLUME 52 [Contribution No. 585, Woods Hole Oceanographic Institution]

UNITED STATE S GOVE RNMENT PRINTING OFFICE • WASHINGTON : 1952

For sale by the Superintendent of Documents, U. S. Government Printing Office, Washington 25, D. C. Price 20 cents CONTENTS

Page Historical !background______37 5 Major theories on orientation______375 Sensory basis of orientation______376 Relation of sensory abilities to environmental patterns______376 Experimental approach______378 Materials and methods______378 Experimental method______379 Procedure______380 Controls______381 Water modification ______--··______382 Temperature______382 Gaseous content______382 N ongaseous chemicals______383 Measurements______383 Water samples______383 Titrations______384 Hydrogen-ion concentration______384 Alkalinity and HCO:r-______384 Carbon dioxide______384 Oxygen______385 Temperature______385 Velocity______385 Aberrations in experimental controL ______385 Experiments______386 Orientative influenceof temperature______386 Relation of temperature response to temperature leveL______387 Orientative influence of CO2------387 Relative orientative influence of CO2 and temperature______388 Orientative influence of pH______388 Experiments using nitrogen and oxygen______389 Nitrogen______389 Oxygen______389 Other factors influencing orientation______390 . Th� 3:1 ratio of the response______391 D1scuss1on______392 Summary______394 Literature cited______395

II FACTORS INFLUENCING THE ORIENTATION OF MIGRATING ANADROMOUS FISHES

BY GERALD B. COLLINS, Fishery Research Biologist HISTORICAL BACKGROUND agement, helpful suggestions, and criticisms dur­ ing this research. The work was done in partial The orientation of migrating fishes has been the subject of investigation and conjecture for many fulfillment of the requirements for the degree of years. Migration paths have been outlined by doctor of philosophy at Harvard University, De­ tagging experiments, and a wealth of valuable partment of Biology. information on the physiology, development, and Thanks are due also to Prof. Alfred C. Redfield behavior of migratory fishes has been acquired and many other members of Harvard University through the persistent efforts of many able in­ and the Woods Hole Oceanographic Institution vestigators; but the means by which the fish are for their advice and aid in securing funds and directed on their migrations are still largely a special equipment. The research during the matter of speculation. summer of 1949 was done while occupying a re­ The purpose of the study presented here was to search fellowship at the Woods Hole Oceano­ investigate the influence of certain physical and graphic Institution. chemical factors upon the orientation of migrating The cooperation of the town of Bourne, Mass., anadromous fish of the genus Pomolobus. In is acknowledged for permission to work in the course of this investigation an effortwas made Herring River and to use town property at to determine experimentally if the migrating fish Bournedale. would orient to differences in the physical and A special debt is gratefully acknowledged to chemical characteristics of water. Director Francis W. Sargent and John Burns The existence of differences in physical and of the Division of Marine Fisheries, Massachu­ chemical factors in natural waters to which fish setts Conservation Department, for their interest might respond has long been known. Slight and aid which made this investigation possible. gradations usually exist in the relative amounts The research in the spring of 1949 was done as an of dissolved gases, in pH, in temperature, and in employee of the Massachusetts Division of Ma­ other physical and chemical characteristics of a rine Fisheries. stream from the source to the mouth. Such The encouragement and aid of Dr. Lionel A. gradients are usually so slight that the differences Walford, Ralph P. Silliman, Clinton E. Atkin­ between points miles apart are still below the son, and others of the Fish and Wildlife Service thresholds of the sensory perception of fish. are also acknowledged. The research in the Much greater gradients are found at stream spring of 1950 was done as an employee of the junctions and at stream entrances into lakes or U. S. Fish and Wildlife Service. into the sea. In these gradients there frequently MAJOR THEORIES ON ORIENTATION are chemical and physical differences between points a few feet apart which are large enough to Currently, two major concepts on the orienta­ be detected by the fish. The existence of these tion of migrating anadromous fishes have gained gradients at crucial points along the migration wide support; and, while not mutually exclusive, paths of anadromous fishes where they might in­ they are, as the frequent clashes in the literature fluence the fish in the selection of a stream has led suggest, not entirely compatible. to speculation on their possible role in directinrr0 One of these theories regards the upstream the migration of the fish. migration of the fish as a purposeful "homing" The author is greatly indebted to Prof. George to the part of the river system in which it was L. Clarke of Harvard University for his encour- hatched or in which it spent the early stages of

375

210217-52-1 376 :FISHERY BULLET'IN OF' THE F1SH AND WILDLIFE SERVICE its life. Memory impressions of the goal or home With fish migration, as with bird migration, are implied and the fish is said to be seeking its there has been much speculation over the possible spawning area. The fish is thought to return to existence of a special sensory perception that is its native stream because that was the place unknown at present. No physical basis for this where it was spawned rather than because the has ever been found nor have such sensory abili­ stream was more accessible or made more attrac­ ties ever been demonstrated. There is evidence tive by immediate environmental conditions. that fish can see, hear, taste, and smell. It is The fact that significant numbers of fish have known that they can respond to tactile and kines- actually been observed to return to the streams . thetic stimuli, react to acceleration and nonrecti­ from which they originated is frequently used as linear motion, and maintain equilibrium. It has evidence for this homing viewpoint. been demonstrated that they can respond to tem­ The other theory is that environmental factors perature and to various chemical substances by control the direction of migration, and therefore means other than taste or smell. In the absence physical and chemical conditions such as tempera­ of any evidence to the contrary, it is reasonable ture, current, amounts of dissolved gases, or odors, to suppose that these sensory abilities are the ones are thought to determine the ultimate destination by which fishare guided while on their migrations. of the fish. These factors .fluctuateand are dupli­ Therefore, it appears logical to explore fully the cated in nature so that migratory fish could go to relation between these known sensory abilities and any stream with these conditions. prevailing environmental patterns before consid­ Those who support the idea of environmental ering hypothetical sensory abilities. control of the direction of migration present evi­ RELATION OF SENSORY ABILITIES TO dence that the fish responds to its immediate en­ ENVIRONMENT AL PATTERNS vironment at each point of its migration and they look upon the migration itself as merely the sum of As the sensory abilities of fishes have been re­ the successive responses. The various species are vealed and delineated many investigations have thought to arrive at their separate destinations been made to examine their relation to migration. 1 because they respond in specificallydifferent wa ys Chidester (1924) pointed out the number and to the existing patterns of environmental stimuli. diversity of these investigations as well as their apparently contradictory evidence. The return of many fish to the stream of their Attention was turned to natural environmental origin is to be expected, according to this view­ patterns which might have a directional in.fluence point, because the patterns of environmental con­ upon fish with the known sensory abilities. The ditions which direct them persist year after year. physical and chemical gradients which exist along Variations of these views may be found ex­ the migration paths of anadromous fishes were pressed under different interpretations of the examined for their possible role in orienting "parent stream" or "home stream" theory devel­ migrating fish. oped by investigators working on salmon migra­ The observations of Ward (1920) in Alaska led tion. An excellent review of major problems and him to believe that temperature was an important controversial questions in salmon migration is factor in the choice of a spawning stream by afforded by the symposium, "The Migration and sockeye salmon, Oncorhyrwhus nerka. He ob­ Conservation 0£ Salmon," published in 1939 by the served that on the upstream migration the fish American Association for the Advancement of pass from swift to slow water and vice versa, that Science. they go from shallow to deep water and vice versa, and that they move from turbid to clear water SENSORY BASIS OF ORIENTATION and vice versa. However, he found that at the Whether the migratory fish has purpose and junction of two streams the sockeye salmon con­ seeks its native stream or whether the fish is en­ sistently chose the colder stream and he concluded tirely directed by immediate environmental fac­ that temperature was the chief orienting factor. tors along its route, its orientation must be 1 For references referred to parenthetically, see Literature achieved by some sensory means. Cited, p. 395. ORIENTAT'ION OF MIGRATING ANADROMOUS FISHES 377 The earlier investigations of Chamberlain respiratory influences. He maintained that the (1907) in Alaska also indicated that temperature "branchiopolarity" of the salmon drives it forward influenced the selection of streams by sockeye and acts as its guide. salmon. However, Chamberlain found that the Chevey, Roule, and Verrier (1927) blamed the sockeye chose the warmer streams. depletion of salmon runs in certain streams on the Foerster (1929) observed the migration of sock­ lack of dissolved 02 in the streams owing to mill eye salmon near Cultus Lake, British Columbia, wastes. Their investigations indicated that and found the sockeye at one time entering the salmon would not enter a stream with a low dis­ colder stream and at another time selecting a solved 02 content. The observation that the shad, warmer stream. He concluded that temperature Alosa alosa, were not affected by the low 02 con­ probably had very little directing influence on the tent of the water pointed to a marked species upstream migration of the sockeye salmon. difference. Foerster also made measurements of the pH and of Russell ( 1934) did not agree with the findings of the dissolved oxygen in the streams, but was un­ Roule and his coworkers. He pointed out that in able to find any correlation between variations in the mouths of certain salmon rivers, such as the these factors and the selection of streams by the Tees and the Tyne, there is a long stretch of sockeye salmon. heavily polluted water in which the 02 content Roule (1933) expressed the opinion that the may fall very low. However, salmon enter and direction of shad, P(J/}'rilosa nilotica rhodanensis ascend these rivers. and Alosa alosa, migration was controlled by tem­ Shelford and Powers ( 1915) using a gradient perature. He pointed out that the migration be­ tank technique, found that herring fry, Olupea gins when the river water, pouring into the sea in palla.si, would orient to differences in tempera­ estuaries, is at a higher temperature than that of ture and dissolved gases. Powers became con­ the sea. Roule related that in one section of the vinced that gradients of CO2 tension exert an Rhone River shad fishermenalways set their traps important influence upon the orientation of mi­ on only one bank of the river, whereas along most grating fishes. Powers and Hickman (1928) of the river they set traps on both banks. Investi­ measured the CO2 tension in lakes, in rivers drain­ gation revealed that the water on one side of the ing lakes, and in rivers which did not drain lakes river was several degrees warmer than on the other in the Fraser and the Columbia River systems. due to the influence of a warm tributary upstream. They found that, in general, lakes and rivers The shad were always found migrating on the draining lakes had a lower CO2 tension ( average warmer side; and they were also observed to turn 0.57 mm. Hg) than rivers not fed by lakes ( average off the main stream into certain tributaries which 1.05 mm. Hg). Further analysis of the data also were warmer than the main stream. Roule's ob­ indicated that typical mountain streams had servations on the behavior of salmon, Salmo salar, higher CO2 tensions than streams of the lowland. during migration convinced him that the salmon Powers (1939) pointed to an observation that were indifferent to temperature. the sockeye salmon, given a choice, will always Roule (1914), observing the selection of par­ choose the fork of a river which drains a lake. ticular streams by the Atlantic salmon, Salmo The chinook salmon, Oncorhynchus tshawytsoha, salar, in its upstream migration to spawning areas, in the same situation apparently is indifferent and · concluded that the proportion of dissolved 02 in moves up either branch. Powers suggested that the water was the dominant factor in directing the sockeye is responding to di:fferences in CO2 the migration. He maintained that the salmon tension in its selection of a stream draining a always choose the water with a higher concentra­ lake. The lack of a similar response on the part tion of dissolved 02 • Roule (1933) suggested that of the chinook salmon was presumably looked the metabolic condition of the fish results in in­ upon as a species difference. creased respiratory activity and it becomes polar­ The publications of these investigators who ized toward the more highly oxygenated water. were seeking to correlate the observed movements Roule used the word "branchiotropism'' to de­ of migratory fishes to gradients of temperature scribe deviations in direction brought about by and of dissolved gases were received with great 378 FISHEiRY BULLETIN OF THE FISH AND WILDLIFE, SERVICE interest. There has been, however, a general re­ rect experimental approach was used. Experi­ luctance on the part of many engaged in fishery ments were undertaken which attempted to meas­ research to accept some of the conclusions put ure the directional responses of migrating fishes of forth. Those investigators who were convinced the genus Pomolobus to certain differences in of a highly developed homing ability in anadro­ wa:ter characteristics including temperature, pH, mous fishes found it difficult to reconcile the idea and amounts of dissolved gases ( 02, N2, and CO2). of migrating fishes responding to environmental The experiments were designed to avoid many gradients with that of homing. Many contra­ complications that have to be considered in labora­ dictory observations made by other workers in tory experiments with live fish. The experiments the field, because of a general tendency to avoid were made in the stream in which the fish were negative reporting, seldom reached the literature. migrating so that the fish would not be handled or Furthermore, fieldworkers who were aware of the subjected to the shock of being removed from their enormous difficulties in making dependable ob­ natural environment. Tests were made on thou­ servations of the movements of fish in large rivers sands of fish, and each fishwas tested only once so and streams, particularly with only intermittent that considerations involving learning could be observations made over a wide area, felt that the ignored. evidence upon which the conclusions were based Inasmuch as the special migratory behavior of was inadequate. anadromous fishes is exhibited for only limited In summarizing previous research in this field periods of time, perhaps only once or twice during it might be said that- the life of the fish, an essential condition of these (1) The ability of fishes to detect differences in experiments was that they were conducted while temperature and in amounts of dissolved gases the fish were actually migrating. The impor­ has been established experimentally. tance of this is realized when one considers how (2) Laboratory (gradient-tank) e:x;periments have greatly the response of a fish to environmental also established that differences in temperature stimuli may differ at various physiological stages and in amounts of dissolved gases can influence the direction of fish movement. of its life. As fingerlings, the fish are found mi­ (3) Differences in water temperature and amounts grating downstream; as maturing adults, they of dissolved gases have been shown to exist in migrate upstream. In those species which survive natural waters at points where they might spawning, the spent fish again migrate down­ exert an important influence upon the direction of migrating anadromous fishes. stream. Frequently, individual fish of the same ( 4) It has been suggested that these physical and species may be found in a stream responding in chemical differences do influence the orientation very different ways to identical environmental of the migrating fish. stimuli. (5) Some field observations of migratory-fish be­ havior appear to show a relation between the The differencesin physical and chemical factors direction of migration and gradients of certain tested were artificially produced in these experi­ physical and chemical factors. These field ob­ ments but the type and magnitude of the differ­ servations are relatively few, inconclusive, and contradictory. ences were within the range commonly found at stream junctions in nature. Thus, it is known that some fishescan orient with reference to certain physical and chemical dif­ MATERIALS AND METHODS ferences in water, but whether migrating anad­ romous fishes actually do orient with reference The experiments were conducted in the Herring to such differences is not known. The following River at Bournedale, Mass., in 1949 and 1950 dur­ experiments were undertaken to explore this ing the spring herring runs. The herring run at question with one type of anadromous fish. Bournedale actually consists of two overlapping runs of closely related species. The earlier run, EXPERIMENTAL APPROACH made up of alewives, Pomolobus pseudoharengus In studying the influence of gradients of physi­ (Wilson), begins about the firstof April and con­ cal and chemical characteristics of water upon the tinues until the end of May. The second and orientation of migrating anadromous fishes a di- smaller run of glut herring, Pomolobus aestivalis ORIENTATION OF MIGRATING ANADROMOUS FISHES 379 (Mitchill), usually starts about the last week of than would be necessary for fishes such as the May and lasts until the middle of June. shad or the salmon. At Bournedale each year There are visible external differences by which during the herring run, more than half a million these two species can be identified. The glut her­ of these fishes enter the small Herring River ring is generally smaller than the alewife, has a ( completely fresh water; average flow less than maller eye in relation to the head, and when ex­ 20 c. f. s.) from the sea water of the Cape Cod amined out of water has darker dorsal pigmenta­ Canal and migrate upstream for approximately tion. The individual variations in size within the a mile to their spawning grounds in Great Herring two groups overlap considerably, and in the water Pond. The experiments were conducted in the each species has the ability to modify its pigment stream a short distance below its entrance into to blend with the background in the matter of a Great Herring Pond. few seconds. Consequently, it is very difficult to EXPERIMENT AL METHOD distinguish bet"een the two species with any de­ gree of certainty without first removing the fish As the fish migrated upstream, they were di­ from the water. Attempts to separate the two rected by wire screens into a shallow experimental species while the experiments were in progress trough. The upstream end of the trough was proved to be impractical. When the studies were divided into two channels of equal size (fig. 2). completed, a careful comparison of the responses As a fishprogressed upstream through the experi­ shown by the fish in experiments conducted at the mental trough, it was presented with a choice be­ beginning of the early run when only alewives tween the two channels. Differences in water were present with those in experiments made near the end of the later run when only glut herring were present, failed to reveal any differences be­ tween the responses of the two species. These anadromous members of the herring family ( Clupeidae) proved to be ideally suited to � 0 experimental purposes. They are small and _,J •RECORDING POINT• IL FOR TEMPERATURE TESTS migrate in enormous numbers through easily ac­ ::E < cessible brooks and streams ( fig. 1) . The con­ w a:: struction of the experimental apparatus and the I­,.,, conditioning of water for experiments with these IL 0 fishescan be done on a smaller, less expensive scale z 0 j:: u w a:: 0 •POINT OF DECISION•

ZONE OF MIXING

SCREEN DIRECTING FISH INTO TROUGH

FrnunE 1.-The. herring run at Bournedale, Mass. This FIGURE 2.-Diagram of experimental trough. Dimensions: pool is immediately below the experimental station. 18 ft. long, 21 in. wide, and 10 in. deep. 380 FISHERY BUIJLETIN OF 'I'HE FTSH AND WILDLIFE SERVICE temperature, in pH, and in dissolved gases between the two channels were created experimentally. The influence of these differences in the physical and chemical characteristics of water upon the orientation of the migrating fish was measured by the number of fish choosing each channel. The experimental trough was 21 inches wide, 10 inches deep, and 18 feet long. It was open at both ends so that when it was alined with the direction of stream flow and partially submerged, the water flowed freely through it. The two channels in the upstream end of the trough were 10 feet long and 10 inches wide. A 10-inch chan­ nel width was chosen because it was approximately equal to the length of the fish and would allow FIGURE: 3.-General Yiew of experimental station, Herring River, Bournedale, i\Iass., 1950. enough room for normal swimming movements A. Entrance for upstream migrants. and turning. The experimental trough could be B. Bypass exit for downstream migrants. 1�aised or lowered in the water in order to create C. Laboratory for chemical determinations. any desired depth of flowregardless of the natural D. Light-control arrangement. fluctuation in the water level of the stream. The flow of water through the trough was maintained actually made for the entire school. There were at a depth of 6 to 8 inches, while the velocity of instances of schools splitting, with all those on the water through the trough varied, depending on left side entering the left channel, and all those stream conditions, from a minimum of 1 foot a on the right side entering the right channel, sug­ second to a maximum of 2 feet a second. gesting that perhaps in these cases spatial rela­ The trough was stained a dull mahogany to tions alone were involved. There were many provide a dark background for the fish so that variations in such group behavior and they were their behavior ,rnuld be as natural as possible. difficult to interpret in a quantitative way. To Exploratory tests the previous year had shown avoid the problem completely, an entrance gate that a light background made the fish extremely was designed. Through this gate (figs. 4 and 5) nervous and excitable when they were in the shal­ the fish were allowed to enter, one at a time, and low water and confinement of the experimental only after the previous one had made its decision trough. The trough was also placed in the stream and was completely out of the trough. several weeks before tests were begun to reduce The entrance gate also served to center the fish the possibility of odors, or other factors that might so that as it entered the experimental trough it be unfamiliar or objectionable to the fish. was subjected to a mixture of the waters of both To prevent light inequalities between the brn channels. Thus, the fiflh started in the center of channels resulting from shadows, the experimental a strong transverse gradient (fig. 2) and almost trough was shaded from direct sunlight by a can­ any lateral movement resulted in its being sub­ vas canopy (fig. 3) . jected to water of a different quality. At the downstream end of the trough a wire­ screen gate was installed to control the entrance PROCEDURE of the fish into the experimental trough. The As the fish progressed upstream in the trough necessity for such a device became apparent during to the point where it had to choose between the preliminary exploratory tests in which the en­ two channels ( the "point of decision" in fig. 2), it trance of the fish was unrestricted. When several usually moved from one side of the trough to the fish entered at the same time, they frequently ex­ other, alternately approaching each channel until hibited a schooling tendency and all followed the it finally entered one. In those tests in which choice of the leader. Their behavior was such that both channels were completely unobstructed the it seemed probable that only one decision was choice was recorded when the fish had completed ORIENTATION OF MIGRATING ANADROMOUS FISHE:S 381

Frci..:B1, :i.-Entrnnce gate. The fish were allowed to lc'11ter, ouP at a time, through a screen gate (shown open). (The white cloth was not presPnt during the tests.)

renchecl, the fish was considered to have made its decision and the result was recorded. vVhen once the fish had passed the recording point if its for­ ward progress ,ras too slo"·, it \Yas urged to con­ tinue on out of the trough by taps on the trough with a stick immediately behind the fish. ·when the fish was completely out of the trough, the entrance gate ,ms raised again to allow another to enter. FIGURE. 4.-Entrance to experirnentnl trough. Before each series of tests, any necessary ad­ A. Entrance gate (closed). B. Retaining pool. justments of experimental conditions were made 0. Retaining-pool gate (open). and water samples and temperature measurements D. Bypass for downstream migrants. were taken. The fish were then allowed to enter (The white cloth on the bottom of the trough was for the trough one at a time. It usually was found photographic purposes and was not present during the convenient to nm approximately 25 to 30 indi­ tests.) vidual tests in succession before the measurements were taken again. The average time for such a its passage through the channel and ,vas entirely series of tests ,vas about 40 minutes, although it clear of the trough. This system of recording fluctuated considerably depending upon the be­ decisions permitted the fish to change its choice havior of the fish. of channels at any stage of its progress by turning back and entering the other channel. During the experiments in which the water was CONTROLS being modified by heating, the reluctance of the Every effort \Yas made to keep conditions, e.g., fish to pass through the heating apparatus in the light, depth of \Tater, rate of flow, turbulence, as upstream ends of the channels, resulted in con­ uniform as possible in both channels so that any siderable delay. To save time a "recording directional response would be due solely to the point" (fig. 2) ,ms chosen below the heating ap­ factor being tested. vVhenever major adjust­ paratus. When this arbitrary point which was ments in the experimental conditions had been 20 inches ( approximately twice the length of the made, before tests with modified water \Yere be­ fish) from the entrance to the channel was gun, a series of control tests \Yas run to ensure 210217-52--2 382 FISHE,RY BULLET'IN OF 'IHE FISH AND WILDLIFE SERVICE that such uniform conditions prevailed. The re­ conditions created by the unequal distribution of sults of these tests are summarized in table 1. heating apparatus, a series of control tests was run with the heaters in place but with the power TABLE 1.-Daily totals of control tests made to ensure that uniform conditions existed in both channels of the turned off. The results of the control tests which experimental trou,gh, 1949 were interspersed with the actual temperature tests are shown in table 3. They indicate that the Number offish entering- presence of the nonoperating heaters had little or Date no effect upon the choice of channels made by the Left Right channel channel fish.

April 30_·····························-··········-- 91 89 TABLE 3.-Control tests during temperature experiments, May L ..·-··-·-·····················-····-·····-- 49 39 1950 2 ..... _ .. _ -··-·······-····-··············-··- 54 48 3...... ·-.... -· ·--·. _-······- -· ..··- ···--··. 68 73 [Immersion heaters in place but power turned off] 7...... -·· -· .. ····-· -··- ..•.-·· _··-..•. ··· - 97 91 8..... ·-··...... -··· ···- 58 49 11...... ····-- 19 15 Number of immer· 12..... _.. ..-· ...... •..•...... ·· ·-. 20 22 Number of fish sion water heaters ]4_ ..•. _ .. _.-·· _ ..• ·-·. _ ...... •. -·.. 71 54 entering- in- 15.. .. •-··-·········-··-······················· 81 83 Date and time 16 .... -·-...... _ ...... ·- ...... _ 49 47 19 .... -···-· ...... -·...... _. 25 29 Left Right Left Right 22 .•...···-·-·· -··· ·-· ...... •.... -···.•. -·· _ 511 49 channel channel channel channel 23_. __ .••.....••.....•.. ·-·······-..•.-··...... 47 58 24.. , ...... -·· ..•.. -····· -·...... -··... _ 24 24 30...... •...... -······ ..•...... 29 30 May 13, 3:45 p. m.•.•...... •.. 16 17 5 5 , , 14, 6:25 p. m...•....•...... 27 25 0 IO TotaL._ ...... -··· ...... _ ...... ____841 _____800 15, 8:00 a. m••.•..••.•.•..• 9 8 0 11 Percent...... 51. 3 48. 7 20, 10:00 a. m ...•..••.•.•.•. 12 12 0 11 24, 2:30 p. m...••.••...•••• 13 14 0 11 24, 3:30 p. m.•••.•.•.•..•.. 14 13 0 11 25, 4:00 p. m...•.•...... 26 27 0 11 26, 9:15 a. m...••.••...... 15 12 0 11 To minimize the possibility of some unrecog­ 26, 11:30 a. m...•...•...... 12 15 0 11 nized factor influencing the choice of the fish, 27, 10:30 a. m...... • 21 30 0 8 27, 11:55 a. m...••.••...... 9 IO 0 8 27, 1:30 p. m .••...•...... 11 15 0 8 control and test channels were alternated between 27, 4:00 p. m•...... 8 7 0 8 27, 5:00 p. m_ ...... 27 34 0 8 each series of tests. Such alternation resulted in 28, 12:40 p. m.•...•...... •.. 13 12 0 8 28, 2:00 p. m...... 13 11 0 8 any "nonalternating" influence being cancelled 28, 5:30 p. m..•..•...... •.. 13 12 3 8 June 3, 8:15 a. m...... 7 5 0 11 out when the figures were totaled. An example 6, 10:30 a. m...... •.... 4 4 0 11 of the pattern of controls and tests is given in 7, 11:15 a. m..•..••..•..•.. ---13 ---15 ---0 ---11 Total ...... 283 298 ------table 2. Percentage...... 49 51 ------

TABLE 2.-Example of the pattern of controls and tests, May 15, 1949 WATER MODIFICATION [Tests listed in chronological order) Temperature Left channel Right channel The modification of the water temperature was

Number Number accomplished by the use of industrial electrical Water characteristic 1 of fish Water characteristic 1 of llsh immersion water heaters. These heaters (32 a.,

ControL ...... 18 ControL...... _ 17 7.5 kw., 230 v.) were of the tubular type and co, ...... 11 ControL...... 31 ControL...... 26 co, ...... 19 shaped in a simple loop so as to fiteasily into the ControL ...... 25 Control...... 27 ControL .. ··-··...... 35 co, ...... 12 channels of the experimental trough (fig. 6). co, ...... 11 Control...... 34 ControL ...... 16 ControL .•...... 16 Each of the 11 heaters was connected separately ControL ...... 33 co,...... 17 co, ...... 9 Control...... _ 26 to the main circuit and could be turned on or off ControL ...... 22 ControL...... 23 co, .. -...... 8 ControL .... _...... 31 at will without interfering with the operation of ControL ...... 31 co, ...... 14 any of the other heaters. The arrangement was

1 Control, water unmodified; CO,, gaseous CO2 added. simple and flexible and made a graded control of temperature possible. During tests involving temperature differences of greater than 1 degree centigrade, it was neces­ Gaseous Content sary to have a greater number of heaters in one The gaseous content of the water was modified channel than in the other. To be certain that by the introduction of specific gases. These gases the response shown was the result of temperature (02, N2, and CO2 ) were bubbled into the water at differences rather than differences in hydraulic the upstream end of the trough directly in front ORIENTATION OF MIGRATING A ADROMOUS FISHES 383

FIGURE 6.-Upstream encl of experimental trough during temperature tests. All immersion water heaters are FIGURE 7.-Device for modifying gaseous content of water. in place for maximum temperature difference. The device contained two separate batteries of aerators ( Shade on left was raised to allow light for photo­ so that the two channels could be modified independently graphic purposes. During the experiments no shadows at the same time. were present.) of one of the channels through a colorless plastic of the channel or channels in which the water was tube positioned horizontally an inch above the to be modified. floor of the trough at right angles to the direc­ The device used to accomplish this ( fig. 7) con­ tion of stream fl.ow. Small holes were bored in sisted of two batteries of porous-stone aerators of the plastic tube at short intervals to allow an even the type commonly used in small aquariums. Each distribution of the carboy liquid throughout the battery of aerators produced a band of very fine channel. bubbles. The incorporation of two separate bat­ MEASUREMENTS teries in the same device made it possible to modify Water Samples independently the water of both channels at the Water samples were collected in wide-mouthed same time. 100-milliliter bottles at the do,vnstream end of each channel. Sample bottles ,,ere stoppered Nongaseous Chemicals while still under water to avoid any gaseous ex­ The nongaseous chemicals used to modify water change with the air. The samples were collected in experiments were first dissolved in a carboy simultaneously in both chairnels and "·ere imme­ of water from the stream. They were then intro­ diately brought into the fieldlaboratory where the duced at a controlled rate into the upstream end chemical determinations were begun at once. 384 FISHERY BULLETIN OF THE FISH AND WILDLIFE, SERVICE Titrations Titrations were done with a Beckman pH meter. 7.0 The water-sample bottles were calibrated with the solid stoppers on so that the exact amount of 6D water in each sample was known without further measurement. The titrations were performed in the sample bottles to avoid any opportunity for 5.0 the modification of the gaseous content of the fREECO2 (PPM) water due to exposure to air in pouring. 4.0 Hydrogen-ion Concentration Measurements of hydrogen-ion concentrations 3.0 were made to within 0.03 pH unit with a Beckman meter. A Hellige pocket color comparator pH 2D . .· . 0 '°"• : • was used for rapid pH measurements (to within . . 0.1 pH unit) during adjustment of valves while ...... , setting up experimental conditions and during a 1.0 .. . . number of the earlier exploratory experiments. 60 62 6.4 66 68 Alkalinity and HCO,-. In the range of pH in which the experiments were made (pH 7 to pH 6) the alkalinity was l.<'IGURE 8.-Free CO, measurements, May 1950. Only measurements of samples of unmodified stream water assumed to equal the concentration of HC03• Methyl-orange alkalinity was measured (Ameri­ and water modified by the addition of gaseous CO, are included. The samples were taken under a wide variety can Public Health Association 1946) by titrating of stream and weather conditions. the sample with H2S04 to an end point of pH 4-'1. Carbon Dioxide the stream produced several effects. It increased the concentration of free CO2 and it decreased the 2 content of the The free, or uncombined, CO pH. It also raised the partial pressure, or tension, water was measured in two ways. Some of the of CO2 in the water. Examples of data from measurements were made directly by titration of individual tests (table 4) illustrate the first two the sample with NaOH (American Public Health of these effects. It will be noted that the concen­ Association 1946) to an end point of pH 8.2. tration of HC03 was largely unaffected. In the Other determinations of free CO2 were made indi­ range of pH in which the experiments were made rectly by measuring alkalinity (methyl orange) (pH 7 to pH 6), carbonates were not present. and pH and then determining CO2 by the graphic Water was modified by addition of K H (P04) method of Moore ( 1939). The measurements were 2 during the tests of the orientative influence made to within 0.1 p. p. m. CO2. Those measure­ of pH. Examples of data from individual tests ments made directly by titrating with N aOH were are given in table 4. In these tests the ch nges slightly lower than the measurements made in­ � in pH were of approximately the same magmtude directly by the graphic method. Therefore, in as the changes in pH in the tests in which CO2 was comparing the water of the two channels only added (although K2H(P04) raised the pH while one method was used for each pair of measure­ CO2 lowered it). The addition of K2H( P04), ments. The range of conditions under which the however, had much less effectupon the concentra­ measurements were made is indicated by figure 8 tion of free CO • which shows CO2 measurements made during the 2 experiments. During this investigation, measurements of CO2 During the series of tests which examined the were restricted to the convenient and widely ac­ orientative influence of CO2, water was modified cepted methods available for measuring the amount of free CO • The importance of measur­ by the introduction of gaseous CO2• Addition of 2 gaseous CO2 to the almost unbuffered water of ing dissolved gases in terms of partial pressures ORIENTATION OF MIGRATING ANADROMOUS FISHES 385

TABLE 4.-l!J(l)amples of chemical measurements made the downstream end in each channel with the during individual tests, June 11, 1950 center of the thermometer at a point 2 inches from ° ° [Stream temperatures, 19 to.19.4 C.] the bottom of the trough ( the level at which the drog fish usually swam) and 2 inches from the wall Channel Water ,!lonyconc:: Alkalinit y Carbon characteristic 1 tration or HCO1 dioxide which divided the trough into two channels.

P.p.m. P.p.m. Velocity Ca CO, 6.40 6. 57 4.043.98 i: gi The velocity of the water flowing through the �i�c::::::::::: 8��iroi_::::::::: 1----1----1------::-:- .17 1 trough was measured by means of an impeller­ Difference ___ ------l====l====i.06 99_ 6. 57 3.88 =: ==:=::: type current meter. Measurements were accu­ Left______Control ------6.18 3. 76 Right ______, _____ c o, ____ ------l- i:: ---t------ii--- rate to within 0.2 feet per second, and represent Difference ____ ------.39 .12 7 - l====l== ==l===:=2.==' 3 the average velocity of the water over a 90-second Left______Control.______6. 52 ------'I. 70 Right______K,H(PO,)______6._ 97 _--_--_--_-_--_--_- __'_I ._62 period of time. The measurements were taken __ 1 _1 1 Difference____ ------= =·45== -=·=-·=·=----= =--=-= ·0 at the downstream end of each channel at a point l = I I =::'8 2 ==:":'I. halfway between the channel walls. Left______K H(PO,)______6.6. 94 57 . ______2' 1.71 Right ______ControL ______.;.·__ _ _ _--_-_----_ --_ _--_- --- 78 1 1 1: -::: Difference______.3 7 ------. 07 ABERRATIONS IN EXPERIMENTAL CONTROL 1 CO2, gaseous CO2 added; control, water unmodified; K2H(PO,), Throughout the experiments, measurements of K,H(PO,) added. 1 Samples titrated with NaOH. physical and chemical water characteristics such as temperature, velocity, pH, and amounts of when considering their physiological effects upon dissolved gases were much more precise than organisms is recognized. The ability of a gas to the experimental control of these characteristics. diffuse through a membrane depends upon the Hydraulic conditions above the head of the ex­ tension of the gas in solution. Therefore, the perimental trough created a periodic eddying tension of the gas is critical where, as in respira­ ( every 5 to 15 seconds) which resulted in a fluctua­ tion, an actual gaseous exchange is made. How­ tion in the rate of flow alternately in each channel. ever, very little is known of the sensory mecha­ During tests in which water characteristics were nism by which fish detect differences of COi; nor being modified, eddying caused a periodic fluctua­ is it known whether CO2 must actually permeate tion in the degree of modification. For example, a membrane to affect the sensory organs. Under during tests in which the water was modified by the circu�stances of these experiments, where heating, eddying resulted in a temperature fluctu­ both channels receive water £rom the same source ation of approximately 0.5 ° C. in the channel being under identical conditions, the CO2 tension would modified. be directly proportional to the amount of £ree Temperature measurements to within 0.1 ° C. CO2 present. Therefore, the relative amounts of were made almost instantaneously, and the extent CO2 in the two channels may be used as indexes ofthe temperature fluctuationwas easily measured of the relative CO2 tensions. by a comparison of maximum and minimum tem­ Oxygen perature readings. However, measurement.s of other water characteristics such as velocity, pH, The amount of dissolved oxygen was deter­ and amounts of dissolved gases were average mined by the standard Winkler method ( Ameri­ measurements. Velocity measurement.s ( read in can Public Health Association 1946) after pre­ terms of propeller revolutions per unit time) rep­ liminary orientation tests (Ellis, Westfall, and resented an average velocity for a 90-second period Ellis 1948) for the presence of interfering sub­ of time. The manner in which the water samples stances proved to be negative. were taken for measurements of pH and amounts Temperature of dissolved gases tended to collect a mixture of Temperatures were measured to 0.1 ° C. by the the water flowing through a channel over a period use of a mercury thermometer held horizontally of time greater than the time for a complete cycle in the water with its long axis in the direction of of eddy fluctuationand so these measurements also the current. The measurements were taken at represent measurements of average conditions. 386 FISHERY BUl.iLETIN OF' THE FISH AND WILDLIFE, SERVICE Such average measurements gave no clear indica­ fied ( see Aberrations in Experimental Control, tion of the degree of fluctuation. p. 385) was probably responsible for the appar­ Some of the other experimental conditions also ently intermediate nature of the response to tem­ could not be controlled completely. The depth perature differences of 0.5° to 1.0° C. The tem­ , of the water in the trough, maintained at 6 to 8 perature difference values given in tables 5 and 6 inches at the downstream end of the two channels, are maximum values. The minimum values were varied from 4 inches to 10 inches at the ends of approximately 0.5° C. less. For example, during the trough when the trough deviated from a hori­ tests at the recorded temperature difference of zontal position. The downstream end of the 0.7° C., for at least part of the time, the tempera­ trough had a tendency to settle as a result of the ture difference was approximately 0.2° C. clogging of the screen entrance gate by floating Therefore, only at recorded temperature dif­ organic debris. Variations of this sort affected ferences above 1.0 ° C. was the threshold difference the channels equally, however, and probably are of 0.5° C. exceeded continually throughout the of little significance. entire test. During the experiments in which water was heated electrically there frequently was a vertical TABLE 5.-Response to temperatttre differences as shown variation in the temperature of the water in the in experiments of May 7 to Jttne 8, 1950 ° ° modified channel due to inadequate mixing of the [Stream temperatures, 11.1 to 22,3 C.]

heated water. This was particularly true when Water temperature differencebetween Entered channel with- channels 1 Number of 1------only a few heaters were in operation. At such decisions ° Warmer Cooler times, the variation was as much as 0.4 C. from (Centigrade) water water the warmer water near the bottom of the channel Percent Percent ° to the cooler water near the surface. The tem­ 0,4 ______98 2 49 51 (====1====0 1=== perature measurements, made 2 inches from the 0,5°______109 57 43 0.6° ______51 61 39 ° 63 37 bottom of the channel ( the level at which the fish 0,7°______180 0.8 ______427 60 40 usually swam), were always of the maximum 0.9° ______471 61 39 1.00______76 63 37 temperature. TotaL ______------1,314 "61 39 In interpreting the data collected during these 0 1.1 0 ______22 77 23 experiments, such variations in experimental con­ 1.2 ------98 80 20 1.30 ------136 76 24 ditions must be taken into consideration, particu­ 1.40 ------100 76 24 1,5°------• 142 77 23 larly when thresholds are concerned. The thresh­ l, 7° ------.. ------.------. 98 79 21 0 80 20 1.80------107 olds for the responses of the fish may actually be 1.9o ------. ------­ 27 78 22 ? o ------83 80 20 lower than those indicated by the data. 1----1-----1---- TotaJ.. __------__--- -______813 '78 22 l====l=====I=== 2.2° ------213 73 27 EXPERIMENTS 2.30------132 73 27 0 83 17 2.4° ------108 ORIENTATIVE INFLUENCE OF TEMPERATURE 2.5 ------28 75 25 0 79 21 2.70------24 Experiments were conducted in which the 3.0 ______12 75 25 migrating fish were presented with a choice be­ TotaJ. ___ ------517 3 76 24 tween waters of two different temperatures. The 1 Temperature differences are maximum values. Minimum temperature differences were approximately 0.5° C, Jess. responses of the fish (table 5) indicated a pref­ • No response. , Response. erence for the warmer water. • See notes on eddying phenomenon, p. 385. The stream temperature during these experi­ Taking the eddying into account, the inter­ ments ranged :from 11.1 ° to 22.3° C. The tem­ mediate nature of the response to temperature perature differences between channels, created by ° ° heating the water of one channel, were varied differences from 0.5 to 1.0 C. should probably from 0.4° to 3.0 ° C. The threshold of the re­ be discounted and considered as due to imperfect sponse appeared to be at a temperature difference experimental conditions. The data would then of approximately 0.5° C. be interpreted as describing a uniform ungraded Periodic eddying which caused a fluctuation in response to temperature differences greater than the water temperature of the channel being modi- the threshold. ORIENTATION OF MIGRATING ANADROMOUS FISHES 387 RELATION OF TEMPERATURE RESPONSE TO the .fish (shown in group A, table 6) was due to TEMPERATURE LEVEL an increase in the threshold of the response to Tabulation of experimental data according to temperature differencesas the temperature level of stream temperature levels ( table 6) revealed a the stream increased. general tendency for the response to temperature ORIENTATIVE INFLUENCE OF CO2 differences between 0.5 ° and 1.0 ° C. to decrease as the temperature level of the stream increased. The migrating fishwere presented with a choice ,vhen the temperature differences between chan­ of waters having different amounts of free, or nels were greater than 1.0° C., the temperature uncombined, CO2. The difference in free CO2 level of the stream had no discernible influence between the waters of the two channels was estab­ upon the re�ponse of the fish. lished by the direct addition of gaseous CO2 to the water of one channel while the other remained TABLE 6.-Relation of temperature response to temperature level unmodified. During some of the tests this pro­ cedure was varied by the addition of the gaseous [ Retabulation of data from table 5. Stream temperature levels include +o.9° 11° 11.9° C. (e.g., C. includes temperature levels to C.)] CO2 to both channels but at different rates.

Entered channel with- Throughout the tests the fish indicated a definite preference for the water with the lower free CO2 Stream'temperatnre level Number of decisions Warmer Cooler content ( table 7) . water water

TABLE 7.-Responses to differences in CO, as shown in Group A: 1 Percent Percent n ° c ______-__ --___ ----. ----- 139 65 35 experiments, May 1---30, 1949, and May "I-June 9, 1950 12° C _. ______. ___ . _____ 178 65 35 13° C ______• 185 65 35 ° 14 c_. _____---- ___ ------274 60 40 15° c_. ______116 59 41 Entered channel with- 16° c ______° 124 55 45 17 C •• ______66 55 45 Free CO, differencebetween channels Number of 18° C ______85 54 46 decisions Lower Higher 19° C ______147 53 47 co, co, Group B: • 11 ° C ____ ------35 80 20 120 c ______43 79 21 Percent Percent 14° c ______-- --. ------238 77 23 >4.0 p. p. m______·------1,120 73 27 15° c ______217 80 20 1.4 p. p. m______------128 77 23 0.6 p. p. m______663 71 29 rn°° c ___ . __ . ______--___ . ---. --. -- 159 75 25 17 C ______290 76 24 0.3 p. p. m______216 69 31 18° c ______198 77 23 0.2 p. p. m ______157 59 41 20° c ______123 82 18 22° 27 78 22 c ______0.7 2.9 NOTE.-Unmodified stream water varied from p. p. m. free CO, to ° p. p. m. free CO2 and pH 6.9 to pH 6.4. Stream temperatures 11.1 ° to 22.3 1 Temperature difference between channels, 0.5° to 1.0° C. C. CO, differences listed include differences ±0.1 p. p. m. of listed differ­ • Temperature difference betweenchannels, 1.1° to 3.0° C. ences except 0.3 which includes only +0.1 p. p. m.

Taking into account once again the eddying con­ Whether the fish in the CO2 experiments were dition ( see p. 385), the data indicate that the re­ responding to differences in free CO2 or to associ­ sponses of the fish were affected by the tempera­ ated differences in HC03 is not actually known. ture level only when temperature differences of The data ( table 7) indicate that if the response threshold magnitudes were concerned. The evi­ was to differences in the amount of free CO2, the dence suggests a possible relation between this threshold of the response lies below differences phenomenon and the type of threshold phenomena of 0.3 p. p. m. If the response of the fish was to described by Weber's law. Weber (1846) be­ differences in HCO3, then the threshold of the lieved the threshold of difference to be propor­ response must be much lower. During most of tional to the intensity of stimulus. Although the the tests in which the differences in free CO2 were ratio of these two factors has since been shown less than 1.0 p. p. m., the differences in HCOi to be variable, a tendency for the threshold of were not even measurable by the method of meas­ difference to increase with an increase in intensity urement used (i. e., they were less than 0.1 p. p. m. of stimulus has been observed. The experiments HC03 as CaC03). It seems more probable that at Bournedale, planned with other purposes in the response of the fish was a response to free mind, did not produce the type of data necessary CO2. to examine this particular aspect of the response The experiments do not indicate whether the to temperature differences. However, it seems response of the fish assuming( that the response

very probable that the decrease in the response of was to free CO 2) was to differences in the amount 388 FISHE,RY BULLETIN OF 'IHE FTSH AND WILDLIFE, SERVICE' of free CO 2 or to differences in CO2 tension. The strate circumstances under which either factor difference in CO2 tension associated with a differ­ could balance or even dominate the other when ence of 0.3 p. p. m. free CO 2 under the conditions the factors were in opposition. The data also of these experiments would be approximately 0.1 suggest that when the two factors are not in op­ mm.Hg. position they may actually augment each other The response of the fish to differences in CO 2 and together provoke a response in a greater num­ ( whether it was to the amount of free CO 2 , to the ber of fish than either factor could produce alone. CO 2 tension, or to associated HCO 3 ) was similar ORIENTATIVE INFLUENCE OF pH to the response of the fish to temperature differ­ ences in that it appeared to be a uniform response In the preceding experiments involving the ad­ to all differences above the threshold difference. dition of gaseous CO2 to the water, the differences in the amount of free CO 2 between the two chan­ RELATIVE ORIENTATIVE INFLUENCE OF CO2 nels were always accompanied by differences in AND TEMPERATURE pH (see table 4). The question arises as to As the evidence was acquired indicating that whether the response shown by the migrating differences in temperature and differences in CO 2 fish is to the differences in CO 2 or to the accom- could influence the orientation of the migrating panying differences in pH. fish, the need for some information on the relative An attempt was made to answer this question influence of the two factors became increasingly experimentally with the migrating fish. It was obvious. Under the controlled conditions of the necessary to use a substance which would, when experiments, where all factors other than the one added to the water, modify the pH of the water used for testing were maintained equal in both to the same degree that it was modified during the channels, the fish might show as great a response CO 2 experiments ( see table 9) without, at the to a relatively minor influence as they would to same time, materially affecting the amount of free an important or dominant factor. To examine CO 2 in the water. It was also necessary that the the relative orientative influence of CO 2 and tem­ substance be one to which the fish would not perature, two groups of experiments were under­ respond by means such as taste or smell. To taken. In one set of experiments the directional avoid the difficulties of determining whether the influences of the two factors were arranged so as response of a fish was to the taste or to the smell to be in conflict. The differences in CO 2 favored of a chemical, or whether its response was to pH the choice of one channel, while the differences in differences created by that chemical, it was neces­ temperature favored the selection of the other. sary to select a substance to which the fish did In a second group of experiments the differences not respond at all. both in CO 2 and in temperature favored the selec­ In one of a series of exploratory tests ( table 10) tion of the same channel. NaOH was used to modify the water. There was The data (table 8) collected indicate that the no response to a difference in pH of 0.1 ( compare relative importance of the two orienting factors with response to pH difference 0.1 created by the depends upon their quantitative relationships. addition of gaseous CO 2 , table 9). However, it By altering the relative amounts of heat and of will be noted that when the difference in pH was CO2 added to the water, it was possible to demon- greater than 1.0 pH unit, the fish favored the

TABLE 8.-Relative orientative influence of CO, and temperature as shown in tests made June 3-11, 1950

Entered channel with- Water Free CO, temperature difference Number of Relation of factors difference between decisions Warmer Cooler Warmer Cooler between channels water and water and water and water and channels higher CO2 lower CO, lower CO, higher CO,

C. P.p. m. Percent Percent Percent Percent ؋° 99 24 76 Opposing______------{ 129 62 38 tHH! �:i!H 169 48 52 1-----1-----1----1----1 Augmenting______2. Oto 2. 4 0. 5 to 2. 0 154 ------83 17

NOTE.-Stream temperatures, 16.9° to 22.2° C. Unmodified stream water, 0.7 p. p. m. to 2.1 p. p. m. free CO,; pH 7.0 to pH 6.5. ORIENTATION OF MIGRATING ANADROMOUS FISHES 389 TABLE 9.-Ea;periments with ao. tabulated, accoraing to TABLE 12.-Measurements of free CO, taken auring the d,ifferences in pH, May 1949 tests shown in table 11 [Samples titratedwith NaOH] Percententering- Number pH differencebetween channels 1 of K2H(PO,) added Unmodified Difference Difference decisions Higher Lower control In free co, lnpH co, co, P.p.m. P.p.m. P.p.m. >1.0()______650 30 70 1.94 1. 96 0. 02 0. 2 0.10______663 29 71 1.62 1. 69 .07 .3 28 0.05 ______114 72 1.71 1. 78 .07 .4 <0.05 ______157 41 59 1.62 1. 70 .08 .4 1.31 1. 45 .14 .6 1 Range, pH 6.0 to pH 6.8. On the basis of the evidence from this experi­ TABLE 10.-Ea;periments with NaOH tabulated, accoraing ment, it would seem reasonable to· conclude that to differences in pH, May 24, 1949 the response shown by the migrating fish in the Percent entering- previous tests was a response to differences in CO2 Number rather than a response to differences in pH. pH differencesbetween channels of Channel decisions Control with NaOH channelI added 2 EXPERIMENTS USING NITROGEN AND OXYGEN > 1.00 _____ ------267 39 61 Nitrogen 0.10 ___ ------104 48 52 During the tests with CO2, particularly those 1 Unmodifiedstream water, pH 6.6 to pH 6.8; low pH; high CO, content. • High pH; low co, content. tests in which the CO2 had been added to only one of the two channels, the possibility had to be con­ modified water. Although CO2 was not measured sidered that the physical presence of many 'bub­ in this test it seems very probable that the dif­ bles in the modified channel might be influencing ference in free CO2 which would be associated with the choice of the fish. To eliminate this possi­ the large difference in pH would exceed the bility nitrogen was used as a control. The gaseous threshold 0.3 p. p. m. nitrogen was bubbled into the water of one of the­ After several acids and bases were tried,· an ex­ cha1.mels in the same manner and at the same rate periment was undertaken using K2H(P04) as the as the CO2 in the previous experiments. The ad­ modifying agent. The data collected ( table 11) dition of the N2 produced no measurable differ­ reveal no indication of a response on the part of ence in the amount of dissolved 02. The data ( table 13) reveal that the fish failed to show any the fish although differences in pH were present. response to the nitrogen or to the presence of the These differences in pH were of the same magni­ many bubbles produced. The nitrogen was then tude as those in many CO2 tests in which the fish used as a control in a series of CO2 tests. The responded. The accompanying table 12 of free -results of these tests show the same response to CO2 CO2 measurements made while the tests were in that was shown in previous tests, again indicating progress shows that during the tests differences of that the presence or absence of the bubbles had no free CO2 between channels were very small, gen­ influence upon the choice of the fish. erally less than the precision of measurement. Oxygen TABLE 11.-0rientative influence of pH as shoicn in tests maae June 10-12, 1950 To investigate the orientative influence of 02, (Differencesin pH Include differences::1:0,1 pH unit] the migrating fish were presented with a choice of waters which contained different amounts of dis­ Enteredchannel Unmod­ Differ­ Number with- solved 02• The water of one of the channels was K2H(PO,) added ified ence in of control pH decisions High pH Low pH modified by the addition of gaseous 0.2• The ex­ periment, however, was severely limited by the pH pH Percent Perce'llt 6.7 6. 5 0. 2 148 49 51 fact that the water of the stream was already more 6.9 6. 5 .4 127 51 49 7.1 6. 5 ,6 141 47 53 than 100 percent saturated. The data in table 13 7.3 6. 5 . 8 20 50 50 1------1-- ---•----,----,------show that the fish did not respond to a relatively TotaL___ , __ ------436 49 I s1 small difference in 02 under these conditions. The 390 FISHE:RY BULLETIN OF THE FISH AND WILDLIFE, SERVICE

TABLE 13.-Emperiments testing orientative influence of with less turbulence ( table 15). CO2 was then omygen and nitrogen added to the channel with less turbulence in order

Entered channel with- to discover the relative influence of turbulence and Number Entered 1 ______Factor of control CO2 • The data indicate that CO2 had the greater decisions channel N, CO, 02 orientative influence. �ded I �ded �ded ------___ --- - _ _ Percent Percent Percent Percent TABLE 15.-0rientatii:c influence of turbulence and 00, 300 50 50 I Nitrogen bubbles ______Nitrogen (control) ____ _ 477 ------71 29 ------Oxygen'------344 ------50 50 Entered channel with­ Nnmber or/----.--­ ° ° Influence of- 1 Stream temperature, 15.4 to 19.1 C. 02 content of unmodified stream decisions Maximum Minimum water, 10.5 to 10.7 p. p. m, 02 difference between channels, 1.1 p. p. m. turbulence turbulence experiments unfortunately provide no informa­ Percent Percent Turbulence______--__ ---- 26 32 68 tion on the possible effect of differences in 02 at Turbulence and CO, ______27 70 130 lower values where differences may be very important. 1 CO, added, 4 p. p. m. OTHER FACTORS INFLUENCING ORIENTATION An exploratory experiment was made which in­ dicated that visual factors may influence fish Although the major experimental efforts of this orientation. The downstream end of one chan­ investigation were concerned with the orientative nel was partially blocked ( see fig. 9). The fish influence of temperature and dissolved gases, ex­ normally swam within a few inches of the bottom ploratory experiments also examined the influence of the trough so that the "obstacle" did not inter­ of other factors. fere in any physical way with their progress. The influence of ·water velocity upon the orien­ However most of the fish entered the channel tation of the migrating fish was explored by sub­ . which ,vas completely unobstructed. To examme jecting the fish to a choice between waters of dif­ the relative influence of this visual factor and ferent velocities. The difference in velocity be­ temperature, the water of the partially blocked tween the two channels was created by placing a ° channel was heated 2 C. The data (table 16) glass plate across the upstream entrance to one of show that under these conditions temperature was the channels restricting the amount of water en- ' . the dominant orienting factor. tering that channel and so reducing the velocity These tests were crude experiments of an ex­ of the water in the downstream end of the channel. ploratory nature. There undoubtedly were some The response of the fish (table 14) indicated that velocity differences involved in the turbulence water velocity could be a factor in fish orientation. tests. The partial block at the downstream end The influence of water turbulence upon the ori­ of one of t.he channels in the tests involving the entation of the fish was explored in a similar man­ ner. The turbulence in one of the channels was visual factor probably created slight differences reduced by placing glass plates, several feet in length, in the center of the channel parallel to . OPEN the channel walls. This produced a flow which WOODEN 11>( II BLOCK11 ( 3 ) was smooth and laminar in appearance. In the I X IO other channel, small glass plates were set at an angle to create eddies which produced a visible tur­ bulence. Most of the fish selected the channel

TABLE 14.-0rientative influence of velocity Water velocity Entered channel with-- - - Number of --- Left decisions Higher Lower Right channel channel velocity velocity

Ft.fsec. Ft.{sec. Percent Percent 1.5 0.7 26 85 15 0.7 1.0 26 65 35 FIGl.'RE 9.-"Visual" factor (exploratory tests). ORIENTATION OF MIGRATING ANADROMOUS FISHES 391

TABLE 16.-0rientative influence of the visual factor TABLE 17.-Inff,uence of sere on the response to 002 and ------temperature Entered channel that was- Percent that entered channel Influenceof- Number of decisions with- Open Obstructed Nnmber 1------­ Factor and sex of decisions CO2 content- Temperatnre- Percent Percent Visual factor ______89 65 35 Higher Lower Hi�hcr Lower Temperatnre and visual factor ______93 40 160 ------1------° co,: 1 Heat added, 2 C. Male ____ . ---. _.. ------54 26 74 ------Female____ . -____.. -. __---. 70 27 73 ------Temperature: in hydraulic conditions between the two channels. Male ______------24 ------79 21 Female______------38 76 24

The experiments were done with relatively small ---�------numbers of fish and the degree of the response should not be interpreted too literally. However, more fish. This, however, did not occur. The the experiments are presented here bcause they response remained approximately 3 :1 even at the do indicate that characteristics of flow, such as maximum attainable differences (>7.00 p. p. m. ° velocity and turbulence, can have a directional free CO2 and 3.0 C.). influence upon migrating fish. They also indi­ The possibility was also considered that some in­ cate that, under some circumstances at least, dividuals might be completely insensible to the visual factors are capable of influencing fish differences in temperature and CO2 which were orientation. being used in the experiments. An experiment was set up in which fish that had previously been THE 3: 1 RATIO OF THE RESPONSE tested were again subjected to the same choice. The explanation for the persistence of the ap­ A trap was placed at the head of each channel proximately 3: 1 ratio in the response of the fishto and a series of CO2 tests were made. The trapped temperature differences and to CO2 differences is fish were then brought back to the entrance of the not readily apparent. The absence of a response experimental trough and the CO2 tests repeated. much closer to a 100 percent response under such The data ( table 18) show that the response of controlled conditions would seem to indicate that the fish that had previously entered the channel only a proportion of the fish were influenced by with the higher CO2 was approximately 3: 1 in the orienting factor. If, for example, half of the favor of the channel with the lower CO2, and those fish were influenced by the testing factor and the fishwhich had previously entered the channel with other half entered the channels at random, the the lower CO2 also exhibited a 3 : 1 response in resulting ratio would be 3 : 1. favor of the channel with the lower CO2. A possible sexual variation in the response of Such evidence strongly suggests that the 3 : 1 the fish was considered. A trap was placed at the ratio is not due to the failure of particular in­ head of each channel and after a series of CO2 dividuals or particular groups of individuals to tests had been run, the fish in each trap were respond to the orienting factor. It seems more examined for sex. This procedure was later re­ probable that the explanation lies in the behavior peated with a series of temperature tests. The patterns inherent in all the fish. If, for example, data (table 17) indicate that the sex of the fish every fish responded to an orienting factor only has no effect upon its response to differences in half the time and acted at random the other half, CO2 and temperature. the result would also be a 3: 1 ratio. Further The possibility of individual variation in the f sensitivity of the fish to temperature diferences TABLE 18.-Retesting the response of fish to 002

and to CO2 differences was examined by gradually Entered channel with- increasing the differences between the two chan­ Fish from channel with- Number of 1------.---­ decisions Higher Lower nels. Had there been significant individual vari­ co, co, ation in sensitivity, the response would have be­ Percent Percent come greater as the difference between channels Lower CO, ______------111 27 73 Higher gradually exceeded the thresholds of more and co, ______------54 30 70 392 FISHERY BULLETIN OF 'I'HE FISH AND WILDLIFE, SERVICE experiments may be necessary to throw some light in the selection of streams leading to lakes or on the nature of the ratio and to learn whether ponds. the phenomenon has any significance beyond the The fundamental nature of the response of the restricted circumstances of this experimental fish to differences in CO2 and to differences in method. temperature is indicated by the manner in which DISCUSSION CO2 and temperature were able to dominate com­ peting orienting factors in the exploratory tests When the results of the preceding experiments shown on page 390. The low threshold values for are compared with the findings o:f other investi­ these responses also suggest their probable gators several interesting possibilities are sug­ importance. gested. One possibility is that the response of the The experiments examining the orientative in­ fish to temperature differences might be a family fluence of pH (p. 388) confirm the conclusion o:f characteristic. The two species of fish, Pomolobus Powers (1930) that pH was largely ineffective as pseudoharengus (Wilson) and Pomolobus aesti­ a :factor influencing the behavior of aquatic ani­ (Mitchill), which at Bournedale showed a valis mals. Whether the response of the fish in the preference for warmer water, are mem­ consistent experiments at Bournedale was to differences in bers of the family Clupeidae. Shelford and free CO. or to the associated differences in HCO P 3 owers (1915) found that herring, Olupea was not-actually established although the thres­ fry preferred warmer water. The obser­ pallasi, hold values involved suggest that the response was vations of Roule (1933) indicated that migrating to free CO • If the response was to the free CO shad Paralosa nilotioa rhodanensis and Alosa 2 2 the influenceof CO2 differencesis limited to waters alosd, which also belong to the :family Clupeidae, with a pH of less than 8.4. This :fact might be selected water o:f higher temperature. used to advantage iri an experiment to determine The threshold o:f the response to temperature to which of the two factors the fishare responding. differences shown by the two species of fish at ° Although the main efforts of this investigation Bournedale (0.5 C.) agrees very closely with that were directed toward examining the orientative reported by Shelford and Powers (1915) for influence of CO2 and temperature upon the mi­ herring fry. This threshold may seem high com­ grating fish, this was not meant to imply that pared to the minimum effective thermal stimulus these are the only major influences which might range for certain fresh-water fishes ( differences be concerned in other situations. The influence ° ° of 0.03 to 0.07 C.) reported by Bull (1936). of differences in dissolved 0 may be of great im­ However, it must be remembered that this temper­ 2 portance when lower 02 values are concerned, ature difference represents the minimum tempera­ which was not the case at Bournedale. Under ture difference which will provoke an uncondi­ special conditions flow characteristics such as ve­ tioned orientative response in the fish. The locity and turbulence may play an important part minimum temperature difference which can be in directing the fish. The role of olfactory mem­ perceived by the fish is probably much less. ory needs to be further explored. The influence Powers and Hickman (1928) presented evidence o:f factors affecting the orientation of the fish in­ to show that rivers draining lakes usually had directly ( e. g., by controlling the depth of swim­ lower CO2 tensions than other rivers (average ming) must be considered. It was largely to difference0.5 mm. Hg). Powers (1939) contended stress the fact that many factors may be CQn­ that by means of these differences in CO2 tension cerned in fish orientation that the exploratory migrating fishcould select certain types of streams. tests involving velocity, turbulence, and visual The results of the CO2 experiments at Bournedale factors were included in this report. would fit well into his argument. The alewife Perhaps one of the most important considera­ and the glut herring both showed a strong prefer­ tions to which the foregoing experiments call at­ ence for water o:f lower CO2 tension (threshold tention is that not only are there many factors difference 0.1 mm. Hg). These fishes usually which may have a directional influence upon the spawn in small ponds and shallow lakes and the migrating fish but also that they must all be con­ choice of water of lower CO2 content would result sidered together. The experiments examining ATION OF MIGRATING ANADROMOUS FISHES 393 the relative influence of CO2 and temperature fled water. It, therefore, seems improbable that demonstrated conditions under which the response the response of the fish in the other CO2 experi­ to temperature differences could dominate the re­ ments was a response to change in chemical sponse to CO2 differences when the factors were conditions. in opposition. The reverse situation · was also Powers and Clark ( 1943), discussing certain demonstrated. The experiments further revealed gradient-tank experiments with brook trout, circumstances under which the two factors were Sr.ilvelinus f. fontinalis, and rainbow trout, Salmo equal and the influence of the one balanced the gairdneri iridus, presented evidence indicating influence of the other. Had an observer at these that the reactions of the trout to CO2 in these experiments been considering only temperature, experiments might be characterized as reaction for example, he would have seen the fish choose to change. They made the observation that "The the warmer water in one group of tests and the immediate response to carbon dioxide tension of colder water in another. During the third group the water depends to a large extent upon the car­ of tests he would have concluded that the fish bon dioxide tension to which the fish is adjusted." were indifferent to temperature. It is quite pos­ The period of adjustment in their experiments sible that many apparently contradictory observa­ was only 10 to 15 minutes. • tions may be explained in this way, e.g., the If this observation were true for migrating fish observations of Ward (1920), Chamberlain then CO2 would probably have little directive in­ (1907), and Foerster (1929), regarding the in­ fluence. The fish adjusted to the CO2 tension of fluence of temperature on sockeye salmon. the sea would be restrained from entering fresh One of the considerations that had to be taken water with a different CO2 tension. At a junc­ into account during the experiments at Bourne­ tion between two streams the fish would tend to dale was that the fish might be reacting to a select the stream with the larger volume in an change in water conditions. This possibility be­ effort to remain in the CO2 tension nearest to the came evident in one of the exploratory tests ex­ one to which it had become adjusted. amining the influence of light intensities. One A possible explanation is suggested by an ear­ channel of the experimental trough was shaded lier statement of Powers ( 1939) : and the other was left exposed to full sunlight. The ova and sperm heads contain a protein (pro­ When the undivided lower portion of the experi­ tamine) containing large percentages of arginine. It re­ mental trough was shaded, the fish chose the quires a vast protein metabolism to obtain the necessary , shaded channel. When the lower part of the arginine. Protein metabolism and especially fasting­ both necessary for the liberation of arginine from the trough was exposed to sunlight, the fish chose the muscular tissue of the salmon-tends to produce acidosis sunlit channel. Thus the fish showed no prefer­ of the blood, i. e., lower the alkali reserve of the blood. ence for either the sunlit or the shaded channel. This is common knowledge. A salmon with a low alkali They simply avoided any change in light reserve blood would find low carbon dioxide tension conditions. water more advantageous. During experiments in which water was modi­ The special physiological state of the migrating fied by the introduction of gaseous CO2 the fish fish might prevent it from becoming adjusted to avoided the modified water. Although the fish its environment. entered the trough in a mixture of water from As experimental evidence is acquired indicating both channels ( fig. 2), and before entering they physical and chemical differences may have a di­ had been subjected to a mixture of both waters in rectional influence upon migrating fish, the the upstream end of the retaining pool (fig. 4), method by which the fish is oriented to these dif­ the possibility that the reaction of the fish was ferences becomes an important consideration, par­ one to change in chemical conditions still had to ticularly when one is trying to relate artificiale x­ be considered. However, in one experiment perimental conditions to situations found in which was made at Bournedale, the CO2 content nature. of the water of one channel was reduced by the There are two methods by which fish might be­ addition of Na OH. In that experiment (se e come oriented in a gradient. The fishmight make p. 388) the fish showed a preference for the modi- a simultaneous comparison of intensities of stimu- 394 FISHEiRY BULL:ETIN OF THE FISH AND WILDLIFE• SERVICE lation by means of symmetrically placed receptors Orientation by means of a comparison of in­ and turn toward (or away from) .the maximum tensities that are successive in time would make stimulation. I£ this was the method of orienta­ it possible £or fish to become oriented in much tion, the orientative influence of gradients would smaller gradients than would be necessary if orien­ be limited to those gradients great enough so that tation was by a simultaneous comparison of in­ the differences between points a few inches apart tensities. The minimum gradient in which a fish ( the distance between paired receptors) are above could orient would depend on the speed of the the threshold of sensory perception of fish. Grad­ fish and the time interval of its sensory memory. ients as high as this do actually exist in nature but The experiments at Bournedale have demon­ usually only as a narrow zone between two bodies strated that one group of migrating anadromous of water which are just beginning to mix. fishes will orient with reference to differences in A second method by which a fish might orient certain physical and chemical water characteris­ in a gradient is similar to the "trial" method of tics created artificially. It is logical to suppose orientation characterized by Jennings (1906) as that the fish will also be oriented by similar dif­ "selection from among the conditions produced ferences occurring along their migration routes J.iy various moveinents." He points out that in in nature. this type of behavior, "Each stimulus causes as a SUMMARY rule not merely a single definite action that may be called a reflex,but a series of "trial" movements, The purpose of this investigation was to exam­ of the most diverse character, and including at ine experimentally the influence of certain physi­ times practically all the movements of which the cal and chemical water characteristics upon the animal is capable." This type of orientation orientation of one type of migrating anadromous Fraenkel and Gunn ( 1940) have labeled "klino­ fish. The migrating fish were presented with a taxis," and defined as "a directed orientation choice between two channels that carried water made possible by means of regular deviations and with differentcharacteristics. The orientative in­ involving comparison of intensities at successive fluence of the water properties in question was points in time," and they pointed out examples of measured by the number of fish selecting each such behavior in many invertebrates. In this channel. The reactions of more than 8,000 fish method of orientation the fish, subjecting itself to of the genus Pomolobus-alewife, P. pseudoharen­ varied conditions by its active movements, would gus (Wilson), and glut herring, P. aestivalis select the most favorable condition or direction. (Mitchill)-were tested as the fish migrated up­ The method involves both movement and com­ stream through the Herring River at Bournedale, parison of intensities of stimulation that are suc­ Mass., toward their spawning area, in the springs cessive in time. of 1949 and 1950. The behavior of the fish as they selected a chan­ The fish were not removed from the stream or nel in the experimental trough at Bournedale handled in any way. Each was tested individ­ strongly suggested that this latter method of ori­ ually and tested only once ( the few exceptions are entation is the one used. As the fish entered the noted). The findings of this investigation were: trough they usually swam from one side to the 1. Presented with a choice of waters having dif­ other, approaching first one channel and then the ferent temperatures, 77 percent of the fish entered other. When once the fish had left the narrow the channel with the warmer water when the tem­ zone in which the water was mixing (fig. 2) it was perature difference continuously exceeded 0.5 ° C. no longer subjected to a gradient. Its sensory The temperature differences examined ranged ° receptors were all subjected to the same intensity from 0.4° to 3.0 C. Water temperatures during of stimulation and, therefore, the current was the experiments varied from 11.1 ° to 22.3° C. only orientative influence directing the movement 2. The response of the fish to temperature dif­ of the fish upstream, unless memory was involved. ferences near the threshold difference decreased Yet the fishrepeatedly swam from one side of the as the temperature level of the water increased. trough to the other and frequently even after en­ 3. Presented with a choice of waters having tering one channel, they turned back and entered different amounts of free CO2 , 72 percent of the the other. fish entered the channel with water of a lower ORIENTATION OF MIGRATING ANADROMOUS FISHES 395 free CO2 content when the free CO2 dMference CHAMBERLAIN, F[REDERIC] M. exceeded 0.3 p. p. m. (0.1 mm. of Hg in terms 1907. Some observations on salmon and trout in Alaska. Report of the Commissioner of Fisheries of CO2 tension). The differences in free CO2 ex­ for the Fiscal Year of 1906 with special papers, U. S. amined ranged from 0.2 p. p. m. CO2 to greater Bureau of Fisheries Doc. No. 627, 112 pp., illus. than 4.0 p. p. m. The amount of free CO2 in the CHEVEY, [P.], [L.] ROULE, and [M.] VERRIER. water during these experiments varied from 0.8 1927. Sur l'interruption de la montee des Saumon par p. p. m. to 7.1 p. p. m. la diminution de la teneur du cours d'eau en oxygene 4. The sex of the fish appeared to have no in­ dissous. Comptes Rendus, L'Acad. Sciences 185 (25) : 1527-1528. Paris. fluence on its response to differences in CO con­ 2 CHIDESTER, F[LOYD] E. tent or in temperature. Hl24. A critical examination of the evidence for physi­ 5. Exploratory experiments indicated that vis­ cal and chemical influences on fishmigration. British ual factors and such factors as velocity and Jour. Exper. Biol. 2: 79-118. turbulence can influence orientation. ELLIS, M. M., B. A. WESTFALL, and MARION D. ELLIS. 6. The fish did not respond to a difference of 1948. Determination of water quality. U. S. Fish and Wildlife Service, Research Report 9, 117 pp. 1.1 p. m. m. 02 created during experiments. The amount of 0 in the water during the experiments FOERSTER, R[USSEL] E. 2 1929. Notes on the relation of temperature, hydrogen­ ranged from 10.5 p. p. m. to 10.7 p. p. m. and ion concentration, and oxygen to the migration of ° water temperatures ranged from 15.4 to 19.1 ° C. adult sockeye salmon. Canadian Field-Nat. 43 (1) : 7. The fish were indifferent to pH differences 1-4. as large as 0.8 pH unit when associated differences FRAENKElL, GOTTFRIED, and DONALD L. GUNN. in free CO2 were less than the threshold 0.3 p. p. m. 1940. The orientation of animals. 352 pp. Clarenden The pH of the water varied from 6.5 to 7.3 during Press, Oxford. these experiments. JENNINGS, H[ERBERT] S. 8. The relative orientative influence of CO2 1906. Behavior of the lower organisms. 366 pp. and temperature, when the two factors were in Columbia University Press, New York. opposition, was shown to depend upon their rela­ MOORE, EDWARD w. tive differences of magnitude. A difference in 1939. Graphic determination of carbon dioxide and ° the three forms of alkalinity. Jour. Amer. Water temperature of approximately 2 C. domin�ted Works Assn., vol. 31, pt. 1, p. 51. an opposing difference in free CO of 2.0 p. p. m. 2 POWERS, EDWIN B. A differenceof CO2 slightly in excess of 2.4 p. p. m. ° 1930. The relation between pH and aquatic animals. balanced the opposite effect of a 2 C. tempera­ Amer. Nat. 64 (693) : 342-366, illus. ture difference. A difference of >7.0 p. p. m. CO2 1939. Chemical factors affecting the migratory move­ ° dominated over a temperature difference of 0.6 C. ments of the Pacific salmon. Publ. Amer. Association 9. The behavior of the fish during the experi­ for the Advancement of Science No. 8, pp. 72--85. ments indicated that the orientation was accom­ POWERS, EDWIN B., and R. T. CLARK. plished by a method of "trial" involving both 1943. Further evidence on chemical factors affecting movement of the fish and a comparison of in­ the migratory movements of fishes especially the tensities of stimulations which were successive salmon. Ecology 24 (1) : 109-112. in time. POWERS, EDWIN B., and T. A. HICKMAN. 1928. The carbon dioxide tensions of the Fraser River LITERATURE CITED and its lower tributaries and of certain tributaries of AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE. the Columbia River. Puget Sound Biol. Sta. Publ. 1939. The migration and conservation of salmon. 5: 373-380. Publ. No. 8, 106 pp. Science Press, Lancaster, Pa. ROULE, LOUIS. AMERICAN PUBLIC HEALTH ASSOCIATION. 1914. Sur l'infl.uence exercee sur la migration de mon­ 1946. Standard methods for the examination of water tee du Saumon (Salmo salar L.) par la proportion and sewage. Ed. 9, 296 pp., illus. New York. d'oxygene dissous dans l'eau des fleuves. Comptes BULL, HERBERT 0. Rendus, L'Acad. Sciences 158: 1364--1366. Paris. 1936. Studies on conditioned responses in fishes. Part 1933. Fishes. Their journeys and migrations. Trans. VII. Temperature perception in teleosts. Jour. Ma­ from the French by Conrad Elphinstone. 270 pp., rine Biol. Assoc. United Kingdom (n. s.) 21 (1): 1-27, mus. W. W. Norton and Co., Inc., New York. illus. 396 FISHERY BULL:E'T'lN OF THE FISH AND WILDLIFE- SERVICE RUSSELL, E[DWARD] S. WARD, BltNRY BALDWIN. 1934. The behavior of animals; an introduction to its 1920. Some features in migration of the sockeye salmon study. 196 pp., illus. E. Arnold and Co., London. and their practical significance. Trans. Amer. Fish­ rRevised 1938.] eries Soc. 50 : 387--426. SHELFORD, VICTOR E., and EDWIN B. POWERS. WEBER, E. H. 1915. An experimental study of the movements of her• 1846. Der Tastsinn und das Gemeingefiihl. In Dr. ring and other· marine fishes. Biological Bull. 28 Rudolph '\Vagner's Handwoerterbuch der physiologie, (5) : 334. Band III, Abth. 2, pp. 481-588. Braunschweig.