Loading...
HomeMy WebLinkAboutAPA2149I,'t lXl&OO~&c mOO&®©@ Susitna Joint Venture Document Number Please Return To DOCUMENTCONTROL The Role of Behavior in the Ecology and Interaction of Underyearling Coho Sahnon (Oncorhynchus kisutch) and Steelhead Trout (Salmo gairdneri)1.2 By G.F.HARTMAN Fish and Game Branch Department of Recreation and Conservation Vancouver,B.C. ABSTRACT Two similar salmonids,coho and steelhead,cohabit many coastal rivers of British Columbia. Field collections reveal that the distributions of underyearling coho and steelhead are similar along the length of these streams.However,the micro-distribution of the two species is different. In spring and summer,when population densities are high,coho occupy pools,trout occupy riffles.In autumn and winter,when numbers are lower,both species inhabit the pools.Nilsson (1956)stated that segregation (such as that shown by coho and trout in spring and summer) may be indicative of competition resulting from similar ecological demands.To test this hypothesis the distribution ·and behavior of coho and steelhead were compared in a stream aquarium at different seasons with gradients of light,cover,depth or depth/velocity,and in experimental riffles and pools.Distributions and preferences of the two species in the experimental environments were most similar in spring and summer,the seasons when segregation occurred in nature,and least similar in autumn and winter,the seasons when the two species occurred together in nature. Spring and summer segregation in the streams is probably the result of interaction which occurs because of similarities in the environmental demands of the species and which is accentuated by dense populations and high levels of aggressiveness.The species do not segregate in streams in winter because certain ecological demands are different,numbers are lower,and levels of aggres- siveness are low.When the two species were together in the experimental riffle and pool environ- ment,trout were aggressive and defended areas in riffles but not in pools;coho were aggressive in pools but less inclined to defend space in the riffles.These differences in behavior probably account for the distribution of trout and coho in natural riffles and pools. The data support the basic contention of Nilsson (1956)and illustrate the role of behavior in segregation produced by competition for space. INTRODUCTION Two SIMILAR species of salmonids occur together in many of the coastal streams of British Columbia.These fish,coho salmon (Oncorhynchus kisutch)and steelhead trout (Salmo gairdneri),resemble each other in several respects. Both coho and steelhead are anadromous.Young coho remain in the streams where they hatch for 1 or 2 years,steelhead may remain in the stream up to 3 years.In addition,they are similar in size and morphology.On the basis of present concepts,they are potential competitors.3 lReceived for publication December 28,1964. 'Based on a thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy at the University of British Columbia. 3The meaning of competition,when used in this paper is,"The demand,at the same time, of more than one organism for the same resources of the environment in excess of immediate supply"(Milne 1961).The meaning of niche is as given in DeBach and Sundby (1963). 1035 J.FISH.RES.BD.CANADA,22(4),1965. Printed in Canada. This document is copyrighted material. Permission for online posting was granted to Alaska Resources Library and Information Services (ARLIS) by the copyright holder. Permission to post was received via e-mail by Celia Rozen, Collection Development Coordinator, on March 11, 2013, from Eileen Evans-Nantais, Client Service Representative, NRC Research Press. This article is identified as APA 2149 in the Susitna Hydroelectric Project Document Index (1988), compiled by the Alaska Power Authority. 1036 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 Young coho and steelhead coexist along the lengths of streams, but within each stream they occur in different microhabitats. Their ecology is given more fully in the text. To introduce the problem, it is sufficient to note that microhabitat segregation is pronounced in the spring when population den- sities are high and breaks down during fall and winter when population levels are low. Nilsson (1956) developed valuable concepts which may provide an ex- planation for the separation of coho and trout during spring and early summer. He postulates that allopatric populations of closely allied species, or species having similar ecological demands, utilize the full range of their ecological potentials under conditions of intense intraspecific competition. Intense in- terspecific competition in sympatric populations of similar or closely related species forces each species to compete only at its "ecological optimum," i.e. under those conditions to which it is best adapted or where it has some com- petitive advantage. It is this tendency of species to utilize only their ecological optima that results in segregation during rigorous interspecific interaction. On the basis of Nilsson's ideas it seems likely that certain environmental requirements and aspects of behavior of young trout and coho may be more similar in spring and early summer than during fall and winter. Accordingly there are two objectives to this study. The first is to obtain and present data from the field study outlining in some detail the period and nature of inter- specific segregation. The second objective is to compare, under partially con- trolled conditions, the environmental responses, preferences, and behavior of young coho and trout, and to determine if segregation occurs when these are most similar. Essentially this involves testing Nilsson's ideas (Nilsson, 1956, 1963) under controlled conditions. If the species are segregated at that period when requirements are similar, attempts will be made to ascertain the role of behavior in this interaction. Newman (1956), Lindroth (1955a), Kalleberg (1958), and Nilsson (1963) have shown how behavior enters into interspecific competition. It is not clear, however, what type of behavioral mechanism func- tions to give one species an advantage over the second species in one situation and for this advantage to be reversed in another situation. The research is described in two sections; the first deals with studies under field conditions, the second deals with studies in a partially controlled environ- ment. Methods and Results are within each section. FIELD STUDY DESCRIPTION OF STUDY AREA Three rivers in the lower Fraser valley of southwestern British Columbia were studied (see insets of the Chilliwack, Alouette, and Salmon Rivers in Fig. 1). CHILLIWACK RIVER The Chilliwack River rises in the Cascade Range in Washington and drains north into Chilliwack Lake (elevation 620 m), thence it flows west HARTMAN: BEHAVIOR OF SALMON AND TROUT FIG. 1. Lower Fraser valley area in southwestern British Columbia showing Alouette, Chilliwack, and Salmon River systems. 1037 into the Fraser River. Figure 2 shows the portion of the river studied and station locations. The river runs through a deep valley in a stable rock channel from Chil- liwack Lake to the region at V-28 (Fig. 2). Large areas of the upper river are covered with extensive log jams (Fig. 3,A). In the middle stretches of the river (V-28 to V-13), the channel bottom is less stable and large log jams are absent (Fig. 3,B). Several large tributaries (Slesse, Foley, Chipmunk, and Tamihi Creeks) enter this stretch of the river. The discharge of these tributaries fluctuates considerably, hence, below V-13 the channel is braided and much FIG. 2. Chilliwack River and location of stations. 10 38 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 FIG. 3. A. Upper Chilliwack River in a rea of large lo g jams, V-36 to V-38. B . Chilliwack River in the region of V-17 .5 . C. Chilliwack River at V-5.5 and V-6. Note braided channel and unstable gravel bars. of the bottom is unstable (Fig. 3,C). There is one log jam near V -5. The low- ermost region of the river travels across flat terrain and the bottom is composed of unstable sand a nd gravel. Much of the channel is modified and dyked (see V-5 to V-2 , Fig. 2). The Chilliwack River drains an area of 1250 km2 • Between 1958 and 1962, highest flows occurred during two periods each year. In the first peak in May or June mean monthly flows ranged from 105 to 184 m 3 /sec . In the second peak, October to January, mean monthly flows ranged from 62 to 96 HARTMAN: BEHAVIOR OF SALMON A N D TROUT 1039 m 3 /sec (data from the Water Resources Division of the Department of North- ern Affairs and National Resources). During 1960, 1961, and 1962, mean monthly water temperatures ranged from a low of 2 C in January to a high of 13 C in August. (Details of tempera- ture and discharge in the Chilliwack, Alouette, and Salmon Rivers a r e recorded in the thesis upon which this publication is based .) ALOUETTE RIVER The South A louette River originates at the west end of Alouette Lake . It drains west into the Pitt River, a tributary of the Fraser. Figure 4 shows stations a long the portion of the river which was studied. The upper river (A-16 to A -9) runs down a shallow valley through big pools and stretches strewn with large boulders (Fig. 5,A). The stream bottom is relatively stable in most areas above A-9. From A-9 to below A -6 the river passes through flat terrain in a channel with an unstable rock and gravel bottom. In this region the river bed has been modified considerably to prevent flooding (Fig. 5,B). Lower sections of the Alouette (A-3 to A-0) lie in meadow- land (Fig. 5,C). The stream bottom is composed of fine gravel, sand, and mud. Log jams are absent along the full length of the river. The Alouette River drains an area of 205 km2 • During 1960, 1961, and 1962, highest mean monthly flows, 5.2-9.8 m 3/sec, occurred in January. Low- est discharges occurred in July and August and ranged from 0.2 to 0.4 m 3 /sec. Although mean monthly discharges in winter are quite low, the Alouette River occasionally freshets violently, changing from flows of 2-3 to 112 m 3 /sec over a period of 5 or 6 days (data from Water Resources Division of Department of Northern Affairs and National Resources). N.ALOUETTE 2 MILES S.ALOUETTE RIVER FrG. 4. Alouette River and location of stations. 1040 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 FrG . 5. A. Alouette River at A -12, channel with large boulders on bottom and with stretches containing large pools. B. Alouette River at A-7. Note gravel bottom and modified channel. C. Alouette River at A-2 in meadow and farm la nd. Mean monthly temperatures, durin g 1960, 1961, and 1962, ranged from 2.8 C in January to 20 C in July. SALMON RIVER The Salmon River rises in low wooded farm l and, northeast of Langley, British Columbia, at an elevation of about 100 m. Figure 6 shows the location of stations along the Salmon River. The upper tributaries of the river lie in small valleys and have stable channels (see S-12 to S-19, and S-10 to S-18, Fig. 6). The middle section of the river lies in almost flat surroundings in a sand and gravel bed (Fig. 7,A). Below S-6 the river meanders through meadows in a mud and sand channel. HARTMAN: BEHAVIOR OF SALMON AND TROUT ___ .. _........1 I~ Ia: I MILE /z I ~ • I ~ I ls-12 ----.._,__._ TRANS CANADA --.._____,~HWAY --------- 5·18 ' FIG. 6. Salmon River showing location of stations. 1041 The overall character of the flow is gentle, with much of the river co nsisting of ripples and p oo ls lying in we ll-wooded areas (Fig. 7,B). The Salmo n Ri ve r drains an area of 83 km2 . Mean monthly flows (based on ·water Resources Division data) reach a peak between November and February. In 1960, 1961, a nd 1962, mean monthly flows r a n ge d from 0.23 m 3 /se c in July and August to 5 .9 m 3 /sec in January and February . During winte r, large short-term fluctuations in discharge occur, e.g. from 0 .8 5 m 3 /sec on J a nu ary 23 to 8.32 m 3 /sec on January 25, 1960 . In the Salmon Ri ver, during 1960, 1961, and 1962, mea n monthly t e m- peratures r anged from 3.2 to 3.8 C in December a nd J a nuary, and from 12.9 to 15.5 C in June, July, and Augu st. The three ri ver s studied contain a variety of types of habita t . These h a bitats r anged from sm a ll, low-elevation tributa ries with gentle flows, to l arge, r apid, and turbulent rivers. Bottom conditions in each stream vary from uns t able sand and gravel to s table gravel or boulders. In the fie ld work dea lt with in the nex t section, fish co ll ections were m a de over the fu ll r a n ge of conditions described for the three s treams . MATERIALS AND METHODS Young coho salmo n and steelhead trout were studied in three lower m a inland rivers of Britis h Columbi a, the Chilliwack River (Fig. 2), the Alouette 1042 JOURNAL F I SHERI ES RESEA RC H BOA RD OF CANADA, V OL. 22, N O. 4, 1965 FrG. 7. A. Salmon River at S-9. B. Riffle and pool areas at S-14. River (Fig. 4), and the Salmon Rive r (Fig. 6). Twelve to 16 stations per river were visited about once a month from November 1959 to March 1962. A routine collection procedure was followed if conditions permitted. Fish were a lways collected by seining in the Salmon River and wherever possible in the larger rivers. In addition C.I.L. "Prima Cord," a detonating fuse, and electrical blasting caps were u sed to collect fish among the large boulders and under log jams. The explosive was detonated in the stream above a set seine net a nd the fish drifted into the seine; the blast a rea (usually small) was searched. Records, kept on a standard data sheet and a sketch map for each station, included number of fish collected and approximate area of stream bottom sampled. Temperatures were taken by Weksler temperature recorders. Stream velocities were calcul ated from the rate of movement of floating objects. Turbidity a nd bottom composition were recorded on a rough quantitative basis . In addition, distribution data based on collections were supplemented with a series of diving observations in the Chill iwack River. In each diving census the number, behavior, and distribution of fish were recorded in three HARTMAN: BEHAVIOR OF SALMON AND TROUT 1043 standard census strips on the stream margin at V -2 8 and V-30. Each strip was 67 m (200 ft) long and about 1 m wide. RESULTS LONGITUDINAL DISTRIBUTION IN STREArviS Highest densities of young steelhead trout and coho in the Chilliwack occurred in the upper reaches of the river (Fig. 8). High densities recorded in Fig. 8 (V-29 to V-38) were not necessarily representative of the entire upper river. However the type of habitat where highest numbers were recorded (large log jams) was characteristic of the upper part of the C hilliwack. In this region, the river bed was more stable and offered better shelter to fish. For this reason numbers of fish at stations V-22, V -28, and V-30 (Fig. 8), where log jams were absent, were higher than those at comparable locations (V-1 7.5, V-13, and V-10) in the lower h a lf of the river (Fig. 8). In the down- stream portions of the Chilliw ack (below V-10) where the bottom was unstable and the channel was braided, numbers of both species were lowest . The rel- ative numbers of trout and coho and the general distribution pattern were the same in winter as in the period from March to September (Fig. 8). At most stations, coho were more abundant than steelhead. During the early summer young coho were captured further downstream than steelhead; but aside from this the two species were distributed in a similar fashion along the length of the river. In the Alouette River, greatest concentrations of coho and steelhead occurred at stations A -9, A -12, A-14, and A-15 (Fig. 9), a ll of which are char- acterized by a cover of heavy boulders (Fig. 5,A). Below A -9 in areas with mud, sand, or unstable gravel bottom (Fig. 5,B and C), densities were low I STEELHEAD OCOHO V·2 -4 5 I 0 5.56 7 I 10 MAR I-SEPT 30 OCT I -FEB 28 ......... ::::::::::::::::::i::::::::::::::::::::::::::::: 250 ···::::::::::::::::::::::::::::::::::::::::;:::;:;:;::;:::::::::::::::. 0 I~ 17.5 I 22 28 1'29 30 32 36 38 I STATION NO 20 30 40 50 KM FrG. 8. Density and distribution of young coho and steelhead in the Chilli- wack River (data combined for period from November 1959 to March 1962). 1044 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 I STEELHEAD 0 COHO MAR 1-SEPT 30 J:oiLo'F"""F: :: =r=:o :~o ~~~ ?; z Q 100 50 0 ~ OCT I-FEB 2B ~ ,:o~ :' ': : ': : : : ;w .. ~ .. ~¥1-•: N~ 0 Q a: UJ a. ~ u: A-0 I 0 2 3 6 7 8 9 I 10 12 14 IS 16 1 STATION NO 20 KM FIG. 9. Density and distribution of young coho and steelh ead in the South Alouette River (data combined November 1959 to March 1962). (Fig. 9). During the winter, density of both species was reduced to a low level, probably due to violent winter freshets which scoured the river periodically. In the period March 1 to September 30, coho fry were distributed further downstream than steelhead. Figure 9 shows that, a lthough relative numbers of trout and coho varied at different stations, both species were distributed together along much of the river. In the Salmon River, unlike the Chilliwack and Alouette Rivers, young coho occurred with young steelhead trout and cutthroat trout (Salmo clarki). It was not possible to identify and separate the two trout species during their first few months of life. Reference to trout in the Salmon River therefore includes some cutthroat as well as steelhead. Highest densities of young trout and coho were recorded in the upper part of the Salmon River (S -9 to S -1 8, Fig. 10) and in its upstream tributary (S-12 to S -19, Fig. 10). This area of the river was characterized by small pools and gentle riffles (Fig. 7, A and B). Much of the shoreline was overgrown and covered with fallen trees. Below S-8 the bottom was composed of unstable gravel or sand and mud; numbers of coho and trout in this area were lower (Fig. 10). In early summer young coho were distributed down the Salmon River into the mud bottom portions of the stream at S-2 and S-3. Coho den- sities were higher than trout at all stations, but both species occurred together over most of the length of the stream (Fig. 10). HARTMA N : BEHAVIOR OF SALMON AND TROUT I TROUT [] COHO APRI-OCT 31 I IOOiL--~--~~--~~---FT=~qp~lliillqillhl¥lli__ ~ ~ z 0 ... < > LLI .J LLI I NOV I-MAR 31 ISO 100 so 0 a: LLI a.. r IJ) IL S-1 2 3 6 8 9 10 13 17 18 STATION NO 0 I 10 I 20 KM FIG. 10 . Density and distribution of young trout and coho in the Salmon River (data combined November 1959 to March 1962 ). Broken bar in upper part of Figure represents 466 coho. 1045 A variety of types of physical habitat w as studied within each of the three rivers . Furthermore, size, bottom, a nd flow conditions differed consid- erably among the streams. Methods of sampling were not the same in all rivers. In spite of such differences in habitat and sampling the two species exhibited comparable distribution patterns in each of the three ri vers (Fig. 8-10). Trout and coho cohabite.d the lower sections o f the streams in low numbers and occurred together in highest numbers in the more stable envir- onments near the headwaters or headwater lakes. MICROHABITATS OF UNDERYEARLING COHO AND TROUT In the Chilliwack and Salmon Rivers, young coho and trout exhibited seasonal changes in choice of microhabitat. In the Alouette River it was not clear whether or not changes in choice of microhabitat occurred a t different seasons. 1046 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22 , NO. 4, 1965 Recently-emerged trout and coho in the Chilliwack occupied three types of microhabitat: shallow water in small bays at the stream margin, small shallow riffles, and small crevices about the inshore portions of lo g jams and large boulders. The distribution changed during summer and early fall. By late fall most coho were located under log jams or under overhanging banks and boulders; steel h ead occupied the rocky areas of the stream margin and the log jams. Figure 11 shows the density of steelhead and coho in relation to log jam cover during 3 seasons. A pronounced seasonal reduction in density occurred in a re as where heavy log cover was absent. During winter those fish utilizing areas where log cover was absent were found only under or among the boulders. In cases where young coho were found among the l arge stones at the stream margin, their distribution did not extend far from shore. Young trout on the other hand were seen and collected among the boulders as much as 8 m from shore. Steelhead were able to occupy a wider variety of microhabitats in the Chilliwack River. The species composition in each of the two micro- h abitats distinguished in Fig. 11 was approximately the same at a ll seasons but steelhead made up a lar&er fraction of the fish taken in the areas where lo g cover was absent than where it was present (Fig. 11). In the Salmon River, recently-emerged coho and trout became segregated, with regard tu microhabitat, during the first 2 months (Fig. 12 ). Trout density COVER PRESENT ABSENT CJ35 ~18 2 SPRING & SUMMER FALL WINTER COHO--+\ STEELHEAD~ FISH PER 100M FIG. 11. Densities of young coho and steelhead in areas where cover (log jams) is present compared to areas where cover is absent. Data from the Chilliwack River, November 1959 to March 1962. HARTMAN : BEHAVIOR OF SALMON AND TROUT 300 200 100 200 I8J COHO D TROUT MAY-JUNE ~ 100 0 JULY-AUG 0 Z~ IOOjL-L~~~~~-L [L SEPT-OCT NOV-DEC D 1001: ---'--'--==-'---''"""-- soj =--07! JAN -FEB 50 j~==: ... =:cp:.:.,.~qJ~=(~~=MAR -APRIL w _J LL .u.. a: z ;;:j _J w z 0 a. z 0 0 :?_ a. u FrG. 12. Densities of young trout and coho in three types of stream habitat , riffle , open channel, and pool, during six periods of the year. Data from the Salmon River, November 1959 to March 1962. 1047 was highest in the riffles and lower in the open channel habitats and pools throughout May-August. In this period coho densities were highest in the pools and lowest in the riffles. The density of fish decreased through the first 10 months (Fig. 12). In the fall and winter, when numbers of fish were lowest, the degree of habitat segregation was reduced. By January and February coho and trout exhibited the same pattern of microhabitat distribution (Fig. 12). SIZE RELATIONSHIPS OF FISH IN THE SALMON RIVER Coho began to emerge in late March, while the first trout emerged in early June. Because of this difference in h a tching time, coho were larger than trout in June and July (Fig. 13). This size discrepancy decreased during late summer and autumn until, by winter, the size ranges were alike (Fig. 13). Virtually all coho migrated out of the Salmon River in May and June, at age of about 14 months. Trout remained in the river 2 years or more. Winter samples of trout could be separated into underyearling and "1 year plus" fish using length frequency plots of all trout (Fig. 14). Fish over 85 mm were designated as 1 year or more in age. 1048 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 1960-61 30 60 90 30 60 90 FORK LENGTH (MM) FIG. 13. Size relationships of underyearling trout and coho in the Salmon River. Smooth ed curves are based on moving averages of three, data before smoothing was plotted in 1-mm len gth intervals. 30 I ~20 lL 0 0 10 z ~ ~ ~ ~ ~ FORK 0 0 0 ! ~ ~ -0 LENGTH {MM) FIG. 14 . Length frequencies of two or more year- classes of trout in the Salmon River. Data are lumped, October to February inclusive, for 3 winters. HARTMAN: BEHAVIOR OF SALMON AND TROUT 1049 SUMMARY AND COMMENTS Field data show that trout and coho occur together along the lengths of the three streams. They are, however, found in different microhabitats within the streams. The differences in microh abitat distribution are most distinct in the small stream, the Salmon River, where trout and coho are seg- regated in riffles and pools in spring and summer, but to a large degree occur together in pools in winter. This case, where segregation occurs at one season but not at another , provides a good situation where Nilsson's (1956) concepts may be tested. However, fish are difficult to observe and environmental conditions cannot be controlled in the natural stream habitat. For this re aso n the investigation was brought into the laboratory where fish could be studied in a stream aqua- rium; the second part of this paper describes the experimental facilities and the methods of comparing the distribution a nd behavior of the two species. Field results are discussed in the light of experimental data. STUDY UNDER CONTROLLED CONDITIONS MATERIALS AND METHODS HOLDING CONDITIONS AND FISH The experimental study was conducted between October 1962 and De- cember 1963 in the Puntledge Park Hatchery at Courtenay, British Columbia. The coho salmon were obtained from Little River, a small stream near Comox; the steel head were from Big Qualicum River near Parksville. All fish were captured with seine nets. Fish used in the first series of observations (November 1, 1962, to February 20, 1963) were captured during October 15-22, 1962. Coho used in experiments between April 19 and October 9, 1963, were seined April 11 -23, 1963. Trout u sed in work from June 9 to October 9, 1963, were obtained May 25, 1963. Size range and mean fork lengths of samples of fish, measured at several intervals during the work, are given in Appendix I. All fish were held in running water in painted plywood troughs, 40 em wide and 220 em long. The troughs were housed in black plastic chambers and illuminated with fluorescent lights. Fish used in spring and fall experiments were held at a 12-hr day length (see Appendices II and III), those used in winter experiments were held at a 12-hr day-subsequently reduced to 8 hr (see Appendix IV). Water used in holding troughs and stream aquarium came from the City of Courtenay mains. Water temperature increased during spring and early summer, declined gradually during autumn and dropped to 0.5 C or less in winter (Fig. 15). Stream aquarium temperature in Fig. 15 will be referred to later in the text. The sharp rise and 3-day temperature peak, December 3-6 (Fig. 15), occurred when a break in the city water main forced the use of an alternate supply. 1050 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 20 10 20 10 20 10 SPRING & SUMMER 7 1<4 21 28 MAY FALL 8 IS 22 SEPT WINTER 18 2S NOV 29 6 16 23 DEC 18 25 JUNE 9 JULY OCT DAILY ··-""-MAXIMUM-.ti1TTilth. -f ·'·-MINIMUM-~ HOLDING STREAM TROUGH KlUARIUM 30 6 13 20 27 JAN FrG. 15. Temperature maxima and minima (daily) in the stream aquarium during seasons in w hi ch the study was made (so lid lin es). Breaks in the line indicate that n o experiments were in progress. Dotted lin es represent temperature maxima and minima in holdin g troughs p lotted for 2 weeks or more previous to experiments. Fish in the t r o u ghs were fe d a diet (by weight) 21 % liver, 65 % drained canned salmon, 8 % brewer's yeast , 6 % pablum, a nd salt (about 1 teaspoonful for 300 g of food). This food mixture was ground into a paste, frozen for storage, and fed in the form of frozen scrapings . Fish were fed once daily and remained hea lthy throughout the study. DESCRIPTION O F STREAM AQUARIUM The stream aquari um was designed to represent a short section of a s mall stream . Dimensio ns of the unit are 6 .3 m long, 2 m hi g h, and 1.2 m wide with an observation area 5 m long a nd 0 . 7 m deep (Fig. 16 ). Most of the aq u a rium is made of l-inch (2.5 -cm ) plywood supported in a framework of 2-inch X !--inch (5.1 X 0 .6-cm) angle iron . The windows are i -i nch (1.6-cm) plate gl ass. Co nstructi on o f the axia l-fl ow pump required a strong, nontoxic, rust-resistant material. This portion of the aquarium was therefore made of !-inch (0.3-cm) welded mild s teel lined with !-inch (0.3-cm) fiber -glass reinforced p l astic. l HARTMAN: BEHAVIOR OF SALMON AND TROUT FrG. 16. Experimental stream aquarium. Details of the drive mechanism are not shown. 1051 Current in the stream aquarium could be maintained at the desired vel- ocity with a variable-speed-drive mechanism. Water level was adjusted with an inlet hose and a series of drain pipes. Water was circulated from the pumps along the tapered duct at the bottom of the unit, up at the end opposite the motor, and along the observation flume back to the pump. The apparatus was lighted from overhead by parallel fluorescent lights running the full length of the observation flume. An observation gallery of black polyethylene sheeting paralleled each side of the tank. Adjustable horizontal slits in the plastic facing the aquarium permitted observation from the darkened galleries without disturbing the fish. DESCRIPTION OF EXPERIMENTAL ENVIRONMENTS Behavior and distribution of fish were compared in a control environment, in four different environmental gradients, and in a riffle -pool environment. The following is a description of these arrangements and some of the condi- tions associated with them. Figure 17 shows lateral aspects and plan views of control and four gra- dients. In the control situation (Fig. 17,a) the water depth (28-29 em), bottom gravel (3-6 em), velocity (22-24 em/sec), and lighting conditions were uni- form along the length of the observation flume. The light gradient (Fig. 17,b) was produced with a series of 10 plexiglass sheets. The first sheet was clear, the remaining nine sheets were coated with progressively more black paint. Light intensity in the gradient was measured with a "Photovolt" model 514 M photometer. Table I shows the average light intensity (lux) along the observation flume . Bottom condition, depth, and velocity were the same as in the control. 1052 J OURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 I ~L*b=d2?Ef1 I I I I I I tl =1 I I I I fl I I I I I 5 4 3 2 SECTION d . CONTROL b. LIGHT c. COVER d. DEPTH e . DEPTH +VELOCITY FIG. 17. Lateral and plan views of control arrangement and light, cover, depth, and depth and velocity grad ients. Light and control condi- tions appear similar except for graded filters u sed to produce li ght gradient. T ABLE I. Average light intensity over the len gth of the observation flume. Photo cell readings taken on the bottom with no water in the aquarium. 5 4 3 2 Section (Meters from 4.75 4.25 3.75 3 .25 2.75 2.25 1.75 1.25 0.75 0.25 upstreamend) 2.60 10.8 23.8 48.6 85.3 157.7 189.0 201 .9 375.8 28 1 .0 aM ean of three readin gs across the aquarium. The cover gr adient consisted of five groups of s t ones (Fig. 17,c). Stones were ele vated above small depressions in the gravel so that each had a 4-to 6-cm cavity under it. The size of stones in each section varied so mewhat as given in Table II. To produce the depth gradient four s heets of plywood were arranged step- wi se in sections 1, 2, 3, and 4 (Fig. 17,d). Screens between the leading a nd trailing edges of the steps kept fish above them, but a llowed an even flow of water so that velocities were near constant at all depths (Fig. 18). The bottom was covered with the same gravel u sed in the control. Illumination was sl ightly higher in the upstream section where the floor panels were e levated closer to the li ght source. The d epth-plus-velocity gmdient was produced by means of a sloping false floor (Fig. 17, e). It was not possible to vary velocity along the length of the observation flume without a ltering depth or width. Details of velocity profiles in horizontal and sagittal planes are shown in Fig. 19. The bottom was covered with the same gravel used in the control. Light intensity on the raised upstream end of the false floor was sli ghtly higher than on the downstream end. Figure 20 represents the riffle and pool environment. This arrangement caused the current to exhibit complex flow patterns which a re described briefly below: 1.2-! :::! : ~ j X ; ..... 6-; 0 : i ! 20 20 0 20 0 0 20 0 20 0 20 0 20 0 A 20 0 20 0 20 0 WATER VELOCITY -CM PER SEC B FIG . 18. Velocity profile in horizontal and saggital planes in depth gradient. Upper figure shows velocities 10 em above bottom, except in section 1, rig ht end, where profile is dra wn for position 5 em above bottom. Lower figure shows conditions in saggital plane (two center readings averaged). Arrows indicate d irection of current, dotted lin es represent screens. 1054 JOURNAL FISHERIES RESEA R C H BO ARD OF CANADA, VOL. 22, N O . 4, 1965 1.2-• .----,.,---.,.-----...,.,-----,.,-----,"""T'""?-....--~ ::l i ! : 20 020 (1.\IS! S! :S: ! ??: 20 0 20 0 20 0 20 0 20 0 20 WATER VELOCITY-CM PER SEC 0 A 0 B FIG. 19. Velocity profile in depth and velocity gradient. Upper figure shows velocities 10 em above bottom. Lower shows velocities in saggital plane (two center readings averaged). Arrows indicate direction of flow, dotted lines represent screens. Pool in section 1 , upper 30 em of water current flows downstream at about 20 em/sec. In lower 10-12 em current circulates upstream along the bottom at 4-5 em/sec. Pool in section 4, velocity at the surface 28-30 em/sec, at 20 em depth about 10 em/sec and at the bottom, current near 5 em/sec circulating upstream. Riffles in sections 2, 3, and 5, current 28-30 em/sec at the surface and 20 em/sec along the bottom. Temperature in the aquarium (Fig. 15) was governed by seasonal changes in the temperature of the water supply, conditions within the building, and energy input from the pump. Tank temperature could usually be lowered by adding fresh water. It could be raised slightly by cutting the input of new water and hence allowing the propeller to heat that already in the aquarium . These measures did not however provide full control. During periods of freshet, b) 0 0 T 0 a CJ 0 Q 0 l 0 (] 0 FIG . 20 . Horizontal (upper) and lateral views of experimental riffle and pool environment. Horizontal arrow indicates direction of current. Vertical arrow indicates water surface. HARTMAN: BEHAVIOR OF SALMON AND TROUT 1055 the water source became turbid and consequently new water could not be run into the aquarium. If the air temperature in the building was high during such freshets the aquarium temperature rose . If it was low the aquarium tem- perature fell (see October 5 and 6, November 9, 10, and 11, 1962, and January 8-10, 12-13, 1963, Fig. 15). Temperatures within the aquarium deviated somewhat from those of the holding troughs early in the winter but approximated them later (Fig. 15). During the spring and summer experiments, water in the stream aquarium ranged from 10 to 16 C and was generally warmer than that in the holding ponds. In the autumn, temperatures in the stream aquarium followed holding trough temperatures and fluctuated from 9.4 to 14.6 C. Figure 15 shows that stream tank temperatures fluctuated seasonally and daily. They corresponded to those of the holding troughs but were gen- erally higher. EXPERIMENTAL PROCEDURE Day length was maintained at 12 hr (0600-1800 hr) in all experiments. This photoperiod was not consistent with th a t of the holding troughs during the winter. It was necessary however to use a 12-hr day in order to give the .fish a 2-hr period of adjustment after the lights came on since the subsequent 10-hr observation period was necessary to obtain adequate records on dis- tribution and behavior. In all experiments fish were handled, fed, and observed as consistently as possible. In each experiment, 40 fish were placed in still water between 2000 and 2200 hr. At 0800 hr the following day, the current was sta rted at 14 em/sec and raised in two steps to 23 em/sec in control, light, and cover gradient. During the first day numbers of fish in each section were recorded at !-or 1-hr intervals. During the second day positions of all 40 fish were plotted 12 -16 times on outline maps of the stream bottom. Behavior was recorded in a series of 10-min observations along the length of the tank. A preliminary study revealed that the two species exhibited com- parable components of agonistic behavior. These were coded so that a sequence of events co uld be recorded approximately and quantified. The following is a list of behavior elements and their code letters; details of behavior will be described later : L -lateral display, F-frontal display, N-nip, C-chase, WW-wig-wag display, TN-threat nip, IM-intention movement, and Fl-flight . A protocol for a behavioral sequence is as follows: Fish" A" Fish "B" "A" "B" L+N L+WW N----~ FL I ______ ___J 1056 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 In this series of events fish "A" displays in lateral posture then nips "B." Fish "B" displays l ateral posture then exhibits a wig-wag display, "A" nips "B" again and "B" flees. The arrow under "FL" indicates the direction of flight. Fish, fed twice daily in the stream aquari um, were given 8-10 cc of brine shrimp frozen into a block of ice. The food was placed above the screen in the upstream end of the tank where it melted during a period of 20 -60 min. Many of the brine shrimp released kept circulating in the current for several hours, thus evoking feeding over a prolonged period. Immediately after feeding fish tended to shift about and temporarily take up new positions. Because of this, observations on behavior were not made until 30 min or more after feeding began. The procedure followed in studying fish in the riffle-pool arrangement was sli ghtly different from that used with the gradients. Coho and steelhe ad were studied separately in two series and then observed in combination in a third series. Table III gives numbers and species of fish used and the time schedule followed during the three series of experiments. To begin each series 20 fish were placed in still water in the aquari um at about 2000 hr. The following day the current was started and increased step- wise to the desired velocity by 0900 hr. On each succeeding day, until the fourth, an additional lot of fish was added at 0800 hr. Feeding and recording of distribution and behavior were as previously described. Experiments on fish in the riffle-pool arrangement were conducted in July and November 1963 with underyearling fish. Experiments in the control and gradients were clone in the spring, fall, and winter. Appendices II-IV give details of gradient experiments. SCOPE OF THE RESULTS Experiments were conducted so that seasonal changes in behavior and distribution could be observed in each species and so that differences between TABLE III. Number and species of fish used in experiments in riffle and pool environment. In Series 1 coho (C) were used alone, in Series 2, steel- head (S) were used a lone, and in Series 3 the two species were combined. Series 1 Serie s 2 Series 3 ---- No. of No. of No. of c s c s c s First day 20 20 10 10 Second day 40 40 20 20 Third day 60 60 30 30 Fourth day 80 80 40 40 HARTMAN: BEHAVIOR OF SALMON AND TROUT 1057 species could be recorded. Seasonal and species comparisons were made on groups of one species of fish at a time. The interaction of the two species was studied in summer and winter conditions in certain experimental arrangements. Distributional data are based on groups of animals. If individuals had been tested singly, the preferred positions may have been different from those inferred from the distribution of a group. The maximum number at a par- ticular point in a gradient may not always represent the preferred position. In spite of this limitation, however, groups of fish were used because field data are based on the behavior of animals in groups. Temperatures in the holding troughs and stream aquarium varied at different sea sons. These temperatures a lso fluctuated within each season (Fig. 15). The day length at which fish were held was shorter during winter than during spring and autumn. The effects of variations in these conditions co uld not be fully evaluated, but physical conditions such as bottom configuration, bottom gravel, depth, water velocity, and light conditions were duplicated in all cases; hence the environment was partially but not fully controlled. RESULTS REPLICATION OF EXPERIMENTS Certain experiments were replicated during the winter and the spring- summer series. Observations on the distribution of coho in the control en- vironment were made twice in April and repeated in June 1963 (Fig . 21). Experiments with each species in the control, the cover gradient, and depth gradient were replicated under winter conditions (Fig. 21 and 22). In general, the duplicate distribution patterns were simil ar (Fig. 21 and 22). Differences between species were consistent in repeated experiments in the control and cover gradient during the winter series (Fig. 21). Repeated observations in the depth gradient produced similar distribution patterns for each species but the differences between coho and trout in the depth gradient were not consistent (Fig. 22). Although repeated experiments in the depth gradients did not give distributions that were identical, they did reveal that each species exhibited characteristic patterns. In the control and cover gradient, trout were distributed in a skewed "U" -shaped pattern, coho in a sigmoid pattern, usually with highest means in the first two sections of the tank (Fig. 21). SEASONAL CHANGES IN DISTRIBUTION Comparison of the data obtained during the three seasonal series of ex- periments indicates that numbers of fish were more uniform along the length of the tank in spring and summer than in fall or winter (Fig. 21 and 22). The greatest differences in sectional averages occurred in fall or winter. Such large variations in the average number of fish per section were a result of the tendency of many individuals to congregate in one portion of the tank during fall or winter conditions. In spring both species were distributed over the whole bottom area of the aquarium. SPRING-FALL 20~SUMMER COHO ....._ WINTER SPRING- 301 Sutv1MER FALL 10 ' STEEU-£AD <>----0 20. ~,~~ ~ hl Vl a: UJ a. 0 z 20 J z 20 0 ~ u UJ Vl 0: UJ a. t~J ~~ li t 1T 10 <f 30 L\.1~;1 COHO .___.. STEELHEAD <>----e WINTER f- I 0 _J I f- 0.. w 0 / 20jl IOl bi··. i/ 31 101~---~ jh.l : .. I J llifl···· .. I)_. ...... ~----J ~~ -----. : t"l ~~~ Lll '.~ ! I ~ -'---:--...,.---r--•-I . 543215432154321 5432 15 432 1 54321 SECTION SECTION FIG. 21. (Left) Distributions of coho and steelhead in "control" a nd "cover" gradients, described in text, during 3 seasons . Dots and circles represent mean number of fish per section over a 1-day period (10-15 observations). Vertical lines indicate range. Section 1 represents the upstream end of the aqua rium. FIG. 22. (Right) Distributions of coho a nd steel head in "li ght ," "depth," and "depth and velocity" gradients, as described in text, during 3 seasons. Circles and dots represent mean number of fish per section (10-15 observations). Vertica lines represent range. Section 1 represents the upstream end of t he aquarium. ,_. 0 Vl 00 '-< 0 ~ to z >-r' 'IJ u; ::r: t:rl ~ Bl to t:rl Cfl t:rl >-to () ::r: ttl 0 >-to t:l 0 'IJ () >-~ t:l >- < 0 r N N z 9 _,.. "' "' '" HARTMAN: BEHAVIOR OF SALMON AND TROUT 1059 Young coho were scattered in the spring and early summer but in autumn and winter they tended to form aggregates, with some social organization, near the upstream end of the aquarium in the control and the cover and light gradients (Fig. 21 and 22). In the depth gradient the seasonal trend exhibited by coho was a shift to the deepest section of the tank during winter. In spring and early summer young steelhead were distributed along the tank in a fashion similar to the coho. In fall and winter many trout were active, i.e. moving about in the control, as well as the cover and light gradients (Ap- pendices III and IV). These fish were usually moving and searching about the screen in section 1. The apparent seasonal shift into the upstream region of the aquarium is indicative of wandering and searching in the upstream end, rather than a preference for it. The numbers of stable steelhead positions in the upstream section was usually half, or less, of the numbers shown in Fig. 21 and 22 in the control, depth gradient, and light gradient. Considering this movement, it appears that steelhead which are not roaming assume a more scattered distribution than coho along the tank in the control and the light gradient. Activity accounts for the apparent seasonal shift of fish into section 1 in the cover gradient. However, the high numbers of fish in section 5 repre- sents a preference for positions under or around the large stones. Steelhead, like coho, show a strong winter preference for the deepest section of the depth gradient. Trout exhibited no seasonal change in distribution in the depth- plus-velocity gradient (Fig. 22). COMPARISON OF SPECIES IN GRADIENTS In experiments conducted under spring and summer conditions, the dis- tributions of coho and steelhead were similar in each of the five experimental conditions (Fig. 21 and 22). However, in autumn and winter experiments, the distribution differences between the species were greater. Environmental preferences, as inferred from experimental distributions, were most similar in spring and early summer, the season in which segregation was most pronounced in the Salmon River. Distributions and inferred preferences were divergent during fall and winter (Fig. 21 and 22), the seasons in which populations over- lapped most in the Salmon River (see Fig. 12). During the season when labor- atory distributions are similar the two species meet and interact in the natural stream environment. In the seasons when experimental distributions are dif- ferent, the two species are most compatible in the natural stream environment. Different environmental responses in the laboratory (i.e. response to cover, Fig. 23, 24) are indicative of the mechanisms that allow the two species to coexist in close proximity in fall and winter. In the cover and depth gradients young fish utilized stones and pool space in a similar fashion in spring and summer but not in winter. Coho, 6-8 weeks old, scattered among the stones which formed the cover gradation (Fig. 23). During spring about one-third (126 of a total of 412) of the positions taken by coho were immediately downstream from stones. Many positions recorded were among the stones. In winter, 126 of 390 positions occupied 1060 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 were immediately downstream from the stones. Fish were, however, recorded at fewer positions among the rocks (Fig. 23) in the winter. Those that were not in the shelter of stones were at positions at the sides of the tank. COHO ~~· >:0 .· 0 .... :.0 ~. ·i a':. · .. ~· .: .r?:.: fo ·: .. ~~.;;-· ~J COVE:L:~A~::~T ~u ... . ·.~ · · · ...... :: . ~~,. ·. ·~ ·::. ,. . .:, .. : ·$1 I ··>o: · -· ·\=i· '· .. .o . 'C). D ·1 MAY 28 jO · ··;a .. ;:o·:· "a·:· . ·:: .. :~p .' :.:· .~.,-:! 1·: ···,:. · .. :·C?· ,. .. ···.D · · .. ,:-::.~·\.-·· :: >.o '·:··· : .. ·~ .. i i D •· ::.0 lQ ·.·• · :: ··o ... · .. :. ·'' · ·· .. o ·· .: .. ·::.: .•i DIRECTION OF FLow !o·· ..... =:·::~ ;:,;·:·~~··;;· ... ~~· }6"<<>~ ... ·,;~.·:.'·~.<~ ... ~ < e: OCT I ;P·· •.· 0 o .. ". ~ •:· .. · 0 ,, ',',,"O:.x: .. '·d ·; lu..... ,.,D ... _.,.. .:o.< .... ; .. ~-·'"~.. . .I 1 . ·~~·Q ...... ~o-·· ~~6 · :\.0·~· · · .. :o 1 lC]. . .. a:::· .. · ... CJ ... ·.: ...... b ... -~~:·.:·.:.![~ .. o,, .: .. ~:.: ... .! JAN 25 5 4 3 2 SECTION STEELHEAD i.' . . :·.0 '.0 . .' · · ·!'(J · ·,. ·. · ·iQ . .0 .··i MAY 30 ttJ ,;·· . '0 · .. ,: . ,: 6'. _: ... : .. :0' "'·. :.: ;.:~ .··. :I I · . ..-:o · :·'CJ ·a =o .. ·.;'P ·I !0 .. J;i\:J. . .. .•C) ..... 7.o·, ..... T ••• a·: ... , ..,] 5 4 3 2 SECTION COVER GRADIENT -PLAN VIEW FIG. 23. (Top) Distribution of coho in cover gradient during 3 seasons. Locations of 40 fish during 10 combined observations are given. Groups of dots represent same fish occupying same location repeatedly, or different fish in same location repeatedly, in this and succeeding figures of this type. Points within the stone outlines represent fish under stones. Dotted lines represent screens at ends of aquarium in this figure and succeeding figures of this type . FIG. 24. (Bottom) Distribution of steelhead in cover gradient during 3 seasons. Locations of 40 fish during 10 combined observations are given. Points within the stone outlines represent fish under stones. HARTMAN: BEHAVIOR OF SALMON AND TROUT 1061 Steelhead, 3-5 weeks old, distributed themselves in the same pattern as the 6-8-week coho (Fig. 24). During spring, one-third of the steelhead positions were immediately downstream from stones. As in the case of coho, the other positions were scattered among the stones, and none was under them (Fig. 24). During the fall, a large number of young trout was active and remained in the upstream end of the aquarium (Fig. 24). One-sixth of the positions re- corded were immediately downstream from stones; only six positions were under them (Fig. 24). In winter one-fourth of the positions recorded were under stones and approximately one-eighth were downstream from them (Fig. 24). It is evident th at the distributions observed in the spring condition would result in a high degree of contact between species if together in a cover gradient. However, in winter the tendency of trout to hide under stones would, to a degree, isolate them from coho which do not do so (Fig. 23 and 24). A second instance of trout and coho using the same space in a different manner, in winter, occurred in the depth gradient. Table IV shows that in June there is a significant difference in the numbers of young coho and trout in the upper and lower halves of sections4 and 5 (chi square tests). Segregation is, however, more pronounced during winter. This increase in segregation is primarily due to a change in the behavior of steelhead. In summer conditions, about one-third of the steelhead were off the bottom where they would be in contact with coho if the species were mixed . In the winter most steelhead were spread over the bottom in the deep section. On the other hand most coho were distributed at the edges near the bottom or in loose aggregations up each side of the deepest section. Individuals of both species defended areas along the downstream edge of each depth zone. A large amount of intraspecific fighting occurred in these areas. When the two species were placed together during summer in equal numbers the amount of intraspecific and interspecific aggression was high (Table V). During winter when trout and coho were placed in the depth gradient together they were segregated spatially as already de- scribed. Intraspecific and interspecific aggression were lower than under summer conditions (Table V). Interspecific fighting was not disproportionately lower as expected on the basis of spatial segregation. However practically all inter- TABLE IV. Numbers of coho (C) and steel head (S) in upper and lo wer halves of sections 4 and 5 in the depth grad ients in early summer and winter (June and J an uary). Upper half Lower half c 124 103 June s 75 145 c 168 161 January s 24 281 1062 JOURNAL F I SHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 TABLE V. Aggressive contests per fish per hundred minutes during June and J anuary 1963. Symbols are as follows: C =coho; S = steelhead; C.C =coho attacking coho ; C.S =coho attacking steelhead; S.S = steelh ead attackin g steelh ead; and S .C = steelhead attacking coho. Date June 28 January 17 c 20 20 Fish s 20 20 C.C 12.7 4.4 Number of contests C .S 6.1 1.3 S.S 6.3 2 .4 S .C 1.3 0.9 Observation min 90 90 specific contests occurred at the downstream edges of depth zones in sections 3 and 4. Few aggressive contests were recorded between the coho, in aggre- gations at the side of the deep section, a nd trout, on the bottom. DESCRIPTION OF AGONISTIC BEHAVIOR Young coho and steelhead display using a number of similar postures and movements. These displays and movements are listed in Materials and Methods. Before any classification and quantification of the behavior compo- nents are attempted, it is essential to understand the basis of the classification. Each of the different components appears in a variety of forms which appear to be related to the intensity of the behavior. In the following description of behavior, each component is described and the variability indicated. It is acknowledged that this type of fish behavior could be classified on a more refined scale by quantifying intensity or duration of co mponents. Such was not feasible in this study because of the number of fish that were observed and the rapidity of the action. Lateral display was described by Fabricius (1953) and Kalleberg (1958). This varied from a simple erection of the dorsal fin , lasting 1 or 2 sec, to a prolonged erection of dorsal and paired fins and a lowering of the basihyal ap paratus for 10 or 15 sec. The dorsal line of the body was either straight or slightly recurved. The criterion for a l ateral display was the erected dorsal fin and the line of the back. Figure 25,A shows characteristic lateral posture of 10-month-old coho, Fig. 25,B shows lateral posture of 8-month-old trout. Figures 25,C and 26,A illustrate lateral displays in 2-month-old coho. Lateral displays were similar for the two species, although the display was usually more obvious in coho which possess large median fins with long, colored edges (Fig. 25,C and 26,A). The frontal display, d escribed by Fabricius (1953) and Kalleberg (1958), varied from a posture in which the back was slightly arched, the dorsal fin compressed, and the basihyal extended for 1 or 2 sec (Fig. 25,B and 26,B), to a posture in which the back was strongly arched, the dorsal fully compressed, and the basihyal well extended for longer (unmeasured) periods. HARTMAN: BEHAVIOR OF SALMON AND TROUT A c FrG. 25. A. Coho , about 10 mont hs old, in lateral threat posture. B. Steelhead, about 8 months old, in latera l posture (see fish on the left). Fish on the right in frontal threat posture of lo w intensity. C. Coho about 2 months old, in lateral threat posture. (All fi gures are traced from photographs.) 1063 In the wig-wag displ ay fish a dopted a lateral posture, usually with me- dian and paired fins well extended, a nd swam with accentuated lateral move- ments with the head down and the body at a 20-30° angle from horizontal (Fig. 26,B and C). In this display the a mount of fin erection varied. The a ngle of body inclination and the amplitude of swimming movement was low in displays that were of short dura tion. In wig-wag displays (inferred to be of high intensity) the fins were erected fully, the angle of inclination was steep (near 30°) and lateral movements were accentuated. The criterion for a wig- wag display was the erected fins, the inclined posture, and accentuated swim- ming movements . As in the case of the lateral displ ay , the wig-wag was more striking in coho than trout because of differences in fin shape and color. 1064 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22 , NO. 4, 1965 FIG. 26. A. Coho, about 2 months old, in lateral threat posture. B . Coho, about 2 months old, in frontal threat posture of low intensity (right). Coho, about 2 months old , in wig-wag posture (left). C. Coho, about 2 months old, in wig-wag (left) and lateral threat posture (right). Chasing involved chase and flight. If one fish darted after another and pursued it past the point from which it fled it was designated as chasing. Pursuits ranged from slow short excursions of 20 or 30 em to long chases of 2 or 3m. Nips were those bites in which there appeared to be a definite contact. In some instances fish only "mouthed" the individual attacked. In other cases the bites were so hard that the animals seemed to be briefly stuck to- gether . Threat nips were nips which were aimed at other fish . Such bites appeared to be inhibited and hence missed conta ct by as much as 20 or 30 em. In some cases a fish swam a short distance and nipped in the direction of another. HARTMAN: BEHAVIOR OF SALMO N A N D TROUT 1065 In other situations they turned their heads and nipped in the direction of a neighbouring fish. Intention movements were responses in which a fish only turned its head quickly toward another or made a short lunge at it. No threat nip was involved. These movements, which were brief, varied as described and were often dif- ficult to distinguish from nonsocial behavior and threat nips. An example of an aggressive bout involving several of the described ag- gressive components was given in the Materials and Methods section . Some contests were short and involved only two or three behavioral components, others were long and involved series of bites and di s plays interspersed among each other and reciprocated between individuals. ANALYSIS OF BEHAVIOR By recording bouts in terms of individual behavior components it was possible to examine the rate of occurrence of various aggressive components as well as complete contests. The breakdown of agonistic behavior into a ll its components made it possible to compare quality and quantity of aggres- siveness in different seasons and species. Figure 27 represents behavioral repertoires of coho and steelhead during 3 seasons. Details of observation time are given in Table VI. The diagrams in the figure show only the rate of occurrence of each component, they do not show how these may be related to each other during aggression. There are three main points illustrated in Fig. 27. First, the repertoires of the two species differ at all seasons. Second, within each species the amount of aggressive behavior decreases seasonall y from sprin g to winter. Third, the quality of behavior exhibited by each species changes seasonall y. The most obvious species differences were the relatively strong nipping and chasing components of the trout as opposed to strong wig-wag, threat nip, and intention-movement components of coho. The lateral display, which often preceded the wig-wag, was stronger in the coho than in the steelhead during spring and a utumn (Fig. 27). Level of aggressive behavior among coho was high in spring, s ummer, and fall but decreased during winter. Among trout aggressiveness decreased progressively from spring to autumn and winter. During spring and winter lateral displays, nipping and chasing were frequent in both species. Nipping and chasing components were very strong in the steelhead (Fig. 27) during spring and s ummer. Behavior composition of the two species was most similar in the spring; by autumn it had diverged. By winter the lateral display components were greatly reduced and equalled by threat nips and wig-wag elements in the coho. The most evident seasonal change in the steelhead configuration was the reduction of the chase component. A higher frequency of elaborate displays and noncontact behavior was evident in the coho. The main behavior elements of the trout were latera l displays, biting, and chasing. H a rtman (1963) showed that young brown trout (Salmo tr u tta Linn.) displayed frequent ly, but nipped less at low water veloc- ities (8 -9 em/sec). At higher velocities (18 -19 em/sec and 28-30 em/sec) 1066 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 COHO L IM TN N ww F STEELHEAD SPRING - SUMMER FALL WINTER ~ ~ z ~ 0 Q IS a: UJ a. 10 Vi iL a: 5 UJ a. UJ 0~ L -LATERAL WW-WlG WAG F-FRONTAL C-cHASE N-NIP TN-n-REAT NIF" I M-INTEHTION t.DIEMENT FIG. 27 . Rate of occurrence of individual components in the agonistic behavior of young coho and steelhead. Data are based on combined observations of fish in the five experimental arrangements used. See text for description of behavior components. they nipped relatively more and displayed less. The mechanical difficulty of holding position in the current with median and paired fins extended appeared to be the main reason that agonistic behavior took on a different character at higher water velocities. A comparison of behavior patterns of coho and steel- TABLE VI. Minutes of observation of coho and steelhead in control and four experimental arrangements during spring, fall, and winter. Coho 750 390 890 Steelhead 410 440 720 Season Spring Fall Winter HARTMAN: BEHAVIOR OF SALMON AND TROUT 1067 head with brown trout suggests that steelhead behavior, involving primarily lateral displays, nips, and chases, is more adapted to rheocrene conditions than the behavior of coho which involves more wig-wag displays and less nipping. Results of field and laboratory studies also suggest that the differences in behavior of coho and steelhead are related to their ecology. DISTRIBUTION IN RIFFLES A N D POOL HABITAT The major difference in distribution of trout and coho in the field was related to riffle and pool habitats. In the stream aquarium certain behavior features of each species appeared adaptive to particular current conditions and an attempt was made in the laboratory analyses to determine whether there were behavior characteristics which conferred advantages on trout in riffles and coho in pools. Distributions in riffle and pool habitats (Fig. 28 and 29) indicate that both species preferred pools or that some environmental regulation of be- havior allowed more individuals to remain in the pools. Both trout and coho had similar distributions in the riffle and pool habitats when the species were separate. Steelhead, however, were more numerous in the riffle areas (Fig. 5432154321 SECTION 543.21 54321 SECTI O N FIG. 28. (Left) Distribution of coho and steelhead at four different densities in riffle and pool environment (July). Solid dots represent the average number of fish per section, species separate. Circles and broken lines indicate the average number of fish per section, species mixed . Scale for the points for species mixed is half that for species separate. FIG. 29. (Right) Distribution of coho and steelhead at four different densities in riffle and pool environment (November). Solid dots represent the average number · of fish per section, species separate. Circles and broken lines indicate the average number per section, species mixed. Scale for the points for species mixed is half that for species separate. 1068 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 28 and 29). When coho and steelhead were mixed in July experiments, density of steelhead was reduced (in relation to the situation where the species was · alone) in six out of eight cases in the pool habitat, and increased in eight out of 12 instances in the riffle habitat. Coho density was reduced (in relation to the situation where the species was alone) in nine out of 12 cases in the riffles and increased in five out of eight instances in pools. Upon mixing, coho densities increased in the pools and decreased in the riffles, and steelhead densities ch anged in the opposite direction in more cases than expected by chance (P < .OS, chi-square test). During winter the effects of interspecific mixing were not clear (Fig. 29). In experiments where the species were mixed, steel head densities decreased in the pool in section 1, and increased in the three riffle sections in all but one instance . However, density of steelhead in mixed groups was higher in the pool in section 4 also. Changes in relative density of coho showed no consistent relation to those of steel head as occurred in July (Fig. 28). AGGRESSIVE BEHAVIOR IN RIFFLES AND POOLS Levels of aggressiveness were higher in riffle habitat than in the pool h abitat when the species were separate (Fig. 30). An exception to this was the case of steelhead under winter conditions. Fighting and displaying occurred more frequently in summer than in winter in the riffle and pool environment as was observed in the experimental gradients. 20 ~ D JUL ::E [] NOV 0 r-- Q 0: I 0 UJ _j a. 0 m r-:r: 0 Vl .;.;.;. 00 a.. ;:,: :·:·:·: ;:;:;:; 0: :;:;:;: UJ t~~~ 00 a. =~=~=~= UJ Vl :·:·:·: _j 1-·:·:·:· u. :;:;:;: Vl g u. Ul 1-a: z 10 0 u - 20 - CvC SvS Frc . 30. Rate of aggressive behavior in riffle and pool habi tats during July and November. Data based on observations made with species separate. C vs C indicates coho attackin g coho and S vs S indicates steelhead attacking steelhead. Coho observed 390 min in July a nd 340 min in Novem- ber; steel head observed 380 min in each period. HARTMAN: BEHAVIOR OF SALMON AND TROUT 1069 Interspecific mixing revealed an environmental effect on behavior which may in a large degree explain why trout maintained themselves in the riffle sections of the aquarium and actually reduced utilization of this space by coho. Figure 31 shows that coho displayed a high level of interspecific and intraspe- cific aggressiveness within pools. Aggressiveness of trout was correspondingly low in the pools (Fig. 31). (Chi-square values indicate that differences in rates of aggressiveness of trout and coho in riffles and pools during July are signif- icant P > .01.) In the riffle habitat of the aquarium coho were not particularly combative; steelhead on the other hand were aggressive (Fig. 31). In addition ' to being more aggressive, steelhead tended to defend temporary territories . A comparison of Fig. 30 and 31 indicates that mixing the two species in a riffle and pool environment had the effect of reducing the level of coho aggressive- ness in riffles and increasing it in pools. The degree of aggressiveness of steel- head in riffles was increased in mixed groups. In mixed groups steelhead fight- ing was more frequent in the pools in November than in July; such was not the case when the species were separate (Fig. 30 and 31). High rates of aggressive behavior in the riffles (species unmixed) resulted in low densities of fish in such a reas. Behavior differences, which were related z ::E 10 0 Q cr LLI Q_ 05 5 lL. 0: LLI a. <f) I- "' w 1-- z 0 u 5 10 I I - - C'IC C'/5 SvS 0 JULY [] NOV ...1 0 0 -!ill a.. - SvC w ...J lL lL a:: FlG. 31. Rate of aggressive behavior in riffle a nd pool habitat during July and November. Data based on experiments in which species were mixed in equal numbers and observed 270 min in July an d 250 min in November. Meaning of symbols as follo ws: C vs C = coho attacking coh o; C vs S = coho attacking steel head; S vs S = steel head attacking steel head; S vs C = steel head attacking coho. 1070 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 to the environment, acco unted for the strict density regulation in riffles. Strong current induces more distinct territorial tendencies as has previously been demonstrated (Kalleberg, 1958). In addition the presence of reference objects induces fish to establish and defend territories (Hartman, 1963). In the riffle sections current was fast and there were reference objects, i.e. large stones. Agonistic activities in these areas were easily induced, hence in Fig. 32 rates of aggression were high at low densities. High densities of fish did not occur because increased fighting led to displacement of fish. In the pools, however, aggression decreased with an increase in fi sh den- sity (Fig. 32). Keenleyside and Yamamoto (1962 ) demonstrated a lmost the 30 z ~ 2 "' "' a: 10 Cl 1.? -< IL 0 30 "' ~ 0: 20 10 0 0 0 oo • 0 POOL 0 RIFFLE JULY""' NOV ..... STEELHEAD COHO ... o~o~ •• # 0 • •• .,. • 20 AO 20 40 FISH PER SECTION U.2 M 2) FIG. 32. Relation of rate of aggressive behavior to density of fish. Coho and steelhead compared in riffle and pool habitats during 2 seasons. Each dot or circle represents the number of a ggressive contests per fish 100 min during 10 min of observation in one section of the aquarium. same effect with Atlantic salmon in small still-water aquaria. Reduction in rate of aggression with increase in numbers was particularly evident in coho (Fig. 32). Certain behavior features probably acco unt for this phenomenon in groups of coho. In a group of coho, competition was strongest for positions near the front. In fighting for positions, fish often swam parallel to each other in lateral threat posture. After swimming parallel for a short distance one fish, usually the dominant individual, darted ahead of the second and per- formed a wig-wag display in which it dropped backward downstream toward the second fish (Fig. 33,A-C). In many cases the upstream fish ended by lit- erally brushing its opponent back with its tail (Fig. 33,C). If the displaced fish remained behind the victor, little more fighting occurred. The wig-wag threat was closely associated with the formation of stable social groups with one to three dominants at the front and several subordinates behind them. I I j HARTMAN: BEHAVIOR OF SALMON AND TROUT FrG . 33 . A. Coho, about 10 months old, in wig-wag posture. Fish at left is displaying and beginning to drop back toward fish at right. B. Both fish dropping down- stream and coming closer together. C . Coho at left still in wig-wag posture, its tail a lmost striking fish at right. At this point the fish at left may wheel and nip the second fish or second fish may flee. 1071 Steel head did not establish stable social groups as did coho. In July ob- servations, aggressiveness decreased with increase in numbers of steelhead. Such a change took place because as numbers went up many steelhead settled to the bottom and became quiet while others began to roam about . These fish were not often attacked. There were usually large, potentially dominant trout; they did not often exhibit the wig-wag threat and did not hold the front positions in any stable groups. SUMMARY OF RESULTS Field observations revealed seasonal changes in the distributional rela- tionships of young coho and trout. Concomitant with these were changes in water temperature, and population density. Laboratory experiments pointed to features of environmental and social behavior which were related to changes occurring in nature. Field and laboratory results are summarized below as an introduction to the Discussion. Field observations apply particularly to the Salmon River. 1072 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 SPRING AND EARLY SUMMER Field observations 1. Species largely segregated in different microhabitats. 2. Coho in pools, trout in riffles. 3. Population density per unit of area is high. 4. Coho relatively large compared to trout. 5. Temperature of water 8.3-17.2 C (time corresponds to laboratory period). 6. Body and fin colors vivid. FALL Field observations 1. Species partially segregated in early fall, coming together more in late fa ll. 2. Coho in pools, trout density about even in riffles and pools. 3. Population reduced in pools, reduced more in riffles. 4. Trout size range approximating that of co ho . 5. Temperature of water 7.2-12.5 C. WINTER Field observations 1. Species exhibit no mi crohabitat segre- gation . 2. Highest density of both species in the pools. 3. Population density reduced further in pools, very low in riffl es. 4. Trout size range approximating that of cohos. 5. Temperature of water 0.3-7 C. Laboratory observations 1. Both species have similar distributions in experimental gradients. 2. Both utilize space in pools and cover in the same manner. 3. Both species exhibit hi g h level of aggression which involves much biting and chasing. 4. Coho larger than trout. 5. Temperature 10-16 C. 6. Body and fin co lors vivid . Laboratory observations 1. Steelhead and coho h ave different distributions in experimental g ra- dients. 2. Species utilize space a nd cover in abo ut the same mann er. 3. Coho aggression high, but less biting a nd chasing is exhibited. Steelhead aggression lower than in spring, relatively le ss c hasing. 4. Temperature 9-14.5 C . 5. Body and fin colors less v ivid. Laboratory observations 1. Species have different distributions in experimenta l gradients. 2. They utili ze space in pools and around cover in different manners. 3. Aggression very low in both species. Coho display components are strong, very little biting and chasing. Steel- head show on ly two components strongly: simple displays a nd biting. 4. Temperature 0.5-7.5 C. 5. Body and fin co lors le ss vivid than in spring and fall. l HARTMAN: BEHAVIOR OF SALMON AND TROU T 1073 DISCUSSION An animal's behavior is adapted to its environment, as is its morphology and physiology. Accordingly , there are both environmental and social res- ponses of coho and trout which relate to their ecology. Certain aspects of the ecology and behavior of coho and trout will be considered before entering the main body of the discussion, which will de a l more directly with interspecific interaction. BEHAVIOR OF STEELHEAD AND COHO I N RELATION TO THE IR ECOLOGY Under natural co nditions coho were most frequentl y distributed in gro ups which were restricted to certain types of habitat. Trout were more scattered a nd appeared capable of utilizing a wider a rray of stream h a bitats. Under e xperimental conditions coho were best adapted to maintaining positions in pools, and trout to holding positions in riffle s. These differences, which are most evident during spring and summer, probably account for differences in population s tability o f the two species in small coastal stre ams. During May and June la r ge numbers of young coho were displaced downstrea m in the Sal- mon River. Chapman (1962) h as shown that such displacement is the result of aggression and competition for space. The emigration of young coho from streams occurred at a season when levels of aggression were highest and when behavior w as least ritualized. Downstream displacement of trout did not occur even though density a nd r ate of aggression were high . Kalleberg (1958) showed that territory size decreased with increased population density of Atlantic salmon fry. Trout in the Salmon River may accommodate for changes in numbers by changing territory size. Coh o, which are restricted to pools, may displace surplus in- dividual s out of the pools. Experimental d ata indicated that such individuals would be unable to maintain positions against trout in the riffles. As a result coho, pushed out of pools, would move downstream to unused pool space or be displaced completely . The direction of retreat following combat may be important in a consideration of the m atter of downstream displacement. Retreating trout in the experimental stream aquarium tended to move upstream o r l a tera ll y . Figure 34 shows that about 25 % of the coho retreats were l ateral a nd 30 % were ups tream . However, the larges t single percentage (about 45) of retreats was downstream. These differences are consistent with differences in amount o f down stream di s place m e n t a nd with the t ype of aggress ive be- havior exhibited by coho and trout. During winter, coho are usually found in dense groups. The tendency to form such gro ups is usually reflected in the winter distributions (Fig. 21 and 22 ). In winter no downstream emigration occurred in the Salmon River. L a boratory studies revealed several behavioral phenomena which would fa- cilitate stability of groups in restricted areas of the stream during the winter. Levels of aggressiveness were lower in both species. The amount of biting and chasing was low in proportion to noncontact aggressiveness. The wig- wag di splay occurred frequently in laboratory conditions a nd was exhibited 1074 JOUR N AL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22 , NO. 4, 1965 (f) 40 f-< w a: f- w a: u. 20 0 ~ ... . . I LATERAL r::-"1 ~ D .·; :~·).: // · . .:· ... . ... .. -~-~~.: COHO STEELHEAD 7;: -::_:-~; I · ... ··:.· .. ·. :::.:.:. ... . · .. · ····· ······ DOWNSTRI:.AM UPSTREAM DIRECTION OF RETREAT FrG. 34. Direction of retreat of co ho and steelh ead follow- in g intraspecific aggressive contests (based on 248 con- tests among steelh ead and 458 among co h o). in contests for position near the front of a group . Fish which were displaced b y others using the wig-wag threat were pu sh ed back into the group of subor- dinates but we re not driven entirely o ut of the group . During winter, trout did not occur in tight gro u ps as did co h o. Behavior components (threat nips a nd wig-wag displ ays) whic h were evident in the gr o up behavior of co ho were exhibited infrequently in l a borator y groups o f trout. Hiding beh avior shown by trout under winter co nditions has adaptive value in protecting them from "scouring" and predation. Coastal rivers o f British Columbi a are fre quently s ubject to freshets, hence, hiding behavior, either in l og j ams or under stones, is advantageou s in maintaining position. Lindroth (195Sb) h as shown that mergansers can t a ke a heavy toll on trout parr populations. Lindroth a nd Bergstrom (1959) demonstrated that mer- ganser s could easily see fi sh in open water and c h ase them tenaciously. The birds even searched actively under the stones for the trout p arr. It m ay be ass umed that fi s h in positions unde r large stones would gain co n siderable protection from s u c h predation. In most cases hiding trout in the C hilliwack River were under rocks 20 -40 e m in diameter. Many hiding fish were found w ell down a mong the stones r ather th a n near the surface. The h a bit of see k- ing s helter i s important in the ecology of yo ung s t eelhead because it offers protection from winter freshets a nd from predation to m a n y fi s h which a re distributed a l o ng the stream margins in otherwise unprotected l ocation s. HARTMAN : BEHAVIOR OF SALMON AND TROUT 1075 The foregoing comments have pointed out some differences in the behavior and ecology of the two species. Although trout occupy a wider variety of stream habitats than coho (particularly in the largest stream) the two species overlap to a large degree in space utilization. INTERACTION OF YOUNG STEELHEAD AND COHO Segregation of natural populations of young coho and trout occurred at the season in which experiments indicated great similarity of environmental preferences. Separation in the wi ld was least pronounced during winter months, when experiments indicated differences in preferences. These two observations considered together support the belief that interspecific competition may be manifested in segregated resource use as suggested by Nilsson (1956, 1963). Interaction, which occurred in spring and summer because both species had simil ar demands, was accentuated by three factors. Population densities (in the stream) were highest in spring and summer (Fig. 12). Levels of aggressive- ness (laboratory) were highest early in the year (Fig. 27). In addition to this the aggressiveness shown involved much biting and chasing. Size differences may have contributed to the effect of segregation . Coho in the Salmon River were larger than the trout in spring and summer (Fig. 13) and could have displaced them from pools. In direct opposition to the above situation, winter populations of coho and trout coexisted to a large extent in the pools. Three main factors contrib- uted to this interspecific compatibili ty. First, spatial distribution and pre- ferences of the two species in the stream aquarium were different in winter. Second, stream population densities were lower in winter (Fig. 12). Third, levels of aggressiveness were lower in winter (Fig. 27). These three factors must contribute substantially to the winter coexistence of coho and trout. There is an apparent paradox in the fact that wild populations of both trout and coho occupy pools at a season when experiments indicate differences in preferences. It should be pointed out, therefore, that both species showed a preference for the deepest section of the depth gradient, which was com- parable to a pool (Fig. 22). However, trout and coho utilized this pool space differently; coho formed groups in open water above bottom, and trout scat- tered across the bottom. In the cover gradient trout occupied space under stones but coho occupied space beside the stones or downstream from them (Fig. 23 and 24). In a stream during the winter both species may make a de- mand on pool space. However, small but important differences in the use of space and cover, such as those described, permit coexistence of both species in a pool within a few inches of each other. As already stated, such coexistence would be facilitated because levels of aggressiveness in both species are low during winter. The previous discussion explains some of the reasons why coho and trout segregate spatially in spring and summer but occur together in winter. However one important question still remains. How do these two species remain m equilibrium in the two distinctive natural microhabitats, riffles, and pools? 1076 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 To gain better understanding of this problem it may be valuable to con- sider the effect of environment on the behavior of young brown trout. Hart- man (1963) showed that young brown trout could be induced to take up and defend positions if presented with simple visual reference points. If the struc- tural complexity of these reference points was increased, the rate of occupancy was increased. Kalleberg (1958) showed that defence of territories by young brown trout and salmon was initiated by running water. Territorial behavior of young brown trout was released by certain stimuli, the effect of which could be reinforced by others. The behavior of choosing and defending ter- ritories appears to be a reactive type of behavior which is governed by a complex of environmental stimuli. Presumably young steelhead in the riffle environment received more stimuli which elicited aggressive behavior and territory defence, than they received in the pools. The responses of young coho to various stimuli were different, hence, they were more strongly moti- vated to defend space in pools and less so in riffles. Such a differential response to environmental conditions is indicated by the differences in aggressiveness in riffle and pool habitats (Fig. 31). Segregation in the Salmon River is probably maintained because of differences in motivational states of trout and coho in the three microhabitats of the stream. If it were not for this differential aggressiveness, coho displaced from pools would be able to eliminate the smal- ler trout from the riffles, thus shifting the balance, in the whole stream, in favor of one species. In concluding the comments on the ecological relationship of these two species it is emphasized that changes in social behavior account, in a large way, for the seasonal change in severity of interaction. Differences in aggres- siveness in riffle and pool environments account for the segregation and the equilibrium of coho and trout in the two microhabitats. COMMENTS ON CONCEPTS OF COMPETITION A number of investigators have reported instances in which competttwn or interaction between species is manifested in segregation (Beauchamp and Ullyott, 1932; Macan, 1961; Connell, 1961). Segregation produced by com- petition among fish has been recorded by Nilsson (1955, 1958, 1960, 1963). Miura (1962) reviewed several cases in which it occurred in competing species of Japanese fish. Lindroth (1955), Kalleberg (1958), and Saunders and Gee (1964) deal with segregation of competing species of stream-dwelling salmonids. In most of the preceding cases each species has a slight morphological, phy- siological, or behavioral advantage over the other in some part of the envir- onment. It is considered necessary to emphasize that similar, competing spe- cies segregate and come into equilibrium in nature since many laboratory investigations on competition, carried out in homogeneous controlled envir- onments, would indicate otherwise (reviews by Crombie, 1947; DeBach and Sundby, 1963). Grinnell (1904), Gause (1934), and DeBach and Sundby (1963) have indicated that species having the same niche cannot occur to- gether without one eliminating the other. DeBach and Sundby (1963) have HARTMAN: BEHAVIOR OF SALMON AND TROUT 1077 recorded a case in which one species of Aphytis eliminates another and is subsequently displaced by a third species. They suggest that the displacement mentioned above illustrates "the competitive displacement hypothesis," i.e. species with identical ecological niches cannot coexist long in the sam e habitat. Because of the way the hypothesis has been stated and because of the varia- tions in its interpretation, the competitive displacement concept has been controversial (Hardin, 1960; Cole, 1960; Patten, 1961; Van Valen, 1960; Mcintosh, 1961). The concept might have been more acceptable if it had stated that in sympatric populations of similar species the level of competitive interaction will increase with the degree of ecological and behavioral similarity. This does not lead to the difficulty of discussing different species with identical niches, although it does still leave the problem of quantifying ecological and behavioral similarity. It is impossible to say how such interaction will be manifest, because competition in the natural environment may a lter the numbers, the growth rate, or the niche of an animal in a particular habitat. Temperate freshwater fish are in general unspecialized and flexible (Larkin, 1956) and hence can a lter their niche, as young trout and coho are presumed to do. This, on the other hand, may not be true of fish in the old freshwater environments studied by Fryer (1959). Highly speciali zed animals such as the parasitic wasps (DeBach and Sund- by, 1963) may be virtually incapable of occupying an altered niche; thus el imination of one species is the necessary outcome of competition when no additional factors control the numbers of both competing species. It is rea- sonable to assume that the amount of specialization as well as the degree of similarity of species will determine the effects of competition. These effects may involve displacement in space, displacement or segregation in food habits, separation in some gradient of environmental condition s, changes in growth rates, or the complete elimination of one species . Species interaction need not be manifested in one type of end result only. In the field of ecology, hypotheses can be postulated more easi ly than they can be tested. The present research adds support for the concepts ad- vanced by Nilsson (1956, 1963). In doing so it has emphasized the role of behavior in the interaction of species and has shown how an imal behavior may enter into certain problems in population biology. ACKNOWLEDGMENTS Many people have helped me during the course of this work. Dr W .S. Hoar h as given advice and stimulation throughout the investigation. I am particularly grateful to him. This study has been fully supported by the Fish and Game Branch of the B .C. Government. I am appre ciative of the encour- agement and support given by R.G. McMynn, former C hief Fish eries Biologist, S.B. Smith, present Chief Fisheries Biologist, and Dr T.G. North cote of the Fish and Game Branch. The work wo uld not have been possible without their interest . Dr I. MeT. Cowan, Dr J. Adams, Dr D. Chitty, Dr P. Dehne!, Dr P.A. Larkin, and Dr C .C. Lindsey read the manuscript and made several 1078 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 valuable criticisms. I appreciate the extensive help given by C.A. Gill during all stages of the work. J. Baehr, K. Calvert, J. Gee, G. Halsey , D. Harvey, P. Kohler, R. Krejsa, R. Leighton, B. Lister, J. Logan, M. Teraguchi, D. Udy, P. Wickett, and D. Wilkie all helped in the field work or some other phase of the study. My late brother Fay helped me in the early stages of the field work; I appreciate the time he gave. I wish to express my gratitude to W. Caulfield and H. Neate who gave help and brought good cheer at Cour- tenay. Finally I wish to acknowled ge the help and the patience of my wife Helen during the last 3 years . REFERENCES BEAUCHAMP, R. S., AND P. ULLYOTT. 1932. Competitive relationships between certain species of fresh-water triclads. J. Ecol., 20 (1): 200-208. CHAPMAN, D. W. 1962. Aggressive behavior of juvenile coho salmon as a cause of emigration. J . Fish. Res. Bd. Canada, 19(6): 1047-10 80. COLE, L. C. 1960. Competitive exclusion. Science, 132 (3423): 348-349. CONNELL, J. H. 1961. The influence of interspecific competition and other factors on the distri- bution of the barnacle Chthamalus stellatus. Ecology, 42(4): 710-723. CROMBIE, A. C. 1947. Interspecific competition . J. Animal Ecology., 16 (1): 44-73. DEBACH, P., AND R. A. SUNDBY. 1963. Competitive displacement between ecological homo- logues. Hilgardia, 34(5): 105-166. FABRICIUS, E. 1953. Aquarium observations on t h e spawning behavior of the char, Salmo alpinus. R ep t. Inst. Fresh water Res. Drottningholm, 34: 14-48. FRYER, G. 1959. The trophic interrelationships and ecology of some littoral communities of Lake Nyasa with especial reference to the fishes, and a discussion of the evolution of a group of rock-frequenting Cichlidae. Proc. Zool. Soc. London, 132(2): 153-281. GAUSE, G. F. 1934. The struggle for existence. The Williams and Wilkins Co., Baltimore. 163 pp. GRINNELL, J. 1904. The origin and distribution of the chestnut-backed chickadee. Auk, 21 : 364-382. HARDIN, G. 1960. The competitive exclusion principle. Science , 131(3409): 1292-1297. HARTMAN, G. F. 1963 .. Observations on beh avior of juvenile brown trout in a stream aquarium during winter and sprin g. J. Fish. Res. Bd. Canada, 20 (3): 769-787. KALLEBERG, H. 1958. Observations in a stream tank of territoriality and competition in juvenile salmon and trout (Salmo salar L. and S. trutta L.). Rept . I nst. Freshwater Res. Drottningholm, 39: 55-98. KEENLEYSIDE, M. H. A., AND F. T. YAMAMOTO. 1962. Territorial behaviour of juvenile At lantic salmon (Salmo salar L.). Behaviour, 19 (1-2): 139-169. LARKIN, P. A . 1956. Interspecific competition and population control in freshwater fish . J. Fish. Res. Bd. Canada, 13 (3): 327-342. LI NDROTH, A. 1955a. Distribution, territorial behaviour and movements of sea trout fry in the River Indalsalven. Rept. Inst. Freshwater Res . Drottningholm, 36: 104-119. 1955b . Mergansers as salmon and trout predators in the River lndalsalven. Ibid., 36: 126-132. LINDROTH, A., AND E. BERGSTROM. 1959. Notes on the feeding technique of the goosander in streams. Ibid., 40: 165-175 . MACAN, T. T. 1961. Factors that limit the range of freshwater animals."fBiol. Rev., 36: 151-198. MciNTOSH, R. P. 1961. Competitive exclusion principle. Science, 133(3450): 391. HARTMAN: BEHAVIOR OF SALMON AND TROUT 1079 MILNE, A. 1961. Definition of competition among animals. Symp. Soc. Exptl. Biol., 15 : 40-61. MIURA, T. 1962 . Early life history and possible interaction of five inshore species of fish in Nicola Lake, British Columbia. Ph.D. Thesis, Dept. Zoo!. Univ. British Col umbia, 133 pp. NEWMAN, M. A. 1956. Social behavior and interspecific competition in two trout species. Physiol. Zool., 29(1): 64-81. NILSSO N, N. A. 1955 . Studies on the feeding habits of trout and char in North Swedish lakes. Rept. Inst. Freshwater Res. Drottningholm, 36: 163-225. 1956 . Om konkurrensen i naturen. Zoologisk Revy, 1956: pp. 40-47. 1958. On the food competition between two species of Coregonus in a North-Swedish lake. Rept. Inst. Freshwater Res. Drottningholm, 39: 146-161. 1960 . Seasonal fluctuations in the food segregation of trout. char and whitefish in 14 North-Swedish lakes. Ibid., 41: 185-205. 1963. Interaction between trout and char in Scandinavia. Trans. Am. Fish. Soc., 92 (3): 276-285. PATTE N, B. C. 1961. Competitive exclusion. Science, 134 (3490): 1599-1601. SAUNDERS, R. L., AND J. H. GEE. 1964. Movements of young Atlantic Salmon m a small stream. J. Fish. Res. Bd. Canad a , 21(1): 27-36. VAN VALEN, L. 1960. Further competitive exclusion. Science, 132(3440): 1674-1675. WATER RESOURCES BRA NCH. 1958-1959 . Dept. Northern Affairs and National Resources. Surface Water Supply of Canada, Pacific Drainage. Water Resources Paper No. 12 8. 388 pp. 1959-1960. Dept. Northern Affairs and National Resources. Surface Water Supply of Canada, Pacific Drainage. Water Resources Paper No. 131. 402 pp. APPENDIX I. Mean fork length and range of samples of fish used in experiments, October 30, 1962, to November 23, 1963. Date preserved Period Mean Sample or measured when used Species" fork length Range size mm mm Oct. 19-20, 1962 Nov. 1, 1962 to c 58.3 40-95 13b Feb. 20, 1963 s 50.9 38-72 20b Jan. 10, 1963 c 68.6 52-97 73 s 60.7 50-75 40 Jan. 19, 1963 c 66.2 47-88 40 s 61.7 47-76 40 Apr. 25, 1963 c 67.5 51-99 56b s 59.3 49-67 40b May 1, 1963 May 7 to c 37.9 35--±2 15b May 25, 1963 Nov. 23, 1963 s 33 .5 26-42 22b June 8, 1963 c 43.4 40-47 16b June 12, 1963 s 39.7 29-49 62b July 19, 1963 c 52.1 38-64 65 " s 42.7 35-56 64 Sept. 2, 1963 c 66.4 50-89 60 " s 56.2 39-80 74 Oct. 8, 1963 c 73.1 52-104 65 " s 72.7 54-102 36 Nov. ~3, 1963 c 76.7 62-93 30b s 73.2 47-112 35b a,: = coho; S = steel head. bMeasurements made on preserved material. 1080 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 22, NO. 4, 1965 APPENDIX II. Deta ils of spring a nd early summer experiments comparing behavior and distribu- tion of co ho and steelhead. Date a nd year 1963 April19-21 " 2 1-23 June 7-9 " 9-11 May 26-28 " 28-30 June 11-13 " 13-15 Jun e 22-24 " 24-26 June 30-July 2 J ul y 2-4 April 28-30 " 24-26 June 16-18 APPENDIX III. Date and year 1963 Sept. 19-21 " 2 1-23 27 -29 29-0ct . 1 23-25 25-27 Oct. 1-3 " 3-5 5-7 7-9 Arran~eme n t In aquarium Control " Cover " Li?,ht D7pth Depth and ~e l~7ity Cover Depth Details of fall Arrangement in aquarium Control " Cover Li7,ht Depth " Depth and velocity " " Specie s (4 0 fish) Co ho " Steelhead Co ho Steel head Coho Steelh ead Steelhead Coho Steelh ead Coho Temperature range (C) 6 .5-8.3 8.5-9.2 11.7-12.0 14 .0-14 .3 10.0-11 .2 13.0-13.3 13.0-13.5 13.4-14 .0 15.2-15.6 14.0-15 .9 14 .0-14.5 14.9-15.4 7 .1-7.5 7 .1-8 .4 15 .0-16.0 experiments comparing behavior steelhead and coho. Temperature S pecies range (40 fish) (C) Steelhead 12 .7-13.3 Coho 13.2-13.4 Steel head 12 .3-13.1 Coho 14 .2-14.5 12.7-13.5 Steelhead 13.6-14 .2 Coho 12.5-13.0 Steelhead 11.6-12.1 Coho 10.1-10.6 Steel head 10.8-11.0 Day length in h old ing pond 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 Average number moving 1.8 2.7 3.9 6.0 3.7 2.2 4.1 1.9 4.8 4.1 9.8 8.4 2.5 1.1 3.5 and distribut ion of young Day length Average in holding number pond moving 12 7.9 12 3.0 12 11.5 12 2.6 12 3 .5 12 11 .6 12 3.9 12 7 .2 12 6.2 12 7.5 HARTMAN: BEHAVIOR OF SALMON AND TROUT 1081 APPENDIX IV. Details of winter experiments comparing distribution and behavior of young steelhead and coho. Arrangement Temperature Day length Average Date and in Species range in holding number year aquarium (40 fish ) (C) pond moving 1962-63 Nov. 5-7 Control Coho 6.9-7.1 12 8.7 " 19-2 1 " " 7.2-7.8 9 6.3 Dec. 4-6 4. 7-4 .9 8 Nov. 11-13 Steel head 7.0-7 .1 12 18.3 " 17 -19 6.2-7.2 12 17.7 Dec. 6-8 5 .2-5 .7 8 19.7 Nov. 30-Dec. 2 Li,~ht Coho 2.7-3.2 9 2.2 Dec. 2-4 Steelhead 3.1-4.0 9 7.4 17-19 Depth and Coho 5.0-5.3 8 5.0 ,, 19-21 vel?~ity Steel head 5.3-5.6 8 10 .6 J~~· 19-21 Cover Coho 2 .0-2.4 8 1.1 23 -25 " " 2.4-2.8 8 2.1 21-23 Steel head 3 .0-3.1 8 7.0 25-27 2 .8-2 .8 8 8.1 8-10 D~pth Coho 1.7-2 .5 8 2.3 12-14 0.5-1.0 8 0.6 6-8 Steel head 3.9-4.4 8 0.7 10-12 0.8-1.0 8 0.9