“““““ 1111:1111; 1911111 .11 :11 .1 I. 1 :1II _.. ' 111.1 1111 111111111111". .11.” .91 1' 11111 ' 11.1 11.1.1“ 11:11' 1 1 11” 11111111 1111111111 111 1111:1111 1 1111' ‘ ,1 1. '1, 111111 111 1' II11111111 1111:- 1- 11111111111111111I111111111111111111111111 - 1 E15111 1111111111111 111 11111 1.111111 111.1 1111 1111111111: ' 111' : 111‘ . .‘E I 11 1111111111111 .111 111111'111111111111111 1 111111111111111 1 111111111111 {11111 1111111 117111111; 1: '111111'11111111111'1111111 . 11111111111111 . M11 11111 11 1111'! 1111111.11111111111111111111'I11 1111111 111111111 111 11111111 11111’111111111111111 I11 1111111111I11111111111111~11111111111 11 1111111115111111111111111111111111111111111111 11.111111 1'11 11111 11.19’111. 111111 1 INI1L1I‘1 ..‘ 11;: . 1111' 1 ?:«:1'111111 1111 1111111111 111111111 111 '. :‘f 111:1111111111111 11.111.311.111; 111111111 111.111 111111111111 .11 11 111111 11111111111111 1111 .1! 11111111111111'11111111 1111 1111.1: 111 111111111111111111111111 1111111 111 11111111 41111111 1111 211’fi11tmr 1 151 11:1111 11:51 1 ‘ .1111 111.1111. '111‘111III111111.11'1I1111111I11 .11 111111111 111111. 111 111 11111111111i13~l1 $11, 1 1111111111111111111111111111111111111111111111111111111111111 1111111111I 1111111111111111111111111111111111'11‘ “111111111 I 1:1I11I1111I1111111111I111'1'11111I11 1111111111111111111111111111111111I‘1I 11111111111111 . 1 111 111111 1111111111 1111111 - . .11 1 .111 1 1 $1111111111111111111111'1111 1111111 11111111111 111111 1111111111111 111111111111 111111111 11111 1.1.1711 1 111 This is to certify that the thesis entitled Effect of Cover and Social Structure on Position Choice by Brown Trout (Salmo trutta) in a Stream. presented by Charles Eugene Bassett has been accepted towards fulfillment of the requirements for Masters . Fisheries degree in Wofessor Date February 9, 1978 0-7639 EFFECT OF COVER AND SOCIAL STRUCTURE ON POSITION CHOICE BY BROWN TROUT (SALMO TRUTTA) IN A STREAM BY Charles Eugene Bassett A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife 1978 ABSTRACT EFFECT OF COVER AND SOCIAL STRUCTURE ON POSITION CHOICE BY BROWN TROUT (SALMO TRUTTA) IN A STREAM BY Charles Eugene Bassett Position choice by adult wild brown trout was observed in controlled-flow stream channels during the summers of 1975 and 1976. In 1975, groups of 25-30-cm (TL) trout were offered pairs of cover types. Occupation of a specific type represented a choice between known stimuli. The response to overhead cover was primarily visual rather than tactile. Response was strengthened by reducing the amount of light reflected from the stream bed beneath cover. Visual re- sponse was the same to a landmark on the stream bed as to one which was 15 cm above the stream bed. Lateral visual concealment is an important function of opaque vertical surfaces. In 1976, groups composed of small (23-28 cm FL) and large (35—41 cm FL) trout were observed from elevated blinds. One large (25 x 122 cm) and one small (25 x 61 cm) opaque overhead cover was provided for each group. Small trout avoided using cover with some large trout. Low current velocity beneath cover seemed preferable to large cover size, even for large trout. The trout avoided feeding in direct sunlight and in the fastest currents. Large trout fed closer to the channel walls than small trout. ACKNOWLEDGEMENTS I would like to extend my appreciation to Dr. Ray J. White, my committee chairman, for the opportunity to undertake this study and for his field assistance and encouragement. Thanks, too, are extended to committee members Drs. Howard E. Johnson and John A. King for their review of the manuscript. This study was made possible through the interest and co- operation of the Michigan Department of Natural Resources. Loan of equipment and valuable advice, particularly by the Grayling staff, greatly facilitated the field aspect of this study. For their assistance in the field thanks are extended to: Al Bourgeois, Peter Cardinal, Charlie Dewberry, Mike Enk, Kurt Fausch, Jim Gruber, Ray Rustem, and Dick Stanford. I am also indebted to the members of the Challenge Chapter of Trout Unlimited for their enthusiastic support. Special thanks are due to Mrs. Ruth Gruich for loaning the binoculars that greatly facilitated observations of trout in 1976, and to Mr. T.W. "Rip" VanWinkle for photographic assistance and hospitality. Finally, I am immeasurably grateful to my family for their moral support throughout the study and for their physical assistance in 1975. ii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . FACILITIES . . . . . . . . . . . . . METHODS RESULTS METHODS RESULTS - 1975 . . . . . . . . . . . Experiment 1 Experiment 2 . . . . . . . . Experiment 3 Statistics . . . . . . . . . - 1975 . . . . . . . . . . . Experiment 1 . . . . . . . . Experiment 2 Experiment 3 . . . . . . . . - 1976 . . . . . . . . . . . Activity Definitions . . . Light Intensity Beneath Cover Statistics . . . . . . . . . - 1976 . . . . . . . . . . . Agonistic Behavior . . . . . Cover Use . . . . . . . . . Preferred Cover . . . Cover Use and Feeding Feeding . . . . . . . . . . Types of Feeding Behavior Drift Food Supply . . Feeding Time and Movement Characteristics of Feeding Stations Size of Trout and Size of Food Miscellaneous . . . . . . . Social Rank Summary and Energy Use . Time of Day Summary . . . . Replicate Summary . . . . . Page xi 18 18 18 19 21 21 21 21 27 37 38 4O 44 44 SO 52 61 64 64 66 73 80 93 94 96 103 103 DISCUSSION - 1975 AND 1976 . . . . . . . Introduction . . . . . . . . . . Statistics.. . . . . . . . . . . Activity Budget . . . . . . . . Social Structure . . . . . . . . Advantages of Dominance . . . . Lateral Surfaces . . . . . . . . Light . . . . . . . . . . . . . Current Velocity . . . . . . . . Application to Stream Management SUMRY C O I O O O O O O O O O O O I 0 LITERATURE CITED . . . . . . . . . . . . APPENDICES C O O O O O O O C O O O O O O A. Tables . . . . . . . . . . . B. Figures . . . . . . . . . . iv Page 107 107 109 110 111 116 118 119 123 124 130 134 139 139 170 Table LIST OF TABLES Discharge and mean cross-sectional depth and current velocity in each channel section in 1975 . . . . . . . Number of sightings of trout under low (10 cm) or high (15 cm) transparent overhead cover devices, or not under cover. Tested only with Group I . . . . . . . Number of sightings of trout under opaque overhead cover devices with dark substrate or light substrate. or not using cover. All devices were 10 cm above the stream bed . . . . . . . .-. . . . . . . . . . . . Number of sightings of trout responding visually or tactually to lateral cover devices with low (at the stream bed) or high (15 cm above the stream bed) dark stripe, or not responding visually or tactually. See text for definitions of "visual" and "tactile" responses . . . . . . . . . . . . . . . . . . . . . . Light intensity under overhead cover devices of various sizes and heights above the stream bed. Measurements were made indoors under a constant light source . . . . . . . . . . . . . . . . . . . . . . . . Derivation of F-values for 3-factor ANOVA with two fixed (F) and one random (R) factor, and unbalanced replication. MS 8 mean square . . . . . . . . . . . . Number of one-on-one agonistic bouts between pairs of social ranks in replicates with 4 trout . . . . . . The number of agonistic bouts in which trout were displaced from cover or feeding stations in replicates Wi th 4 tron t O O O O O O O O O O C O O O C O O O O O I Mean percent of observation session time and movement, and mean rate of movement in agonism by small trout in replicates with two large trout and with no large trout I O O O O O O O O O I O O O I O O O O O O O O O Page 22 23 24 39 42 47 48 50 Table 10 11 12 13 14 15 16 17 18 19 Page Current velocity and depth at large and small covers in each replicate. "Free-choice" time and number of observation sessions spent under cover by each social rank indicates preference as described in the text . . 54 Time (min) and number of observation sessions spent under cover by each social rank alone and in all possible combinations with other ranks during sessions when positions of all 4 trout were known. Preferred covers indicated by an asterisk . . . . . . 55 Time available to use the "non-preferred" cover when dominants were using only the "preferred" cover. Percent utilization indicates the extent to which ranks 2, 3, and 4 took advantage of opportunities to avoid dominants under cover. This only includes sessions when positions of all 4 trout were known . . 58 Observed frequencies and derivation of expected frequencies of using small and large covers for combinations of social ranks . . . . . . . . . . . . . 60 Location of feeding stations used by each social rank and location of overhead covers in replicates with 4 trout. Percent of total feeding time spent on the west and east sides of the channel, and in the middle of the channel shown. Time spent feeding from cover is not included . . . . . . . . . . . . . . . . . . . 62 Mean percent of the volume (cc) of trout stomach contents and drift net contents as isopods, mayflies, and caddisflies. Miscellaneous items were not included in the total volume . . . . . . . . . . . . . 71 Summary of daily feeding activity by brown trout in this study . . . . . . . . . . . . . . . . . . . . . . 75 Mean time (min) spent per feeding station by each social rank during morning, midday, and evening in replicates with 4 trout. This includes time spent feeding from cover . . . . . . . . . . . . . . . . . . 78 Mean distance (m) moved per feeding station by each social rank during morning, midday, and evening in replicates with 4 trout. This includes movement associated with feeding from cover . . . . . . . . . . 79 Mean current velocity (cm/s) at feeding stations used by each social rank during morning, midday, and evening in replicates with 4 trout. This includes cover feeding stations. Current velocity is weighted by the percent of total feeding time spent at each station .. . . . . . . . . . . . . . . . . . . . . . . 81 vi Table 20 21 22 23 24 .25 26 27 Mean water depth (cm) at feeding stations used by each social rank during morning, midday, and evening in replicates with 4 trout. This includes cover feeding stations. Depth weighted by percent of total feeding time spent at each station . . . . . . . . . .,. . . . Observed and expected percent of feeding time spent at current velocities greater than 19.5 and 24.0 cm/s in replicates with 4 trout. Expected values equal percent of section area in which velocities exceeded 19.5 and 24.0 cm/s (Appendix Figures 33-86). This includes cover feeding stations . . . . . . . . . . . Time (min) spent at feeding references by each social rank in replicates with 4 trout. Replicate IA2 did not have feeding references. Percent of total feed- ing time at references in parentheses . . . . . . . . Location of feeding stations used by each social rank and location of shade in replicates with 4 trout. Percent of feeding time spent on the west and east sides of the channel and in the middle of the channel shown. Parentheses indicate the shaded side. Time spent feeding from cover is not included . . . . . . Summary of the mean percent of observation session time and movement, and mean rate of movement in each activity by social rank in replicates with 4 trout. Means with different superscripts are significantly different (p < 0.05) . . . . . . . . . . . . . . . . . Relationship between metabolic rate (Y = mgO /h) and swimming speed (X = cm/s) for each social rank plotted from equations relating metabolic rate and body weight (g) at several swimming speeds for rain— bow trout (Rao, 1968). Standard metabolism occurs when swimming speed is zero . . . . . . . . . . . . . Estimated energy use (mgO /h and mgO /kg/h) above standard metabolism (Table 25) for each social rank. Values were derived from time spent in each activity, mean swimming speed associated with each activity (Table 24), and equations relating metabolism to swimming speed (Table 25) . . . . . . . . . . . . . . Summary of the mean percent of observation session time and movement, and mean rate of movement in each activity by time of day in replicates with 4 trout. Means with different superscripts are significantly different (p < 0.05) . . . . . . . . . . . . . . . . . vii Page 82 84 86 89 97 99 101 104 Table 28 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 Summary of the mean percent of observation session time and movement, and mean rate of movement in each activity by replicates with 4 trout. Means with dif- ferent superscripts are significantly different (p<0.05)..................... Daily maximum and minimum water temperatures for the East Branch of the AuSable River at the DNR Grayling Field Office during summer 1975 . . . . . . . . . . Daily maximum and minimum.water temperatures for the East Branch of the AuSable River at the DNR Grayling Field Office during summer 1976 . . . . . . . . . . Sex, social rank, and the weight and fork length of brown trout in each replicate at the beginning and end of each 3-week experiment . . . . . . . . . . . Time and number of observation sessions devoted to each replicate during morning, midday, and evening . Time trout spent in each activity during morning, midday, and evening. Values are cumulative over all social ranks and replicates .. . . . . . . . . . . . Movement trout devoted to each activity during morning, midday, and evening. Values are cumulative over all social ranks and replicates . . . . . . . . Number of one-on-one agonistic bouts won and lost by each social rank in replicates with two small trout Number of one-on—one agonistic bouts won and lost by each social rank in the replicate with two large and one small trout . . . . . . . . . . . . . . . . . Number of one-on-one agonistic bouts won and lost by each social rank in replicates with two large and two small trout . . . . . . . . . . . . . . . . . . Mean percent of observation session time spent in agonism by each social rank during morning, midday, and evening in replicates with 4 trout . . . . . . . Mean percent of observation session movement spent in agonism by each social rank during morning, midday, and evening in replicates with 4 trout . . . Mean rate of movement (cm/s) associated with agonism for each social rank during morning, midday, and evening in replicates with 4 trout . . . . . . . . viii Page 105 . 139 140 141 142 143 - 144 145 145 146 147 148 149 Table — Page A13 Mean percent of observation session time spent under cover by each social rank in replicates with 4 trout. Time spent feeding from cover is not included . . . . 150 A14 Mean percent of observation session movement associated with cover use by each social rank during morning, midday, and evening in replicates with 4 trout. Movement associated with feeding from cover is not included . . . . . . . . . . . . . . . . . . . . 151 A15 Mean rate of movement (cm/s) associated with cover use by each social rank during morning, midday, and evening in replicates with 4 trout. Movement associated with feeding from cover is not included . 152 A16 Number of observation sessions in which each social rank used large cover, small cover, and no cover and the probability of using small and large covers in replicates with 4 trout. Sessions in which each rank used cover alone and in all possible combinations with other ranks are included in sums . . . . . . . . . . 153 A17 Mean distance (m) to the nearest cover from feeding stations used by each social rank during morning, midday, and evening in replicates with 4 trout. Dis- tance weighted by percent of total feeding time spent at each station. Cover feeding stations are included ,154 A18 Mean percent of feeding time spent feeding from cover by each social rank during morning, midday, and evening in replicates with 4 trout . . . . . . . . . . . . . 155 A19 Number and volume (cc) of organisms in subsamples of drift net contents on each sampling day. Subsamples equal 5% of the total drift net contents . . . . . . 156 A20 Total estimated number and volume (cc) of organisms collected in drift nets on each sampling day . . . . 160 A21 Number and volume (cc) of organisms in stomachs of trout in each group in the morning (0800 h) and evening (2000 h). Ranks 1 and 2 are large trout, ranks 3 and 4 are small trout (Appendix Table A3) . . 161 A22 Mean percent of observation session time spent feeding by each social rank during morning, midday, and evening in replicates with 4 trout. This includes time spent feeding from cover . . . . . . . . . . . . 164 ix Table A23 A24 A25 A26 Mean percent of observation session movement devoted to feeding by each social rank during morning, midday, and evening in replicates with 4 trout. This includes movement associated with feeding from cover . . . . . Mean rate of movement (cm/s) associated with feeding by each social rank during morning, midday, and evening in replicates with 4 trout. This includes movement associated with feeding from cover . . . . . Mean distance (m) to the nearest wall from feeding stations used by each social rank during morning, midday, and evening in replicates with 4 trout. Dis- tance weighted by percent of total feeding time spent at each station. Cover feeding stations are not included . . . . . . . . . . . . . . . . . . . . . . Schedule of artificial feeding and response of large trout in replicate IIIAl to grasshoppers. Positions that grasshoppers landed in the channel are indicated by coordinates (longitude, latitude, see Appendix Figure 83) . . . . . . . . . . . . . . . . . . . . . Page 165 166 167 168 Figure 10 LIST OF.FIGURES Controlled-flow channels used in 1975 with positions of current deflectors and screen barriers . . . . . . Cross-sectional view of a channel showing gravel sloped against the walls . . . . . . . . . . . . . . Cover devices used in 1975. (a) Opaque overhead cover the same size as the transparent devices used in experiment 1. (b) Overhead opaque cover with "dar " substrate. (c) "High"- and "low"-stripe lateral cover devices. Both were buried to the level of the base of the low stripe . . . . . . . . . . . . Location of cover devices in one channel section in 1975. Placement was identical in all 4 sections . . Controlled-flow channels used in 1976 with positions of current deflectors, screen barriers, and elevated blinds . . . . . . . . . . .7. . . . . . . . . . . . Large overhead cover . . . . . . . . . . . . . . . . Location of overhead covers and feeding references (-) during each 3-week experiment . . . . . . . . . . (a) A feeding reference. Only the shaded part was above the stream bed. Top view of current velocity around a feeding reference (b) 2.5 cm and (c) 5 cm above the stream bed. Numbers indicate the mean dif- ference (cm/s) from background velocity at 3 references . . . . . . . . . . . . . . . . . . . . . Side view of the downstream observation blind show- ing the enclosed ladder and high channel walls. A window faces the channel on both sides of the blind . Mean distance to the nearest cover from feeding stations used by each social rank in replicates with 4 trout. Distance weighted by the percent of feeding time spent at each station. Cover feeding stations are not included . . . . . . . . . . . . . . xi Page 11 12 15 16 28 29 32 33 36 65 Figure Page 11 Estimated total volume of drift food organisms that entered channels A and B on each sampling day. Volumes were estimated from sub-samples equal to 5% of the total volume of organisms collected . . . . . 68 12 Mean volume (cc) of food in the stomachs of large and small trout removed from the channels at 0800 h and 2000 h (EDT). n = number of stomachs checked . 70 13 Mean distance to the nearest wall from feeding stations used by each social rank in replicates with 4 trout. Distance weighted by the percent of total feeding time spent at each station. Cover stations are not included . . . . . . . . . . . . . 88 Bl Location of trout sightings for Group I in test of lateral cover devices. L = low-stripe device, H = high—stripe device. Drawn to scale . . . . . . . . 171 82 Location of trout sightings for Group II in test of lateral cover devices. L = low-stripe device, H = high-stripe device. Drawn to scale . . . . . . . . 173 83 Depth (cm) and current volocity (cm/s) in section Al. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout posi- tions. Distance between numbers on the vertical axis is l m. Mean depth—21.1 cm, mean velocity- 21.9 cm/s. Velocity was measured 5 cm above the stream bed . . . . . . . . . . . . . . . . . . . . . 175 B4 Depth (cm) and current velocity (cm/s) in section A2. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout posi- tions. Distance between numbers on the vertical axis is 1 m. Mean depth-23.5 cm, mean velocity- 20.7 cm/s. Velocity was measured 5 cm above the stream bed . . . . . . . . . . . . . . . . . . . . . 177 ES Depth (cm) and current velocity (cm/s) in section Bl. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout posi- tions. Distance between numbers on the vertical axis is l m. Mean depth-19.2 cm, mean velocity- 21.6 cm/s. Velocity was measured 5 cm above the stream bed . . . . . . . . . . . . . . . . . . . . . 179 xii Figure 86 Page Depth (cm) and current velocity (cm/s) in section B2. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout posi- tions. Distance between numbers on the vertical axis is l m. Mean depth-22.6 cm, mean velocity- l7.4 cm/s. Velocity was measured 5 cm above the stream bed . . . . . . . . . . . . . . . . . . . . . 181 xiii INTRODUCTION The brown trout (Salmo trutta) is popular among stream fisher- ‘men. It tolerates higher water temperatures than the native brook trout and therefore provides fishing in streams that become too warm for brook trout. Several studies have demonstrated the brown trout's ability to maintain populations where other trout are severely de- pleted by heavy fishing pressure (Thorpe et al., 1947; Shetter, 1950; Cooper, 1952; Lemmien et al., 1957; Marshall and MacCrimmon, 1970). The brown trout is particularly difficult to catch because it hides beneath log jams, undercut banks, ledges, and other types of cover during the day (Butler and Hawthorne, 1968; Brynildson et al., 1973). Abundance of yearling and older brown trout has been strongly correlated with amount of hiding cover in streams (Lewis, 1969; Wesche, 1976; Enk, 1977). Fishery managers have taken advantage of the brown trout's association with cover by building artificial cover in streams to in- crease carrying capacity. Results have generally been favorable. White (1975) reviewed evaluations of artificial cover addition on 6 streams containing brown trout for which several years of population measurements were available. In every case, abundance of brown trout or mixed populations of brown and brook trout increased, sometimes dramatically, following cover addition. The greatest benefits accrued to the largest size classes. Cover requirements of adult brown trout are not clearly under- stood. Fishery managers have had to rely largely on intuition gained from fishing experience to guide decisions on the design and place- ment of artificial cover in streams. Consequently, most cover devices have been built to simulate undercut banks, and log jams, often at great expense (White and Brynildson, 1967; White, 1975), without knowledge of the specific physical characteristics that make such areas attractive. Manipulation of cover changes other aspects of brown trout living space as well. Changes in streamflow patterns and the availability of gravel after cover addition changes the avail- ability of feeding and spawning microhabitat (Hunt, 1969). Recently the fisheries literature has provided some informa- tion to aid in improving living space for stream salmonids. These studies indicate that position choice reflects responses to phototactic, thigmotactic, and rheotactic stimuli. Characteristics of the food supply and the social environment are also very important. A brief review of the influence of these factors follows. Light is an important regulator of the activity of brown trout. Peak locomotory activity of wild age-11+ brown trout occurred during darkness and was independent of food availability (Chaston, 1968). Re— versing the light conditions so that the dark period occurred at what was previously the light period reversed the response so that activity was still greatest during the dark period. Brown trout avoid incident light by using shade produced by objects above or below the water sur- face (Hartman, 1963; Butler and Hawthorne, 1968; Baldes and Vincent, 1969; Jenkins, 1969; DeVOre, 1975). Large brown trout, in particular, avoid bright light and generally remain inactive during the day. Im- mature and mature brown trout show an attraction to dark backgrounds (Hartman, 1963; Ritter and MacCrimmon, 1973). Rainbow trout show the same response (MacCrimmon and Kwain, 1966; KMain and MacCrimmon, 1967). The relative importance of incident and reflected light has not been studied. A strong thigmotactic response has also been attributed to brown trout in flowing water. This response has been confounded with response to reduced current velocity close to solid objects in most studies, e.g. Hartman (1963) and Baldes and Vincent (1969). Clearer evidence for thigmotaxis has been found. Brown trout used overhead cover devices with clear plastic streamers fastened under both sides significantly more than covers without streamers (DeVore, 1975). Clear streamers presumably offered very little resistance to the current and little or no visual stimulus. Serchuk (1976) concluded that movements of adult brown trout along the walls of a pumped— storage resevoir demonstrated thigomotaxis as they were not affected by changes in light level or current velocity. Current velocity may exceed the importance of other physical factors in determining position choice by brown trout in streams as it is the dominant factor determining energy expenditure (Vincent, 1969). Response to current velocity in this study is considered separately from response to cover, but in the stream the two usually cannot be separated. Objects below the water surface that provide shade and thigmotactic stimuli also usually create pockets of low current velocity. Adult brown trout tolerate a wide range in current velocity but seem to prefer velocities less than 15 cm/s (Baldes and Vincent, 1969; Wesche, 1976). Food availability obviously is important in determining the suitability of streams for trout production. Feeding on drifting food from a more-or-less fixed station is characteristics of stream- dwelling salmonids. This permits maximum.food intake with minimum energy expenditure (Allen, 1969a). Wild brown trout about 24 cm long in stream aquaria fed at the depth of maximum prey density and pre- ferred the largest familiar items available (Ringler, 1975). In the natural stream environment, prey density and size are determined by streamflow patterns and velocity. Fast currents carry more food items per unit of time (Elliot, 1967) and larger items (Chapman and Bjornn, 1969) past feeding stations than do slow currents. The pattern of flow concentrates food in areas characterized as principal lines of drift. Brown and rainbow trout choose feeding stations in relatively slow currents overlain by these principal lines of drift (Jenkins, 1969). Drift food supply and its influence on position choice cannot be divorced from streamflow characteristics. Temporal patterns of position choice by brown trout in streams may be influenced by the diel periodically of invertebrate drift (Elliot, 1970). But lack of correlation between drift availability and stomach contents in some cases (Warren et al., 1964; Chaston, 1969), indicates that reliance on drift is not complete. Bottom foraging and feeding on fish seems to be the primary sustenance for brown trout larger than about 40 cm long (Jenkins, 1969; Waters, 1972). The social environment affects position choice of trout by influencing the expression of territoriality. Territoriality is a characteristic of most stream-dwelling salmonids (Allen, 1969b). Two types of social organization have been described for brown trout: territorial mosaics, and hierarchies. A territorial mosaic is where the stream bed is divided into many adjoining defended areas. This pattern is typical among juveniles (Kalleberg, 1958; Hartman, 1963). Each trout is consistently dominant only within its own fixed territory. Qualifications for mosaic formation seem to be uniformity in body size, physical homogeniety of the environment, and lack of movements (Jenkins, 1969). Hierarchies tend to form in groups of trout containing a mixture of body sizes in a heterogeneous environment. Social rank is related to body size and individual agressiveness, but not to geographic location (Jenkins, 1969), i.e. social rank does not change after trout move to a new position. The highest ranking trout occupies the most upstream position under cover (Butler and Hawthorne, 1968; Butler, 1975) and at feeding areas in the open (Jenkins, 1969). The upstream position seems to have feeding advantage. Subordinates may move to better positions vacated by dominants but always yield these positions when the dominant returns. In territorial mosaics and in hierarchies a space or "social force field" (McBride, 1964) surrounding the trout is defended. The area immediately upstream from the fish is defended most persistently (Butler, 1975). The size of the social force field depends on the motivation of the fish and the physical characteristics of the stream channel. The social force field is smaller when trout use cover than when they feed (Butler, 1975). Streambed features that provide visual isolation from competitors also reduce space requirements (Kalleberg, 1958; Hartman, 1963). Large trout defend more space than small trout (Allen, 1969b). General objectives of this study are: (1) To define characteristics of high quality hiding cover for adult brown trout. (2) To describe the daily activity of two sizes of adult brown trout, with particular attention to the importance of cover use. (3) To investigate the relative importance of the physical characteristics of cover and social rank on choice of feeding and cover positions. Specific hypotheses concerning cover use are: (1) The response to overhead cover is visual rather than tactile. (2) Trout respond to light reflected from the stream bed be- neath overhead cover. (3) Visual isolation is an important function of cover offer- ing lateral vertical surfaces. (4) Lateral vertical surfaces are preferred if they have a distinct visual landmark low rather than high in the water column. (5) Large (older) adult trout use cover more than small (younger) adult trout. (6) Socially dominant trout feed closer to overhead cover than subordinates. (7) Socially dominant trout use the largest overhead cover available. (8) Socially dominant trout prevent subordinates from using overhead cover. (9) Trout feed on the side of the stream channel with over- head cover. FACILITIES The study was conducted at the Field Office of the Michigan Department of Natural Resources in Grayling, Michigan during the summers of 1975 and 1976. The site was formerly a hatchery; raceways served as controlled-flow stream channels. Water was supplied to the channels from the East Branch of the AuSable River. The East Branch supports a mixed population of brown, brook, and rainbow trout with brown trout the predominant species. Daily ranges in water temperature during 1975 and 1976 parts of this study are presented in Appendix Tables Al and A2. The physical characteristics of the experimental environment and the methods employed differed between years so they will be described separately. METHODS - 1975 Two channels, 3.4 m wide and 61 m long, were used for the experiments in 1975 (Figure 1). Each channel was divided into two 27-m sections. Substrate in each channel ranged from one-inch dia- meter gravel at the upstream end to sand at the downstream end. The channel sides were concrete slab that sloped approximatley 100° to the water surface. Gravel sloped against the walls eliminated the toe of the wall as lateral cover (Figure 2). Prior to the beginning of the experiments, the substrate in each channel was raked and smoothed to create a uniform cross-section throughout. Average depths and current velocities are presented in Table 1. Discharge measured at the midpoint of each section with a pygmy Gurley current meter was 0.158 cms in channel A and 0.144 cms in channel B. Constant discharge was maintained throughout the study by regulating the level of the head pool above the channels. Current deflectors placed just below the entry sluices to each channel dispersed flow evenly in the channel cross-section (Figure 1). Table 1. Discharge and mean cross-sectional depth and current velocity in each channel section in 1975. Mean depth (cm) Mean velocity (cm/s) Discharge (cms) A1 18.9 24.9 0.158 A2 23.1 21.9 0.158 Bl 22.9 19.8 0.144 82 22.5 ' 20.4 0.144 10 Figure l. - Controlled flow channels used in 1975 with positions of current deflectors and screen barriers. 11 Head Pool ' Current Deflector- Screen Barrier —— Screen Barriers Screen Barrier-— Sefiiu‘ Section A1 81 l l".— M Section Section A2 82 5 Meters Figure l 12 at 5" fi 3 45 Figure 2. - Cross-sectional view of a channel showing gravel sloped against the walls. 13 Screens with one-inch mesh blocked the head and foot of each section. The screens were cleaned daily to prevent buildup of debris that would have created an irregular pattern of flow, which in turn, might have influenced position choice of the trout. Trout responses to overhead and lateral cover devices were tested. Overhead cover consisted of 25 x 61-cm plates of single- strength (1.6 mm) glass supported by plexiglas holders mounted on 6-mm-diameter threaded, galvanized rods (Figure 3a). These cover de- vices were thinner than those used by DeVore (1975) which he found had very minor influence on current velocity under the cover. Lateral cover consisted of glass plates supported in a vertical position by the plexiglas holders and rods which were buried horizontally just below the surface of the stream bed (Figure 3c). Five pairs of evenly-spaced cover devices were placed in each section, each pair representing two cover options. Within each pair the devices were randomly assigned to the right or left side of the channel, 1 m from the wall (Figure 4). Three experiments were con- ducted with 3 different pairs of cover types. Experiments 2 and 3 were repeated with two groups of trout. The subjects of the experiments were 25—30—cm (TL) wild brown trout. Two groups of 20 trout were used. Group I was captured in the AuSable River mainstream about 8 km east of Grayling using a boat- mounted 250 volt D.C. electro-fishing unit. Group II was captured with a backpack shocking unit in the East Branch of the AuSable River about 2 km above the DNR Field Office. Following capture, the trout were held for 5-7 days in a channel lacking overhead cover, but in 14 Figure 3. - Cover devices used in 1975. (a) Opaque overhead cover the same size as the transparent devices used in experiment 1. (b) Overhead opaque cover with "dark" sub- strate. (c) "High"- and "low"-stripe lateral cover devices. Both were buried to the level of the base of the low stripe. '16 v0 9 Adeiers 4L“ Figure 4. - Location of cover devices in one channel section in 1975. Placement was identical in all 4 sections. 17 which they had access to the walls. The holding period permitted some recovery from the effects, if any, of electro-fishing. Five trout from each group were randomly assigned to each of the 4 channel sections. Each group was given 3 days to acclimate to the cover options available. Three to 5 days of observation followed. Then the trout were seined from the channels and transferred back into the holding area where they were confined for 18-24 h in a 3.7 x 1.2-m net enclosure while the next set of cover devices was installed. The entire procedure was repeated for later experiments. The first experiment with Group II was proceeded by a lO-day acclimation period because high, turbid water prevented observations according to the original schedule. Reassignment of trout to channel sections between experiments reduced the chance that prior social experience and familiarity with the section would influence choice of cover devices. Observations were made 4 times daily at 1000 h, 1200 h, 1400 h, and 1600 h (EDT). A mirror fastened to the end of a 24m pole permitted viewing trout under cover. The mirror and pole did not seem to dis- turb the trout. The observer approached the channels in a crouched position from downstream and facing into the sun when possible to avoid frightening trout with a moving shadow. Tall grasses bordering the channels also concealed the observer. Trout rarely fled when approached, even though they seemed to be aware of the observers pre- sence. Trout that moved were recorded as having the pre-movement position. 18 Experiment 1 The choice was between transparent, overhead glass plates suspended 10 cm ("low“) or 15 cm ("high") above the stream bed. This was tested only with Group I. Experiment 2 The choice was between opaque, overhead glass plates with "dark" or "light" substrate beneath. Glass plates painted black were suspended 10 cm above the stream bed. A glass plate the same size as the overhead plate, either painted black or left transparent, was placed on the stream bed directly under each overhead plate (Figure 3b). The painted plates were positioned with the unpainted side up. Experiment 3 The choice was between lateral glass plates with a chm wide black stripe at the stream bed ("low") or 15 cm above the stream bed ("high") (Figure 3c). The plates were vertical and parallel to stream flow. Use of overhead cover was defined as occupation of a position where any part of the trout's body was under cover. Use of lateral cover was defined two ways: (1) visual use - occupation of a position between a cover device and the wall or between a cover device and the longitudinal centerline of the channel, but not in direct contact with the cover device. Use of positions more than one body length upstream or downstream from a cover device was not considered use; and (2) tactile use - direct contact with any cover device. 19 'Statistics As individual trout could not be reliably identified, obser- vations of the trout in each section were combined for a group total representing (the number of trout per section) x (number of observa- tion periods/day) X (number of days observed). A flood midway through the study resulted in loss of trout and a change in the number of observation days per experiment, so the total number of sightings/ group/experiment was not constant. Predators or vandals also removed some trout. Cover preference was of primary interest in each experiment. The 3 options available to each trout - use of one of the cover types provided, or the other type, or no cover,were not independent, as use of one precluded use of any other. Data was therefore expressed as a derived variable (% of sightings at one cover type minus % sightings at the other cover type). Preference was determined by placing a confidence interval on the mean difference from all of the observation days combined. When the confidence interval did not in- clude zero, the null hypothesis of "no preference" was rejected. Use of derived variables may result in non-normal distribu- tions and great repetition of some values due to the higher probability of occurrence of some percentages than others (Sokal and Rohlf, 1969). Great repetition of percentages did occur in some sets of data. Tests for non-normality and heterogeneous variance were applied prior to testing hypotheses. Non-normality and heterogeneous variance as determined by the Shapiro-Wilk W-test and the F-max test, respectively, were corrected with the arc-sine transformation (Gill, In press). 20 Several days of unusually high and turbid water created cir- cumstances suspected of causing substantial differences in responses to cover between the two groups. Group differences in total cover use and cover preference were tested with a one-way analysis of variance. The effect of mean cross-sectional current velocity on sectional differences in response to cover was tested with a correlation analysis. Degrees of freedom in analysis of variance and correlations were derived from (number of days the experiment was repeated) X (number of channel sections). RESULTS - 1975 Experiment 1 Trout did not prefer either the high or the low transparent overhead covers (p < 0.5; Table 2). Mean cross-sectional velocity at the channel section midpoint was negatively, but weakly correlated with total cover use (r = -0.449, p < 0.05). Experiment 2 Trout in both groups preferred overhead cover with black sub- strate to overhead cover with light substrate (p < 0.01; Table 3). Total cover use was significantly lower for trout in Group II (p < 0.05). Less total cover use by Group II trout may be related to the fact that the water was turbid during the second experiment. Mean sectional current velocity was not correlated with total cover use (p < 0.5). Experiment 3 Trout in both groups did not show visual preference for either the low- or high-stripe lateral cover devices (p < 0.5; Table 4). Tactile use of lateral cover was significantly greater for the low- stripe device (p < 0.001). Total cover use, visual and tactile, both types of devices, was negatively, but weakly correlated with mean sectional current velocity (r = -0.440, p < 0.05). Positions close to the channel walls were used much more than positions permitting trout to touch lateral cover devices (Appendix 21 22 Table 2. - Number of sightings of trout under low (10 cm) or high (15 cm) transparent overhead cover devices, or not under cover. Tested only with Group I. Under low Under high transparent transparent Not under Day, cover cover cover Section Al 1 l 1 l8 2 0 0 20 3 0 0 20 4 1 2 17 Section A2 1 8 l 11 2 5 4 11 3 4 4 12 4 4 4 12 Section Bl l 8 8 6 2 5 4 16 3 4 6 8 4 4 6 13 Section 82 l 4 1 15 2 0 1 19 3 4 0 l6 4 l 1 18 Total 43 43 154 23 Table 3. - Number of sightings of trout under opaque overhead cover devices with dark substrate or light substrate, or not using cover. All other devices were 10 cm above the stream bed. Under cover Under cover with dark with lighta Not under Grogp Day substrate substrate cover Section Al I l 8 11 l 2 ll 9 0 3 11 9 0 II 1 5 3 12 2 8 4 8 3 8 4 8 4 4 3 8 Section A2 I l 6 11 3 2 12 7 1 3 ll 7 2 II 1 7 12 l 2 8 11 l 3 8 9 3 4 10 5 0 Section Bl I 1 11 2 3 2 15 l 0 3 12 3 1 II 1 3 2 15 2 4 5 11 3 10 4 6 4 ll 3 1 Section 82 I 1 7 5 8 2 l3 3 4 3 12 7 1 II 1 5 8 7 2 9 6 5 3 7 9 4 4 6 6 3 Total 242 169 117 aTransparent glass plates permitted light stream bed to show through. 24 Table 4. - Number of sightings of trout responding visually or tactually to lateral cover devices with low (at the stream bed) or high (15 cm above the streambed) dark stripe, or not responding visually or tactually. See text for definitions of "visual" and "tactile" responses. Type of response Visual Tactile Gropp» Day Low-stripe High-stripe Low-stripe High-stripe None Section Al I l 1 1 0 0 18 2 1 4 0 0 15 3 O 4 4 0 12 II 1 0 0 0 0 20 2 2 0 0 0 18 3 l 1 0 0 13 4 4 1 0 0 15 5 l 1 0 0 18 Section A2 I 1 1 6 2 1 10 2 1 4 2 1 12 3 1 3 0 3 13 II 1 0 2 7 0 11 2 1 2 4 1 12 3 2 2 3 l 7 4 5 8 3 0 4 5 3 3 6 0 8 Section Bl I l 0 l 4 3 12 2 1 4 0 0 15 3 0 0 2 l 17 II UianNH NI—‘Ol—‘N HNOOH OOHU‘IO OOOOO ommmm Table 4 . II Total wNH U'Iiwal-J (cont'd.) relate NMWHA 48 l-‘OO #01th 77 25 Section 82 U1 bah OOOi—‘O 57 ON 00000 17 26 Figures 81 and 82). Use of positions on the sloped gravel or at the base of the sloped gravel accounted for 54% of all sightings. Tactile use of low- and high-stripe cover devices accounted for only 13% of all sightings. METHODS - 1976 The experiments were run in two concrete channels, 3.7 m wide and 45 m long. Each channel was divided into two 15.4-m sections separated by a 4-m buffer zone (Figure 5). Substrate in each channel was gravel. Channel walls were vertical 1.2 m high. Gravel sloped against the walls eliminated them as sources of lateral tactile stimuli. Precautions taken in 1975 to create nearly uniform depth, current velocity and channel cross-section within each section, as well as to remove uncontrollable sources of cover were followed in 1976. Mean depth was slightly greater and mean current velocity slightly lower in the downstream sections (Appendix Figures 83-86). Discharge was maintained at about.0.181 cms in channel A and 0.156 cms in channel 8 for the duration of the experiments. A low dam at the head of each channel dispersed flow evenly across the channel cross- section. Screens with one-inch mesh blocked the head and foot of each section. Screens were cleaned daily. A small (25 x 61 cm) and a large (25 x 122 cm) black glass overhead cover was provided in each section. Large cover consisted of two small covers placed end to end (Figure 6). Both sizes of cover were supported 10 cm above the stream bed. One cover was placed 5 m downstream from the upper screen, the other was placed 5 m up- stream from the lower screen. Both covers were placed the same 27 28 803“ :E 5 E 22 . i z: i i i 6 Current Deflector i-——' —-—--' Sennnamnmr _——————a. L______. lawn Sankm . SanMn A1 . 81 land SomenemanM/ Dnr \\ Chanel ‘F" i_i I SmnMn Smfihn A2 , 82 Saeenamner '----’ --—-* Roai 5 «no 5 hAeters Figure 5. - Controlled flow channels used in 1976 with positions of current deflectors, screen barriers, and elevated blinds. 29 Figure 6. - Large overhead cover. 30 distance from the wall. Three cover positions were tested in each section: 0.5 m from the west wall; mid-channel, and 0.5 m from the east wall. Within a section, each position was tested for 3 weeks with a different group of trout. Covers 0.5 m from the walls did not extend over the sloped gravel. Cover positions were not the same in all sections during any single 3-week experiment (Figure 7). Six "feeding references" were placed in each channel section before observing trout Groups II and III in an attempt to provide low velocity feeding stations not immediately adjacent to the channel walls. The references were 5.0 x 7.5-cm black plexiglas plates supported vertically from the substrate perpendicular to the direction of flow (Figure 8a). They were placed at longitudinal transects 1.5, 2.5, and 3.5, and at the latitudinal transects 1 m upstream from the upstream edge of each cover for Group II, and l m downstream from the downstream edge of each cover for Group III (Figure 7). Use of feeding references was defined as feeding within 0.25 m of the reference, for at least 15 sec. Reduced current velocity farther than about 0.25 m upstream or downstream from references could not be detected with a pgymy Gurley current meter (Figure 8b, c). An unconfined rainbow trout approximately 30 cm long in the East Branch of the AuSable River frequently fed just downstream from a branch stub protruding above the stream bed that was less than one-half the size of the plexiglas references provided in the channels. Therefore, I assumed that the references were large enough to provide sufficient velocity buffer to attract even the largest trout used. 31 Figure 7. - Location of overhead covers and feeding references (-) during each 3-week experiment. Experiment I Experiment I I Experiment I I I A1 32 Section A2 .fl. B1 82 Figure 7 33 stream flow > a 1 s )1 i“ ’ : s3 . 63:23::3833828338883 9.9898888328883338! ) b «.5 o 45 -9.0 -93 .94 o 4.5 c -3.o o o -u 4.5 3 a -33 o 3.0 -3.o o 15 21.5 ‘32:: 49 go centimeters Figure 8. - (a) A feeding reference. Only the shaded part was above the stream bed. Top view of current velocity around a feeding re- ference (b) 2.5 cm and (c) 5 cm above the stream bed. Numbers in- dicate the mean difference (cm/s) from background velocity at 3 references. 34 Three groups of 16 wild brown trout were observed. Each group consisted of 8 small (23-28 cm FL) and 8 large (35-41 cm FL) trout (Appendix Table A3). All were captured by electro-fishing in the North Branch of the AuSable River near Lovells, Michigan. Trout were transported to the study site in an aerated tank. Following anaesthetization with tricaine methane sulfonate (MS-222), each trout was weighed, measured, and given a distinctive cold-brand (Everest and Edmondson, 1967) just under the dorsal fin on both sides of the body to allow individual identification. Two large and two small trout were randomly assigned to each channel section. Time from capture to transplantation in the channels did not exceed 6 h. Each group was given 10-14 days of acclimation. This was followed by 5-10 days of observation. At the end of each experiment all trout were weighed and measured, then sacrificed to determine sex and to examine stomach contents. Trout were sacrificed at 0800 h and 2000 h (EDT) to compare day and night food consumption. I assumed that food in the digestive tract anterior to the junction of the pyloric cacae and small intestine indicated consumption during the previous 12 h. Organisms were identified to order and their volume measured by water displacement in a lO-ml graduated cylinder. One 24 h drift net sample was gathered during the acclimation period for each of the 3 groups. The entire cross-section of each channel was sampled simultaneously. Nets were emptied at 6-h inter- vals and organisms were identified to order and measured volumemetri- cally as above. This provided a rough estimate of the relative avail- ability of drift food in each channel. 35 Trout were viewed from elevated blinds. The floor of each blind was 5 m above water level. Each blind was placed so that the channel section on either side could be viewed. The high channel walls and encldSed ladder leading up into the blinds assured approach without detection by the trout (Figure 9). Observations were made at irregular times. Both average length of observation periods and total observation time varied be- tween channel sections and groups. I attempted to observe each section at least once during every daylight hour, but observations were concentrated during the morning hours when the trout were most active. Observations were lumped into 3 time categories: (1) morning - pre-sunrise to 4 h after sunrise, (2) evening - the 4 h preceding sun- set and post-sunset, and (3) midday - all other hours. Low light level prohibited observations more than one-half hour before sunrise or after sunset. Trout activity was classified as: (l) agonistic, (2) cover use, (3) feeding, and (4) miscellaneous. Time to the nearest 16 sec, and movement to the nearest 0.25 m were recorded for each activity for each trout in a section. Numbers painted at 0.5-m intervals on the channel walls and two white cords equally spaced and running the length of each section provided a grid for identifying positions of the trout. At the beginning of an observation session, each visible trout was identified, using 7 x 50 binoculars with polarizing filters. Activity type, position, and time were recorded vocally on tape. 36 Figure 9. - Side view of the downstream observation blind showing the enclosed ladder and high channel walls. A window faces the channel on both sides of the blind. 37 Trout using cover were not visible unless very close to the edge of a cover device. Individuals that remained under cover for the entire observation session were recorded as using cover, but their vposition could not be specified. As trout moved, the time and new position and the new activity, if it changed, were recorded. All tape- recorded information was transcribed onto data sheets for analysis. Distance moved was determined by measuring distance between coor- dinates on scale drawings of the channel sections. Over 60 h of observation time were accumulated (Appendix Table A4). Percent of total observation time in each time period was: morning--45%, midday--31%, evening-~24%. Mean length of observation sessions was 21.3 min. Activity Definitions Cover Use: Time spent with any part of the trout's body directly under cover, provided no feeding activity occurred, was con- sidered cover use. Movement associated with cover use was measured from the last position held in the open not clearly associated with one of the other 3 activity categories. Movement associated with other activities was defined similarly. Feeding: Feeding time was that spent in positions from which drift or benthic prey were consumed at least once during 15 min of observation. Continuous swimming time which was interrupted at least once during 5 min to consume prey was also recorded as feeding time. Occasionally, occupation of a position not under cover for periods longer than 15 min without food consumption was recorded as feeding time if that individual fed from that position during a subsequent 38 observation session. I expected that trout would not occupy the shallow open water unless motivated to feed. Time under cover was considered feeding time beginning with the occurrence of a foray. from cover to intercept drift food or benthos. Feeding time was de- fined as ending 15 min after the last observed foray from cover. Time spent under cover following movement from a feeding position in the open was considered cover use time if no more feeding activity occurred within one minute. Movement from cover to a position where food was consumed was recorded as feeding movement. Agonistic: Time involved in competitive interactions con- stituted agonism. Social rank determinations were based on the out— come of one-on-one agonistic bouts in which one of the interacting individuals was displaced from a feeding or cover position. Dis- placement was assumed to indicate inferior social status. Miscellaneous: Time and movement involved in activities not included in the definitions of cover use, feeding, and agonistic be- havior were lumped into the miscellaneous category. Light Intensity Beneath Cover After completing the outdoor experiments with trout, I measured light intensity beneath opaque covers indoors under a constant light source (Table 5)1. Doubling the width of covers used in this study to 50 cm, without changing cover length, reduced light intensity 3-fold. Doubling cover length to 122 cm, without changing width, did not measurably reduce light intensity beneath the device. Light intensity lWeston Model 856 R8 selenium photovoltic cell and Keithley Instruments Model 150 B microvolt ammeter. 39 Table 5. - Light intensity under overhead cover devices of various sizes and heights above the stream bed. Measurements were made indoors under a constant light source. Light intensity Cover Cover under middle of dimensions (cm) height (cm) cover (ft-c r 6%) 25 Slab 10 0.0916 25 61a 15 0.2916 50 61 10 0.0332 25 122b (5 cm 10 0.0916 (measured gap, see Figure 12) beneath middle of upstream half) C 30 30 30 3.8076 61 61C 30 1.2416 91 91c 30 0.7332 Incident light - 15.0808 aCover sizes and heights used by DeVore (1975). bCover sizes and heights used in this study. cCover sizes and heights used by Butler and Hawthorne (1968). 40 beneath low cover (10 cm above the stream bed) was 1/3 that under high (15 cm) cover. Beneath large square (91 x 91 cm) cover, light in- tensity was 5-fold lower than beneath small square (30 x 30 cm) cover. Statistics Two factors were expected to have major influence on position choice by brown trout: social rank and time of day. A third factor, replicate (a group of trout in a channel section within one of the 3- week experiments) was of secondary interest as a random influence. The effect of each of these factors on trout behavior was evaluated with a 3-factor analysis of variance (ANOVA) adapted for use on sets of data with unequal replication between factor levelsz. Daily records of the percent of total time and movement spent in each activity, and the rates of movement (cm/s) associated with each activity were used to calculate the mean percent of time and movement and the mean rate of movement for each trout. These means were the data points used in the ANOVA. Use of mean values in treat— ment combination cells precluded distinguishing variance due to the 3-way interaction of social rank, time of day, and replicate from random error variance. It was also impossible to estimate the com- ponent of variance contributed by the random factor, for the same reason. Three-way interaction variance was assumed to be zero. De- grees of freedom were derived from (number of social ranks) x (number of times of day) x (number of replicates). Social rank and time of 2SPSS computer program, Michigan State University Computer Center. 41 day were considered fixed factors, replicate was considered a random factor. Calculation of F-values is summarized in Table 6. An assumption of ANOVA is that residual errors are distributed independently (Sokal and Rohlf, 1969). Random assignment of trout to sections was an attempt to minimize non-independence. But randomiza- tion cannot correct for the effects of social factors operating within newly formed groups of trout. Follow-the-leader behavior would tend to reduce the range in behavior between trout in a replicate from what it would have been had each trout been observed in isolation (in— dependently). Conversely, avoidance of dominants would tend to in- flate the range in behavior. Either situation weakens the validity of the usual F-test of significance. Lack of independence is inherent in data from groups of animals. Results of ANOVA must be interpreted with this in mind. All data sets were tested for normality and homogeneous variance prior to hypothesis testing. The Kruskal-Wallis non-parametric one- way ANOVA was applied to sets of data that did not meet these assump— tions. Contrasts between factor level means were done with Tukey's w-procedure where replication was not badly unbalanced and variance was homogeneous, or with Scheffé's procedure where replication was badly unblanced and/or variance was heterogeneous. The Student's t- test was also used to contrast two means where variance was homogeneous (Gill, In press). One and two-factor ANOVA and simple regression were applied as needed. Vandals removed several trout during acclimation periods so there were only 6 replicates with 4 trout during the observation days. 42 Table 6. - Derivation of F-values for 3-factor ANOVA with two fixed (F) and one random (R) factor, and unbalanced replication. MS = mean square. Source of Degrees of a Variation freedom F-value A Social rank (F) 3 MSA/MSAC 8 Time of day (F) 2 MSB/MSBC C Replicate (R) 5 MSC/MSE AB Social rank x Time of day 6 MSAB/MSE AC Social rank x Replicate l3 MSAC/MSE BC Time of day x Replicate 9 MSBC/MSE E Residual errorb l4 aFrom Gill, (In press). bCannot be distinguished from 3-way interaction (ABC). 43 Replicates with fewer than 4 trout were considered separate from these replicates as the social environments were not comparable. Hereafter, replicates will be designated by experiment (Roman numeral) and channel section. RESULTS - 1976 Precent of total cumulative observation time and movement in each activity category was (see Appendix Tables A5 and A6): Activipy % Time % Movement Agonistic ‘ 0.9 39.5 Cover use 54.6 13.0 Feeding 34.8 44.6 Miscellaneous 9.7 2.9 Cover use time was actually greater than indicated here as time spent feeding under cover was put in the feeding category. Correcting for this, percent of time spent in cover use was 63.0%. The importance of miscellaneous activity is inflated, as 83% was con- tributed by only 4 trout. Movement in each activity was not propor- tional to time. Agonism involved much movement. Replicates from which vandals had removed some trout were observed infrequently, as they provided less information on social interactions. Except where stated otherwise, results apply only to replicates with 4 trout. Agonistic Behavior Agonism usually consisted of an agonistic act by the initiator and one by the receiver. Agonistic acts included various types of 44 45 agressive approaches, counter-approaches, direct attacks and sub- missive withdrawals. The sum of agonistic acts associated with one conflict period constituted an agonistic bout. Three general types of one-on-one agonistic bouts were ob- served: approach-withdrawal, approach-counter-approach-withdrawal, and direct attack-withdrawal. Approach-withdrawal: This type of agonistic bout was by far the most common. The initiator swam rapidly, either upstream or down- stream, toward the receiver or drifted slowly downstream tail-first toward the receiver, resulting in withdrawal of the receiver without physical contact. The initiator often took up the position from which the receiver had been displaced, but occasionally returned to the previously occupied position or to a new position. The initiator often chased the receiver several meters from the site of the initial conflict following a head-first approach. Those receivers approached tail-first were chased much less frequently following displacement. Tail-first approaches were made exclusively by large trout. Regard- less of the type of approach, subordinate receivers usually began flight while the initiator was still 1-2 m away. Rarely, an initiator apparently terminated an approach before the would-be receiver responded. In those cases the initiator was always subordinate to the receiver. Approach-counter approach-withdrawal: This type of bout occurred only 3 times. Following an approach by the initiator, the receiver either withdrew a short distance and then returned to displace the initiator, or did not withdraw at all and began chasing the initiator 46 while still one or more meters away. All counter approaches occurred following high-speed entry of cover by the initiator. Whether the intent of the entering individual was agonistic or cover-seeking in nature was questionable but the response of the cover occupant was clearly agonistic. These were the only occasions observed when rank-1 and rank-2 trout were displaced by subordinates. It seems likely that the initial exit from cover was in response to the sudden appear- ance of a disrupting or frightening stimulus in confined quarters rather than avoidance of a specific individual. Direct attack-withdrawal: Direct attacks involved nips and head- first bumps to the body and fins of the receiver. This was observed only in replicate IIIBZ. The receiver was an injured trout that was larger than the initiator. These bouts often began as the approach- withdrawal type but became more violent when the receiver withdrew to a corner of the channel section. Twice the receiver pressed sideways against the lower screen in a head-down position, exhausted, follow— ing repeated attacks. On several occasions one or more trout followed or circled another, or swam slowly side by side. At least one of the large trout was always involved. These activities were also included in the agonistic category even though there was no apparent initiator or receiver, and displacement was not apparent. Lateral threat displays, wigwag displays and other agonistic acts described for salmonids by several authors (Newman, 1956; Kalleberg, 1958; Jenkins, 1969) were not observed. The blinds were 47 too far from the trout to permit viewing the more subtle agonistic acts that may have occurred. A total of 187 bouts were observed (Appendix Tables A7, A8 and A9). One-hundred and sixty-five of these occurred in the 6 replicates with 4 trout. In these replicates, rank-3 trout were in- volved in the greatest number of bouts (96) and rank-4 in the fewest (70). Ranks 1 and 2 were both involved in 82 bouts. More conflicts occurred between trout differing by only one social rank than between trout differing by two or three ranks (Table 7). Aggressiveness was greatest during feeding periods. More bouts resulted in displacement from feeding positions than in displacement from cover (Table 8). This occurred even though less time was spent feeding (32.36%) than using cover (65.74%). Table 7. - Number of one-on-one agonistic bouts between pairs of socials ranks in replicates with 4 trout. Ranks of interacting trout 1+2 1+3 1+4 2+3 2+4 3+4 Total Number of bouts 34 29 19 32 16 35 165 Percent of total 20.61 17.58 11.52 19.39 9.70 21.21 100.01 The social hierarchy was very stable in all replicates. Dominants lost only 2.14% of all bouts. Body weight was the primary determinant of social rank (Appendix Table A3). Rank 2 in replicate IIAl weighed 15% more than rank 1 but clearly lost the single bout observed. No injury was apparent, but failure to feed and use cover in addition to the agonistic loss suggests that behavior had been 48 Table 8. - The number of agonistic bouts in which trout were dis- placed from cover or feeding stations in replicates with 4 trout. Trout displaced from Feedinga Replicate Cover station Total 1A2 5 5 10 IIAl 8 6 l4 IIA2 13 18 31 1181 4 34 38 IIBZ 15 7 22 IIIAl 29 21 50 Total 74 91 165 aIncludes feeding from cover. 49 affected by capture and handling. This trout was excluded from statistical analyses. Rank 2 in replicate IIIBZ was badly bruised by vandals. It weighed 28% more than rank 1 but lost all 11 of its bouts. It, too, did not feed and was excluded from statistical analyses. The effect of sex on social rank could not be determined due to the masking affect of body weight. Mean percent of observation session time devoted to agonism increased from rank 1 to rank 3, but rank 4 spent less time in agonism than ranks 2 and 3 (Appendix Table A10). Differences be- tween ranks were nearly significant during the morning (p = 0.1) but not during midday or evening sessions. A significantly greater por- tion of time was spent in agonism during the morning than during midday or evening (p < 0.01). Replicates did not differ signifi- cantly (p < 0.1). Mean percent of observation session movement devoted to agonism was nearly the same during each time of day (Appendix Table All). Social ranks also did not differ significantly (p < 0.5). But during midday and evening, rank 1 devoted a significantly greater portion of total movement to agonism than ranks 2, 3, and 4 considered together (p < 0.001). During these two time periods, most movement of rank 1 was associated with agonism. Replicates did not differ significantly (p < 0.5). Mean rate of movement (cm/s) in agonism was greatest for small trout (Appendix Table A12) but differences between ranks were not significant (p < 0.5). Times of day (p < 0.5) and replicates (p < 0.5) did not differ significantly. 50 The presence of dominants seemed to inhibit agonism.among low- ranking trout. The intensity and duration of agonistic bouts was greater among small trout in replicates lacking large trout than in replicates where two large trout were present. This is reflected in time and movement (Table 9). Table 9. - Mean percent of observation session time and movement, and mean rate of movement in agonism by small trout in replicates with two large trout and with no large trout. Rate of Number of movement replicates % Time % Movement (cm/s) Two large trout 6 0.60 20.61 14.88 No large trout 2 3.26 58.93 44.27 Cover Use The highest ranking trout using cover always occupied the most upstream position. Usually the tip of the snout was just down- stream from the leading edge of the cover. Subordinates moved to the upstream position following exit of higher ranking trout. As dominants entered cover, subordinates moved to the lower end of the cover or moved to the other cover. If a dominant entered cover occupied by two or more subordinates, the lowest ranking trout was most likely to move to the other cover. Occasionally trout displaced from cover remained in the open and began feeding rather than moving to the other cover. Most agonistic bouts associated with cover use occurred as trout entered an occupied cover. If a bout did not occur immediately, the occupants usually coexisted for the duration of the session. 51 Trout were much less aggressive when using cover than when feeding. Chases originating from cover were shorter than those associated with displacement from feeding positions. All ranks used both covers in their section and seemed to displace subordinates just as readily from their preferred cover as from their non-pre- ferred cover. Trout rarely moved from cover to chase approaching subordinates, even when subordinates approached from upstream. This contrasted sharply with behavior at feeding stations where trout that approached dominants closely from the side or upstream were usually chased. Mean percent of observation session time spent under cover was highest for rank-l trout (Appendix Table A13). Rank 1 used cover significantly more than ranks 3 (p < 0.01) and 4 (p < 0.01), but did not differ from rank 2 (p < 0.4). Cover use was greatest for all ranks during midday. Mean midday cover use was significantly greater (p < 0.01) than morning use but was not significantly different than evening use. Ranks 1 and 2 usually remained under cover well into the evening, but ranks 3 and 4 often left cover to feed once the channel was mostly shaded. Rank 2 in replicate IIAl and rank 4 in replicate IIIAl were excluded from statistical analyses. Both spent most of their time in a small area around coordinates 1.25, 12 (Appendix Figure 83) where substrate irregularities reduced current velocity and produced sur- face turbulence that may have been responded to as overhead cover. Occupation of that position could have been related to motivation to feed or motivation to use cover. 52 Cover use was significantly greater in replicate IIBZ than in replicates IA2 (p < 0.05) and IIA2 (p < 0.01). Use in replicate IIAl was significantly greater than in IIA2 (p < 0.05). Mean replicate cover use time was not significantly correlated with either mean sectional current velocity (p < 0.5) or depth (p < 0.5). Mean percent of observation session movement (Appendix Table A14) associated with cover use was significantly lower during the evening than during morning (p < 0.05) and midday (p < 0.05). Dif— ferences between social ranks and replicates were not significant (p < 0.5). Differences in rate of movement (cm/s) (Appendix Table A15) also were not significant between social ranks (p < 0.5), times of day (p < 0.5) and replicates (p < 0.5). Preferred Cover Total time spent by trout at each cover in its section re- flected responses to other trout as well as responses to the physical characteristics of the cover. I assumed that subordinates would occasionally not use their preferred cover due to avoidance of dominants under cover. Only time spent at each cover when dominants did not occupy either cover, i.e., when "free choice" was possible, was used to determine individual preference. This included time spent feeding under cover, as there seemed to be no reason for the cover preferred for feeding to be different than the cover preferred when not feeding. Rank 1 always had free choice by definition. Free choice occurred less frequently proceeding from ranks 2 to 4, so that preference could not be determined for most of these trout. 53 Preferred cover was defined as the cover at which a trout spent at least 66% of its total "free-choice" cover use time. A pre- ference judgement was based on the results of at least 3 observation' sessions, and at least 20 min of total observation time. Preference was determined for 8 rank-l trout, 3 rank-2 trout, and one rank-4 trout (Table 10). Rank-2 and rank-4 trout preferred the same cover as rank 1 in their section. Hereafter, "preferred" will refer to the cover preferred by rank 1. Cover size did not strongly influence preference. Five of 8 rank-l trout preferred small cover (Table 10). Relative current velocity seemed to be the most important stimulus involved in cover choice. Depth had no apparent influence on cover choice. Five rank—1 trout preferred the cover with the slowest current velocity beneath. Only two preferred the cover with the highest current velocity. One of these was a large cover and one was a small cover. In the remain- ing replicate where current velocity was the same under both covers, the large one was preferred. Size of cover clearly influenced patterns of association be— tween trout. Coexistence between rank 1 and one subordinate was most frequent under the preferred cover regardless of its size except in replicate IIAl (Table 11). But in replicates where large cover was preferred, its use by rank 1 and one subordinate was proportionally greater than use of small cover by rank 1 and one subordinate in replicates where small cover was preferred. Ninety-one percent (n a 43 observations) of all multiple occupancies occurred under large cover in replicates where large cover was preferred. In replicates 54» o o o o o o .o.mm «v o.m~ o.mH .m o o o o o o o.v~ a m~.om o.m~ A Noam m~.o~ A 0.0 a «m.oo m co.~m um o.m~ v.m~ «m o o m.m m mh.n~ m mh.ba v o.m~ ~.m~ A amHH co.mm m o.m a o.v a com.oo~ and 0.0m o.m~ .4 o o m.m a o o mn.o N mh.m~ o.v~ w ~4HH o o o o o o am.oom «he o.o~ m.m~ m o o o o o o 6.3 m 93 6.8 .3 3: HH macho o o m.o H mn.m~ m.m~ A o o o o m~.om ~.m~ m «mm o o o.o m o.o~ o.m~ A o o «mh.mn av mm.- v.n~ cm ~mu o o o.« A m.n A o o m.- v.o~ m c o o o am.av N cm.~v an m~.o~ o.n~ «A New 0 o o o 0.9 a m.n~ o.vm a o o o o o.vm m m.h~ m.ma m dam I C u H macaw Aces. .mno Aces. .moo Assn. .ono Acqe. .nno .30. Am\mov coda concealer use? .0: made .0: ease .0: made .0: Eamon made iwmon v xcom m sea: N scam A zoom 0io~o> use chum muco>o e:oo«0:o:eoum= meaunduouocuego uo>ou n .uxou on» :4 oonuuoooo no oocoueuoua nauseaosu xceu deacon some >3 uo>oo woos: ucomo occaouom C seabe>uomno no Mensa: one oauu soouozoioeuhs .oueOAAQou some :« nuo>oo Hanan use pound on cameo use auwuo~o> oceanso i .od ounce 55 .ndeoce> mo oueowmmou easy noun oo>osou one: uoouu Hado .oon seesaw one o>one 36 m .uo>ou no outed! ozu woos: consumes >ua00ao>o Eeouuocsoo .Auean i m leeuumczoo .emuea i A aeouumms .omuMH i A Buchanan .HaoEu i we .uo>oo sesame on: go: out nucecqaoo cos: we: :ooqozo ooumxe o.~ H m~.om v o.m~ m.o~ a o o o.on m m~.o~ o.m~ m NmHHH campus c.4n ~ o.o~ m.m~ m o.mm H m.s~ m.- o Nessa o.” a .m.m~ ~ .m~.m- . .m o.m~ m.- .4 m~.~ H o.m a m~.m- o m>.m~ m.- m Hausa HHH ozone A.O.u:oov On munch 56 c.5ma NH mh.~m m mN.vva 9N o.Nm o o.o- VA o.mm o mN.mHm ON m.om h mN.oN~ ma mN.VmN MN o.AN~ ca mn.mm OH #6009 o c o o o c o o o o o o o o o o o o m.m H o o o o v+m+N+H o o o o c c o o o o o o o o o o o o o o o o m.~ A v+m+N o o o o c o o o c o o o o o mN.« a o o o o o o o c v+m+~ o o o c o.o~ a o o o o o o c o o o o o mN.o~ N o o o o v+Ni~ c.« A o o mh.s N o o o o o o 0.9 A o o o a mN.m~ a c o o o n+N+H o o o o o o o o o.m A o c o o m.NN a mN.v A o o mh.0m v m.m a v+m o o o.h A o o o o o o o o o o o o o o o o o o o o V+N o o o o m.m N o o o o mh.vN N n.0n N o o m.n a mN.o H o o o.mN a m+N o o o o o o o o mN.NN n o o c o o o c o mN.mn v mh.n A o.o~ a v+~ o o o o m.~ A o c o o o o o.nv~ m mN.N N o o o o m.oN v o o m+d mh.nN H m.mN n m.nN o o c c o o o mb.o N o o m.m~ N mN.om m o o o.v a N+~ o o o o a o mh.vv m mN.o~ n o.o~ H o o o.H a m.m~ N MN m o o o.n a v no.0N v o O c.m~ v o a m.n H mh.wu N o.nv N c.n~ N m.mo o mn.m~ N m.No o m.n a m o.mv m o o m.om o O.H H o o m.h A mh.mv N c.h a mh.hu v o o o o vn.m a N m.vo n mN.m~ A m.N~ N mN.o N 0.00 o o o mN.mv o m.v H mh.v N nh.mN v m.n H m.NN N A :«a .0: :«E .0: :NE .0: :NB .0: :NE .0: :Na .0: :«a .0: :«E .0: ca: .0: :«E .0: :«E .0: :ME .0: when: HtHNH NmHH HmHH Ntnn H‘HH Nflh ~OO e e a e e e no «.53.— 0uoONHm0m oueomwmom . u0>oo macaw u0>ou pawog .xoquouoe :e an 600:0«0:« muo>00 oouuououm .:30:x one: usouu v age we occuuuooa :05: occunmee mcwuso cacae nozoo sud: oceauecaoaoo odowmmom an: :« 0:: ocean xceu neaoou son: an u0>oo 000:: ocean 0:0«nmoe ceaue>uouo0 no sense: 0:: .:«5. mafia I .HH manna 57 where small cover was preferred, only 56% (n = 36 observations) of all multiple occupancies occurred under small cover. The extra space under large cover permitted more trout to coexist. Failure to coexist with a dominant under cover could be the result of: (l) displacement immediately following an agonistic bout, (2) learned avoidance following several agonistic bouts in the more distant past, and (3) preference for cover not used by the dominant.3 An evaluation of the effect of (2) on cover use was attempted by determining the extent to which trout took advantage of opportunities to avoid dominants under cover. Time spent under the non-preferred cover was expressed as a percentage of the total potential time avail- able to use the non-preferred cover when only the preferred cover was occupied by dominants (Table 12). Time spent under the preferred cover when dominants occupied only the non-preferred cover was ignored as occupation of the preferred cover by a subordinate may have re- flected preference for that cover rather than avoidance of a dominant. Data were too limited to permit evaluation of relations between specific pairs of ranks. The data in Table 12 therefore indicate avoidance of all dominants simultaneously. Avoidance of dominants under cover increased from rank 2 to rank 4. Variability between replicates was so great, however, that differences between ranks were not significant overall (p < 0.2). Much of this variability was related to the size of the preferred 3 All trout were assumed to prefer the same cover but even rank-l trout occasionally used the "non-preferred" cover. 58 Table 12. - Time available to use the "non-preferred"a cover when dominants were using only the "preferred" cover. Percent utiliza- tion indicates the extent to which ranks 2, 3, and 4 took ad- vantage of opportunities to avoid dominants under cover. This only includes sessions when positions of all 4 trout were known. Rank Size of ._ 2 3 4 pre- % % % ferred Time utiliza- Time utiliza- Time utiliza- Replicate cover (min) tion (min) tion (min) tion IA2 large 28.75 83 40.5 35 15.0 87 IIAl small b b 108.0 94 16.5 100 IIA2 large 69.0 0 30.0 0 96.5 46 1181 small 68.75 7 70.75 85 15.5 100 11132 small 31.5 21 34.5 100 2.0 0° IIIAl large 72.0 51 171.75 24 b Mean 32.4 56.3 83.2 a I "Non-preferred" cover is the cover used least by rank-l trout. These trout did not use cover. c Non included in the mean. 59 cover in the section. Ranks 3 and 4,considered together,avoided dominants significantly less (p < 0.01) in replicates where large cover was preferred than in replicates where small cover was preferred. However, cover size did not influence the degree of avoidance of rank 1 by rank 2: these two ranks coexisted as well under small cover as under large cover. A separate analysis also indicated that relative social rank influenced trout association under cover. Observed frequencies of associations under cover by specific combinations of social ranks were compared with the frequencies that would be expected if associations were solely by chance. Use of large cover was considered separately from use of small cover as large and small covers were not equally attractive, i.e., the probability of using large cover differed from the probability of using small cover (Appendix Table A16). Ranks 1 and 2 used cover more frequently than subordinates and were therefore expected to coexist under cover by change more frequently than com- binations including ranks 3 and 4. This was taken into account by calculating the probability of using cover separately for each social rank and then using these values to derive expected frequencies. Ex- pected frequencies are derived in Table 13. They provide a measure of relative expectations of coexistence under cover. Time spent under cover and frequency of cover use were highly correlated (r = 0.913, p < 0.05). Observed use of large cover by combinations of two or more social ranks was almost significantly different than expected (p < 0.1). Observed use of small cover was not significantly different than Table 13. - Observed frequencies and derivation of expected fre- quencies of using small and large covers for combinations of social ranks. Rank a of . Probability of b Chi Signi- . . c . cover ‘ Obs. uSlng cover Adj. Exp. square flcance users freq. together pxob. freq. value level Large cover 1+2 10 0.26 0.104 4.784 5.687 1+3 10 0.351 0.14 6.44 1.967 1+4 6 0.3 0.12 5.52 0.042 2+3 5 0.286 0.114 5.244 0.011 2+4 0 0.244 0.098 4.508 4.507 p < 0.1 3+4 8 0.33 0.132 6.072 0.613 1+2+3 2 0.162 0.065 2.99 0.328 1+2+4 2 0.138 0.055 2.53 0.111 l+3+4 1 0.187 0.075 3.45 1.739 2+3+4 1 0.152 0.061 2.806 1.162 l+2+3+4 1 0.086 0.034 1.564 0.199 Small cover 1+2 10 0.184 0.207 4.968 5.097 1+3 1 0.104 0.117 2.808 1.164 1+4 3 0.102 0.115 2.76 0.022 2+3 4 0.128 0.144 3.456 0.087 2+4 1 0.126 0.142 3.408 1.702 p < 0.3 3+4 1 0.071 0.078 1.872 0.406 l+2+3 3 0.050 0.056 1.344 2.039 l+2+4 1 0.049 0.055 1.32 0.079 l+3+4 0 0.028 0.031 0.744 0.739 2+3+4 0 0.034 0.038 0.912 0.91 l+2+3+4 0 0.013 0.015 0.36 0.361 aProbability of using cover together is the product of the proba- bilities that each social rank used cover alone (Appendix Table A16). bSum of adjusted probabilities within a cover size category equals 1. cExpected frequency is the product of the adjusted probability and the sum of observed frequencies (large cover = 46; small cover 23). 61 expected (p < 0.3). However, it should be noted that a large portion of the total chi-square value was contributed by the rank-l-and-Z combination (Table 13). Rank 2 used cover with rank 1 more than ex- pected regardless of cover size. Rank 4 used cover with rank 2 less than expected regardless of cover size. All other combinations were relatively as frequent as expected. To summarize this section, cover use patterns were influenced by social rank, cover size, and probably by current velocity under cover. Individual choice of cover was determined at least as much by current velocity beneath as by cover size. Low current velocity seemed preferable to large size. Attractiveness of cover in regard to current velocity and other unmeasured physical features in the vicinity of cover was not sufficient to fully compensate for space limitations imposed by small cover size. Large cover received at least as much total use as small cover even in replicates where small cover was preferred. Much of the heavy use of large cover in replicates where small cover was preferred was due to avoidance of rank 1 by ranks 3 and 4 and by avoidance of rank 2 by rank 4. Trout differing by only one rank co-existed nearly as well under small cover as under large cover. Cover Use and Feeding There was a tendency for some trout to feed on the side of the channel with cover (Table 14). But, ANOVA on the difference be- tween the percent of time spent feeding on the west and the east sides of the channel showed that differences between replicates with dif- ferent cover locations were not significant (p < 0.5). Time spent 62 0 0 0.000 00000>0 0.00 0 0.00 000002 0.00 0 0.00 0 0.0 0.00 0.00 0 0.0 0000002 0 0 0 0.000 00000>0 0.000 0 0 000002 0.000 0 0 0 0 0.000 0.000 0 0 0000002 0 Na :00uo0m .HH macaw 0m :owuo0m .HH @5000 He :00uo0m .HHH @5000 0 0.00 0.00 0 0 0.000 0 0 0.00 00000>m 0.000 0 0 0 0 0.000 000002 0.0 0.00 0.00 0.00 0.0 0.00 0.0 0 0.00 0000002 0 0.00 0 0.00 0.00 0.0 0.0 0.00 0.0 0.00 00000>0 0.00 0.0 0 >00002 0.00 0.0 0.0 0.000 0 0 0000002 0 0 0 0.000 0 0 0.000 00000>0 0.000 0 0 0.00 0 0.00 000002 0.00 0 0.0 0 0 0.000 0000002 N 0.0 0 0.00 00000>0 000002 0.00 0 0.00 0.000 0 0 0000002 0 0d :O0uo0m .HH @9000 N: :00uo0m .HH @0000 N4 :00000m .H @0000 ummm MTHUQAZ UmQS umwm mwfififiwz .Hmwz Hmmm MGHOUNZ umw3 >66 x :0000000 mc00000 :0000000 m:0o00m :0000000 m:0000m 0Ewo M 0000 000002 0003 .0 m COHHMUOH H0>CU .000000:« 00: 00 H0>oo Eoum mc0o00m uc0mm 0509 00::020 0:0 no 00000 0000 0:0 #003 0:0 :0 0:0mm 0E0» m:0o00m 0000» Mo 0:0000m .:3o:m 00::ono 0:0 «0 000008 0:» :0 0:0 .usouu v :003 m0uoo0am00 :0 mu0>oo 000500>o no :0000000 0:: xcou 000000 5000 >0 000: 0:000:00 0:0000m no :0000000 i .va 00009 63 .m.N 00000m:00 :0 003 :000000 0:00000 0:0 00 00:0 00::000 0:0 00 000006 000 :0 0:00000 0000000:00 0003 050080 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.0 0.00 0002 0.000 0 0 0.00 0 0.00 00000>m 0.00 0.00 0.0 000002 0.00 0.0 0 0.00 0.00 0.00 0000002 0 0.0 0 0.00 0.00 0.00 0 0.00 0 0.00 00000>0 0.00 0.00 0 0 0.000 0 000002 0 0.0 0.00 0.00 0.00 0 0.0 0.00 0.00 0000002 0 0.0.0coov 00 00005 64 feeding under cover was not included in the calculations, as this would have biased results in favor of the side of the channel with cover. Large trout did not feed closer to cover than small trout (Figure 10; Appendix Table A17). Distance between the feeding station and the nearest cover, weighted by the percent of feeding time spent at the station did not differ significantly between ranks, (p < 0.1) or times of day (p < 0.5). Mean percent of feeding time spent under cover was signifi— cantly less (p < 0.05) for rank 4 than for ranks l, 2, and 3 considered together (Appendix Table A18). Rank 4 may have been particularly "afraid" to feed under cover used by dominants. But, low-ranking trout commonly fed from the lower end of cover when dominants occupied the upstream end of cover without agonistic conflict. Mean percent of feeding time spent under cover was highest during midday for all ranks, but extremely high individual variability nullified statistical dif- ferences between times of day (p < 0.5). Differences between replicates were nearly significant (p < 0.1). Feeding Types of Feeding Behavior Three types of feeding behavior were observed. By far the most common type was drift feeding from a more-or-less fixed station. Short vertical and lateral forays to intercept surface and sub-surface drift food organisms were followed immediately by return to the station. Movements greater than 1 m from the station rarely occurred except to capture unusually large items that were artificially 65 201 :1 1.5‘ b 1_, 0 q > O 0 z 1} ‘~> 2 a m C C E 1.0+ 3 0 I 0 I”, C ‘2 . .2 ‘D c 1 D O a 1 0.5‘ J 0 1 2 3 4 Social rank Figure l0. - Mean distance to the nearest cover from feeding stations used by each social rank in replicates with 4 trout. Distance weighted by the percent of feeding time spent at each station. Cover feeding stations are not included. 66 introduced. Constant swimming maintained the trouts' body in a stationary position above the stream bed. Less often, trout remained on the stream bed between forays. This occurred most.among large trout and when food was not abundant. Drift feeding while cruising about the channel was the second most common type of feeding behavior. It consisted of upstream swimming interrupted at irregular intervals by short vertical or lateral movements to consume drifting food. This occurred most often as trout moved between feeding stations during times when drift food was very abundant. When drift food was scarce, movement between feeding stations occurred without feeding. Feeding on benthos was the least observed type of feeding be- havior. Trout feeding on benthos cruised slowly about the channel showing little affinity to particular areas. The stream bed was bitten repeatedly, followed by conspicuous jaw movements and gill flaring to void detritus and fine sediment. Much of the stomach con- tents of these trout consisted of fine gravel and organic matter. The assumption that trout would use positions not under cover only when motivated to feed was supported by almost all observations. The only exceptions were rank 2 in replicate IIAl and rank 4 in replicate IIIAl which exhibited abnormal behavior by rarely or never using cover and rank 2 in replicate 11132 which was injured. Drift Food Supply The volume of drift food entering both channels during 6-h drift net sampling periods was twice as great at night as during the day (Figure 11; Appendix Tables A19 and A20). Peak volume of drift 67 occurred during the first half of the night (2000-0200 h). Daytime volumes were similar for both 6-h periods (0800-1400 h and 1400-2000 h). Drift volume was over twice as great in channel A as in channel B on all 3 sampling days. Mean volume per food item in channel A was slightly greater than in channel B (Appendix Tables A19 and A20). Drift volume was about the same on the first two sampling days, but lower on the last day in both channels (Figure 11). Caddisfly (Tricoptera) larvae and pupae (especially Brachycen- Eggs) and aquatic isopods (Asellus) were the most common drift food organisms comprising 34.9% and 30.6%, respectively, of total drift net volume summed over the 3 sampling dates in both channels (Appendix Table A19). Fish (9.6%) and mayfly (Ephemeroptera) nymphs (7.0%) were the third and fourth largest groups represented. Fish were rare numerically, however (0.4%). The tiny mayfly Tricorythodes was extremely abundant on the last two sampling dates but drift net mesh size was too large (1.5mm) to retain most of them. The availability of this potential food source, as well as any others of similar or smaller size was underestimated. Composition of the drift did not differ much between channels or sampling dates, with the notable exception of mayfly nymphs which increased in volume from 2% to 5%, to 19% of the total on the first, second, and third sampling dates, respectively. Diel composition was quite constant overall, though the absolute and relative contribution of ants (Hymenoptera) was greatest during mid-day. 68 150- - m - 8/7 7/15 - 7/16 6/28' 6/29 Estimated total volume (cc) 0! dtitt load (2000 - 0200) (0200 - 0800) (0800 -1400) (1400 - 2000) Hours sampled Figure 11. - Estimated total volume of drift food organisms that entered channels A and B on each sampling day. volumes were estimated from subsamples equal to 5% of the total volume of organisms collected. 69 Drift Food Utilization Trout consumed drift food roughly in proportion to its diel availability. Mean volume of food in the stomachs of large and small trout examined in the morning (0800 h) was over twice that in stomachs examined in the evening (2000 h) (Figure 12; Appendix Table A21). Morning and evening means were not significantly different within the large size group (p < 0.2), but were nearly significant within the small size group (p < 0.1). Mean volume of food in stomachs of small trout was over twice that in stomachs of large trout. Differences were significant in both the morning (p < 0.05) and the evening (p < 0.02). Data were not sufficient to permit a statistical comparison of food consumption in channel A with consumption in channel B. There were no apparent differences despite the large measured difference in drift food availability. Perhaps much of the food consumed originated in the channels, rather than drifting in from upstream. Isopods and adult mayflies composed a greater portion of stomach content volumes than of drift net volumes (Table 15). Caddis- fly larvae and pupae were not as well represented in stomachs as in drift nets. As mentioned above, large mesh size was probably res- ponsible for the low capture rate of mayfly adults. 70 Large trout Smell trout (181 l139 9 oz Mean volume (colol load In trout etomeche 0.4‘ , "n11 r_._1. 0.2'‘ ":6 OJ (0800 h) (2000 h) Time trout removed from channel Figure 12. - Mean volume (cc) of food in the stomachs of large and small trout removed from the channels at 0800 h and 2000 h (EDT). n = number of stomachs checked. 71 Table 15. - Mean percent of the volume (cc) of trout stomach contents and drift net contents as isopods, mayflies, and caddisflies. Mis— cellaneous items were not included in the total volume. Isopods Mayflies Caddisflies (All instars) (Adults)(Nymphs) (Larvae and pupae) Stomachs 56.8 9.3 0.9 6.7 Drift nets 31.0 2.3 7.2 36.1 Small trout as a group relied more heavily on isopods than large trout. Discounting miscellaneous items, isopods were the largest constituent by volume of stomach contents in 75% of the small trout, but is only 36% of the large trout (Appendix Table A21). Caddisfly larvae and pupae, mayfly adults, adult beetles, (Coleoptera), or snail and clams (Mollusca) constituted the major portion of the stomach con- tents in other trout. The mean volume of individual food items was similar for large (i = 0.0064 cc) and small trout (x 8 0.0090 cc). Means do not include items in the miscellaneous category for which numbers could not be reliably determined. Mean volume of individual food items was some- what lower in Group-II trout (i = 0.0088 cc) than in either Group-I (i 8 0.0190 cc) or -III, (E = 0.0111 cc) trout due largely to the pre- dominance of Tricerythodes mayflies on the examination date for Group II. Mean volume of individual food items consumed at night (i = 0.0171 cc) was nearly twice that of daytime food (x a 0.0091 cc). The assumption that food in the digestive tract anterior to the junction of the pyloric cacae and small intestine was a reliable indicator of consumption during the previous 12 h was supported by 72 examination of stomachs from individuals that were known to have fed near the beginning of the 12-h period. In trout whose stomach con- tents were checked at 2000 h Tricorythodes mayflies that were consumed between 0800 h and 1000 h were concentrated in the lower stomach and pyloric cacae: few had moved into the intestine. All major food items were of similar size and digestability so that the importance of specific types should not have been biased by this method of analysis. Additional support for this assumption comes from calculations of the rate of gastic evacuation for brown trout based on data pre- sented by Elliot (1972) for 20-30-cm (TL) brown trout that were fed a variety of aquatic organisms similar in size to most of the foods available in this study. Mean water temperature during the 12 h pre- ceding the time that stomachs were examined gave estimated times for complete gastic evacuation ranging from 10 h during the daylight period for Group I, to 15 h during the same period for Group III. The overall mean was 13.2 h. Elliot found that rates of gastric evacua- tion were not significantly different for different sized food organisms of the same taxon, for different sized meals, for mixed or multiple meals, or for different sized trout within the size range tested. His equations are therefore readily applicable in this study. Food supply in the channels was not adequate to meet the energy demands of most trout, especially the large ones. Mean weight loss over 3-week experiment periods was 8.5% for large trout and 3.7% for small trout (Appendix Table A3). Mean weight losses among large trout were 11.5%, 7.7%, and 6.5% for Groups I, II, and III, respectively. Corresponding losses among small trout were 6.0%, 1.7%, and 3.2%. 73 Mean group weight losses did not parallel trends in drift net food volume. Among the 3 groups, weight loss among Group-II trout was intermediate despite the occurrence of the highest drift food availability during that period. Of course, this assumes that drift samples accurately indicated drift food availability throughout the experiment. Volume of benthic food organisms in the channels was not determined but presumably it increased, over the course of the experiments. Stream bed disturbances during initial channel prepara— tion undoubtedly destroyed some of the initial standing crop which was probably replaced through immigration (Waters, 1972). This may ex- plain why weight losses were conspicuously greater among Group-I trout than in those following. Of the more important drift food items only Tricorythodes nymphs should have decreased substantially in abundance over the study period as ecdysis to the adult stage occurred daily throughout the last two experiments. Emergences of various caddisflies and Baetis mayflies occurred late in experiment III, but these were too late and too minor to have significantly reduced the availability of benthos or drift food in the channels. Feeding Time and Movement Mean percent of observation session time spent feeding in- creased from rank-1 to rank-4 trout (Appendix Table A22). Rank 1 fed significantly less than ranks 3 (p < 0.01) and 4 (p < 0.01) but did not differ significantly from rank 2 (p < 0.2). The greatest dif- ferences were during midday and evening when small trout commonly fed for short periods but large trout usually stayed under cover and did not feed. 74 Most feeding occurred during the morning at the peak of the Tricorythodes mayfly emergence and imago fall. Mean percent of time spent feeding was significantly greater during the morning than during midday (p < 0.01) and evening (p < 0.05). The midday mean was nearly significantly less (p<=0.l) than the evening mean. Small trout (ranks 3 and 4) fed about as much during the evening, once the channels were fully shaded, as during the morning. But large trout usually did not begin feeding until 10-20 min after sunset when light level became too low to permit individual identification. On several evenings after sunset, small and large trout left cover to feed within a 5-min period. There was not a sudden change in light level that might have caused such a sudden change in behavior. The opposite behavior, i.e., sudden termination of feeding as light level increased before sunrise, was not observed. But pre-sunrise observations were not made as often as post-sunset observations so that similar behavior may have occurred. Drift food abundance was not monitored frequently enough to permit comparisons of food availability at comparable light levels in the morning and evening but visual inspec- tion of the water surface with binoculars indicated no sudden change in food availability after sunset or before sunrise. These observations suggest that there is a light threshold in- tensity below which cover is not required. However, that the abundant Tricorythodes mayflies released feeding behavior until late morning, even on sunny days, indicates that the typical negative response to high light intensity shown by adult brown trout is modified when food is abundant. The daily pattern of feeding activity is summarized in Table 16. 75 Table 16. - Summary of daily feeding activity by brown trout in this study. Description of Time period feeding activity Dawn to beginning of Sporadic feeding by all trout. Tricorythodes emergence. During Tricorythodes Peak feeding by all trout. emergence and imago fall. End of Tricorythodes Sporadic feeding, mainly by small activity to sunset. trout. After sunset to dusk. All trout feeding. There were large differences between some replicates in per- cent of time spent feeding (Appendix Table A22). Significantly more feeding occurred in replicate IIIA2 than in replicate IIBZ (p < 0.05). Replicates IA2 and IIBZ also differed significantly (p < 0.05), and the mean difference between replicates 1181 and IIBZ was nearly significant (p < 0.1). The source of differences between replicates was not differences in drift food availability. As shown above, drift food was twice as abundant in channel A as in channel B in experiment II. But mean percent of time spent feeding in replicate IIAl was less than in replicate IIBl. In fact, all social ranks in every time period in replicate IIBl spent proportionately more time feeding than their social counterparts in replicate IIAl. All ranks in replicate IIBZ fed less, on the average, than counterparts in other replicates. Each trout in a replicate was a fairly reliable indicator of the be- havior expressed by other trout in its replicate. If one trout fed more than average for its social rank, then the same was often true of the other trout in its replicate. 76 Differences in feeding activity between replicates were strikingly apparent during the morning Tricorythodes activity. Visual inspection indicates that availability of Tricorythodes was similar in all channel sections. Yet on most mornings all 4 trout in replicates IIA2 and IIBl fed heavily, while no trout or only the small ones fed in replicates IIAl and IIB2. The same occurred in experiment I: feeding activity was consistently greater in section A2 than in section AI. Sectional differences in depth and current velocity did not correspond to differences in feeding activity. It seems likely that social factors determined behavior within replicates. This applies to cover use patterns as well. As presented earlier, small trout in replicates without large trout were more active agonistically than small trout in replicates with large trout. This was not true of feeding activity. Mean per- cent of time spent feeding by small trout was 43.7% and 37.2% in replicates with and without large trout, respectively. Mean percent of observation session movement devoted to feed- ing increased from rank-l to rank-4 trout (Appendix Table A23). Mean percent of movement in feeding was roughly proportional to mean per- cent of time, spent feeding. Rank 1 differed significantly from rank 3 (p < 0.05) and rank 4 (p < 0.01) but did not differ from rank 2 (p < 0.2). Differences between times of day were not significant (p < 0.25). Replicates differed significantly overall (p < 0.01) but contrasts between means were not significant. This statistical re- sult occurred because the sum of variation between all replicates was quite large but differences between any two replicates were relatively 77 small; also the Scheffé procedure is quite weak in detecting mean differences (Dr. J.L. Gill, Statistician, Dept. Dairy Science, Michigan State University, pers. comm.). Rate of movement (cm/s) during feeding was greatest during the morning and least during midday (Appendix Table A24). The midday and morning means differed significantly (p < 0.05). Fewer feeding station changes per unit of feeding time (Table 17) rather than fewer and/or shorter movements per feeding station (Table 18) was the reason for less midday movement. The number of minutes spent per feeding station during midday was significantly greater than during the morning (p = 0.005) (Table 17). Large trout (ranks l and 2) changed feeding stations less frequently than small trout (ranks 3 and 4) but these mean differences were not significant (p < 0.5). Rate of movement was significantly greater in replicate IIBZ than in all other replicates considered together (p < 0.05). This may reflect the fact that mean current velocity in section 82 was slower than in the other sections. Trout could swim against the current with less energy expenditure and thereby increase the effective availability of food. To summarize, feeding activity was greatest at night and during the morning when drift food was most abundant. Small trout spent more time feeding than large trout and consumed more food than large trout. Absolute rate of movement associated with feeding was similar for all trout. High food availability resulted in more fre— quent changes in feeding stations for all trout. This appeared to be the result of more frequent agonism. There were considerable 78 . H0fl>0£0£ HUSH 0”»de Ho.e Hm.HH v mm.m mo.mH mchm>m om.o ~m.~H m mm.o Ho.eH smoeHz mm.o mm.mH m mm.v mm.oH qucuoz Ho.a mo.mH H mm com: Om sum: was «0 msHa xcmm m~.a mH.mH m me.HH m.o~ o.HH m mm.m~ v.H~ mo.m~ m mm.mH mm.mH mH.m HaHHH om.~ mm.m o.mH v.m mn.m me.m Hm.n mn.o mmHH m.v v~.m MH.cH He.¢ m.HH m.m 6.6m m.m «H.h m.m om.oH HmHH mm.m Hm.~H mH.oH mm.HH o.mH so.mH om.oH mH.m m.mH m.mH HH.~ me.m mo.m mm.OH maHH mm.m om.mH mm.>H o.m m o.- o.mH mn.mm m om.m mv.mH a o.vm Hm hooves mswsuoz .He>oo Scum msfiooom Momma mafia moosaocu macs .uzouu v and: mwuoofiamou cw mcasm>o coo .mooows .mcfisu0E mcauso xsou HoHOOm some an coHuoum mcfioeom Hem usomm AsHEV mafia com: I .hH magma .H0H>mnon HoEHocnm vm.m NH.m m He.m mm.v smeon mo.m no.m m mm.~ mv.m mchuoz mm.~ mm.m H mm emu: am new: mom mo mafia xcmm mm.~ mo.m m 6.m mo.m o.m m mm.mH mm.m mm.m m mm.HH mo.n «4.6 HaHHH m~.m HH.6 m~.o 6.6 mv.o ms.m em.m mm.m mmHH vo.H mv.m mm.H mv.m mH.m mo.H o.e HH.v mm.o Hm.v om.m HmHH H.m em.m m.H mm.H mm.HH Ne.H o.m mH.H o.H oo.q s.v om.m ma.v mv.> NHHH mo.m mm.m mm.m ma.m m o.H o m~.m m m.m mm.v mm.~ HaHH mm.m mm.m mq.m m.m me.o mm.o mm.m mm.H m~.m mm.o mv.H ma.e maH am saw: v m m H e m m H v m m H «umonmum oumoaamom xsom xsmm xsmm (mswcm>m >mooaz mswcuoz .Ho>oo scum mcaooom saw: poumHUOmmm usoae>oa mmoaaosfi many .usonu v nuwz moumoHHmou sfl msficm>o poo .>MQQHE .msacuoa mcfiuso xcou Hmaoom some an :oHuoum mcwpoom Hem ou>05 Hay oocmumwo com: I .ma mHnma 80 differences in feeding behavior between individuals and between replicates. Differences between replicates were not strongly related to relative food abundance or physical characteristics unique to particular sections. Social factors seemed to be responsible for dif- ferences between replicates. Characteristics of Feeding Stations Mean current velocity at feeding stations, measured 5 cm above the stream bed was 20.3 cm/s (Table 19). Mean depth at feeding sta- tions was 24.4 cm (Table 20). Velocity and depth readings for in- dividual trout at each feeding station were weighted by the percent of total feeding time spent at that station. Mean current velocity at feeding stations was approximately equal to mean sectional current velocity. Mean depth at feeding stations was about 2.5 cm deeper than mean sectional depth. Feeding station current velocity averaged lowest in section 82 which had the lowest mean current velocity. The greatest mean depth at feeding stations was in section A2 which had the greatest mean depth. Social ranks did not differ significantly in mean current velocity (p < 0.5; Table 19) or depth (p < 0.5; Table 20) at feeding stations. However, rank-1 trout on the average used feeding stations with slightly lower current velocity and less depth than ranks 2, 3, and 4. There was not a significant difference in either current velocity or depth at stations used at different times of day (p < 0.5). In calculating mean current velocity at feeding stations, 6.10 cm/s was subtracted from the velocities shown in Appendix Figures B3-B6 for stations within 0.25 m of feeding references. This value .Hofl>o£mn HoEuocndm 81 n.m m.mH v s.m n.o~ uchm>m e.m m.H~ m m.m s.om smoeHz m.m ¢.o~ N o.m m.mH mchuoz m.m ~.mH H om cam: om cam: New «0 made xcom ~.v H.om m H.m~ o.v~ m.m~ m m.mH v.vH H.5H m m.H~ v.vH m.H~ Hm moppfiz madame: .coflumum some um ucmmm 05H» msflooom Hmuou mo useouom on» an puunmwoz mH hufloon> uceuuau .mcoHumum mcflcoow Ho>oo mooaaocw mafia .usouu v cue: moumoHHmoH :fl mcwco>u pom .mmoowfi .msHGHOE mswuao xcmu HMHOOm some ha mom: chHumum mcwowom um Am\Eov >uflooHo> usouuso one: I .ma oHnoB .H0H>mson HmEHosnmo 82 o.m o.mm v m.m m.vm mchm>m R.H m.qm m a.v o.mm smoeHz m.~ e.mm m o.v a.mm mchuoz m.m ~.- H mm coo: om com: woo mo oEHB xcmm m.v m.~m m h.mm h.m~ ~.om m m.h~ o.mH h.mH m h.v~ N.HN m.mm HeHHH m.m o.vm o.mm N.om h.om N.~N ~.Hm m.m~ NmHH m.m m.~m o.Nm m.vm m.Hm m.vm o.mm N.mm h.mm h.v~ h.mH HmHH v.m ~.om h.vm m.o~ N.om h.mm m.mm n.mm o.m~ 0.0m n.0m o.om m.H~ h.mm meHH o.H o.mm m.mm o.mm ~.Hm o.mm m.m~ ~.m~ m.m~ ~.Hm HmHH m.m o.v~ «.mm h.o~ m.mm m.v~ m.m~ m.vm N.h~ o.m~ ~.m~ h.mH de am com: v m m H v m m H v m N H oumomeom mumOHHmom xsmm xsmm xcmm mswcu>m mmooflz mcwsuoz .cowumum some as Human mafia mswcoeu Houou mo unmouom an pounmwo3 nuom .mcofiumum mcwooow He>oo moonaocw mane .usonu o saws mouooflHmoH :H mcwcu>o one .mooofls .msHGHOE mcfiuso xsou Heaven some an com: mcoHuoum mcaoomm us Have spawn Hone: new: I .om manna 83 approximates the reduction in current velocity experienced by trout feeding within 0.25 m of a reference (Figure 8). Most trout avoided the highest current velocities in their section. The observed percent of feeding time spent by each trout at stations with current velocities higher than 19.5 cm/s or 24.0 cm/s was compared with the expected percent of feeding time spent at those stations with a chi-square analysis (Table 21). Expected values were based on the hypothesis that use of a particular range of current velocities would be proportional to the area of the channel section in that velocity range if trout had no velocity preference. The values 19.5 cm/s and 24.0 cm/s correspond to contours on the sectional current velocity maps (Appendix Figures B3-B6). Use of feeding stations with current velocities exceeding 19.5 cm/s was significantly less than expected in replicates IIA2, IIBl, 1182, and IIIAl (p < 0.001 for all). Use of stations with current velocities greater than 24.0 cm/s was significantly less than expected in replicate IIAl (p < 0.001). Mean current velocity, was higher in section Al than in sections A2 and B2. Rank 1 avoided high current velocities more consistently than all subordinate ranks. Use of feeding stations with velocities greater than 19.5 cm/s was significantly less than expected for rank 1 (Table 21). While rank 1 did not differ significantly from subordinates on an absolute scale (Table 19) it may have had a slight energetic ad- vantage relative to the small range in velocities available for feeding stations. 84 .oen Bowman 0;“ o>ono EU m oousmmozo .HHoo.o v my emuomoxm can» mmmH sHucmoHMHcmHmo .usmoHuHcmHm nozn .HoH>onmn Hmauocndm omH.Hm mm m cm. HH mH o.v~ omH.mv on o mm mm mo m.mH H.~o no we so mm m m.mH HmHH OMH.mm m o m o o o.vm one.vm we mm Hv mm «H m.mH msHH oom.mo mm Hm o m o o.om n on om 00H m OOH m.mH HéHH n m Hm mm mm o o.v~ n we mm mm on o m.mH de wsHm> mean“ HHa v xcmm H gene . a scam H scam Human. can» mumoHHmum mumsquHno touoomwm oe>ummno Houomum mEHu mcHowom wo useouom omuHooHo> .mGOHuoum mcHooom Ho>oo moosHocH mHna .HomImm mousmHm xHocommdv m\eo o.v~ can m.mH omeoxe mmHuHoon> £0H£3 :H mono soHuoom mo ucmouom Hmsoo mosHm> oouoomxm .usouu v :qu moumoHHmou sH m\Eo o.v~ one m.mH can» Houmonm meuHoon> Hummuso um ucomm 08H» mcwommm mo unmoumm wouommxe use oo>uomno I .HN oHnoB 85 The attractiveness of overhead cover or unmeasured features of the stream bed under cover encouraged some trout to feed at stations with relatively high current velocity. The preferred cover (large) in replicate IA2 was frequently used as a feeding station. Current velocity under this cover was 27.43 cm/s. Throughout the study trout rarely fed at stations in the open with such a high vel- ocity. Mean current velocity at feeding stations was higher for trout in replicate IA2 than in replicate IIA2. Use of the preferred cover as a feeding station in replicate IA2 was partially responsible for the difference between replicates. Replicate IA2 also lacked feeding references which were present in replicate IIA2. Feeding references provided additional low velocity feeding stations in replicate IIA2. Use of feeding references was highly variable between rep- licates (Table 22). Over 37% of total cumulative feeding time was spent at feeding references in replicate IIBl, 10% in replicates IIA2 and 1182, and only 1% and 4%, respectively, in replicates IIAl and IIIAl. Relatively high current velocity in section Bl may have made the references particularly attractive there but this explanation is not consistent with the observation that use of references was very light in section A1 which had a similarly high mean current velocity. Percent of feeding time spent at feeding references increased from rank-l to rank-3 trout. Use of references by rank 4 was lower than use by ranks 2 and 3 (Table 22). Feeding references provided current velocities in mid-channel comparable to velocities along the 86 .H0H>mnon HMEHocndo Amm.m v Am.hv A o V o Amm.m v m.©H ANH.m V m.HN AHv.mv m.m HflHHH Av¢.OHV 0.0m A o v o ANm.va o.mN AHm.H v o.H A ov o NmHH Abm.bmv m.bwm Ao.omv m.ovH Avm.mvv m.om Avm.omv m.HN Avm.mv o.m HmHH iqv.oH. n.ooH A o e o imo.H~. m.Ha imH.mH. o.m~ A oo o maHH AmH.H V o.v Am.N.v o.v A ow o m m A 0v 0 HdHH mHmuou N H oumoHHmom ouMOHHmmm xsmm um oEHu msHooom Hmuou mo unmoumm .mumenusoumm :H moosoumwmu .moocouoweu mcHoeem o>o£ uo: oHo NflH muoOHHmom .usouu v nqu moUMOHHQoH :H xsou HoHoom some an moosouuwou mcHooeu um ocean Acflev QEHB I .NN oHnme 87 walls but this rarely induced large trout to feed away from the walls. That large trout did occasionally use feeding references, however, in- dicates that the size of the area of reduced current velocity created by a reference was sufficient to accomodate a large trout. Cover location may have influenced choice of feeding refer- ences. Use of mid-channel references was by far the greatest in replicate IIBl which was the only replicate among the 5 that held 4 trout that had mid-channel cover (Table 22). Without replication this is only speculative. Large trout showed a strong attraction to the channel walls. Ranks 1 and 2 fed significantly (p < 0.01) closer to the channel walls than ranks 3 and 4 (Figure 13; Appendix Table A25). Distance from the nearest wall to each feeding station was weighted by the percent of total feeding time spent there. Feeding activity was greatest on the shaded side of the channels (Table 23). ANOVA was performed on the difference between the percent of feeding time spent on the west and east sides of each section. Zero difference indicates no preference. A positive value indicated preference for the west side while a negative value indicated preference for the east side of the channel. The evening mean was positive and significantly different than the midday (p < 0.005) and morning (p < 0.005) means which were negative. Social ranks did not differ significantly (p < 0.5) but differences between replicates were nearly significant (p < 0.1). Preference for the east side of the channel during midday as well as during the morning reflects the fact that most feeding during midday occurred early in the period while the area along the east wall was still shaded. 88 1'0. I—I I—l 7; I 3' 2 O " - o _ ._—, C 2 —- 0.5‘ E O 0 c II o '3' 6 c d O 0 2 j 0 1 2 3 4 Social rank Figure 13. - Mean distance to the nearest wall from feeding stations used by each social rank in replicates with 4 trout. Distance weighted by the percent of total feeding time spent at each station. Cover stations are not included. 89 OOH O O mm b O v 0 O om O v mm m om O OH N OOH O O H NmHH vv O mm hm mm 0 mm Hv MN v mm HH O NN mo O mm vH O m O 0 VO N O O OOH H HmHH O O OOH OH m Oh v mm m m mm m O 00 m b m O O OOH OOH O O HO O m N m O hm ON O Nb H Nm mooon mchuoz .omosHocH uo: mH Hm>oo Eouu mcHommw usmmm oEHB .oon pounce may muoOHosH mononucoumm .szonm Hoccono esp mo OHooHE on» :H can Hessmno on» no mmon ammo one Hows mnu so usomm OEHu mcHoomm mo unmouom .usouu o nqu mOUMOHHQmu :H oomnm mo coHumooH Use xcmu HoHOOm some an pom: msoHuoum OCHomom mo soHpmooq I .MN oHnma 9O .oooosm HHHum mos HHm3 ammo onu mcon memo may cwnz oonusooo OCHOOOH >MOOHE umoz Q .AwmImm moHDmHm xHocmmme momv m.N oosuHmsoH so coHuoum msHooom mo ucHom Honcho vN ON on Nb OOH OOH hm Hm OOH NN OOH HN mH Nm mm OOH OH on On vv coo: N H HGHHH A.O.ucoov MN OHQMB 91 Note that values reported for percent of feeding time spent on each side of a channel do not exactly correspond to the percent of feeding time spent in the shade. That is, at certain times of the day the "shaded" side of the channel was only partially shaded. The tendency for some trout to frequently dart into sunlit water from shaded feeding stations and the general high mobility associated with feeding would have made attempts to exactly quantify time spent in the shade very difficult. This does not invalidate this analysis how- ever, as trout rarely fed in direct sunlight when shade was available. "Feeding on the shaded side of the channel" and "feeding in the shade" are synonymous for practical purposes. During later afternoon and early morning when more than half of the channel was shaded it was possible for trout to feed on the "unshaded" side of the channel when in fact they were feeding in the shade. This was most often true of small trout which often fed just inside the shadow edge. Therefore, small trout actually spent more time feeding in the shade than in- dicated in Table 23. Large trout fed close to the walls even when the entire channel was shaded. During the evening, trout seemed to avoid the sunlit east wall even when the water at the base of the wall was shaded. Competition for feeding stations was not severe. Trout occasionally moved to stations vacated by dominants but generally each trout spent most of its feeding time in a characteristic position, or set of positions, that was different from the positions that other trout used the most. These characteristic positions were used even when dominants were not feeding. The only notable exception to this 92 was in replicate IIBl where ranks 3 and 4 competed strongly for two feeding references. Most feeding activity occurred in the lower 2/3 of the sections. Only about 11% of total feeding time was spent in the upper l/3 of the sections that held 4 trout. The reason for this was probably that current velocity was highest in the upper 1/3 of the sections. Trout also may have avoided feeding where overhead cover was not in sight. Two trout were very unusual in that they spent almost all of their time in the upper l/3 of section Al. Rank 4 in replicate IIIAl spent 97% of its feeding time in this area near coordinates 1.25, 12 (Appendix Figure B3). Rank 2 in replicate IIAl apparently did not feed but spent nearly 100% of its total observation time in the same area. This area was attractive because a few pieces of large gravel located immediately upstream produced a pocket of relatively low current velocity (15.24 cm/s) and surface turbulence which, in effect, provided some overhead cover. The effects of the gravel were not noticed until current velocity was measured at the end of the last experiment. To summarize this section, current velocity and light level strongly influenced choice of feeding positions. Current velocities over section means were avoided. All trout favored positions in the shade. The location of overhead cover had little, if any, in- fluence on choice of feeding positions (see "Cover Use and Feeding" section). Large trout showed a strong attraction to the channel walls. Competition for feeding positions was generally weak so that social rank had little influence on choice of feeding positions. 93 Size of Trout and Size of Food It was surprising that large trout which experienced severe weight losses, and therefore must have been hungry, ate less than small trout. Even during periods when as many as 10 to 15 spent Tricorythodes mayflies drifted past a feeding station per minute rank-l trout in replicates IIAl and IIBl often ate fewer than one per minute while small trout ate nearly every mayfly that drifted near their station. Rank-2 trout in these replicates fed more actively, though usually not with such intensity as the small trout. Large brown trout in streams rely heavily on crayfish, fish, and other large food items (Brynildson et al, 1973; C.E. Bassett, pers. obs.). To test the hypothesis that large trout in the experi- mental channels were not feeding as much as small trout because large food items were scarce I artifically introduced grasshoppers averaging 3 cm in length to replicate IIIAl during 7 observation sessions. Grasshoppers were directed to the desired position in the channels from the observation blind by blowing them through a plastic soda straw. Thirty-one grasshoppers were introduced to the channel (Appendix Table A26). Thirteen were consumed by large trout, 16 by small trout. Both large and small trout feeding in the open never refused a grasshopper, sometimes moving 2-3 m from their feeding stations to intercept one. Seven grasshoppers were consumed by large trout that had not indicated feeding activity during the preceding several minutes and were under cover. On two other occasions large trout under cover moved toward a grasshopper but did not eat it. 94 During a morning feeding session rank 1 twice moved from cover to con- sume a grasshopper and then moved to a shaded feeding position near the east wall for about 1 min before returning to cover. Grass- hoppers introduced during midday and evening did not induce rank 1 to move to a feeding station in the open, though small trout did this occasionally. Large trout fed on grasshoppers while ignoring small, naturally- occurring food items. Ninety-seven percent of the stomach contents by volume of rank 1 checked over 12 h after the last feeding test consisted of grasshopper remains (Appendix Table A21). Nineteen per- cent of the stomach contents of rank 2 consisted of grasshopper re- mains. That stomachs were checked in the early morning immediately following the heaviest feeding period (2000 - 0800 h) indicates that large trout ignored most small food items that were naturally avail- able. These results suggest that large trout would have eaten more food if large food items had been available. Miscellaneous Three types of activity were included in the miscellaneous category: (1) pressing sideways against channel walls or screens, (2) occupation of a stationary position in the open, and (3) swimming about the channel that was not associated with feeding or agonistic activity. Current action along the channel walls, particularly near screens, caused the gravel to slump in some spots, giving trout access to the walls. This situation was quickly corrected when it occurred, so that miscellaneous time was minimal. 95 Only two trout pressed sideways against screens. Rank 2 in replicate IIIAl did this for 15 sec on the morning of the first observation day. No stimulus that might have caused this fright response was apparent. Covers had been destroyed by vandals 3 days earlier, causing all of the trout to press against the screens or each other. Covers were replaced the following day, and within 24 h all trout once again used cover. Rank 2 was the only trout to re- exhibit the fright response after observation sessions began, and only on this single occasion. Rank 2 in replicate IIIBZ had been injured by vandals. It spent 72% of its observation time pressed against the lower screen, the walls, or small rocks along the walls. This behavior was largely associated with avoidance of rank 1, especially during and following violent attacks. Rank 2 successfully used cover only when rank 1 also used cover. Attempts by rank 2 to occupy either cover when rank 1 was feeding inevitably resulted in an agonistic bout with forced return to a wall or screen. As mentioned earlier, 2 trout rarely or never used cover and spent a large portion of their observation time in a small area. Rank 2 in replicate IIAl spent over 99% of its observation time in a single position, and never indicated feeding activity. Physical appearance was normal. All of this time was assigned to the mis- cellaneous category. Rank 4 in replicate IIIAl presented a special problem in that it spent over 92% of its observation time in an area that served as a feeding station and as cover, due to the presense of surface turbulence. Motivation to use cover could not be reliably 96 separated from motivation to feed necessitating use of the mis- cellaneous category. Times when food consumption was actually ob- served were recorded as feeding time, but since these data were not directly comparable to feeding data for other trout they were omitted from statistical analyses. All miscellaneous movement was in the form of continuous swimming about the channel. This movement may have been associated with "surveillance" of the food supply as it occurred when other trout were feeding . But as motivation to feed was not clearly the cause of this movement it was assigned to the miscellaneous category. Social Rank Summary and Energy Use Time and movement in each activity category is summarized by social rank in Table 24. Rank 1 and rank 4 were on opposite ends of the behavioral spectrum. Rank-l trout used cover more and fed less often than subordinates. To determine if dominance was an advantage, energy output was estimated for each social rank from swimming speeds and time spent in each activity. Energy gained in food consumption could not be validly compared among ranks as the absence of large food items put large trout at a disadvantage. Large trout could not capture enough small food items to compensate for the energy required to capture them. Even large trout that fed heavily on Tricorythodes mayflies lost as much weight as large trout that rarely fed. Dominance could be an advantage in the experimental environment only if it permitted trout to conserve energy by swimming less than sub- ordinates. Table 24. - Summary of the mean percent of observation session time and movement and mean rate of movement in eaCh activity by social Means with different superscripts are significantly different (p < 0.05). rank in replicates with 4 trout. Rank Activity 1 2 3 4 Mean Agonistic 0.26a 0.65a 0.75a 0.53a 0.55 Mean Cover usec 81.63a 69.12ab 54.26b 54.93b 64.98 % Feedingd 15.09a 29.97ab 43.47b 44.04b 33.14 time e Miscellaneous 2.98 1.41 1.81 0.03 1.56 Total 99.96 101.15 100.29 99.59 Agonistic 37.71a 23.82a 24.01a 18.66a 25.8 Me an C a a a a Cover use 21.6 22.0 14.82 12.98 17.62 % Feedingd 40.86a 52.63ab 62.66b 70.03b 56.54 movement e Miscellaneous 1.5 0 0 0.3 0.45 Total 101.67 98.45 101.49 101.97 Mean Agonistic 31.35a 29.19a 26.55a 30.09a 29.29 rate Cover useC 0.15a 0.55a 0.57a 1.52a 0.72 Of d a a a a movement Feeding 0.77 0.8 0.8 0.95 0.84 (cm/s) Miscellaneous 6.86 0 0 16.57 5.86 c . . Does not include feeding under cover. dIncludes feeding under cover. e . . . . Miscellaneous items were not tested statistically. 98 Oxygen consumption provides an indirect measure of energy expenditure in calories. The precise relation between calories and oxygen consumed depends on the type of fuel metabolized. Most trout lost weight so that stored corbohydrate (liver glycogen) and fat were probably important sources of energy (Hill, 1976). Overall, diet was similar for each social rank. It has been assumed that the fuel metabolized was the same for all trout and that relative oxygen con- sumption provides a good approximation of relative energy expenditure among social ranks. The relationship between swimming speed and oxygen consump- tion for brown trout in this study (Table 25) was estimated from rain- bow trout data reported by Rao (1968). Beamish (1964) examined the relationship between standard metabolism and weight for adult brown trout but did not provide data on the influence of swimming speed on oxygen consumption. The similarity between brown and rainbow trout in phylogeny and body shape suggests a similar metabolism-~swimming speed relationship for the two species. Rao's rainbow trout differed from the brown trout in this study in that they were smaller (54-135 9) and of domestic origin. 99 Table 25. - Relationship between metabolic rate (Y = mgOz/h) and swimming speed (X = cm/s) for each social rank plotted from equa- tions relating metabolic rate and body weight (g) at several swim- ming speeds for rainbow trout (Rao, 1968). Standard metabolism occurs when swimming speed is zero. Standard Rank Mean weight (g) Regression equation metabolism (mgOb/h) 1 547.5 log Y = 1.6353+0.0098X 43.18 2 473.6 log Y = 1.5869+0.0097X 38.63 3 227.3 log Y = 1.3415+0.0089X 21.96 4 145.6 log Y = l.l926+0.0085X 15.58 Extrapolating from immature to mature fish might lead to errors in the calculated difference in energy expenditure between social ranks. This is based on the observation by Brett (1965) that the slope of the regression line relating standard oxygen consumption to body weight is shallower for sockeye salmon in juvenile size classes than for adults. At routine and active levels of metabolism the relationship was the same for juveniles and adults. The regression equation relating oxygen consumption to swimming speed would assume a steeper slope than it should to properly reflect the relationship for adult fish only. If this is true for rainbow trout, the differences between social ranks in relative energy expenditure (mgOz/kg/h) above standard metabolism (Table 26) are slightly high. Extrapolating from domestic to wild fish probably does not introduce serious error to estimates of energy expenditure. The re- lationship between swimming speed and oxygen consumption is nearly identical for wild and hatchery rainbow trout (79-398 9) at water 100 temperatures within the range prevailing in this study (Dickson and Kramer, 1971). Oxygen consumption estimates are based on a water temperature of 15 C which is close to the average temperature of the water in 1976 (Appendix Table A2). Energy expenditure was estimated by the product (mean per- cent of time spent in an activity category) x (mean rate of oxygen consumption associated with that activity). The latter was derived from the swimming speed--oxygen consumption relationship (Table 25). Most feeding involved constant swimming to maintain a stationary posi- tion in the current. To account for this source of energy use, mean current velocity in the channels (20.73 cm/s) was added to swimming speeds in the feeding category. No corrections were necessary in the other categories as upstream swimming was as common as downstream swimming. Rank-l trout expended less energy in excess of maintenance requirements than any other social rank (Table 26) when oxygen con- sumption is expressed in terms of body weight (mgOZ/kg/h). Rank-4 trout expended the most energy. Even without compensating for dif— ferences in body weight, rank 1 used less energy above standard metabolism than ranks 2 and 3. The important question is whether these differences in energy use reflect differences between ranks in the availability of pre- ferred food or rather are a direct result of social status. Most of the difference between ranks occurred in the feeding category. Total energy use (mgOZ/kg/h) above standard metabolism was roughly pro- portional to percent of time spent feeding. Percent of time spent 101 mo.e~ m.mmH mmm.m mv.mH H6908 AHH.o o mo.o Amo.o o no.o moo.o Ho.o msooomHHoomHz Ao~.~mo mm.v~ Amm.mmo mo.me mo.m mv.oH mcHooom Amm.m o mm.H Amm.mvo mo.oo m~.o Ho.o om: uo>oo Ame.H . mv.o Asa.o o mo.H so.o mH.o oHomHooma a Nv.v~ mm.mHH mm.m eH.em H6808 A o o o A o o o o o mooocoHHoomHz Am~.mmo -.m~ Amo.vmo mm.mo .m~.m Ho.oH moHooom Avm.~ o Ho.o Amm.vvo Ho.mm oH.o mo.mH om: uo>oo AaH.m o mm.o Amo.H o m~.H NH.o m~.o oHomHooma m om.oH mo.em 8m.a om.ov Hmuoa A o o o A o o o o o moooomHHoomHz Amm.mmo mm.mH Amv.ooo mm.mm m~.> me.mH ooHooom AHH.v o mo.o on.mmo so.>m mm.o mo.am om: 9o>oo Aom.m o mv.o Aoo.H o mo.H Ho.o mv.o oHomHoooa N mm.m mo.hm mm.o mo.ho Hmooe AHo.e o mm.o AHH.m o mn.~ H~.o m.H moooomHHoomHz Aoo.ooo mv.e AH~.mNo vm.oH no.4 mm.oH ooHooom Aeo.m o mm.o AmH.veo 6.86 NH.o am.mm 065 90506 Aho.~ o mm.o Amv.o o mv.o NH.o m~.o oHumHoooa H n\mx\NOOE £\mx\moms :\NOOE £\NomE >uH>Huo< xcom ouoosoum m>on< oumosmum o>on4 EmHHOQouoz .AmN OHQMBV omomm mcHEEH3m on EmHHonoums msHuoHon msoHumsoe osm .AoN OHnoev >HH>Huoo some nqu omuoHOOmmm omwom mcHEEHzm some .>uH>Huom some cH ucomm mEHu EOHH oo>Huoo ouo3 moDHm> xsou HmHOOm some now AmN OHQoBO fimHHonoums osmosoum o>ono An\mx\NomE can £\Nomsv om: monoco oODMEHumm I .oN oHnoa 102 feeding by large trout, in particular, may reflect the scarcity of large food items. The agonistic and cover use categories are better indicators of the effect of social status. Rank 1 used conspicuously less energy in agonism than all subordinates, even though it was in- volved in at least as many bouts as ranks 2 and 4 (Table 7). Rank 1 also used conspicuously less energy in cover use than all subordinates. This could partially reflect the small amount of time rank 1 spent feeding as movements from feeding stations to cover were included in the cover use category. But, energy expended in cover use by rank 1, and other trout, was not proportional to the percent of time spent feeding as would have occurred if movements from feeding stations to cover were the only source of energy loss associated with cover use. The fact that rank 1 was the only rank not subject to displacement from cover probably permitted less energy use. Much of the movement be- tween covers may have been the result of unseen agonism. The results of this analysis suggest that dominance permitted rank—1 trout to use less energy than subordinates. Energy use estimates demonstrate clearly that the primary source of daytime energy loss for stream-dwelling trout is feeding. Overall energy use above standard metabolism in feeding was 92.6% of total energy use (Table 26). By limiting the amount of time spent feeding, large trout conserved a substantial amount of energy. Agonism was by far the most energy-consuming activity per unit of time spent, but it occurred too infrequently to be an important source of energy loss. Cover use was the least energy-consuming activity. All ranks devoted about the same portion of total energy use (mgOz/kg/h) above standard metabolism to each activity. 103 Time of Day Summary Time of day influenced behavior (Table 27) via changes in light intensity and drift food abundance. Cover was used most during midday when light level was highest. Most feeding occurred during the morning when drift food was most abundant. Agonism was also most frequent and intense during the morning as intense feeding then led to more total movement and greater aggressiveness. The por- tion of total movement devoted to feeding was greatest in the evening. Movement associated with cover use was minimal at this time. Trout often did not leave cover to feed until late evening and once in the open, they usually stayed there. At low light levels there is apparently little need for cover. Replicate Summary There were significant differences between replicates in per- cent of time spent using cover and feeding, and in the rate of move- ment associated with feeding (Table 28). Trout in replicate IIB2 used cover more, fed less, and moved more when feeding than trout in all other replicates. Current velocity was lower in section B2 than in all other sections and this might have encouraged greater movement when feeding. If a threshold density of food drifting past a feeding station were required to release feeding behavior as has been observed for brown trout (Ringler, 1975), then slow current velocity might have resulted in that threshold being reached less often, with less feeding as a consequence. However, this probably is not the explanation here as feeding activity seemed to be independent of drift food density in replicate 1182. Cover use has been observed to decrease at reduced 104 .>HHoOHumHuoum ooumou no: mneB mEouH msoocoHHmomHZO .uo>oo Eoum OCHUOOH moosHocH U .uo>oo Scum msHoomm oosHocH uoc moomo o 5N.o mo.N omsoocMHHoomHz Am\Eov novm.o QH¢.O mmN.H omcHoomm unoao>oe th.o omh.o m v.0 com: Ho>ou mo mums omm.vm mNm.ON oHo.Hm UHHchom< cmoz NH.mm NN.ooH o.HoH Hence 0 Ho.o 0.0. omsoocmHHmomHz omm.m> mom.ev 8mm.Hm omoHooom oooao>oz nH.m ohm.NN 0mm.mm 00m: Ho>ou w ovo.om mmm.Hm www.mN UHumHsomo coo: mm.HOH $0.0m mH.HOH Hence Nm.H wo.H NH.H omSOoGMHHoomHz oomm.mm noo.- «mo.mv omoHooom oEHo nomm.oo mmm.mh nNm.mm 00m: Ho>oo a m m.o mom.o ooo.H oHumHooma coo: moHcm>m mooon mecuoz huH>Huom mum mumHuomHmmom acouumeo nqu memos Io>os mo wuou some one .usqu>OE one mEHu conmmm :oHum>Homno mo Homouom some on» we humEEsm I .hN mHnma .Amo.o v do uoouoHMHo sHuooonHomHm .uoonu v :qu mOHMOHHQmH :H zoo mo mEHu an huH>Huom some :H uses 105 .xHHmOHumHumum ooumou non ono3 mEmuH mnoonMHHoomHz m .nm>oo Scum mnHomom mooSHonHo .Ho>oo Scum mnHowmm monHonH uon moon o mm.o 66.H o o m.H o mmnomanHoomHz Am\Eov . . . . . . mnH om noEm>OE nmm o omH N 365 o n o o 9N6 o a m o m .o m u mNm.o ohm.o 6mm 0 ome.o mmH.o mmN.H oomn no>ou mo moon va.mN omH.oN omN.Nm th.mN 6mm.Hv omN.mH OHHmHnomn new: No.5m mm.NOH 6m.HOH Nm.om mm.NOH om.NOH Hmuoe mo.o Hh.o o O mo.H o mmnoonMHHoomHz whN.vv on.o6 on.mo moN.6m 66H.Oo ohH.Hm omnHooom unoso>os omo.HN 6mm.oH mMN.HN ohv.m mNO.mH on.mm oomn uo>oo w oNo.Hm ohm.om oNo.vH omv.Hm omh.mN 6mm.hN OHHmHnomn new: Hv.mm mm.NOH mH.OOH MH.mm c.00H Nm.mm Hopes v.0 m.v mH.H mm.o mo.o no.H wmnOOnMHHoomHz onmNO.om 0mm.mH onmmH.mm m m.mv N.mN nomH.mm omnHooom mEHu onomh.ho 6mm.Hm onooN.mo o>m.nv nonv.vh onoo.mm oomn Ho>oo w mH~.H om~.o mmm.o mm6.o ovw.o o66.0 0HumHnome :60: HeHHH mmHH HmHH maHH HHu0< ouMUHHmom .Amo.o v do ooouoooHo >HunMOHMHanm mum umHHomummnm unouowwHo SUHS mnmoz .unouu v nuH3 wouMOHHmon >n >HH>Huom some nH unofi Im>OE mo ouou noes one .unoEo>OE onm 08H» nonmom noHum>uomno mo unoonom nme any «0 huoaanm I .mN anme 106 current velocities (Hartman, 1963). This has been explained as a re- action to reduced stress at lower current velocities. But this also contrasts with behavior in replicate IIB2 where cover use was the greatest observed despite the lowest current velocity among the 4 sections. Social factors seem to be responsible for differences be- tween replicates. DISCUSSION--1975 AND 1976 Introduction The results are of value only to the extent that they reflect brown trout behavior in the wild. Simplification of the environment was essential in this study to determine the causation of behavior, but attempts were made to minimize unnatural sources of stress and to permit ample time for acclimation to the test environment. Human disturbances were kept to a minimum, though total extent of such dis- turbances were uncertain, as Uninvited visitors could enter the study area. Occasional human disturbances had no doubt been experienced by all trout in their natural streams which are heavily fished. Most physiological effects of handling and anethesia with MS-222 disappear within a few hours (Wedemeyer, 1970; Houston, et al., 1971). Trout that showed highly unusual behavior were not included in statistical analyses (rank 2, replicate IIAl; rank 4, replicate IIIAl). Trout densities in the study channels were higher than the average densities in northern Michigan streams, but were within the range of natural densities. Density per total area of water surface in the experimental stream was about 10.7 g/m2 in 1975 and 23.3 g/m2 in 1976. Trout density, (brook and brown), in the North Branch of the AuSable River from which the 1976 test trout were captured averages about 11.4 g/m2 (1972-1976), and density of brown trout in the size classes used in this study averages 4.3 g/m2 (William Buc, Fish 107 108 Habitat Biologist, Michigan DNR, pers. comm.). Density within the area of stream suitable as living space may be much higher than density expressed in terms of total water surface area. Allen (1969b) calculated that the average density of several species of stream salmonids over a wide size range was 1.7 g/m2 surface area, but was 25.5 g/m2 in living area ("territories"). Electro-fishing (AuSable and Pigeon Rivers, Michigan, C.E. Bassett, pers. obs.) and snorkling Observations (Devore, 1975) confirm that trout densities in the experi- mental channels were not higher than can occur naturally in the wild. Weight losses during 1976 indicated that food availability was lower than in the native stream. Conflicting evidence in the litera- ture makes it difficult to assess the effects of low food abundance on time spent feeding. Low food abundance may decrease time spent feeding if the threshold food density required to release feeding is reached less often (Ringler, 1975). Conversely, low food abundance may increase time spent feeding as hunger increases (Ware, 1972). This would also tend to increase agonism as trout are most aggressive when they feed. Results of the feeding experiment indicate that large trout would have eaten more had larger food items been available. How- ever, other investigators have found that large brown trout fed in- frequently during daylight in natural environments where availability of large food items was presumably very good (Butler and Hawthorne, 1968; Jenkins, 1969). I infer that the difference between large and small trout in percent of time spent feeding in this study mainly re- flects inherent characteristics of these size classes and is not an artifact of the limited food supply. 109 Statistics The F-test of significance in ANOVA assumes that test animals behave independently from each other (Sokal and Rohlf, 1969). Trout in groups do not behave independently--there was evidence for follow- the-leader behavior and avoidance behavior in this study. To the statistician, this is a problem as variance due to inherent charac— teristics of each trout and variance due to interactions between trout cannot be separated and quantified. Consequently, the alpha level of significance that corresponds to the assumption that trout behave independently cannot be determined. But the significance level ob- tained in ANOVA still reflects what actually occurred in the population sampled--this is the prime concern to the ethologist and biologist. Speculation as to how the significance level might have differed had trout been observed in isolation is relevant here only because trout population densities may be so low in some streams that social inter- action is infrequent during the non-breeding season. Indirect evidence for follow-the-leader behavior was strongest during feeding periods. Differences between social ranks in percent of time spent feeding, therefore, may actually be more significant than the alpha level indicates. Avoidance of dominant trout was associated with cover use but probably had little effect on the percent of time that each trout spent under cover. Avoidance of dominants under cover apparently determined which cover some trout used, but not the percent of time spent under cover. Trout were never excluded from both covers simultaneously in replicates included in statistical analyses. 110 In all activity categories, differences between rank-1 and rank-4 trout in percent of total movement and rate of movement were probably inflated by avoidance behavior. There was not clear evidence that the response of trout to each other within ah activity category differed between times of day. Differences between times of day pre- sumably would have been significant at the alpha levels obtained even if each trout had been observed in isolation. Replicate was considered a random factor in ANOVA as there was no reason a priori to expect large differences in trout behavior be- tween replicates. The physical environment was similar in each channel section. Usually the component of variance contributed by a random factor is estimated and expressed as a fraction of total variance. Mean differences between levels of a random factor are not normally contrasted as the difference cannot be attributed to a particular variable. The significance of mean differences between replicates was tested here for two reasons: (1) use of mean values in treatment combination cells precluded estimation of the random component of variance, and (2) consistent differences between group behavior (4 trout) in some replicates suggested that differences in the social en- vironment caused those differences. Activity_Budget This study agrees with earlier reports that adult brown trout spend most of their time under cover during daylight (Butler and Hawthorne, 1968; Brynildson et al., 1973; DeVore, 1975). This is especially true of trout larger than 35 cm (TL). Cover use, per se, involved almost no energy expenditure over maintenance demands. 111 Feeding was the second most time-consuming activity of trout in this study. It accounted for over 90% of the total estimated day- time energy expenditure. Over 90% of the energy loss involved with feeding resulted from maintaining a stationary position in the current (Table 26). Availability of low velocity feeding stations clearly could influence growth of stream salmonids. Agonism composed less than 1% of the time budget of all trout, and only 2-4% of total energy expenditure. It was the most energy- demanding activity per unit of time spent, however, so a samll in- crease in agonism could produce a large increase in energy expenditure. Factors that disrupt social stability could lead to more agonism and poorer trout growth. One such factor might be frequent stocking. Large trout expended only about 1/3 as much energy (mgOZ/kg/h) above maintenance demands as did small trout (Table 26). Large trout conserved energy by feeding only when drift food was very abundant or when large food items were provided artificially. In an environment where food items were too small to compensate for the energy required to capture and consume them and emigration to better areas was impossible, remaining motionless under protective cover was the best strategy for survival. The food supply was nearly sufficient for small trout. It was to their advantage to feed more often than large trout . Social Structure This study also agrees with earlier reports that social hierarchies are well-developed among adult stream salmonids (Newman, 1956; Jenkins, 1969; Butler, 1975). Each trout defended a space 112 surrounding itself rather than a fixed geographic area. This "social force field" (McBride, 1964) was defended successfully only against smaller individuals so that a social hierarchy developed. Social rank decreased from the largest to the smallest trout. The social force field was largest during feeding periods and smallest during cover use. Social hierarchies among fishes have been described as based on either "nip-dominance" or "nip-right" relationships (Braddock, 1945; Greenberg, 1947; Mryberg, 1972). In the nip-right hierarchy, dominants deliver attacks but are never attacked by lower ranked fish. In the nip-dominance society, dominants deliver more attacks than they receive from lower ranked individuals, but the outcome of a particular bout is not strictly predictable. Brown trout in this study demonstrated a nip-right hierarchy. Subordinates displaced higher ranked trout but this occurred only 4 times and probably resulted from mistaken identity. A high level of social stability has also been reported for other populations of adult stream salmonids (Newman, 1956; Jenkins, 1969). Apparent dominance reversals, (i.e. fish "losing" bouts to previously sub- ordinate fish) have been attributed to unsettled dominance early in hierarchy formation, partial territory, and mistaken identity (Greenberg, 1947). The 10-14-day acclimation period should have per- mitted stabilization of social interactions. Social hierarchies of brown trout have been observed to stabilize within 5 days (Jenkins, 1969). Partial territory involves defense of a single geographic area against all subordinates (Greenberg, 1947). A partial territory 113 holder frequently resists dominants so that determination of social rank can be difficult. Dominants that are persistent eventually dis- place partial territory holders. Persistent defense of a single area did not occur in this study. Each social rank typically moved between a few favored positions which were readily yielded to dominants. Mistaken identity of low-ranking trout by dominants seems to be the most likely cause of apparent dominance reversals in this study. Rank-1 trout exited from cover as rank-2 trout darted under the same cover on 3 occasions. However, rank 1 returned to cover and dis- placed rank 2 to the lower end of the cover almost immediately in each case. Once, rank 2 left cover as rank 3 entered. In this case rank 2 moved to the other cover and remained there. These were the only times observed where higher ranking fish were displaced by subordinates. The explanation for mistaken identity seems to be that in each case the approach to cover was very rapid. Rapid approach was also characteristic of trout about to attack subordinates. A conditioning experiment demonstrated that fish respond to behavior type, e.g. dominant or submissive, in addition to other physical aspects of in- dividual appearance (McDonald et al., 1968). Even the highest ranked trout in each replicate had probably experienced dominants in their native stream and would have been conditioned to submit to dominant be- havior. Thus, a subordinate motivated to use cover could be mistaken for an attacking dominant if it acted like one. If this hypothesis is correct, then all 4 cases of reversed dominance may be considered "accidents". The social structure of brown trout then conforms to the strict definition of a "nip-right" hierarchy. 114 Social structure reflects the distribution of required resources in the environment. Where resources are clumped in a few areas, competition for them favors hierarchy formation. Brown trout competed for feeding positions at some feeding references but the nearly uniform distribution of drift food and current velocity minimized competition for feeding stations. Most agonism associated with feed- ing was more the result of a general increase in movement and aggres- siveness associated with feeding rather than defense of a favored station. Conversely, competition for the upstream position under cover was very apparent and probably was the basis for hierarchy forma- tion. Trout tended to avoid some dominants under cover. Rank-4 trout, in particular, seemed to avoid dominants under small cover when that was possible (Tables 11, 12 and 13). In the wild, where emigration is possible, avoidance may influence the number and size of trout that inhabit an area of stream. This could be tested. On the other hand, subordinates were frequently "tolerated" by dominants under cover, especially under large cover. Several times, rank-l trout permitted small trout to feed from the lower end of their cover without agonism. Overall, coexistence was a more prominent aspect of cover use behavior than avoidance. A small area of cover will accommodate many more trout than the same sized area of feeding hab- itat. Reduction of aggression is commonly held to be the primary function of social hierarchies (Rowell, 1974). Brown trout in this study directed most aggressiveness towards fish ranked only one level 115 lower. This has been observed in other fish populations as well (Greenberg, 1947; Newman, 1956; Jenkins, 1969; Mryberg, 1972) and represents a reduction in aggression over the condition immediately following formation of a new group where aggression is directed with equal intensity towards all other fish (Jenkins, 1969). Small trout learn to avoid agonistic bouts by remaining downstream and distant from larger trout. Subordinate brown trout often yielded positions to approaching dominants before the dominant approached close enough to initiate a chase. The presence of dominants seems to inhibit the aggressiveness of low ranking hierarchy members towards their own subordinates. Agonistic bouts between small trout in replicates lacking large trout were more frequent and more intense than bouts between small trout in replicates with two large trout (Table 9). Jenkins (1969), cited two cases where removal of a dominant was followed by increased aggres- siveness by the highest ranking trout that remained. Social hierarchies appear to reduce agonistic activity among brown trout by reducing the aggressiveness of trout toward dominants and subordinates. This discussion so far has centered on the results of repulsive forces between trout. Attractive forces were also apparent in this study. This was shown in the tendency for stressed trout to aggregate in fright huddles and in other forms of synchronous behavior. Synchronous behavior provided evidence for social facilitation which is the tendency for the performance of a pattern of instinctive be- havior by one animal to act as a releaser for the same behavior in another (Thorpe, 1963). The stimulus "dominant feeding" was not 116 essential to elicit feeding in lower ranking trout but seemed to en- courage it. Small trout in replicates where large trout usually fed during the morning (IA2, IIA2, IIBl) fed more during that period than small trout in replicates where dominants rarely fed at that time (Appendix Table A22). Advantages of Dominance Dominance confers the right to occupy the most upstream posi- tion under cover. An advantage of this position might be first opportunity to consume drift food--especially large food items. Butler (1975) reported that dominant trout feeding under the upstream end of cover fed primarily on large food items, while subordinates feeding farther downstream consumed small items. Dominant trout ob- served by Jenkins (1969) fed at the most upstream position in favored feeding lanes. These fish frequently had the highest volume of drift food in their stomachs, but occasional extreme values among lower ranking trout prevented detecting any statistically significant dif- ferences. Dominance seems to have permitted rank-l trout in this study to use less energy in cover use and agonism than subordinates (Table 26). Rank 1 was the only rank not subject to displacement by other trout. Rank-l trout also used less energy in agonism than subordinates. This occurred even though rank 1 was involved in at least as many bouts as ranks 2 and 4. Subordinates may have yielded positions to rank 1 more readily than to lower ranking dominants. The feeding positions held by dominant trout in Jenkins' (1969) study were closer to cover than subordinates' positions. This would 117 seem to have obvious survival value. But ranks 1 and 2 in this study did not feed closer to overhead cover than subordinates. Position of overhead cover seemed to have very little influence on choice of feed- ing positions by most trout. Of much greater influence were the channel walls. Large trout fed half as far from the walls as from overhead cover. Small trout fed about the same distance from the walls as from overhead cover (Figures 10 and 13). Differences between Jenkins' study stream and the stream used in this study may explain the conflicting results. In Jenkins' com- plex natural environment, response to relative current velocity and spatial patterns of drift food abundance could have easily been con- founded with response to cover. Overhead cover, prominent lateral surfaces, and favored feeding stations in Jenkins' stream were mostly near the outside of meander bends. Flow resistance of the streambank at meander bends produces low velocity stations. Principle lines of drift occur close to the low velocity areas here. Dominant trout may have defended these positions due to the combination of low current velocity and high drift food abundance, not because they were close to cover per se. In this study, cover did not produce a velocity buffer and principal lines of drift did not exist. Choice of feeding stations was potentially based on proximity to overhead cover and lateral sur- faces, which could be clearly distinguished from each other. I con- clude that dominance does not imply feeding close to overhead cover. The advantages of dominants' positions cannot be determined without simultaneous consideration of all factors that potentially influence survival and growth. 118 Lateral Surfaces Large brown trout fed about half as far, on the average, from the channel walls as did small trout (Figure 13). Small trout occa- sionally moved to wall positions vacated by large trout, but more often entirely different mid-channel positions were used. I infer from the lack of competition between large and small trout for wall feeding positions that large trout have a greater behavioral need for lateral surfaces than small trout. There was very little evidence that use of feeding positions close to the walls by large trout reflected stronger preference for low-velocities or for shade than small trout. Greater use of wall positions by large trout than by small trout suggests that thigmotaxis becomes better developed as brown trout age but the mechanisms involved are not clear from this study. Greater use of wall positions by large brown trout could also reflect more experience with predators. Alexander (1976) reported that a substantial portion of the brown trout population in the North Branch of the AuSable River, Michigan in an area just upstream from the source of the brown trout used in 1976 was removed by avian predators. Close association with the streambank may provide large trout with a good source of large terrestial food items, as well as protection from predators. The relative amounts of visual use of low- and‘high-stripe lateral devices in 1975 was considered to indicate the importance of the visual proximity of a distinct landmark to the stream bed. Re- sults indicate that a visual reference 15 cm above the stream bed is just as attractive to brown trout as the same-sized reference on the stream bed. "Tactile" use of lateral cover devices was considered 119 to indicate the importance of lateral visual concealment. Only trout pressed sideways against the low-stripe devices were visually isolated from their surroundings. Preference for the low-stripe devices (tactile use) indicates that visual isolation is an important function of lateral surfaces to adult brown trout. Size of the visual stimulus offered by lateral cover also determines its attractiveness to brown trout. Use of posi- tions close to the channel wall was over 3 times as great as use of positions that permitted trout to press sideways against lateral cover devices (Appendix Figures B1 and B2). This occurred even though direct contact with the walls was not possible. The large visual stimulus offered by the walls was apparently more attractive than the small visual and tactile stimuli offered by lateral cover devices. Tactile stimuli were most important to trout that were stressed. Prior to acclimation, trout pressed sideways against screens or other trout in fright huddles rather than using the shade of over- head cover. In experiments with only lateral cover devices, frightened trout pressed against other trout rather than the lateral cover de- vices. This was surprising, as the tactile and visual stimulus offered by the low-stripe devices seemed comparable to that offered by the bodies of other trout. Attractive forces between brown trout become very strong in stress situations. Light Light intensity affected brown trout position choice very strongly in this study. Trout avoided direct sunlight by using posi- tions shaded by overhead cover or by the channel walls. Several trout 120 fed on the east side of the channel in the morning and on the west side in the evening to stay in the shade. Rank 1 in replicate IIB2 left cover to feed only before sunrise and on a cloudy morning. Several large and small trout consistently did not begin intense evening feeding until after sunset. This was not visibly associated with a sudden increase in the abundance of surface drift food. In a preliminary experiment, fine organic sediment deposited on the stream bed during the day was commonly swept away in several well- defined areas at night. This was clearly the result of trout activ- ity and suggests that cover is not used as such when light level is low, though it may still serve as a landmark for orientation. Less use of opaque overhead cover by trout when water was turbid (1975) may have been the result of less light penetration and, therefore, less need for shade. The response to overhead cover is largely a response to relative light level. DeVOre (1975) suggested that brown trout pre- ference for opaque overhead cover "low", rather than "high" in the water column was related to its closer overhead visual proximity to the trout's stream bed position. Failure to prefer transparent over- head plates low in the water column in this study confirms that the response to overhead cover is primarily visual rather than tactile. But an alternative interpretation of DeVore's results is that pre- ference for low covers was related to the lower light level under- neath them (Table 5). This was indicated in this study by preference for overhead cover with dark rather than light substrate. Dark sub- strates reflect less light than light substrates, and result in lower 121 light intensity under the cover. Stewart (1970) concluded that use of overhead cover by wild rainbow trout was related to light level under- neath. Use of overhead cover increased with increasing structure size, decreasing structure height above the substrate, and decreasing percent holes punched into the structure. Dominant brook, brown, and rainbow trout observed by Butler and Hawthorne (1968) preferred square overhead cover that was large (91 x 91 cm) or medium-sized (61 x 61 cm) rather than small (30 x 30 cm). A large (40.2 cm TL) brown trout used the shade of large and medium-sized overhead covers 80.6% of the time when located in the shade, but only 4.0% of the time when in direct sunlight. Clearly the response here was to relative light level rather than to size of overhead cover. Brown trout in this study did not prefer large (long) over- head cover to small (short) overhead cover. This might seem to con- tradict the results just cited (Butler and Hawthorne, 1968), but might be explained by the relationship between cover shape and light in- tensity beneath cover. Light intensity beneath long (25 x 122 cm) cover in this study was the same as that beneath short (25 x 61 cm) cover (Table 5). Therefore, both types of cover were equally attractive to trout in terms of light intensity beneath the covers. In contrast, Butler and Hawthorne (1968) observed a strong preference by brook, brown, and rainbow trout for large (91 x 91 cm) and medium (61 x 61 cm) square covers rather than small square (30 x 30 cm) cover. Light intensity was lower under the larger covers (Table 5), probably accounting for the consistent preference by their trout for larger covers. A consistent preference for large cover by brown trout 122 in this study would have been expected had the choice been between a wide (50 cm) and a narrow (25 cm) cover of the same length (61 cm). Light intensity is much lower beneath wide than narrow covers 61 cm long (Table 5). A study in progress is to investigate in more detail the influence of cover shape on light level and brown trout preference (J.C. Gruber, in preparation). Somewhat confusing is the observation by DeVore (1975) that brown trout showed no preference when offered opaque overhead covers with either clear or dark plastic streamers along both sides. Dark streamers presumably lowered the light level under cover and would have been expected to increase its attractiveness. Dark streamers also ob- structed lateral vision. DeVore suggested that obstruction of lateral vision could be an undesirable feature of cover because it might inhibit detection of food and predators. But brown trout in the wild frequently use positions where tree roots, and the stream bank obstruct lateral vision on one or both sides (E. Branch AuSable River, Pigeon River, Michigan; Stony Kill, New York, C.E. Bassett, pers. obs.). Lateral concealment from predators would seem to be a very desirable feature of cover even if it does obstruct vision. Brown trout in both parts of this study showed a strong attraction to lateral surfaces. Trout usually responded to drift food while it was still well upstream from their position. Lateral obstruction of vision, therefore, should not inhibit food detection. Perhaps the lateral tactile reference offered by DeVore's streamers was so attractive that the reduction in light intensity produced by dark streamers was relatively insignificant. The majority of the 123 evidence indicates that the primary function of overhead cover for brown trout is concealment from high light intensity. This in turn provides protection from predators. Current Velocity The experimental environment in both parts of this study was designed to minimize the influence of current velocity on position choice. Mean current velocity at feeding positions was nearly the same for all social ranks. Velocities greater than sectional means were avoided (range of means 20.4—24.3 cm/s). Correlations between percent of trout using cover devices and mean sectional current veloc- ity in 1975 were very weak or absent. Conversely, choice of cover in 1976 seemed to be influenced by current velocity. Only two of 8 rank-1 trout preferred cover with the highest current velocity underneath. Lacking large differences in light intensity under cover the trout may have been particularly sensitive to small differences in current velocity on the order of 3-9 cm/s. Minor irregularities in stream bed topography under cover also could have influenced cover choice. Trout in both years consistently used positions beside or immediately downstream from pieces of twig or stones protruding less than 2.5 cm above the stream bed. Use of shallow depressionsixlthe stream bed was common in 1975 in the absence of overhead cover. Similar behavior has been observed among brown trout in other studies (Jenkins, 1969; DeVore, 1975). In 1976 two rank-l trout preferred overhead cover with the highest current veloc- ity. This may have reflected attraction to stream bed irregularities too minor to affect velocity measurements with a pygmy Gurley meter. 124 Application to Stream Management This study and the review of literature in the introduction indicates that proper cover microhabitat for adult brown trout in streams has the following characteristics: (1) Low current velocity. Observations by Wesche (1976) in- dicate that velocities near zero are preferred. (2) Low light intensity. Except when drift food was very abundant, trout spent most of their time from sunrise to sunset under cover. Light intensity approaching nighttime levels seems to be pre- ferred; that is, the darker it is under cover, the better. (3) Overhead surfaces close to the stream bed. Distance between overhead cover and the stream bed should just exceed the body depth of the largest trout likely to use the cover. (4) Lateral surfaces at the stream bed. This provides visual con- cealment from predators and competitors and offers tactile reference. (5) Placement close to principal lines of drift. This minimizes energy use in movement between cover and feeding stations and min- imizes exposure to predators. Characteristics of proPer feeding microhabitat for adult brown trout in streams are: (1) Low current velocity closely adjacent to a principal line of drift. (2) Shade from direct sunlight. (3) Nearness to the stream bank or other extensive lateral sur- face. This is especially important for trout over 30 cm long. 125 The influence of cover location on choice of feeding positions probably depends on size of stream. In the confined channels of this study, trout could not move more than about 6 m from overhead cover, and there was no tendency for trout to feed very close to cover. Pro- minent channel walls and extensive shade during times of peak feed- ing activity offered sufficient security. I infer that in small streams where shade from bank vegetation is usually abundant and where trout are never far from the stream banks, utilization of potential feeding stations is not often limited by location of overhead cover. Cover location probably has more influence on choice of feeding sta- tions in wide streams where a relatively large portion of the channel is sunlit for much of the day. The shade offered by overhead cover along the banks in large streams represents much of the total avail- able shade for feeding. Low-ranking trout that could not defend available feeding stations close to cover would have to either expose themselves to predators in the sunlit water or not feed at all until shaded stations became available. Either situation would limit trout biomass. Creation of additional overhead cover in the middle of large streams should encourage better utilization of the food supply by offering additional shaded feeding stations, provided current velocity and drift food abundance were also suitable at these locations. Habitat manipulations that produce a deeper, narrower channel may be especially successful (Hunt, 1971; White, 1975) because drift food supply is concentrated, cover and all potential feeding stations are brought closer together, a greater portion of the channel is shaded, and the stream bank becomes relatively more prominent. These factors 126 encourage maximum food intake with minimum energy expenditure and minimize exposure to predators. Carrying capacity of cover is apparently related to its length. Trout coexist indefinitely under cover as long as subordinates remain slightly downstream from dominants. This study suggests that an approximation to cover capacity might be that each trout requires a length of cover equal to its body length. Therefore 6 linear meters of overhead cover could support 20 30-cm trout. Electrofishing ob- servations in the AuSable River system indicate that this is a reason- able figure. Increasing cover width tends to make cover more attrac- tive by reducing light level, but probably does not increase the number of trout that can coexist under it unless more lateral visual concealment is provided (Kalleberg, 1958; Hartman, 1963). Chapman (1966) suggested that stream salmonid populations are limited by the spatial demands of feeding and that cover is not as important. This is consistent with the results of this study. Space requirements of feeding trout in 1976 were roughly 6.5 m2 or about 2.5 linear meters of channel length per trout. This estimate is based on the distance at which trout reacted aggressively toward subordinates during periods of peak food abundance. Small trout seemed to have about the same feeding space requirements as large trout. Six linear meters of feeding habitat would support 2 or 3 trout compared to 20 trout of the same size for the same length of overhead cover. Wide streams might have two or more principal lines of drift which would permit a few more trout to feed in this length of stream. Streams where drift food abundance is higher than in the channels used in this 127 study also might support a few more trout per unit length of feeding habitat as feeding space requirements are apparently reduced where food is very abundant (Mason and Chapman, 1965). But, it is readily apparent that creating more cover space can increase stream carrying capacity (numerical abundance of trout) during the summer only to a limit imposed by the availability of space suitable for feeding. The fact that abundance of brown and brook trout is closely related to amount of cover in streams (Boussu, 1954; Lewis, 1969; White, 1975; Enk, 1977), where studied, does not contradict this. Rather, it shows that there is often not enough cover to support the number of trout that could fully utilize the available feeding habitat. The suggestion by DeVOre (1975) that creation of large covers near good feeding habitat might cause over-utilization of the food supply seems unlikely, at least over a long period. Large covers might initially attract more trout than the adjacent feeding habitat could support. But agonism would disperse lowbranking trout into areas where food abundance was insufficient to meet growth or maintenance demands--or into areas readily accessible to predators (Ondera, 1962; Symons, 1968). These conditions would cause emigration and mortality that would tend to bring the population back into equilibrium with the availability of high quality feeding habitat. The problem of too many trout for the available feeding habitat would persist only if the population in the stream as a whole were so unstable that immigrants continued to replace trout lost via emigration and mortality. Fre- quent over-stocking might cause this. 128 Stream habitat manipulation involves consideration of all constituents of trout living space including cover, feeding, and spawning microhabitats. Attempts to increase the availability of one type of microhabitat must not result in a shortage of another type. Stream morphology and the spatial distribution of required micro- habitats can probably be manipulated to encourage more efficient utilization of stream resources by trout. Position choice results from responses to a few simple physical stimuli, but social factors can modify these responses. Social factors have the most influence where limited availability of essential microhabitat leads to strong competition. Further study is needed to determine the influence of the social environment on spatial demands of wild trout. It may be possible to modify hierarchy size or the body size composition of hierarchy members to minimize energy wastage in agonism and to en- courage more efficient use of the food supply. This might be achieved with special fishing regulations and habitat manipulation. Living space requirements at night also deserve much study. Indirect evidence in this study suggests that brown trout do not use cover at night. This needs to be verified. Two other questions must be answered: (1) Will brown trout feed farther from cover (overhead and lateral) at night than during the day? (2) Are feeding space requirements smaller at night than during the day? This is especially of interest for adult brown trout as they are nocturnal feeders. 129 If the answer to these questions is "yes", then distribution of cover in the channel cross-section does not potentially limit use of feed- ing microhabitat as much as daytime observations have suggested. Winter living-space requirements of adult stream trout also must be defined before habitat can be managed comprehensively. The primary function of instream structure that provides concealment from direct sunlight during the summer seems to be protection from severe winter ice and floods on some streams (Ondera, 1962; Hartman, 1963; Hunt, 1969; Bustard and Narver, 1975). SUMMARY Daytime position choice by wild, adult brown trout was ob- served in controlled-flow stream channels during the summers of 1975 and 1976. In 1975 25-30-cm (TL) trout in groups of 5, with 4 groups tested simultaneously, were offered pairs of cover types. Occupa- tion of a specific type represented a choice between known stimuli. The response to overhead cover was primarily visual rather than tactile. In contrast to previously demonstrated preference for opaque overhead cover 10 cm rather than 15 cm above the stream bed, the trout showed no preference when offered transparent overhead cover devices 10 cm and 15 cm above the stream bed. The trout responded to light reflected from the stream bed beneath opaque overhead cover. Cover with dark stream bed beneath was preferred to cover with light stream bed beneath (p < 0.01). Visual isolation is an important function of lateral vertical surfaces. The trout preferred to press sideways against lateral cover devices with a S-cm—wide black stripe on the stream bed rather than devices with the stripe raised 15 cm above the stream bed (p < 0.001). Trout that were not pressed against lateral devices showed no pre- ference; this indicated that the visual reference offered by the stripe on the stream bed was no more attractive as a landmark than the stripe above the stream bed. 130 131 In 1976, 6 groups (replicates) consisting of 2 small (23-28 cm FL) and 2 large (35-41 cm FL) trout were observed from elevated blinds. Five other groups of 1-3 trout were also observed. One small (25 x 61 cm) and one large (25 x 122 cm) opaque overhead cover was provided for each group. The trout developed a strong social hierarchy. Heavier trout held higher rank. Each trout defended a space around it- self which moved as the trout moved rather than defending a fixed area of stream bed. Space upstream from the trout was defended more persistently than space downstream from the trout. Feeding trout de- fended much more space (about 6.5 m2) than trout using cover (about 1 m2). The function of the social hierarchy appeared to be reduction of aggressiveness towards subordinates and dominants. Agonism be- tween small trout in the absence of large trout was more frequent and intense than when large trout were present. In groups of 2 large and 2 small trout, most aggressiveness was directed toward trout of the next lowest rank. Rank-l trout used cover more than rank-3 and -4 trout (p < 0.01) and fed less than rank-3 and -4 trout (p < 0.01). Estimated total energy use (mgOZ/kg/h) above standard metabolism was lowest for rank-l trout (8.25) and highest for rank-4 trout (27.02). This appeared to be related to relative social status. Superior status permitted rank-l trout to use less energy in agonism and in movement between covers. Feeding accounted for 92.6% of the total estimated energy use by all trout combined. Over 90% of this was expended in maintaining a stationary position in the current. Agonism was the most energy-consuming activity per unit of active time. 132 Cover use patterns were influenced by current velocity be- neath cover,by cover size, and by social rank. Individual cover pre- ference was based at least as much on current velocity beneath cover as on cover size. Five of 8 rank-l trout preferred the cover with the lowest current velocity beneath; 2 preferred the cover with the fastest current velocity beneath. Only 3 rank-l trout preferred the largest cover available. Light intensity was the same under the large and small covers because the covers were of equal width. Social rank strongly influenced choice of cover because com- petition for the upstream position under cover was strong and the area of cover was small. Subordinates preferred the same cover as rank-l trout, but they often did not express that preference due to avoidance of dominants. The attractiveness of low current velocity beneath cover did not compensate for space limitations imposed by small cover size. Large cover received at least as much total use as small cover even in replicates where small cover was preferred. Much of the heavy use of large cover in replicates where small cover was preferred was due to avoidance of rank-1 trout by rank-3 and -4 trout and avoidance of rank-2 trout by rank-4 trout. Rank-3 and -4 trout considered together avoided dominants more in replicates where small cover was preferred than in replicates where large cover was preferred (p < 0.01). Trout differing by only one social rank coexisted about as often under small cover as under large cover. Current velocity and light level strongly influenced choice of feeding positions. The trout avoided feeding at stations where current velocity exceeded the channel section mean (p < 0.001; range 133 of means 20.4-24.3 cm/s). Mean current velocity and depth at feeding stations was nearly the same for all social ranks. All social ranks fed more on the shaded side of the channel than on the sunlit side of the channel (p < 0.005). High-ranking trout did not feed closer to overhead cover than subordinates. But rank-1 and -2 trout fed closer to the channel walls than rank-3 and -4 trout (p < 0.01). Large trout rarely used mid-channel feeding stations with current velocities comparable to wall-side stations, and small trout usually did not move to wall-side stations vacated by dominants. The trout spent less time per feeding station during the morn- ing when surface drift food was most abundant than during midday and evening (p = 0.005). This resulted from more agonism during the morn- ing than during midday and evening (p < 0.01). The nearly uniform distribution of current velocity and drift food in the channels (in contrast to more concentrated distribution in natural streams) minimized competition for feeding stations. Consequently, social rank had only minor influence on choice of feeding stations. This study and a review of the literature indicated that the characteristics of proper cover microhabitat for adult brown trout in streams are: (1) low current velocity, (2) low light intensity, (3) overhead surfaces close to the stream bed, (4) lateral surfaces at the stream bed, and (5) placement close to principal lines of drift. Characteristics of proper feeding microhabitat for adult brown trout are: '(1) low current velocity, (2) shade from direct sunlight, and (3) nearness to extensive lateral surfaces. LITERATURE CITED LITERATURE CITED Allen, K.R. 1969a. Distinctive aspects of the ecology of stream fishes: a review. J. Fish Res. Ed. Canada 26:1429-1438. . 1969b. Limitations on production in salmonid populations in streams. Pages 3-18 in_T.G. Northcote, ed. Symposium on Salmon and Trout in Streams. H.R. MacMillan Lectures in Fisheries, Univ. Brit. Col., Vancouver. Baldes, R.J. and R.E. Vincent 1969. Physical parameters of micro- habitats occupied by brown trout in an experimental flume. Trans. Am. Fish. Soc. 98:230-238. Beamish, F.W.H. 1964. Respiration of fishes with special emphasis on standard oxygen consumption II. Influence of weight and temperature on respiration of several species. Canadian J. 2001. 42:177-188. Boussu, M.F. 1954. Relationship between trout populations and cover on a small stream. J. Wildl. Mgt. 18:229-239. Braddock, J.C. 1945. Some aspects of the dominance-subordination re- lationship in the fish Platypoecilus maculatus. Physiol. 2001. 18:176-195. Brett, J.R. 1965. The relation of size to rate of oxygen consumption and sustained swimming speed of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Ed. Canada 22:1491-1501. Brynildson, O.M., V.A. Hacker, and T.A. Klick 1973. Brown trout life history, ecology and management. Pub. 234, Wis. Dept. Nat. Res., Madison. 16 pp. Bustard, D.R. and D.W. Narver 1975. Aspects of the winter ecology of juvenile coho salmon (Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Ed. Canada 32:667-680. Butler, R.L. 1975. Some thoughts on the effects of stocking hatchery trout on wild trout populations. Pages 83-87 in_W. King, ed. Proc. Wild Trout Management Symposium. Trout Unlimited, Inc. Butler, R.L. and V.M. Hawthorne 1968. The reactions of dominant trout to changes in overhead artificial cover. Trans. Am. Fish. Soc. 97:37-41. 134 135 Chapman, D.W. 1966. Food and space as regulators of salmonid popula- tions in streams. Am. Naturalist 100:345-357. Chapman, D.W. and T.C. Bjornn 1969. Distribution of salmonids in streams with special reference to food and feeding. Pages 153-176 in_T.G. Northcote, ed. Symposium on Salmon and Trout in Streams. H.R. MacMillan Lectures in Fisheries, Univ. Brit. Col., Vancouver. Chaston, I. 1968. Influence of light on activity of brown trout Salmo trutta. J. Fish. Res. Bd. Canada 25:1285-1289. . 1969. Seasonal activity and feeding pattern of brown trout (Salmo trutta) in a Dartmoor stream in relation to availability of food. J. Fish. Res. Bd. Canada 26:2165—2171. Cooper, E.L. 1952. Rate of exploitation of wild eastern brook trout and brown trout populations in the Pigeon River, Otsego County, Michigan. Trans. Am. Fish. Soc. 81:224-234. DeVore, P.W. 1975. Daytime behavioral responses of adult brown trout (Salmo trutta) in stream channels. M.S. Thesis. Michigan State Univ., E. Lansing. 38 pp. Dickson, I.W. and R.H. Kramer, 1971. Factors influencing scope for activity and active and standard metabolism of rainbow trout (Salmo gairdneri). J. Fish. Res. Ed. Canada 28:587-596. Elliot, J.M. 1967. Invertebrate drift in a Dartmoor stream. Arch. Hydrobiol. 63:202-237. . 1970. Diel changes in invertebrate drift and the food of trout Salmo trutta L. J. Fish. Biol. 2:161-165. . 1972. Rates of gastric evacuation in brown trout Salmo trutta L. Freshwat. Biol. 2:1-18. Enk, M.D. 1977. Instream overhead bank cover and trout abundance in two Michigan streams. M.S. Thesis. Michigan State Univ., E. Lansing. 127 pp. Everest, E.E. and E.H. Edmundson 1967. Cold-branding for field use in marking juvenile salmonids. Prog. Fish-Cult. 29:175-176. Gill, J.L. Design and analysis of experiments in the animal sciences. volume 1. Iowa State Univ. Press, Ames, Iowa. (In press). Greenberg, B. 1947. Some relations between territory, social hier- archy, and leadership in green sunfish (Lepomis cyanellus). Physiol. 2001. 20:267-299. 136 Hartman, G.F. 1963. Observations on behavior of juvenile brown trout in a stream aquarium during winter and spring. J. Fish. Res. Bd. Canada 20:769-787. Hill, R.W. 1976. Comparative physiology of animals an environmental approach. Harper and Row, New York. 656 pp. Houston, A.H., J.A. Madden, R.J. WOods, and H.M. Miles.l97l. Some physiological effects of handling and tricaine methane sulfonate anesthezation upon brook trout Salvelinus fontinalis. J. Fish. Res. Ed. Canada 28:635-642. Hunt, R.L. 1969. Effects of habitat alteration on production, standing crops and yield of brook trout in Lawrence Creek, Wisconsin. Pages 281-312 in_T.G. Northcote, ed. Symposium on Salmon and Trout in Streams. H.R. MacMillan Lectures in Fisheries, Univ. Brit. Col., Vancouver. . 1971. Responses of a brook trout population to habitat de- velopment in Lawrence Creek. Tech. Bul. 48 Wis. Dept. Nat. Res., Madison. 35 pp. Jenkins, T.M., Jr. 1969. Social structure, position choice and micro- distribution of two trout species (Salmo trutta and Salmo gairdneri) resident in mountain streams. Anim. Behav. Monogr. 2:57-123. Kalleberg, H. 1958. Observations in a.stream tank of territoriality and competition in juvenile salmon and trout (Salmo salar L. and S, trutta L.) Inst. Freshwat. Res. Drottningholm Rept. 39:55-98. Kwain, W. and H.R. MacCrimmon 1967. The behavior and bottom color selection of the rainbow trout, Salmo gairdneri, Richardson, exposed to different light intensities. Anim. Behav. 15: 75-78. Lemmien, W.A., P.I. Tack, andW.F. Morofsky 1957. Results from planting brown trout and rainbow trout in Augusta Creek, Kalamazoo County, Michigan. Mich. Agr. Exp. Sta. Quart. Bul. 40:242-249. Lewis, S.L. 1969. Physical factors influencing fish populations in pools of a trout stream. Trans. Am. Fish. Soc. 98:14-19. Marshall, T.L. and H.R. MacCrimmon 1970. Exploitation of self-sustain- ing Ontario stream populations of brown trout (Salmo trutta) and brook trout (Salvelinus fontinalis). J. Fish. Res. Bd. Canada 27:1087-1102. Mason, J.C. and D.W. Chapman 1965. Significance of early emergence, environmental rearing capacity, and behavioral ecology of juvenile coho salmon in stream channels. J. Fish. Res. Bd. Canada 22:173-190. 137 McBride, G. 1964. A general theory of social organization and be- havior. U. Queensland Fac. Pap. Vet. Sci. 1:75-110. McCrimmon, H.R. and W. Kwain 1966. Use of overhead cover by rainbow trout exposed to series of light intensities. J. Fish. Res. Bd. Canada 23:983-990. McDonald, A.L., N.W. Heimstra, and D.K. Damkot 1968. Social modifica- tion of agonistic behavior in fish. Anim. Behav. 16:437-441. Mryberg, A.A. 1972. Social dominance and territory in the bicolor damsel fish, Eupomacentrus partitus. Behav. 41:207-231. Newman, M.A. 1956. Social behavior and interspecific competition in two trout species. Physiol. Zool. 29:64-80. Ondera, K. 1962. Carrying capacity in a trout stream. Bul. Freshwat. Fish. Res. Lab., Tokyo 12:1-41. Rao, G.M.M. 1968. Oxygen consumption of rainbow trout (Salmo gairdneri) in relation to activity and salinity. Can. J. Zool. 46: 781-786. Ringler, N.H. 1975. The mechanism of selective predation in drift feeding brown trout (Salmo trutta). Ph.D. Dissertation. Univ. of Michigan, Ann Arbor. 209 pp. Ritter, J.A. and H.R. MacCrimmon 1973. Effects of illumination on behavior of wild brown trout (Salmo trutta) and rainbow trout (Salmo gairdneri) exposed to black and white backgrounds. J. Fish. Res. Bd. Canada 30:1875-1880. Rowell, T.E. 1974. The concept of social dominance. Behav. Biol. 11:131-154. Serchuk, F.M. 1976. The effects of the Ludington pumped-storage power project on fish passage through pump-turbines and on fish be- havioral patterns. Ph.D. Dissertation. Michigan State Univ., E. Lansing. 130 pp. Shetter, D.S. 1950. The relationship between the legal-sized trout population and catch by anglers in portions of two Michigan trout streams. Pap. Mich. Acad. Sci. Arts Lett. 34:97-107. Sokal, R.R. and F.J. Rohlf 1969. Biometry. W.H. Freeman and Co., San Francisco. 776 pp. Stewart, P.A., 1970. Physical factors influencing trout density in a small stream. Ph.D. Dissertation. Colorado State Univ., Fort Collins. 88 pp. In Dissertation Abstracts International 31(7):4401-B. Abstract no. 71-2432. 138 Symons, P.E.K. 1968. Increase in aggression and in strength of the social hierarchy among juvenile Atlantic salmon deprived of food. J. Fish. Res. Bd. Canada 25:2387-2401. 'Thorpe, W.H. 1963. Learning and instinct in animals. Harvard Univ. Press, Cambridge. 558 pp. Thorpe, L.M., H.J. Rayner, and D.A. Webster 1947. Population deple- tion in brook, brown and rainbow trout stocked in the Black- ledge River, Connecticut in 1942. Trans. Am. Fish. Soc. 74: 166-187. Vincent, R.E. 1969. The tolerance of water velocity by trout as a basis for enhancement of the stream fishery. Proc. West. Assoc. State Game Fish Comm. No. 49:188-190. Ware, D.M. 1972. Predation by rainbow trout (Salmo gairdneri): the influence of hunger, prey density, and prey size. J. Fish. Res. Bd. Canada 29:1193-1201. Warren, C.E., J.H. Wales, G.E. Davis, and P. Doudoroff.l964. Trout production in an experimental stream enriched with sucrose. J. Wildl. Mgt. 28:617-660. Waters, T.F. 1972. The drift of stream insects. Am. Rev. Entomol. 17:253-272. Wedemeyer, G. 1970. Stress of anesthesia with M.S. 222 and benzocaine in rainbow trout (Salmo gairdneri). J. Fish. Res. Bd. Canada 27:909-914. Wesche, T.A. 1976. Development and application of a trout cover rating system for IFN determinations. Pages 224-234 in_J.F. Orsborn and C.H. Allman, ed. Proc. of the Symposium and Speciality Conference on Instream Flow Needs. Volume II. American Fisheries Society, Bethesda, Maryland. White, R.J. 1975. In-stream management for wild trout. Pages 48-58 in_W. King, ed. Proc. Wild Trout Management Symposium. Trout Unlimited, Inc. White, R.J. and O.M. Brynildson 1967. Guidelines for management of trout stream habitat in Wisconsin. Tech. Bul. 39. Wis. Dept. Nat. Res., Madison. 65 pp. APPENDICES APPENDIX A TABLES Table A1. - Daily maximum and minimum water.temperatures for the East Branch of the AuSable River at the DNR Grayling Field Office during summer 1975. Temperature C Temperature C Date High Low Date High Low 8/11 17.6 12.9 8/27 16.2 13.4 12 17.6 14.2 28 13.4 11.5 13 18.8 16.2 29 12.9 11.8 14 17.6 14.3 30 12.9 12.9 15 17.9 14.0 31 12.9 12.3 16 18.5 14.0 9/ 1 14.0 12.9 17 18.5 15.1 2 14.0 12.4 18 16.0 12.6 3 10.4 9.0 19 14.0 11.8 4 12.0 9.5 20 12.9 11.5 5 12.9 10.0 21 14.0 12.3 6 13.4 11.8 22 14.0 12.9 7 11.8 10.6 23 12.3 11.5 8 11.5 10.4 24 15.1 12.3 9 11.5 8.4 25 17.0 14.8 10 11.8 9.0 26 17.0 14.6 11 12.9 11.5 12 11.8 9.2 a . . Water temperature was not measured during the first 7 days of the study (8/4 - 8/10). Table A2. - Daily maximum and minimum water temperatures for the East Branch of the AuSable River at the DNR Grayling Field Office during summer 1976. Temperature C Temperature C Date High Low Date High Low 6/22 18.2 14.6 7/24 19.9 15.7 23 19.0 16.0 25 18.8 15.1 24 16.2 14.8 26 18.5 15.7 25 19.3 14.6 27 20.4 16.0 26 19.6 15.1 28 15.7 15.1 27 19.3 15.1 29 18.5 14.3 28 16.2 14.8 30 19.0 15.1 29 15.4 14.3 31 18.5 15.7 30 14.6 13.4 8/ 1 16.2 13.4 7/ 1 16.0 12.3 2 17.1 12.3 2 15.7 13.1 3 16.5 13.2 3 18.5 13.4 4 15.7 13.2 4 18.8 14.0 5 14.8 14.0 5 19.9 14.6 6 16.2 12.3 6 19.6 14.8 7 16.2 11.8 7 16.8 16.0 8 17.1 12.3 8 18.8 13.4 9 17.9 13.4 9 18.5 14.3 10 16.0 14.0 10 20.2 16.0 11 18.2 14.0 11 21.8 17.9 12 20.2 16.8 12 19.3 15.4 13 18.8 16.8 13 17.9 13.4 14 15.7 15.1 14 20.4 14.8 15 15.1 11.8 15 20.2 16.2 16 16.0 11.2 16 17.6 16.2 17 16.8 11.8 17 17.4 13.2 18 17.1 12.6 18 17.9 13.2 19 19.6 15.1 19 18.5 14.8 20 20.2 16.2 20 17.6 16.2 21 20.2 16.2 21 18.2 13.7 22 19.6 15.4 22 19.0 14.3 23 19.0 14.8 23 20.2 16.2 24 18.2 14.0 25 16.8 14.0 141 Table A3. - Sex, social rank, and the weight and fork length of brown trout in each replicate at the beginning and end of each 3-week experiment. w h eig t (g) Fork length (mm) Social Replicate Begin End Begin End Sex rank 1A1 640 560 392 392 M 1 550 480 380 380 F 2 228 210 268 268 F 3 IA2 465 440 350 350 F 1 418 354 339 339 F 2 228 215 276 276 M 3 124 126 229 229 M 4 181 242 223 267 267 F 1 166 153 249 249 M 2 182 172 158 258 258 Immature a 141 133 234 234 M a IIAl 535 485 357 357 M 2b 455 430 359 359 F 1 264 253 290 290 M 3 120 119 228 228 F 4 IIA2 620 590 388 388 M 1 565 530 382 382 F 2 187 187 255 255 F 3 153 140 238 238 M 4 IIBl 600 y50 389 389 F 1 440 410 359 359 M 2 242 240 288 288 F 3 190 192 263 263 F 4 I182 610 555 386 386 M 1 550 495 359 359 F 2 181 180 245 245 M 3 153 140 238 238 M 4 IIIAl 535 485 348 348 M 1 395 380 340 340 M 2 190 188 260 260 M 3 137 134 234 234 F 4 IIIA2. 200 176 262 262 M IIIB2 284 213 276 276 M 2C 153 156 241 241 M 1 aNo agonistic bouts observed, rank could not be determined. bAbnormal behavior but no injury was apparent. c . Injured. 142 Table A4. - Time and number of observation sessions devoted to each replicate during morning, midday, and evening. Mean Morning Midday Evening Total length of Repli-a Time No Time No Time No Time No observations cate (min) obs. (min) obs. (min) obs. (min) obs. (min) IAl 69.5 4 56.5 5 46.0 4 172.0 13 13.23 IA2 62.0 4 109.0 5 90.0 4 261.0 13 20.08 IBl 34.5 3 52.0 3 21.0 3 107.5 9 11.94 IB2 30.0 3 33.0 3 21.0 3 84.0 9 9.33 IIAl 188.0 9 122.0 7 91.0 6 401.0 22 18.23 IIA2 224.5 9 120.5 7 90.0 5 435.0 21 20.71 IIBl 185.0 9 152.0 7 82.5 6 419.5 22 19.07 IIB2 197.0 9 91.0 6 66.0 5 354.0 20 17.70 IIIAl 399.0 7 268.0 6 182.0 5 849.0 18 47.17 IIIA2 118.0 4 42.0 2 81.0 4 241.0 10 24.10 IIIBZ 111.0 4 75.0 4 93.0 4 279.0 12 23.25 Total 1618.5 65 1121.0 55 863.5 49 3603.0 169 21.32 aAll trout in IIIBl were removed by vandals. 143 .uw>ou scum ocflwmmw movSHocHn .um>oo scum msflwwmm mvzaocfi #0: macaw Ho.ooa mh.mmvma o.ooH mm.ommm 0.00H mm.mmmm o.ooa mm.m0>m mamuoe mm.m mh.oama mv.m o.m¢~ mm.HH m.Hmv o.m mm.mam msoocmaamomwz hh.¢m mm.mvmv h.Hm o.mam vo.mm m.mmm gm.¢v mb.mmmm nmcwcmmm Ano.mwv vo.vm mm.mmmm mm.mm mm.mH>H mm.vw mh.mNm~ mm.mv mm.mmmm mom: Hw>oo Hm.o m.maa mm.o G.FH mm.o m.om mm.H o.mb oaumwcomg a 5: m 5: a 5: w 5: 5339a mamuoa mcwcm>m waowz mswcho: mafia mum mmDHm> .mcficw>m cam .mmuMOflHmmH can mxcmu HmHoom Ham Hm>o m>wumanfiso .mmccfis .mcflcuoa mcwusv >uw>wuom sumo CH ucwmm usouu mafia I .mm manna 144 .Hm>oo scum maficmmm mmcnaocH n .H0>oo EOHM mswvomm mwzaosfl no: mmoam Ho.OOH m.ommv o.ooH m.mmm o.ooa mh.~mm Ho.ooa mm.mamm mamuoa mm.~ mb.mma o o vo.m o.Hm mm.m m>.vm msom:ManomHz mm.vv m~.vwma 5.0m o.mmm N.hm mm.ham mm.vv o.mmma nmcwnowm O.MH m~.mom mH.o 0.0m Hv.na m.m¢H mo.ma mh.HmN moms Hw>ou mv.mm mm.omha mH.hm m.ham mh.ae o.mmm m.mm mh.mvHH oaumwcomm w E w E w E w E %Ufi>«uo¢ mamuoe mcwcm>m mecwz mcflsuox unmEm>oz .mmuMOfiHQmu can mxcmn AmwOOm Ham Hm>o 0>Humasfiso mum mosam> .mcfl:m>o can .mmnoflfi .ocw:HOE mswuao >ua>wuom 50mm on kuo>mc anon» usmfiw>oz I .md manna 145 Table A7. - Number of one-on—one agonistic bouts won and lost by each social rank in replicates with two small trout. Social Total rank of Social rank of winner number of loser Replicate 1 2 losses I31 0 1 I82 a 0 IIIBZ 0 I31 8 2 IB2 a 19 IIIBZ 11 Total number of wins 19 0 19 aNo bouts observed. Table A8. - Number of one-on-one agonistic bouts won and lost by each social rank in the replicate with two large and one small trout. Social Total rank of Social rank of winner number of _loser Replicate 1 2 3 losses 1 1A1 0 0 0 2 IAl 1 0 1 3 IAl l l 2 Total number of wins 2 1 0 3 146 Table A9. - Number of one-on-one agonistic bouts won and lost by each social rank in replicates with two large and two small trout. Social Total rzgk Social rank of winner hugger loser Replicate 1 2 3 4 losses IA2 0 0 0 IIAl 0 0 0 l IIA2 2 0 0 3 IIBl O O O IIBZ O O O IIIAl l 0 O IA2 2 O O IIAl 1 0 O 2 IIA2 5 O O 32 I181 8 o o IIB2 2 0 O IIIAl 13 l O IA2 0 3 0 IIAl 5 0 O 3 IIA2 5 4 o 60 IIBl 4 3 O IIBZ 0 3 O IIIAl 15 18 0 IA2 0 2 3 IIAl 4 0 4 4 IIA2 10 3 2 70 IIBl 2 4 l7 IIB2 6 8 IIIAl 0 1 Total number of wins 79 50 36 0 165 147 .Hofi>m:mn Hmfiuocndm mm.o mm.o v m~.o ~.o mcacm>m mm.o ma.o m mv.o om.o sauna: mo.a mo.o N H.H mo.a mcacuoz mm.o o~.o H mm :mm: mm saw: how no mafia xcmm mm.a H~.H m me.o o mv.o m mm.a m~.o Ha.o m mm.m mm.m mm.o HaHHH mm.o m~.o o o o o o o om.a mm.o mm.o ma.o NmHH an.o ~m.o mm.o va.o o o o o Hm.a mm.a mm.a mm.o HmHH -.o mv.o o~.o mm.o «.0 «.0 mv.o H~.o om.o 6.0 om.o o~.o o.a Hm.o maHH mm.o v~.o o o a o ~.o m N.o hm.o vm.o m -.o HaHH mv.o oo.o mm.o mm.o va.o o mm.a No.0 mm.o mm.o mm.o mm.a am.o o NaH am cum: 4 m m a m m a v m N a mumoaammm mumoaammm xamm xcmm xcmm mcwgm>m INcpoflz msachoz .uzouu v cu“: moumoaammu :H madcw>m can .hmppwfi .mGH:HOS mcwusp xcmu Hafioow scam an Emficomm cw ucmmm wad» :owmmwm cowum>uwm£o mo unmouwm com: I .oa¢ magma 148 mm.om qo.om mcacm>m mo.m~ mm.Hm sauna: e.ma mm.m~ meaauoz om cam: amp mo mafia .Ho«>m:mn Hmsuocnflm vh.NH wo.ma v v>.mH Ho.vm m Nv.mH mm.mm N mN.®m Hh.hm H Gm cam: xcmm mm.hm mm.am m vv.m~ o o.ooa m mm.h~ H5.ha 0H.NN m mo.hm Hm.mm hm.om adHHH um.am hm.om o o o o o o o o vs.am no.5m mm.v~ «.ma mmHH mm.mH N0.¢H mm.mH o o c o o o o mo.mm mo.mm mh.mm mm.mH HmHH m>.om av.dm Hm.ha 56.6H 0.0m o.om mv.HH mn.~ mv.mv o.m> mn.om mo.~m mv.v~ nm.ma maHH mm.vm m>.m~ o o m o o mo.ov m o.ooa m~.ma ma.vv m Hm.m HmHH m¢.>H mm.>~ Ho.m~ mm.ma o.m~ o n>.mm mv.mv Hm.va mm.mo He.m~ m~.Hm mm.a~ o mmH am saw: v .m m a v m m a v m m H mumoaammm mumoaammm scum scam scum mswcw>m kappa: mcflcuoz .uaouu v suaz mmumoaammu :H mcfico>o can .wmvvafi .mcwgnoe mawuac xcmu Hmaoom 5060 an Emflsoam ca ucmmm uswso>os coammwm nodum>ummno mo unmouom com: I .Hdd manna 149 .Hofi>m:mn Hmsuosnmm am.qH qw.mm v H.6H mm.q~ mchm>m em.Hm qv.m~ m ma.m~ ~m.om smocH: mo.- mm.n~ m am.eH Ho.om mcHauoz mm.mH vo.om H Om :00: I. am :00: how no QEHB xamm mm.mm mv.m~ m m.vH oa.e~ v.6H m mv.oH m.on m.mv m av.H~ no.- mm.o~ HaHHH mm.v mH.m~ o o o. o o o o Hv.s~ HH.m~ 6H.o~ vo.Hm mmHH am.HH m~.mm «.mm o o o o o o mm.om mm.mv mH.wm mm.vm HmHH mo.Hm s~.m~ v.vH m.~H v.om ¢.om 6H.v m.~H v.vH mm.m ~m.¢¢ mm.mm mm.em «H.Ro maHH H~.s~ mm.Hv o o m o o «.mm m ma.mm m~.m~ om.vH m mm.om Hm wanna: mcflcuoz xcmu Hmfioom comm mom .uaouu v nuaz mwumofiamou :H mcwsm>m can .hmppfifi..mcacuoa mcwhsp Emwcomm sag? pmumwoommm Am\eou unmaw>ofi mo mama com: I .Nad manna 150 .uow>mnmn Hmsuocndd N6.NN 66.66 6 6N.NN 66.66 6cho>m 66.HN 6N.66 6 N6.6N 66.66 >666Hz 66.6N HH.66 N 6N.6N N6.m6 6chuoz 66.NN 66.H6 H 66 :66: 66 666: >60 mo mafia xcmm 66.oN 66.66 6 66.66 6.66 66.N6 6 66.66 66.66 6.66 6 66.66 66.N6 HH.66 H6HH HH.HN 66.H6 6.66 6.66 6.66 6.66 o.ooH o.ooH o.ooH o.ooH 66.66 66.N6 HN.66 6N.N6 NmHH H6.6N 6N.66 66.66 66.66 6.66H 6.66H 66.66 66.66 HN.66 o.ooH 6.6H 6.N6 66.66 66.6N H6HH 66.6N 66.66 66.66 66.66 6.66 6.66 N6.N6 N6.6N 66.66 HH.66 HN.66 66.NN 66.6H 66.6N NmHH 6N.6H 66.66 66.66 66.66 6 66.66 HN.66 66.66 6.66 66.66 HN.N6 6 N6.N6 H6HH 66.6N 66.66 66.6N H6.6H 6N.66. o.ooH 6N.Nm NH.NN N6.NN 66.66 66.66 H6.66 66.66 H6.6N N6H am :66: 6 6 N 6 6 H 6 N H wbmoHHmmm oumoHHmmm x666 6:66 6:66 mawcm>m hmppflz mewsuo: .Qmpsaocw no: ma Ho>oo scum mcwpomm usomm mafia xcmu Hmfloom some >8 uw>oo Home: ucmmm ofifiu GONmmwm coaum>ummno mo unmouom :60: I .mad manna .unouu 6 £963 mmumowamou :6 151 .Hofi>6:6n H6Euocn<6 66.6H 66.NH 6 66.6 6H.6 6:Hcm>m 6.6H N6.6H 6 66.NH 66.NN 666662 66.6H 6.NN N 6.NH 66.6N 66choz 66.6N 6.HN H am :66: am :66: N66 no 6569 xc6m mH.hH mo.HN 6 06.0H o o 6 mo.mv mm.mm hm.ha 6 om.mH 6.6m mo.H6 HaHHH H6.6N 66.6H 6 6 6 6 6 6 6 6 66.6N 66.N 66.6 66.66 NmHH m6.0H m~.HN mm.m 66.6H o o 6H.6m mo.vm o 0 mm.@ 66.6m «6.66 Nv.o~ HmHH mH.0H >6.m 66.6 mm.m o o 0.0H mm.om mm.Hm o mm.6 mm.o vm.mH m6.m deH Hm.m~ mo.mH o o 6 o o o 6 o uh.mm v~.om 6 6H.66 acHH No.6H mm.m~ mH.6H «6.6m o o om.~H mm.mH mm.oo No.66 mm.mm mo.NH 66.Hm mo.h~ N4H 66 :66: 6 6 N H 6 6 N H 6 6 N H. 6666HH666 mu6owamwm xcmm xz6m xc6m mcflso>m >6wpfiz mcwcuoz .6665H0Q6 no: 66 Hm>oo scum mcflpmmm zuwz 6696600666 ucwfim>oz .uaouu v 5&63 66u606dmmu :6 mcfiso>o 6:6 .m6cuaa .msacuofi mcflusp 3666 666006 £060 in mmfl HQ>OO gflg UOHMflOOmmM Hflgwxrg QOflmmmm GOfluMKfHQMQO HO UGOOHa 50: I oVHfl GHQ? 152 .HoH>6£0£ H6EHOGQ¢6 66.6 N6.H 6 66.6 6H.6 66H66>m 66.H 66.6 6 66.H 66.6 6666Hz 66.6 66.6 N 66.6 6.6 6666662 6.6 6H.6 H mm :60: am :60: M60 mo 0869 xc6m 66.6 N6.6 6 6 6 6 6 66.6 66.6 66.6 6 66.6 6N.6 66.6 H6HHH NN.6 66.6 6 6 6 6 6 6 6 6 N.HH 66.6 6H.6 6H.6 N6HH 66.6 6N.6 66.6 6H.6 6 6 6N.6 6H.6 6 6 6N.6 66.6 6.6 6H.H H6HH 66.H 66.6 N6.6 66.6 6 6 66.6 H6.6 6N.H 6 66.6 6H.6 N6.N 66.6 N6HH 6N.6 6H.6 6 6 6 6 6 6 6 6 66.6 66.6 6 N6.6 H6HH 6H.N 6N.H 66.N N6.H 6 6 66.6 6H.H 66.6 66.6 6H.6 HH.6 6.N 6 N6H 66 :66: 6 6 N H 6 6 N H 6 6 N H 6666HH666 6666HH666 x666 666m 6666 maficm>m 66666: mcfichoz .6065H0:6 no: 66 um>00 scum unflpmwm £U63 60u6fioomm6 ucofim>oz .uaouu v :u63 mwu6066mmu :6 mcwcm>0 0:6 .>66063 .OG6:HOE 5:6H50 x66“ H66006 £060 an 065 u0>00 nuflz 60u6600666 Am\fiov ucwam>05 mo 0u6u :60: I .mad 0HQ6B 153 Table A16. - Number of observation sessions in which each social rank used large cover, small cover, and no cover, and the probability of using small and large covers in replicates with 4 trout. Sessions in which each rank used cover alone and in all possible combinations with other ranks are included in sums. Proba- Proba- bility bility of using using of Large Small cover large Small Large cover small Rank cover and no cover Sum cover cover and no cover Sum cover 1 48 37 85 0.565 32 53 85 0.386 2 29 34 63 0.46 30 33 63 0.476 3 46 28 74 0.622 20 54 74 0.27 4 26 23 49 0.531 13 36 49 0.265 154 .Ho«>mswn Hmeuocnmm Hs.o am.H v Ho.o em.H mchm>m mm.o om.o m no.0 om.o smoon mo.H H~.H m HH.H nv.H mcHauoz mv.H mb.H H mm com: Gm coo: moo mo mafia xcmm mo.H mo.H m am.o mm.o MH.H m mH.H no.~ mh.~ m oo.o mo.m o.~ HaHHH mm.H mo.H mo.H no.0 o o o o HH.H mm.H nv.o a.¢ mmHH «a.o om.H oH.m mm.o o Hn.H m.o o vo.H vo.H vo.H em.H HmHH mp.o mm.H mH.m mm.H mo.H mm.~ Hm.H mH.H o Hv.H Ho.H ov.H mH.H mcHH mm.o h~.o mo.o «H.o m o mm.o o m o em.o o m o HaHH mm.o mm.H m.- mn.H o.m o ma.o o nv.o o vm.H Hv.H v~.o mo.~ mmH om saw: v m m H w m m H v m N H mamoHHmmm mumoaamwm xcmm xcmm xsmm mcficm>m hoppwz mewswoz .omo9H0:H mum mcofiumum mcwpmwm nm>oo .soflumum comm um ucmmm mafia msflvomu Hmuou mo usaouom an counmfloz oocmumwo .usouu w nuas mmumoflammu CH mcwsobm can .hmoofia .mGHGHOE mcfiuzo xcmu Hmwoom 50mm an coma macaumum mcwcmmm sown um>ou ummumm: map on AEV mocmumwo cum: I .bad manna 155 .uoH>mnon Hmauocndm N.ON H.OH v O.mm O.v~ mchm>m «.mm O.mv m O.mv O.OO smOOHz v.vm H.~m m m.Hm m.~m Ochuoz 0.0v m.mv H am now: am com: woo mo mEaB xcmm O.OH H.OH m mm Om m N OH O m hm OH OH HOHHH H.O~ m.Om O OO mm OO mm mm mmHH H.Om O.mm O OO O HO OOH O OH Om m HmHH O.Om «.mm O O O O OOH O OOH NH H O Om NaHH m.Om H.Ha av mm OOH O OOH a NO OOH m OOH HOHH 5.Om «.mm O O O mm OOH OO O mm mm O NOH Om saw: v m N O m m H O m w H muOOHHmmm muOOHHmom xcmm xcmm xcmm \mcwco>m wanna: mcwcuoz .usouu v nua3 mmumofidmmu cw mcfico>m can .mmvowa .mcwcu03 mcfiusp xcmu HMHUOm some >3 u0>oo scum mswomou ucmmm mew» mcwpomw mo ucmonmm com: I .mH< manna 156 ~.O O OO0.0 ON ON0.0 H Om.O OH OO0.0 O 0.0 mm +OO.O O O0.0 OO mmusumssH O O O O O O O O O O O O ONO.O O OO.O m OOHOOO muwumoowue O O OOO.O O O O + m + H ON0.0 O +OOO.O O +OO.O O mnmssz + m + H OO0.0 O O O + m + H + H +O~0.0 m OOHOOO wumumouofionmm O O O O + H O O O O O O O O + H OomOHHoz O O ON0.0 O + m + H O~.O O O O + O O O mumumocmsam O O O O + H O O O O O O O O O O muwumosom O O O O O O O O ON.O N O O H.O H O O mmssunoHOumo O O O O O O O O O O O O O O OO.O m magnumssH O O O O O O O O + N O O ++ O + H muHOOO mumumomaoo + m + N O O O O + H ++ N ++ O +OmO.O O mumumHsm: O O O O O O + H O O O O +O~O.O O OOO.O m OOOOHOOEO H.O OH 0.0 pm H.O OH ~.O Om 0.0 OH OO0.0 NO OOm.O Om 0.0 pm «OomomH .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz :oxms m O O O O chcan Hmccmno Hmccmno Hmscmso : OOOO. : OOOH ; OOOO : OONO powumEo mum: mafia lean. nOOOO .Ome I umm mumz .mucmucoo um: umaup Hmuou may no wm Hmswm monEmmnsm .amv mcHHQEmm comm co mucousoo no: umwuc mo mmHmEmmnsm Ga mamwcmmuo mo Aoov masao> can Hmnfisz I .aam mdnme 157 mNo.o moo.0 ++ mmo.o on V5 H000mI-l 0‘ :4 HOW) mod. mNo. o o o + + +» + o + + o 000 NMO0H0 on N 000 mNo.o o +000+0 + + 0 + 0 m V000fl'0 N N 0H0 000++0 o mNN.o Asomc n OOOH .OH\O I o + 0 ON mNo.o o o NH mo.o mNo.o 00000010 +00 + 0 o mH.o mm H.O o o H o o mNo.o ma H000 N 0 NH now 0 o H 0 N ++00+0 H.O m.o mumz 00+ ma mv NMHOH0 In th.o mNO.o N.o mNo.o mNo.o mnm.o N.o mNo.o o 0 0H vN vH mm ON 0 o m®.H moo.o mo.o 00+0 ++ ++ moo.o mm.o hm mmHDuMEEH o muasoa cumuQOOHua Om msoaaz H nuanpd mumumoumEmsmm momnaaoz mumumocmfihm snowmosom mmwnusOHmumo mousumEEH 61000-10 OH muazpé mumumomaoo mm mumumflfimm O OcomHnmsc OO mcomomH o msomcmaaoomfiz o mHDuOEEH o muaoom mumumfia H.O.ucoov mad 0HQMB 158 + H H.O m o o N.O h + H H.O m mo.o v mNH.o OH mmHDHMEEH O O O O O O O O O O O O O O O O OOHOOO mumumoofiua OO.O NN O0.0 OH H.O OH O0.0 NH ONH.O OO N.O OO O0.0 ON ONH.O OO mnmsxz + O + O + O + H O O O O O O O O OOHOOO mnmumoumamnmm O O O O O O O O O O O O O O O O momsHHo: ON0.0 OH + N ON0.0 ON O O ON0.0 O O O O0.0 OH O O muoudocmsxm + .O + H + N + H O O O O O O + H mumumoeom O O O O O O O O O O O O 0.0 O O0.0 H OOOOuOOHOumo + H + H O O + H O O ON0.0 N O O O O OOOOOOEEH O O O O O O O O + O O O + H + H OOHOOO muwumomaou ONO.O O + O + N + H + H O O O O + N mumumHsmm O O O O O O O O OOO.O O O0.0 O OO.O N OO.O N OOoOHOOEO O O H.O OH + H H.O OH O0.0 O ON0.0 OO ON0.0 O N.O OO moomomH “somOAOOOOO .O\O I umm muwz O O O O O O O.O H OH.O H O O O O ONO.O O msoocmHHoomHz + O + O + N + O ON0.0 O + O + O ++ O OOHOOOEEH O O + H O0.0 N O O + H + H ONO.O O +ONO.O O OOHOOO mumumwa A.O.u:OUV mafl magma 159 .00 mmoo.o on Hmavm pmeUHmsoo Han .wHQmusmmoE uoz+ mNo. + m + m o m mNo.o v + N + H mc. 0+0 000% ++o NNO 0+0 00-10 + N + m o o mmo.o m + v + m mmo.o m msomGMHHmomHz + m mousuMEEH + N mszod mumumHo A.v.ucoov de mHQdB 160 Table A20. - Total estimated number and volume (cc) of organisms collected in drift nets on each sampling day. Time nets Channel A Channel B emptied (EDT) Number VOlume Number Volume 6/29 0200 h 3180 34.332 1060 13.666 0800 ha 1640 23.664 680 11.832 1400 h 1140 11.832 540 3.5 2000 h 1340 21.832 560 7.5 7/16 0200 h 5440 36.5 1100 21.5 0800 h 2740 29.0 2720 11.0 1400 h 1480 22.0 1320 2.0 7/15 2000 h 1020 13.332 480 3.666 8/7 0200 h 2020 27.166 1300 15.5 0800 h 2320 16.332 1360 6.166 8/6 1400 h 800 8.0 1000 4.166 2000 h 1060 6.5 1360 3.332 aThe 95% confidence intervals on this sample were :15% numbers, and 193% volumes. Confidence intervals were not determined for other estimates. 1&51 ONH.O HH ONO.O O O O O O O O O O O ONO.O N O O O O O O O O H.O O a N O H.O HH OH.O O O O O O O O O O O O O O O O O O O O O H.O HH O N O ONO.O OH N.O O O O O O O O O ONO.O O OO.O O O O O O O O O O OO.O O O N O OO.O HH 0.0 O O OO.O H O O O O O O O O O O O O O O O O N.O OH O N O OO.O O O O O O O O O O O O O ONO.O N O O O O O O O O ONO.O H N N 4 ONO.O O ONO.O O O O O O O O O O O O O O O O O O O O O O O H N O : OOON I H msouo O.O NO H.O O O O O O O O O O O OO.O N O O O O O O O O OO.O OO N H O OOO.H OOH N.O O O O O O O O O ONO.O H OO.O N ONO.O H ONO.O H OO.O H O O O.H OOH H H O OO.O ON OO.O O O O O O O O O O O O O O.O OH O O O O O O N.O OH O H < OH.O N OO.O O O O O O O O O O O O O O O O O O O O O OO.O N N H O N.O O OO.O O O O O O O O O O O O O H.O N O O O O O O OO.O N H H O n OOOO I H Onouo .Ho> .oz .Ho> .Ho> 6: .Ho> .oz .Ho> .oz .Ho> .oz .Ho> 6: .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .oz .Ho> .02 x H Hnuoa. 0.032,. .30» 602.39. upomHH—mat cumu muoumooHoo auouOHHam: JOSH 3164 mzmaxz muHsoc mcodcmu c a Imosuuo Amoco-.3: QNSQOUHNB muoumnioaonmm a : coach. a: mOOU-HO: : .3‘ 0.33. X25935 usouu HHmam 0.3 v as... n oxcou 530.3 @323 nun m can H 333. .2 OOON. ocho>o can. 2 come. 9393... «cu :H 95.3 zoom .5 ”50.3 no ozone—Cum :H EamHHBmuo no .03 m§Ho> «EM .3952 I .Hmc 332. Table A21 (cont'd.) Group II - 0800 h 13 0.525 0.25 41 0.525 0.25 20 0.45 0.1 0 0 0 5 0.025 11 0.025 0 0 30 0.25 0 0 0 0 2 1 0.05 0 0 ... 19 0.15 0 23 0.35 3 4 25 0.383 O 0 0 1 0.025 0 1 4 0.133 3 0.075 0.1 0 0 13 0.025 3 0.025 3 0.025 53 0.55 20 0.15 1 0.05 0.45 68 1.125 0.45 23 0.608 0 0 0.1 0 3 4 2 + 3 Group II - 2000 h b 0.1 162 19 0.375 26 0.35 0 0 0 0 1 0.025 1 0.025 0 0 0 0 22 0.125 18 0.15 3 4 0.2 0 1 0.05 104 0.308 0 0 0 O + 1 0 0 0 100 0.15 110 0.2 0 0 0 12 0.1 29 0.2 0 110 0.2 0.25 138 0.65 1 0.05 0 0 125 0.25 110 0.2 3 4 2 2 144 0.583 0 0 0 2 0.05 2 0.025 163 .00 nmoo.o Hmsvo cu 0052mmo usn .oHnnusmnms uoz+ .cocHEuouou on be: pHsoo mamuH anochHHoomHE no nonszzu . muzmsw v n .nocHauouoo on no: vHsoo xcwu HcHoomm o hmm.o mmn.0 mmm.o mH.H nmm.o mem.o o n NN om QH m o o mm.c o o o H.O o N.O o mm.o mH.o o N.O NH00 N.O mNN.o 0hf~0 0000 0000 0000 o o z OOON I 0000 a come I o o o O HHH dsouo o o O O HHH msouo 0000 0000 0000 0000 0000 004-4- 0 6 Id I4 0000 0000 mNo.o N N 4 mNN.o HN v H 4 mm.o 00 n H t no.0 m N H c + H H H < A.v.u:00. HN< andE 164 .HOH>m:wQ HmEMocndm OO.HN OO.OO O OO.ON OO.NO OOHOO>O HO.HN OO.OO O OO.NN OO.NN OOOOH: O0.0N OO.ON N 0.0N OO.OO OOchoz NO.HN. O0.0H H mm cum: am now: haw Mo oEHB xcmm OO.OH NO.OO O OO.OO O.OO OO.O m O.OH HO.OO ON.OH m OH.NO OO.OO OO.O HOHHH OO.HN O0.0H O.ON O.OO O O O O O O N.OO OO.ON OO.ON OO.O NOHH OO.ON OH.OO OO.HO O0.00 O O OO.OO OO.OO ON.OH O OH.OO OO.OO OO.HO OO.HO HOHH OO.ON O.OO OO.OO OO.OO O.ON O.OO H0.00 OO.OO HO.OH ON.OH OO.OO OH.OO HO.OO O.OO NOHH NN.OH N.ON OH.OO OO.OO m OO.OH ON.OH ON.HN O O OO.OO NN.HO O HH.HH HOHH O0.0N OH.OO OO.OO OH.OO ON.OO O OO.OO OO.ON OO.ON O OO.OO 0.00 O0.00 O.NN NOH Om cum: O O N H O O N H O O N H OOOOHHmOm mHMOHHmum xcmm xcmm xcmm mchm>m mmvon mchuo: ucmmm oSHu mmcaHocH mHne .Hm>oo scum mcHommw .usouu v £UH3 wouMOHHmmu :H mcho>o can .amcoHE .msHsuoE mCHusc xcmu HMHOOO zoom an mchme “comm 05H» GOHmmmm :OHum>Hmmno mo unmouom :60: I .NN< mHnms 165 .uoH>m:mn Hmauocndm OO.OH OO.OO O OO.ON OO.OO OOHOO>O H.HN OO.NO O H0.00 OOOOH: OO.NN OO.NO N OO.OH OO.HO Ochuoz OO.NO HO.OO H am am com: amp xcmm NO.ON ON.OO m HN.OO 0.00H O.ON m OO.ON NO.OO OO.ON a NN.OO OO.OO OO.HN HOHHO OH.HO NO.OO O.OOH O.OOH O O O OO.OO NO.OO OO.OO OH.HN NOHH OO.NN H0.00 O.OO OO.OO O O NO.OO OO.NO O.OOH O NO.OO ON.OO OO.OO OO.OO HOHH OO.ON OO.OO NH.OO H0.00 O.ON O.OO ON.OO O0.0N OO.NN O.ON ON.OO ON.OO ON.OO OO.ON NOHH NO.OO OH.OO O.OOH O.OOH m O.OOH O O OO.NO ON.NO m OO.OH HOHH OO.ON OH.HO OO.OO OH.HO O.OO O.OO OO.OO HO.ON O HO.OO OO.OO OO.OO OO.OO NOH Om cam: O O N N H O O N H OumuHHmwm muOOHHmum xcmm Hana xcmm mcHsm>m NMGQHE mchuoz Im>os mmcsHocH mHna .uw>oo Scum mcHooom nqu omumHOOmmm usmfi .uaouu v :uH3 mmumoHHmmu :H mchm>m can .hmcoHE .mchHoE mcHusv xcmu HmHoom sumo >3 mCHcmom o» cmuo>m© ucoEw>OE COHmmmm coHum>ummno mo ucooumm com: I .MNm mHnma 166 .H0H>mnon Hmeuosndm O.H NO.O O OO.O OO.O OOHOO>O OO.O HO.O N ON.O HO.O OOOOHz O0.0 HO.O H OH.H ON.H OOHOuoz H0.0 OO.O H am com: om com: amp mo wEHB xcmm OO.O OO.O O O.O OO.O NN.H O 0.0 OO.O OO.O O OO.H OO.O NO.H HOHHH OO.H OH.N OO.O O.H O O O O O O ON.O OO.H OO.O ON.H NOHH O.O OO.O OO.O NO.O O O OO.O OH.O OO.O O OO.O ON.H OO.H OO.O HOHH OO.O O.O OO.O OO.O OO.O O.O NN.O ON.O OO.O OO.O NO.H OH.H OO.O NO.O NOHH OO.O NO.O OO.O NO.N O OO.O O NO.O O O O.H NO.O O OH.O HOHH O.O 0.0 O0.0 NO.O H.O O OO.O NN.O NO.O O OO.O OO.O N.O OO.O NOO Om OOOz O O N H O O N H O O N H OuOoHHmOm HmmmmmmmMWII chm NOON chm mchw>m mmoon mcHsuoz cmumHUOOmm unoEo>OE mwcsHocH mHnB .um>oo Scum msHoomw nuH3 .usouu v nqu mouMOHHme :H mchm>m cam .amooHa .mchuoa mcHuso xcmu HOHUOO sumo an mcHoomw :qu UmumHUOmmm Hm\Eov usmfio>os mo mama com: I .de oHQwB 167 .uoH>mnmn Hmfiuocndm OO.O NO.H O O0.0 H0.0 OOHOO>m OO.O O.H O NO.O OO.O OOOOHz HN.O OO.O N OO.O OO.O OOchoz ON.O OO.O H am com: am com: amp mo wSHB xcmm OO.O O.H O O.H OH.H OH.H O OO.H HO.O H0.0 O O.H OO.O OO.O HOHHH H0.0 O0.0 OH.H O0.0 O O O O O O O0.0 NO.O NN.O O0.0 NOHH OO.O OO.O N0.0 O.O O O H.H OO.H O O OH.H OO.O OO.O OO.O HOHH OO.O H0.0 OO.O OO.O OO.O OO.O NO 0.0 OO.O O O0.0 OO.O OO.O OO.O NOHH O0.0 OO.H NO.H NO.H O O OH.H O O O OO.H O O O HOHH OH.O OO.O OO.O 0.0 OO.O O NO.O O OO.O O OO.H OO.O O0.0 OO.O NOH OO OOO: O O N O O N H O O N H OOOOHHOOO OOOOHHmom chm chm xOOm mchm>m II; wwonz Imecuoz ha oouanm3 mocmumHQ .pmosHocH uoc mum msoHumum msHowww Hw>oo .QOHumum comm um oEHu mchwwm Hmuou mo unwouom .usouu v nqu meOOHHmoH :H UGHcm>m new .hMUGHfi .mchnoa NCHHsG xcmu HmHoom 3060 an coma mcoHumum mcHommm Scum HHmz umoumwc mnu ou H50 mosmuch cum: I .mmd mHnt 168 Table A26. - Schedule of artificial feeding and response of large trout in replicate IIIAl to grasshoppers. Positions that grass- hoppers landed in the channel are indicated by coordinates (longitude, latitude, see Appendix Figure B3). Position Time grasshopper Date (h EDT) landed W trout 8/22 0835 1.5, 7 0836 2 , 7 ‘ None. Both large trout under unknown 0839 2 , 7 cover. 0945 2 , 12 Rank 1 under large cover (1.5, 5), none. Rank 2 moved from feeding station at 3.75, l to eat grasshopper at 2, l. 0946 2 , 12 Rank 1 moved from large cover to eat grasshopper at 2, 10, then moved to feeding station in shade at 3.5, 6 for two minutes, followed by return to large cover. Rank 2 under unknown cover - none. 0959 3.5, 8 Rank 1 under known cover, none. Rank 2 feeding at 3.5, 6, ate grasshopper there. 1006 2 , 7 Rank 1 moved from large cover to eat grasshopper at 2, 5, then moved again to feeding station at 3.5, 6 for one minute before returning to large cover. 1420 1.5, 7 None. Rank 1 under small cover, rank 2 under large cover. 1421 1.5, 11 Rank 1 none. Rank 2 moved from large cover to eat grasshopper and immediately returned to cover. 1426 1.5, 7 None, both trout under covers indicated above. 1427 2.25,7 Rank 1 moved to edge of small cover but did not eat grasshopper. Rank 2 under large cover, none. 1428 1.5, 7 1432 2.25,7 None. Both large trout under covers 1434 2 , 7 indicated above. 1435 2 , 7 Table A26 (cont'd.) 8/23 8/24 1436 1000 1001 1005 1007 1751 1752 1753 1754 1800 1557 1601 1605 1607 1932 1934 2 , 11 2.25,7 1.75,7 2 , 7 1.25,11.5 1.75,7.5 1.75,7.5 169 Rank 1 under small cover, none. Rank 2 momentarily left large cover to eat grasshopper. None. Rank 1 under unknown cover. Rank 2 under small cover. Rank 1 briefly left large cover to inspect grasshopper but did not eat it. Rank 2 under unknown cover. None. Rank 1 under large cover, rank 2 under unknown cover. Rank 1 briefly left large cover to eat grasshopper. Rank 1 under unknown cover. Rank 2 feeding at 1.5, 1 ate all 4 grasshoppers. Rank 1 under large cover, none. Rank 2 moved from small cover to eat grass- hopper then returned to small cover. Rank 1 briefly left large cover to eat hopper. Rank 2 under small cover, none. APPENDIX B FIGURES 170 Figure Bl. - Location of trout sightings for Group I in test of lateral cover devices. L = low-stripe device, H = high-stripe device. Drawn to scale. AN _—P O 0 171 SECTION . 0 av- o o -‘-"'5 o .0 0 .fi - O qp O —-.r- o ‘ O qp ‘ —'-3 “P O. . O b. O . . t . —-.r- I c—h- O —-'- 0 Figure Bl 172 Figure 32. - Location of trout sightings for Group II in test of lateral cover devices. L a low-stripe device, H - high-stripe device. Drawn to scale. an —P o‘QOa °. . O 0 O _P —r- .0 III—'- 0 O .0. O .O: —. 0 .- 0 o 0 ‘TP . I —. O O 0 .fi- 0 1"]. 8 I 82 ’ “'2? ——- fl 0 O a o Figure 82 174 Figure BB. - Depth (cm) and current velocity (cm/s) in section Al. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout positions. tween numbers on the vertical axis is l m. cm, mean velocity-21.9 cm/s. above the stream bed. Distance be- Mean depth-21.1 Velocity was measured 5 cm SECTION A1 176 Figure B4. - Depth (cm) and current velocity (cm/s) in section A2. Numbers along horizontal and vertical axes refer to co- ordinates used to identify trout positions. Distance between numbers on the vertical axis is l m. Mean depth-23.5 cm, mean velocity-20.7 cm/s. Velocity was measured 5 cm above the stream bed. 178 Figure B5. - Depth (cm) and current velocity (cm/s) in section Bl. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout positions. Distance be- tween numbers on the vertical axis is l m. Mean depth-19.2 cm, mean velocity-21.6 cm/s. Velocity was measured 5 cm above the stream bed. 180 Figure B6. - Depth (cm) and current velocity (cm/s) in section B2. Numbers along the horizontal and vertical axes refer to coordinates used to identify trout positions. Distance be- tween numbers on the vertical axis is l m. Mean depth-22.6 cm, mean velocity-17.4 cm/s. Velocity was measured 5 cm above the stream bed. Figure B6 HICHIGQN STRTE UNIV. L W 3129310 Rh 3 RIES WIN Q! I 233652