DIEL RHYTHMS OF BEHAVIOR IN JUVENILE PINK SALMON (Oncorhynchus gorbuscha Walbaum)
by
JEAN-GUY JOSEPH GODIN B.Sc. (Honors), St. Francis Xavier University, 1973
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
in
THE FACULTY OF GRADUATE STUDIES (Department of Zoology)
We accept this thesis as conforming to the required standard
THE UNIVERSITY OF BRITISH COLUMBIA
June, 1979
cj Jean-Guy Joseph Godin, 1979 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study.
I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.
Department of Zoology The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5
Date ii
ABSTRACT
Anadromous pink salmon undergo several migratory movements between different habitats during their life history. These migrations are accurately timed on a seasonal basis. Annual rhythms or seasonally-timed events may result from interactions between daily rhythms and annual changes in environmental factors. Therefore, knowledge of daily behavioral rhythms in pink salmon may improve our current poor understanding of the seasonal timing of its migrations.
Hence, the objective of this study was to investigate, in a seasonal context and mainly under laboratory conditions, diel rhythms of ecologically-relevant behavior in juvenile pink salmon, and their timing mechanisms.
Fry emergence from a simulated gravel redd in fresh water was mainly nocturnal below 13°C. Diel emergence timing was synchronized with the onset of night, but was affected by temperature in a non-linear manner. Temperature affected negatively the duration of the intra-gravel alevin stage and the rate of emergence. Nocturnal emergence was considered an anti-predator adaptation.
Fry exhibited mainly nocturnal rhythms of swimming activity and of vertical distribution during the first week after emergence.
However, a gradual shift from a nocturnal to a diurnal swimming activity rhythm occurred 7 to 13 days after emergence, when wild fish are residing in estuaries and adjacent coastal waters. Coincident with this shift was an increasing tendency of the fry to swimnnear the water surface during iii
the day. This suggested a weakening of their negative phototactic response during this period. Thereafter, the fish usually displayed diurnal rhythms of swimming activity and nocturnal rhythms of vertical distribution. The ontogenetic shift in the phase of the activity rhythm and in photobehavior was considered adaptive for schooling and feeding during the day.
Wild fry fed mainly during daylight hours in littoral areas of two marine bays. However, their feeding rhythms varied among study sites. Laboratory experiments showed that hunger level affected fish feeding rate and ration consumed positively. Fish fed continuously on live copepods under idealized laboratory conditions. During a 12-h session they rapidly (< 30 min) filled their stomachs with prey; thereafter, they maintained their stomachs full by feeding at a rate that balanced the rate of evacuation of prey from the stomach. Hence, I concluded that pink salmon have flexible feeding activity rhythms, which may permit opportunistic exploitation of prey, and feed at a relatively low hunger threshold. This feeding strategy may explain in part their relatively high growth rates in nature.
During the periods corresponding to their juvenile coastal and pelagic ocean phases, the fish exhibited generally diurnal rhythms of swimming activity and of aggression, and nocturnal rhythms of vertical distribution in response to simulated seasonal photoperiodic and temperature changes. These rhythms were synchronized with the artificial light-dark (LD) cycle throughout most of the year. Some parameters of these rhythms varied on a seasonal basis, but not according to the iv
Aschoff-Wever model. Mean swimming speed, the degree of diurnalism of the swimming activity rhythm, and the timing of the daily peak of the rhythms were affected by daylength. Hence, photoperiod might be an important proximate factor that pink salmon use to time their oceanic migration on a seasonal basis.
Some data suggested the existence of an endogenous, circadian activity rhythm, and thus a daily "clock", in pink salmon. However, this remains uncertain. The free-running period of their activity rhythm was not related negatively to constant light intensity, as predicted by the Circadian Rule. The LD cycle affected directly swimming activity (masking), rather than entraining an endogenous circadian system. Since the activity rhythm of pink salmon does not possess a strong endogenous component, it is doubtful that the seasonal timing of its migrations results from interactions between a circadian clock and seasonal changes in environmental factors. However, the flexibility and inter-individual variability of their behavioral rhythms may be adaptive responses to the instability and heterogeneity of the marine environment. V
TABLE OF CONTENTS Title Page ABSTRACT ii TABLE OF CONTENTS v
LIST OF TABLES viii LIST OF FIGURES xi ACKNOWLEDGEMENTS . xix CHAPTER I. GENERAL INTRODUCTION 1 CHAPTER II. GENERAL MATERIALS AND METHODS A. Fish and incubation procedures 5 B. Holding conditions 6 C. Experimental conditions 6 D. Statistical procedures 7 CHAPTER III. EMERGENCE FROM A SIMULATED REDD INTRODUCTION 8 MATERIALS AND METHODS 9 RESULTS A. Temporal pattern of emergence 16 B. Diel timing of emergence 23 DISCUSSION 30 CHAPTER IV. ONTOGENY OF DIEL RHYTHMS OF SWIMMING ACTIVITY AND OF VERTICAL DISTRIBUTION INTRODUCTION 37 MATERIALS AND METHODS A. Fish 38 B. Experimental tank 38 C. Experimental procedure 40 D. Estimation of rhythm parameters 43 RESULTS A. General behavior of fish 44 B. Swimming activity 45 C. Vertical distribution 54 D. Relationship between swimming activity and 58 vertical distribution
DISCUSSION 59 vi
TABLE OF CONTENTS (cont'd)
Page
CHAPTER V. TEMPORAL PATTERNS OF FEEDING BEHAVIOR
INTRODUCTION 64
MATERIALS AND METHODS A. Pattern of feeding behavior in the field 66 B. Pattern of feeding behavior in the laboratory 69 1. Experimental apparatus 69 2. Feeding experiments 71 3. Prey-capture success and prey 73 biomass consumed, 4. Gastric evacuation rate 75
RESULTS A. Pattern of feeding behavior in the field 76 B. Pattern of feeding behavior in the laboratory 83
DISCUSSION 92
CHAPTER VI. ANNUAL CHANGES IN THE DIEL PATTERNS OF SWIMMING ACTIVITY, AGGRESSION, AND VERTICAL DISTRIBUTION
INTRODUCTION 102
MATERIALS AND METHODS A. Fish 104 B. Experimental tank 105 C. Experimental procedure 106 D. Estimation of rhythm parameters 107
RESULTS A. General behavior of fish 110 B. Swimming activity 110 C. Aggressive behavior 127 D. Vertical distribution 134 E. Relationship between swimming activity, 142 aggression, and vertical distribution
DISCUSSION 148
CHAPTER VII. TIMING OF THE DIEL SWIMMING ACTIVITY RHYTHM OF INDIVIDUAL FISH
INTRODUCTION 164 vii
TABLE OF CONTENTS (cont'd)
Page
MATERIALS AND METHODS A. Fish and holding conditions 167 B. Description of the activity channels 167 and activity monitors 1. Activity channels 167 2. Activity monitors 170 3. Evaluation of activity monitors 172 C. Experimental procedures 173 1. Experiment 1, Circadian Rule 173 2. Experiment 2, Catching the 174 free-running rhythm 3. Experiment 3, Phase-shifting of 174 the endogenous rhythm D. Statistical procedures 176
RESULTS A. Experiment 1, Circadian Rule 176 B. Experiment 2, Catching the free-running 184 rhythm C. Experiment 3, Phase-shifting of the 186 endogenous rhythm
DISCUSSION 189
CHAPTER VIII. GENERAL DISCUSSION AND CONCLUSIONS 200
REFERENCES 214
APPENDICES 243 viii
LIST OF TABLES
Mean ± SD daily water temperature experienced by sibling alevins in gravel from median hatching time to the time of 50% emergence for each of six simulated redds.
Regression coefficients (b) and their 95% confidence limits (CL) for Probit Y = a + b logjg ^, where Y is theecumulative percentage of fry emerging from a simulated redd at different mean temperatures, and X the number of days elapsed since median hatching time (Day 0).
Geometric mean (95% confidence limits) age at emergence (days after median hatching time = mean duration of the alevin stage in gravel) for pink salmon fry at different mean redd temperatures. Geometric mean ages at emergence for sibling fry emerging during light and dark of the experimental LD cycle are also shown for each redd, and are compared using the t-test.
Period lengths of the diel rhythms of swimming activity and of vertical distribution for six groups of fry tested at different times after emergence. Period length of the cycle of water temperature for each experiment is also given. All period lengths are significantly different (P < 0.05) from random "noise", except where indicated specifically by an asterisk.
Spearman rank correlation coefficients (rs) for comparisons between values of 1) water temperature and swimming speed, 2) water temperature and the index of vertical distribution, and 3) swimming speed and the index of vertical distribution. Swimming speed, vertical distribution, and water, temperature were recorded simultaneously every two hours for each of six groups of fish. The fish were observed at different times after emergence.
Absolute and relative prey biomasses (dry weight) and estimated daily food ration in the stomachs of juvenile pink salmon collected during diel sampling series in Departure Bay and Hammond Bay, British Columbia. All values, except daily rations, are means (± SD).
Diets of juvenile pink salmon in Departure Bay and Hammond Bay, British Columbia on 3 May and 20-21 May 1977, respectively. ix
LIST OF TABLES (cont'd) Page Table 8 Comparison of 1) the relative evacuation rate 87 of calanoid copepods from the stomachs of juvenile pink salmon during the first hour after "satiation" with 2) the mean prey-snapping rate of the fish between 0900 - 1900 hours (hours 1-11), expressed as a percentage of the mean prey-snapping rate at 0800 hours (hour 0). The regression equation for stomach evacuation rate
is Y = 9.375 e -0-104 where Y is the dry weight of prey remaining in the fish's stomach, and t is the time (hours) after the initial feeding period.
Table 9 Absolute and relative mean dry prey biomasses in 88 the stomachs of juvenile pink salmon after a 12-h feeding session, and estimated mean ration consumed by the two fish groups during this period.
Table 10. Spearman rank correlation coefficients (rs) for 121 comparisons between values of 1) water temperature and swimming speed, 2) water temperature and rate of aggression, and 3) water temperature and the index of vertical distribution. Swimming speed, aggression, vertical distribution, and temperature were recorded simultaneously every two hours for groups of fish tested during four- to seven-day experiments at different times of the year. Table 11 Period lengths of rhythms of swimming activity, 122 aggression, and vertical distribution for different groups of fish tedted at different times of the year. Period length of the cycle of water temperature in each experiment is also given. All period lengths are significantly different (P < 0.05) from random "noise", unless indicated otherwise by an asterisk.
Table 12 Spearman rank correlation coefficients (rs) for 143 comparisons between values of 1) fish swimming speed and rate of aggression, 2) swimming speed and the index of vertical distribution, and 3) rate of aggression and the index of vertical distribution. Swimming speed, aggression, and vertical distribution were recorded simultaneously every two hours for groups of fish tested during four- to seven-day experiments at different times of the year. X
LIST OF TABLES (cont'd)
Table 13 Spearman rank correlation coefficients (rg) for comparisons between pairs of the three types of behavioral rhythm recorded. Coefficients are given for the level, period, mean daily activity time (a), and mean phase-angle difference (I'max) of the rhythm. All correlations are non-significant, except where indicated specifically by asterisks.
Table 14 Number, mean (± SD) total length, and mean (± SD) wet weight of juvenile pink salmon tested under three different intensities of constant light in the Circadian Rule experiment.
Table 15 Period lengths of free-running swimming activity rhythms of individual pink salmon recorded in constant illumination (LL) of different intensities for a period of 10 days, Period lengths of the rhythms recorded during the first five days (Days 1-5) in LL and the second five days (Days 6- 10) in LL are presented for each fish. The diel activity pattern of each fish was determined in LD 12:12 for one day before being subjected to LL.
Table 16 Percentage of the total number (n = 38) of individual pink salmon exhibiting four different types of free-running activity pattern in constant light condition (LL) for 10 days.
Table 17 Period lengths of the swimming activity rhythms of individual pink salmon recorded for five days in LD 12:12 (600:0.2 lx), followed by five days in LL (600 lx), and finally for another five days in LD 12:12 (600:0.2 lx) at 10°C. All periods are significantly different (P < 0.05) from random "nmise", unless indicated otherwise by an asterisk.
Table 18 Period lengths of swimming activity rhythms in individual pink salmon recorded in LD 12:12 (600:0.2 lx) and 10°C for five days (Days 1-5), and then in a 12-h phase-delayed LD 12:12 (600: 0.2 lx) cycle for six and a half subsequent days (Days 6-12). The predominant phase of the activity rhythm is indicated in parentheses. xi xii
LIST OF FIGURES (cont'd)
Page
Figure 8 Relation between daily mean (± 95% confidence 28 limits) time of emergence of sibling fry from a simulated redd during one 24-h period (see Fig. 7) and daily mean temperature. The broken line denotes the onset of the scotophase (night).
Figure 9 Oblique view of the experimental tank showing 39 the observation section, delimitated by a baffle and a screen at opposite ends. Dotted lines were drawn with a grease pencil on.the front viewing glass and the back wall as aids in recording behavior. See text for more details.
Figure 10 Daily rhythms of mean swimming speed for groups 46 of pink salmon fry from Day 1 to Day 37 after emergence. Each of six groups was observed independently and sequentially for five consecutive days during this period. Light and dark horizontal bars indicate day (L) and night (D), respectively.
Figure 11 Relation between daily D/N ratio for swimming 47 speed of pink salmon fry and age after emergence. Arrows indicate the times when one group of six fish was replaced by another group of six fish in the experimental tank.
Figure 12 Relation between daily mean swimming speed of 49 pink salmon fry and age after emergence. Meaning of the arrows as in Fig. 11. Regression
line A is Y =2.48 + 0.12 X (rs = 0.82, P < 0.001), and regression line B is Y = 0.13 + 0.004 X
(rs = 0.79, P < 0.001).
Figure 13 Periodograms of rhythms of locomotor activity and 51 of vertical distribution recorded simultaneously in LD 12:12 for each of six groups of fry from Day 1 to Day 37 after emergence. Each periodogram is based on five days of observation on each fish group. P is the major periodic component (in hours) in the time-series data record. All P values are significantly different (P < 0.05) from random "noise".
Figure 14 Daily rhythms of vertical distribution in a water 55 column for groups of pink salmon fry from Day 1 to Day 37 after emergence. Each of six groups was observed independently and sequentially for five consecutive days during this period. Meaning of light and dark horizontal bars as in Fig. 10. xiii
LIST OF FIGURES (cont'd)
Page Figure 15 Relation between daily D/N ratio for the index 56 of vertical distribution of pink salmon fry and age after emergence. Meaning of the arrows as in Fig. 11.
Figure 16 Map of Departure Bay and Hammond Bay, Vancouver 67 Island, British Columbia and vicinity. Fish were collected at one littoral station in each bay. Stations are indicated by 1 and 2 in Departure Bay and in Hammond Bay, respectively. PBS denotes the Pacific Biological Station and Ck. the creeks.
Figure 17 Top and side views of the 83-cm diameter 70 circular Plexiglas tank used in laboratory feeding experiments. Locations of water inlets and outlets are indicated by arrows. The bottom of the tank is divided into 10-cm2 grids. Prey were introduced into the center of the tank via a Tygon tube, and fish were filmed with a video camera located directly above the tank. Figure 18 Variation in mean (± 95% confidence limits) 77 relative dry prey biomass in the stomachs of pink salmon fry collected in Departure Bay and Hammond Bay, British Columbia with time of day. and state of the tidal cycle. Means with dissimilar symbols above them are significantly different from each other (P < 0.05, Mann-Whitney U-test). All samples consisted of 20 fish, except where indicated near the mean. Open, solid and hatGhed horizontal bars represent hours of daylight, darkness and twilight, respectively. Times of high (+) and low (+) tides are also indicated.
Figure 19 Relation between dry prey biomass in the stomach 80 of wild, individual pink salmon and dry body weight of the fish. Regression line A is Y = 0.18 + 0.04 X (r = 0.45, P < 0.001) and line B is Y = -0.15 + 0.02 X (r = 0.39, P < 0.001). xiv
LIST OF FIGURES (cont'd)
Figure 20 Variations in the relative abundance of prey found in juvenile pink salmon stomachs in Departure Bay and Hammond Bay, British Columbia with time of day and state of the tidal cycle. Prey (categories) shown were the most common ones consumed by fish. Other prey did not show consistent tidal or daily variations. Nf and Np denote the number of fish stomachs examined for each time of day and the total number of prey found in fish stomachs, respectively. Relative abundance of a prey category in fish ':' raiac. stomachs at each time of day is given as a percentage of the total number of prey (Np) found in the stomach of all fish (Nf) at that time. Horizontal bars and arrows are as in Fig. 18.
Figure 21 Mean (± 95% confidence limits) number of snaps directed at prey per fish per 10 min (solid line) and mean (±95% confidence limits) swimming speed of fish (broken line) since the onset of feeding at hour 0 (0800 hours) for one group of five fish previously deprived of food for 24 h (A) and for another group of five fish deprived for 72 h (B). Fish fed on live copepods that were available throaghout the 12-h feeding session.
Figure 22 Rate of decline in the relative dry prey biomass (G.M. ± 95% confidence limits) in the stomachs of juvenile pink salmon at 11.4 (± 0.37) °C. The upper curve is drawn from the lower regression line of logarithmic transformed data. Broken lines in the lower panel denote 95% confidence limits on Y.
Figure 23 Cumulative number of prey (N) consumed per fish for group A and group B fish during a 12-h feeding session. Regression line A is N = 140.2 t 0,51+ and line B is N = 151.2 t °-64.
Figure 24 Cumulative mean number of snaps (N) directed at prey per fish for group A and group B fish during the 10-min feeding period at 0800 hours (hour 0). Regression line A is N = 12.29 t 0•7^ and regression line B is N = 61.18 t/(2.40 + t). XV
LIST OF FIGURES (cont'd)
Page
Figure 25 Average daily pattern of mean swimming speed 111 recorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines are 95% confidence limits about the mean. Light and dark horizontal bars are as in Fig. 10.
Figure 26 Annual changes in the D/N ratio of swimming 113 speed, aggression, and vertical distribution calculated for different groups of fish at different times of the year.
Figure 27 Relation between the D/N ratio for swimming 114 speed (m fish -1 10 min -1) and the duration of the experimental photophase. The regression
line is log10 Y = -0.19 + 0.03 X (r = 0.59, P < 0,01).
Figure 28 Annual changes in the geometric mean level of 115 swimming speed (A, B), rate of aggression (C), and index of vertical distribution (D) calculated for different groups of fish at different times of the year.
Figure 29 Relation between geometric mean swimming speed 117 and body length for different groups of fish. The regression line is Y = 1.80 + 0.87 X (r = 0.88, P < 0.001).
Figure 30 A. Relation between geometric mean absolute 118 swimming speed and duration of experimental photophase. Regression line A (•) is Y = -19.98 + 2.11 X (r = 0.79, P > 0.05), line B (#) is Y = 22.92 - 0.70 X (r = -0.67, P < 0.05), and line C (A) is Y = 3.78 + 2.25 X (r = 0.91, P < 0.01). B. Relation between geometric mean relative swimming speed and duration of photophase. The regression line is Y =0.09 + 0.01 X (r = 0.41, P < 0.05). xvi
LIST OF FIGURES (cont'd)
Page
Figure 31 A. Relation between geometric mean absolute 119 swimming speed and temperature. Regression line A (•) is Y = 4.90 + 0.60 X (r = 0.36, P > 0.05), line B (• ) is Y = 40.87 - 2.31 X (r = -0.81, P < 0.01), and line C (A) is Y = -3.46 + 3.33 X (r = 0.94, P < 0.02). B. Relation between geometric mean relative swimming speed and temperature. The regression line is Y = 0.14 + 0.003 X (r = 0.09, P > 0.05).
Figure 32 Mean daily activity time (a) for swimming 124 activity (A) and for aggression (B) in relation to the duration of experimental photophase, and a for vertical distribution (C) in relation to the duration of experimental scotophase.
Figure 33 Variations in mean phase-angle difference for 125 swimming activity (A), aggression (B), and vertical distribution (C) with the duration of photophase. Regression line A is Y = 6.77 - 0.86 X (r = -0.57, P < 0.001, one-tailed), line B is Y = 0.18 - 0.21 X (r = -0.24, P > 0.05), and line C is Y = 13.71 - 0.48 X (r = 0.41, P < 0.05, one-tailed). LON and LOFF denote the daily times of lights-on and lights-off, respectively.
Figure 34 Average daily pattern of the mean rate of 128 aggression recorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines and horizontal bars are as in Fig. 25.
Figure 35 Relation between the D/N ratio for mean rate of 129 aggression and the duration of the experimental photophase. The number besides each value denotes the month of the year when the observation was made.
Figure 36 Variation in the geometric mean rate of 131 aggression for different groups of fish with mean experimental temperature. The regression line is Y = -0.14 + 0.02 X (r = 0.39, P < 0.05, one-tailed). xvii
LIST OF FIGURES (cont'd)
Figure 37 Relation between geometric mean rate of aggression for different groups of fish and the duration of the experimental photophase. The regression line is Y = -0.008 + 0.005 X (r = 0.18, P > 0.05).
Figure 38 Average daily pattern of the mean index of vertical distribution in a water column recorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines and horizontal bars are as in Fig. 25.
Figure 39 Relation between the D/N ratio for mean index of vertical distribution and the duration of the experimental photophase. The number besides each value denotes the month of the year when the observation was made.
Figure 40 Relation between geometric mean index of vertical distribution for different groups of fish and duration of the experimental scotophase. The regression line is Y = 3.04 + 0. 18 X (r = 0.12, P > 0.05).
Figure 41 Daily rhythms of swimming speed (A), rate of aggression (B), and index of vertical distribution (C) recorded simultaneously for seven consecutive days in LD 13:11 and 11,9°C for pink salmon group 12. Periodograms show the major periodic component (P) of each rhythmic function. All P values are significantly different (P < 0.05) from random "noise".
Figure 42 Diagram illustrating the top and side views of an activity channel. Broken lines indicate transparent pieces of Plexiglas. Only the inlet and outlet tubes of the charcoal filter are shown, b, surface and bottom Plexiglas bridges; f, filter 1, fresnel lens; Is, light source; p, phototransistor; pp, darkened Plexiglas plate with clear slit; r, aluminum light reflector; w, clear Plexiglas window. xviii
LIST OF FIGURES (cont'd)
Page
Figure 43 Differences in the percentage of the fish 177 that were day-active in LD 12:12 for one day (A) and in the proportion of calculated periods that are significantly different from random "noise" (B) with different intensity of illumination. Also the proportion of the four types of activity pattern observed under the three constant light intensities are given (C). The numbers above the bars are the numbers of fish examined under each light intensity.
Figure 44 Variations in the free-running period (A) and 180 in the mean swimming speed (B) with intensity of constant illumination (log scale). Open and closed circles in panel A denote non-significant and significant periods, respectively. Regression line A is
Y =27.23 - 0.85 log10 X (r = -0.19, P > 0.05)
and line B is Y = 33.41 + 18.10 log10 X (r = 0.64, P < 0.001, one-tailed).
Figure 45 Examples of the four types of free-running 181 activity pattern of juvenile pink salmon recorded in LL. Type 1 (A), Type 2 (B), Type 3 (C) and Type 4 (D) patterns are shown, and the periodogram of each pattern is given. The fish were subjected to a LD 12:12 cycle during the first day. Thereafter, open and shaded horizontal bars indicate the subjective day and night, respectively, in LL.
Figure 46 Two examples of the activity pattern of pink 187 salmon recorded during the 15-day "Catch the free-running rhythm" experiment. For Days 1-5 and Days 11-15 light and dark horizontal bars indicate the day and night phases of the LD 12:12 cycle, respectively. For Days 6-10 open and shaded bars indicate the subjective day and night, respectively, in LL. Periodograms are given for each of the 5-day sections of the data record.
Figure 47 Two typical examples of the diel swimming activity 190 pattern of pink salmon recorded for five days prior to a phase shift in the LD 12:12 cycle, and for another six days after the phase shift. Periodograms are given for each of these two sections of the data record. Light and dark horizontal bars are as in Fig. 10. xix
ACKNOWLEDGEMENTS
I am grateful to my supervisor, Dr. W. S. Hoar for his generous moral and logistic support during the course of this study and constructive criticism of earlier drafts of this manuscript. Accordingly, I dedicate this thesis to him on the occassion of his official retirement from the Department of Zoology at the University of British Columbia. Many thanks are due to Dr. C. Groot at the Pacific Biological Station, Nanaimo, B.C., who made research facilities available to me and provided helpful criticism during this study. I also thank members of my supervisory committee for critically reading an earlier draft of the thesis.
I appreciate the assistance of Dr. M. Healey and Mr. F. Jordan in the diel sampling of fish in the field. F. Nash and J. Thompson wrote some of the computer programs used in this study. C. Armstrong, G. Atkinson, M. Ilich, and J. Rabeneck helped in the construction of laboratory equipment. Last but not least, I owe special thanks to my wife, Heather for her lasting patience and moral support during the course of my graduate studies.
This work was supported jointly by the Federal Fisheries and Marine Service and by an NRC grant-in-aid to Dr. W. S. Hoar. I was supported by scholarships from the National Research Council of Canada, the Killam Foundation, and the Salmon Research Society of B.C., and also by a Teaching Assistanship from the University of British Columbia and a Research Assistanship through an NRC grant-in-aid to Dr. W. S. Hoar. I am grateful for this financial support. 1
CHAPTER I. GENERAL INTRODUCTION
Biological oscillations are ubiquitous and have been documented in organisms ranging from unicellular algae to man and at several hierarchial levels of biological organization (e.g., Aschoff 1960, 1965, Chovnick 1960, Menaker 1971, Bunning 1973, Palmer 1974, 1976, Thorpe 1978). The periods of these oscillations tange from the milliseconds of neuronal cycles to the years of population-cycles. Such biological rhythms may be considered adaptations to periodic environmental conditions. The most fully documented of biological
rhythms is the daily or diel rhythm,:. which is normally synchronized or entrained by the daily cycle of illumination in nature. Little is known about daily rhythms and their timing mechanisms in fishes compared to plants, insects, and higher vertebrates (Schwassmann 1971a).
Pink salmon (Oncorhynchus gorbuscha Walbaum), one of the six species of Pacific salmon, has an invariable 2-year life cycle, unlike other salmonid species (Hart 1973) . The fish generally spawn in coastal streams of western North America in September and October. The female buries her eggs in the gravel substrate of the stream. The embryos hatch in winter and the alevims emerge from the gravel in the spring; the timing of these events is temperature dependent. Free-swimming fry migrate downstream toward the estuary immediately or shortly after emergence. Juvenile fish remain in nearshore marine areas throughout their first spring and summer. Thereafter, they migrate offshore into deeper coastal waters, and are found in the open North Pacific ocean by late'fall. The fish move southward to 2 overwinter in the southern part of the Gulf of Alaska (Neave 1964,
Royce et al. 1968). The coastal spawning migration of maturing adults returning to their natal rivers in British. Columbia occurs primarily from June until September, and after a variable holding period at the mouth, of their Lome river they migrate upstream, spawn, and die.
Pink salmon undergo at least seven major, habitat changes during their life history. These are 1) from the embryo in the egg to the alevin in the gravel, 2) from the alevin in the gravel to the fry in the river, 3) from the river to estuarine and inshore coastal waters as juveniles, 4) from inshore coastal waters to deeper offshore coastal waters as juveniles, 5) from offshore coastal waters to the open ocean as juveniles, 6) from the open ocean back to coastal waters as maturing adults, and 7) from coastal waters to their spawning grounds as matured adults. These habitat changes occur in a well-timed sequence, and result from a number of movements or migrations (Royce et al. 1968).
The invariable 2-year life cycle of pink salmon and the relatively accurate seasonal timing of its migrations suggest that this species possesses an annual biological "clock" or timing mechanism. Hoar (1965, 1976) and Poston (1978) reviewed evidence suggesting the existence of an endogenous annual rhythm, which is timed by photoperiod and modified by temperature, regulating morphological, physiological, and behavioral changes associated with smoltification in salmonids. Annual rhythms or seasonally-timed events may result from the interactions between daily rhythms and 3 annual changes in environmental factors , - especially daylength
(Gwinner 1973, Menaker 1974). Therefore, knowledge of daily behavioral rhythms in pink salmon may contribute to a better understanding of the seasonal timing of its migrations.
The objective of this study was to investigate, in a seasonal context, diel rhythms of ecologically-relevant behavior in juvenile pink salmon, and their timing mechanisms. The.general strategy adopted was to record diel behavioral rhythms in the laboratory under various controlled light and temperature conditions.
These rhythms were recorded during the period corresponding to the time of gravel emergence of wild fry until the onset of migration of maturing fish from the open ocean toward the coast.
Firstly, the effect of water temperature on the diel pattern of fry emerging from a simulated "nest or redd" was tested. Behavior associated with gravel emergence results in a change from a gravel habitat to an open-water, stream habitat. Secondly, I examined ontogenetic changes in the diel rhythm of swimming activity and of vertical distribution in a water column for groups of fry after emergence, when they normally migrate downstream to the estuary. The diel rhythm of feeding activity of wild fry in the littoral zone of two marine bays was also investigated. In an attempt to gain a better understanding of the factors affecting the timing and the
intensity of feeding behavior, the daily pattern of fish feeding
activity was recorded under controlled laboratory conditions. Seasonal 4 changes in the diel patterns of swimming activity, aggression, and vertical distribution were also investigated during the period beginning when wild fry normally reside in estuaries and adjacent coastal waters and ending when maturing fish begin their migration from the ocean toward the coast. Lastly, I examined the timing mechanisms of the diel rhythm, of swimming activity. Specifically, the existence of a daily biological clock and the role of the daily light-dark (LD) cycle in the entrainment of the rhythm of swimming activity were examined. 5
CHAPTER IX. GENERAL MATERIALS AND METHODS
A. Fish, and incubation procedures
Experimental fish used in this study originated from mature adults captured by seining in the Eve River, Vancouver Island,
British Columbia (50°15'N, 125°55'W) in September of 1974-77.
Generally, the eggs from three to seven females were pooled and fertilized in the field with the pooled sperms of a similar number of males. The method used to collect the gametes and fertilize the eggs is described by Huet (1972). For experiments reported in
Chapter III and in Chapter IV, eggs of individual females were fertilized separately with the pooled sperms of five to nine males, and later incubated separately. The zygotes were transported in buckets, placed on ice in styrofoam coolers, to the Rosewall Creek
Hatchery on Vancouver Island. Transit time was about four hours.
At the hatchery the zygotes were enumerated volumetrically (and later electronically), dosed in a germicide (Weseodyne) solution for 10 min, placed in trays of a Heath incubator, and incubated in constant darkness and in running ground or well water at ambient temperature.
After hatching the alevins were transported in darkened fry cans to the Pacific Biological Station in Nanaimo, British Columbia
(49°13'N, 123°57'W). Upon arrival the alevins were placed into small gravel incubators, similar in design to the one described in
Chapter 111, with running fresh water at ambient or controlled temperature. Fresh water originated from the city water supply, but was dechlorinated and filtered to 5 um before fish were exposed to it. 6
B. Holding conditions
Immediately after emerging from the gravel incubators the
fish were placed into 484-L fiberglass holding tanks with running sea water of natural salinity and temperature. Sea water was pumped from
a depth, of 21 m in nearby Departure Bay, and passed through sand filters before entering the holding tanks. Each, holding tank was equipped with a light-proof plywood hood. Illumination was provided by an
incandescent bulb (General Electric Ltd.) fixed to the underside of the hood. Appendix Figure 1 shows the power spectrum of this light. The
LD cycle, without twilight periods, was regulated by an interval timer.
Light intensities and LD cycles varied according to experimental
procedures. Generally, the fish were fed ad libitum with a mixture of
small pieces of frozen zooplankton, liver, and Oregon Moist Pellets
(OMP), a prepared fish food, five to eight times daily. Feeding times were randomized every day. The fish stocks were periodically examined
for diseases by the Fish Health Group at the Pacific Biological Station,
and were always found to be in a healthy condition.
C. Experimental conditions
All experiments reported were conducted at the Pacific
Biological Station during 1974-78. The LD cycle and temperature
regime were controlled, but differed among the different experiments.
The LD cycle was controlled by an interval timer, but no twilight periods were interposed. Illumination was incandescent. To
facilitate night observations, light intensities during the scotophase
(night) of the experimental LD cycle were similar to natural light 7 intensities occurring during dusk and early evening periods. All light intensities were measured at the water surface, unless specified otherwise. Water temperature was manually controlled by mixing water of different temperature, originating from main lines, with the aid of ball and diaphram valves. Water temperature during experiments reported in Chapter VII was thermostatically controlled.
Fish were not fed during the experimental periods because preliminary observations showed that feeding on external food enhanced their swimming activity. Davis and Bardach (1965) and Byrne (1968) found that feeding fish at a particular time of day affected the timing of their swimming activity rhythms.
D. Statistical procedures
All data were coded on IBM cards, and analyzed by a Xerox
530 computer. Parametric and non-parametric statistical tests were from Sokal and Rohlf (1969) and Siegel (1956), respectively, unless stated otherwise. The significance level for rejecting a true null hypothesis was set at 5%. 8
CHAPTER III. EMERGENCE FROM A SIMULATED REDD
INTRODUCTION
As in other salmonids, eggs of pink salmon are buried in river gravel by females in late summer and fall. The resulting embryos and alevins develop within the gravel, and emerge later from
it as free-swimming fry (Hart 1973). During the spring, apparently
immediately or shortly after emerging from the gravel redds, pink
salmon fry actively or passively migrate seaward. In most rivers
this migration occurs mainly at night, with the diel peak one to
four hours after sunset (Pritchard 1944, MacKinnon and Brett 1955,
Neave 1955, Roppel 1956, Ali and Hoar 1959, Hunter 1959, Andrew and
Geen 1960, McDonald 1960, Kirkwood 1962, Smirnov and Kamyshnaya 1965,
Ishida 1966, Kobayashi and Harada 1966, Vernon 1966, Coburn and.
McCart 1967, Bakshtanskii 1970, Smirnov 1975).
Several authors have stated that the downstream migration
of pink salmon fry occurs immediately or shortly after emergence
(Neave 1955, 1966, Hoar 1958, Andrew and Geen 1960, Ishida 1966,
Kobayashi and Harada 1966, Smirnov 1975). Assuming that nocturnal
seaward migration occurs immediately or shortly after emergence,
the diel pattern of emergence should be mainly nocturnal, and should
show a diel peak within the first four hours following sunset. The
objective of this portion of the study was to test this prediction 9 by examining the diel pattern of sibling fry emerging from a simulated redd exposed to an artificial LD cycle. The influence of temperature on the diel pattern of emergence was also examined.
MATERIALS AND METHODS
During the fall of 1976 and 1977, three similar-sized adult female pink salmon (Appendix Table 1) were obtained from the Eve River, Vancouver Island, British Columbia. The eggs from these females were fertilized separately with pooled sperms of several males from the same river (1976: X ± SD = 52.7 ± 4.2 cm total length, 1.41 ± 0.17 kg wet weight, n = 5; 1977: 64.1 ± 4.3 cm, 2.57 ± 0.70 kg, n =9). No ripe females were found in this river in 1977 similar in size to those captured in 1976. Zygotes originating from each female were incubated in darkness in separate Heath trays with running ground water (incubation temperature: X ± SD =7.1 ± 0.48°C in 1976; 8.0 ± 0.31°C in 1977). About five days after median hatching time, 1,000 sibling alevins (newly hatched fish larvae with large yolk sacs) from each tray were introduced under dim red light into each of six simulated redds (Fig. 1). The mean weight of the alevins and their mean yolk weight are listed in Appendix Table 2.
Sibling fish in each redd experienced a different temperature regime from hatching to emergence (Table 1). An attempt was made to maintain water temperature constant in each, redd, However, daily temperature fluctuations occurred. Water temperature in each redd o
Inflow
Fig. 1. Simulated "redd" with front side removed to show the gravel. See text for further details. 11
Table 1. Mean ± SD daily water temperature experienced by sibling alevins in gravel from median hatching time to the time of 50% emergence for each of six simulated redds.
Brood Redd Temperature
year number, (°C)
1976 1 3.4 + 0.56
1977 2 5.0 + 0.35
1976 3 7.9 + 1.08
1977 4 9.9 + 0.45
1976 5 12.3 + 1.80
1977 6 15.0 + 0.55 12 was measured with, a mercury thermometer at least three times daily.
Temperature readings taken daily for each, redd were averaged separately for that day; this average temperature is henceforth referred to as daily mean temperature. Daily mean temperatures of each redd were averaged separately over the days of the alevin phase
(from hatching to gravel emergence); this grand mean daily redd temperature is listed for each redd in Table 1, and is henceforth referred to as mean temperature or simply temperature.
Alevins instead of eggs were introduced into the redds because I wanted to restrict the effect of experimental temperature on the behavior of the alevins, which move about within the gravel. Hence, sibling embryos from each female experienced a similar thermal regime before hatching, but experienced a different one, as alevins, after hatching.
The redd, modified from that of Dill (1970), consisted of a water-tight 1.91-cm plywood box (51 x 45 x 38 cm) painted flat black and filled to about 25 cm in depth with clean gravel (X ± SD pebble diam = 1.75 ± 0.68 cm, composed of 83% "round" and 17% "crushed" pebbles). The box was insulated on the outside with sheets of styrofoam. Water entered the redd through the side of the box (Fig. 1), and was distributed through a circular (30-cm diam) perforated manifold located on the bottom of the box. This provided the redd with upwelling freshwater, filtered to 5 um, with a flow
-1 rate of 4.0 ± 0.2 L min , and an apparent bulk water velocity of 13
-1 1.74 ± 0.09 cm min . This flow is well within the recommended range for the incubation of pink salmon eggs and alevins in gravel incubators (Bams and Simpson 1977). Periodic sampling showed that dissolved oxygen levels in the water ranged from 96 to 105% of air saturation, and that un-ionized ammonia levels in the water were
-1 less than 0.15 ug L in all redds. An outflow pipe provided a water depth of about 8 cm above the gravel surface.
Since salmonid fry are known to re-enter the gravel after their initial emergence (Stuart 1953, Neave 1955, Hoar 1958, Smirnov 1975), I attempted to prevent this. Accordingly, a clear polyethylene plastic sheet was placed on the gravel surface and sealed at the edges of the redd. A number of evenly spaced tubes (2 cm diam x 3 cm high), fixed over perforations on the plastic sheet, projected above the gravel surface. This plastic sheet did not prevent fry from emerging from the gravel and subsequently emigrating out of the redd. Preliminary visual observations showed that newly-emerged fry swam along the underside of the plastic sheet until it located one of the perforations in the sheet; this was achieved usually within three minutes after emergence. The fry then swam vertically up through the plastic tube into the water column above the plastic sheet. The tubes on the sheet acted as one-way release mechanisms for emerging fry (Fig. 1). This procedure was highly successful in preventing fry from re-entering the gravel (Appendix Table 3). 14
Each redd was exposed to 12 hours of incandescent lighting (ca. 470 lx) alternating with 12 hours of "darkness" (ca. 10 lx). Appendix Table 4 shows incident light intensities penetrating to various depths in the gravel. These light intensities would likely be different if the redd contained water. Alevins were introduced into the gravel via long PVC pipes extending from above the water surface to within about 3-4 cm of the bottom of the gravel. Within two hours all alevins had vacated the pipes and entered the gravel interstices; the pipes were then removed. The fish were allowed to emerge spontaneously from the gravel, and were collected at the outflow in a plastic container from which they were removed twice daily at the times of lights-on (0800 hours) and lights-off (2000 hours). At these times, emerged fish remaining in the water column above the plastic sheet covering the gravel (< 5% of total emerged) were removed with a dipnet. Near or on the modal day of emergence from each redd, fry were collected hourly for 24 hours. The modal day of emergence was estimated in advance based on preliminary observations. Emerged fry were anesthetized with MS-222 and preserved in 10% formalin for later measuring and weighing.
Mean duration (days) of the alevin stage was considered to be the mean age of fry at emergence. This was computed for each redd as follows. Each day of emergence was rescaled to the number of days elapsed since the median date of hatching for that group. This number henceforth is referred to as day of emergence. Each day of emergence was then divided in two; the first half (eg., day 100.0) 15
was assigned to diurnally emerging fry and the second half (e.g., day 10Q.5) to nocturnally emerging fry, This procedure was followed as fry in the latter period emerged later in the 24-h day, which arbitrarily began at 0800 hours (= lights-on). Each part-day number (e.g., 100.0, 100.5) then was transformed to logarithms to normalize the temporal pattern of emergence of fry from each redd. In addition, diurnally and nocturnally emerging fty were considered separately. Hence, the mean age of fry at emergence was expressed as X = XI (f l°gio X)/N> where f is the number of fish emerging per half-day, X the half-day of emergence, and N the total number of fry emerging from each redd.
Generally, the fry remained in formalin for about 100-150 days prior to measuring and weighing. Each fish was removed from the preservative and its fork length (from tip of snout to the most anterior point in the fork of the tail) determined to the nearest 0.1 mm using a binocular microscope. The fish then was placed in an aluminum dish, dried to constant weight in an oven at 60°C, and subsequently weighed to the nearest 0.1 mg. A daily sample of up to 20 fish was used to obtain length and weight measurements. On the days when 10, 50 and 90% of emergence was completed for each redd, the daily sample of 20 fish was used to determine the amount of yolk remaining. This was done by making a mid-ventral incision along the body of the fish, and removing with forceps the formalin-hardened yolk, which then was placed on a pre-weighed aluminum disc. At redd temperatures from.3.4 to 7.9°C the small yolk remnant was sometimes 16 viscous. Its complete'removal from the fish's body cavity presented some difficulties, resulting in unknown errors of measurement. The fry and its excised yolk then were dried to constant weight in an oven at 60°C, and subsequently weighed separately to the nearest 0.1 mg.
All daily samples of the emergence period were pooled for each redd. The means for dry fish weight (fish body including yolk) and fork length were calculated from data normalized by logarithmic transformations. For each redd a weighted mean yolk reserve in the fry was computed from pooled data for fish emerging at the 10, 50 and 90% percentiles. This weighted mean represents the mean amount of yolk remaining in individual fry at the time of emergence. Mean dry body weight of emerged fry from each of the six redds was computed by subtracting the weighted mean yolk weight from the mean dry fish weight.
RESULTS
A. Temporal pattern of emergence The temporal pattern of emergence of sibling fry from the simulated gravel redd was generally unimodal for all temperatures except at 12°C, where some bimodality was noticeable (Fig. 2). At all temperatures the emergence pattern departed significantly from normality (Chi-square test). The skewnesa of the patterns is not consistent. The maximum number of sibling fry emerging during any 17
° 4™f~1—I t I r**r~r—i—i—I—i—i—I—i—i—r—i—r—i—r—i—I—i—r—i—I—r—i—ri I I—I—I 10 22 24 36 38 30 33 34 36 36 40 42 44 46 4C 50 53 54
Day of emsrgsnce
Fig. 2. Histograms of the frequency of sibling fry emerging from simulated redds in relation to days after median hatching time. Distributions shown are for mean temperatures of 3.4, 5.0, 7.9, 9.9, 12.3 and 15.0°C. The triangle indicates the day during which emerging fry were sampled hourly for 24 hours. 18
24-h period never exceeded 20% of the total number emerging, except at:7.9°C. A significant positive correlation (r =0.57) between daily mean temperature and daily percentage of fry emerging was found at 3.4°C, but at no other temperature.
Figure 3 shows the cumulative percent of sibling fry emerging in time at each temperature. All curves are sigmoid, but differ significantly from each, other in their slopes (F = 179.4; d.f. = 5, 122; analysis, of covariance). Although the shape of the curves for 9.9 and 12.3°C are somewhat similar, their slopes -(Table 2) are significantly different from one another (F = 481.4; d.f. =1, 42). s The slope of the sigmoid curve is an index of the rate of sibling fry emerging from the redd. A significant negative correlation (r = -0.89) existed between mean experimental temperatures and their associated regression coefficients for the rate of emergence (slope of the cumulative percent emergence curve). Therefore, with one exception (5.0°C), the rate of emergence declined as mean experimental temperature increased (Table 2), owing in part to the positive relationship observed between the duration of the emergence period and mean experimental temperature (Fig. 4). This indicates that the temporal synchrony of sibling fry emerging from the same redd declined with increasing temperature. Mean age of fry at emergence (mean duration of the alevin stage) also declined with increasing mean experimental temperature (Fig. 4, Table 3). 12.3C 15.0 C i 7.9C 5.0 C
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Day of emergence
Fig. 3. Relation between time (day after median hatching time) and cumulative percentage of sibling emerging from simulated redds at different mean temperatures. 20
Table 2. Regression coefficients (b) and their 95% confidence limits (CL) for Probit Y = a + b logj^g x, where Y is the cumulative percentage of fry emerging from a simulated redd at different mean temperatures, and X the number of days elapsed since median hatching time (Day 0).
Mean temperature (°C) b ± 95% CL
3.4 121.24 ± 7.12
5.0 69.06 ± 2.81
7.9 99.06 ± 2.85 9.9 42.83 ± 2.70 12.3 36.51 ± 1.79 15.0 14.47 ± 0.85 (0 140 -, CA r 35 >* >. CO T3 CO B -a h 30 o 120 3 k_ A3 a> o h 25 c • oo H a
CO J >- 0 0 i_ 3 r 1 1 1 1 1 1 1 1 1 1 o 3 6 7 8.9 10 11 12 13 14 15
Mean temperature (°C) Fig. 4. Relation between mean redd temperature and A) mean age of fry at the time of gravel emergence (days after median hatching time), and B) duration of the emergence period (days). Regression line A is Y = 146.8 - 7.52 X (r = -0.97, P < 0.001) and line B is Y = 11.2 + 1.26 X (r = 0.73, P < 0.05). Table 3. Geometric mean (95% confidence limits) age at emergence (days after median hatching time = mean duration of the alevin stage in gravel) for pink salmon fry at different mean redd temperatures. Geometric mean ages at emergence for sibling fry emerging during light and dark of the experimental LD cycle are also shown for each redd, and are compared using the t-test.
Mean Age at emergence Redd temperature Age at emergence Light Dark
number (°C) (days) (days) (days) t d.f.
1 3.4 131.2 131.1 131.2 0.43 840 (131.1 - 131.3) (130.8 - 131.5) (131.0 - 131.4) * 2 5.0 107.0 106.6 107.1 989 (106.8 - 107.2) (106.2 - 106.7) (106.8 - 107.4) 2.11
3 7.9 78.5 78.6 78.5 0.79 941 (78.4 - 78.6) (78.3 - 78.8) (78.3 - 78.6) 4 9.9 62.0 62.2 62.0 0.49 1000 (61.9 - 62.2) (61.5 - 62.9) (61.8 - 62.2) ** 5 12.3 61.9 63.8 61.4 989 (61.7 - 62.2) (63.5 - 64.2) (60.9 - 61.6) 7.84 ** 6 15.0 37.5 39.3 35.8 9.91 991 (37.2 - 37.8) (38.8 - 39.8) (35.3 - 36.3)
*, P < 0.05; **, P < 0.001 23
Estimated fish survival, within the simulated redd, from hatching to emergence was greater than 80% at all temperatures (Appendix Table 5). No significant relationship (r = 0.66) existed between percentage survival and temperature. Percent survival was overestimated (100.2%) for redd number 4 (9.9°C). This is attributed to an error in counting the 1,000 newly hatched alevins introduced originally into that redd. The cumulative number of thermal units (degree Celsius-days) experienced by the alevins from hatching to 50% emergence was not constant over the range of temperatures examined, nor did it vary linearly with mean temperature (Appendix Table 5). Mean lengths and weights of emerged fry and their mean yolk remnants are listed in Appendix Table 6.
In summary, the distribution of sibling fry emergence in time was generally unimodal, but differed from a normal distribution. The rate of emergence decreased as experimental temperature increased. This means that the temporal synchrony of sibling fry emerging from the same redd declined with increasing temperature.
B. Diel timing of emergence Figure 5 summarizes the diel emergence pattern of sibling fry at different mean experimental temperatures. At all temperatures, except 15.0°C, about 80% of the sibling fry emerged during the night. The ratio of day to night emerging sibling fry differed significantly from a random 1:1 ratio (Chi-square test) for all temperatures except 3.4 5.0 7.9 9.9 12.3 15.0
Mean temperature (°C)
Fig. 5. Frequency (percent) of sibling fry emerging from simulated redds during the day (open bars) and the night (dark bars) of a LD 12:12 cycle at different mean temperatures. The number above the bars for each redd temperature indicates the number of emerged sibling fry. 25
15°C. In general, diurnally emerging fry emerged later compared to nocturnally emerging siblings (Fig. 6).. The mean age at emergence of diurnally emerging fry was greater than that of nocturnally emerging
siblings for four of the six experimental temperatures (Table 3);
this difference was significant only for the two highest temperatures.
The more detailed diel pattern of hourly emergence, obtained for one day near or on the modal day of emergence for each of the six redds (see triangles in Fig. 2), reveals that significantly (Chi-square test) more fry emerged during the night than during the day at all daily mean temperatures except 15.1°C (Fig. 7). A peak of fry emergence occurred between 1000-1200 hours of the photophase in redds having a daily mean temperature of 4.0 and 4.9°C. Such a peak was absent from the diel pattern of the other redds. The daily peak of emergence oecurred within the first three hours after the onset of darkness at all daily mean temperatures except 4.0 and 15.1°C (Fig. 7).
The daily mean time of emergence during the diel cycle occurred about one hour before the onset of darkness at 4.0°C (Fig. 8). As the daily mean temperature increased, daily mean time of emergence occurred progressively later into the night, reaching its latest time at about three hours after lights-off at 9.9°C. The mean then occurred progressively earlier in time with further increases in temperature. At 4.0 and 15.1°C daily mean time of emergence occurred about one hour before dark. Daily mean time of emergence corresponded closely to the daily peak of fry emergence at all temperatures except the two extreme 26
Day of emergence
Fig. 6. Histograms of the frequency of diurnally (open bars) and nocturnally (dark bars) emerging sibling fry in relation to days after median hatching time. Distributions shown are for mean temperatures of 3.4, 5.0, 7.9, 9.9, 12.3 and 15.0°C. Triangles are as in Fig. 2. 27
15. I C N = 103
Time of day (h) Fig. 7. Frequency distributions of sibling fry emerging from different redds during the day-(open bars) and the night (dark bars) of a LD 12:12 cycle on or near the modal day of emergence. Distributions are shown for redds at daily mean temperatures of 4.0, 4.9, 8.0, 9.9, 12.2 and 15.1°C. In ascending order these temperatures are for redds 1 through 6. 0200-1
2400H tt)0) (0 -C a o o 2200H o CO
2000 0) E in 1800H (0 JC a o o .c 1600J a.
T 'T• 5 11 13 IS n 1 7 Mean temperature (°C)
Fig. 8. Relation between daily mean (± 95% confidence limits) time of emergence of sibling fry from a simulated redd during one 24-h period (see Fig. 7) and daily mean temperature. The broken line denotes the onset of the scotophase (night). 29 ones (see triangles in Fig. 7). Hence, the diel timing of fry emergence varied in a non-linear manner with, temperature. With time of day held constant statistically, no significant correlation was found between the frequency of fry emerging during a particular hour of the diel cycle, and the water temperature during that hour, for any of the redds except the one at:9.9°C (r , = 0.45, partial y 1 • 2 product-moment correlat ion, Steel and Torrie 1960).
Fry emerging during the day were, on the average, longer than siblings emerging during the night of the same diel cycle in only one out of eight instances, but the diurnally emerging fry were heavier in five out of eight instances (Appendix Table 7). However, differences in length and weight between diurnally and nocturnally emerging sibling fry were not significant.
In summary, emergence was mainly nocturnal at all temperatures except 15°C. Generally, diurnally emerging fry emerged later, on the average, than their nocturnally emerging siblings at temperatures above 7.9°C. The onset of darkness appeared to be the major time cue for emergence at all temperatures except 15.0°C. In general, the daily peak of fry emergence occurred within three hours after the onset of darkness. However, temperature affected the diel timing of emergence in a non-linear manner. 30
DISCUSSION
The distribution of sibling fry emergence in time was generally unimodal, but differed from a normal d istribution. At all temperatures examined, except 15°C, about 80% of sibling fry emerged during the night, and the daily peak of fry emergence generally occurred within three hours after the onset of darkness. These findings conf irm that emergence occurs shortly before nocturnal seaward migration in wild fry, as postulated previously by several investigators (Neave 1955, 1966, Hoar 1958, Andrew, and Geen 1960, Ishida 1966, Kobayashi and Harada 1966, Smirnov 1975).
Predominantly nocturnal emergence from natural and simulated redds has been reported for sockeye salmon, Oncorhynchus nerka (Heard 1964, Bams 1969) and for rainbow trout (Salmo gairdneri) at 15.0°C (Dill 1970). Conversely, more trout emerged during the day than during the night at 10.0°C; but the total number of emerged fry was only 36 (Dill 1970). No significant preference for either diurnal or nocturnal emergence was observed for the redd at 15.0°C in the current study. A similar observation was reported for coho salmon (Oncorhynchus kisutch) at temperatures ranging from 7.8 to 11.7°C (Mason 1976) and for Atlantic salmon (Salmo salar) at 10 and 15°C (Dill 1970). These observations indicate that the LD cycle is an important synchronizer of diel gravel emergence in rainbow trout, sockeye salmon, and pink salmon, but that the synchronizing effect varies with temperature. The reasons for its weak influence on the 31 diel timing of emergence in coho salmon, Atlantic salmon and pink salmon at some temperatures are unknown. In the current study the onset of darkness appeared to be the portion of the LD cycle that timed daily emergence at temperatures below 13.0°C.
The thresholds for photopic and scotopic visions, based on histophysiological criteria in pink salmon fry, are about 1.0 and 0.001 lx, respectively (Ali 1959). Since these light intensities occurred within 7 to 9 cm of the gravel surface in the current study
(Appendix Table 4), it is possible that pink salmon alevins could have perceived diel changes in incident light intensity at these depths in the gravel. Hence, the fry may have synchronized their diel motor activity rhythms with the diel LD cycle as they ascended through the gravel and approached its surface. This suggestion is supported by the following observations. Dill (1970) observed diel motor activity rhythms, which were synchronized with an artificial
LD cycle, in alevins of Atlantic salmon and rainbow trout about five days after hatching in gravel-free activity chambers. Dill and
Northcote (1970) noted that coho salmon alevins in an artificial redd appeared to be more active during the night than during the day.
My data suggest that as much as 20% of pink salmon fry emergence from natural and artificial redds is diurnal at temperatures ranging from 3.0 to 13.0°C; this range approximates the normal range of intragravel temperatures in many streams and rivers on Vancouver
Island, British Columbia during the normal incubation period for pink 32 salmon (W.P. Wickett, personal communication). Heard (1964) observed that sockeye salmon fry emergence from natural redds was. mainly nocturnal, although about 15% of the fry emerged during daylight hours. However, because these diurnally emerging fry likely re-enter the gravel, or hold position in the water current during light (e.g., Neave 1955, 1966, Hoar 1958, Coburn and McCart 1967,
Bakshtanskii 1970, Smirnov 1975), they occur infrequently in fry traps in streams or in the collecting devices of gravel incubators.
On the average, diurnally emerging pink salmon fry emerged later than their nocturnally emerging siblings at temperatures equal to or greater than 7.9°C. Mason (1976) observed a similar pattern of later emergence in sibling coho salmon fry. This coincided with a gradual increase in light preference in siblings of similar age denied gravel experience. Bakshtanskii (1970) reported that pink salmon fry migrations in some Soviet rivers are composed primarily of nocturnal migrants during the early portion of the migration, but of diurnal migrants during its later portion. He also reported that nocturnal migrants were relatively more photonegative than diurnal migrants. McCart (1967) noted that daytime downstream migration of sockeye salmon fry occurred on the average later in the season than nocturnal migration^ when temperatures exceeded 10°C. These observations suggest that the increased tendency of the fry of pink, coho and sockeye. salmon to emerge and migrate downstream during light, as the emergence period progresses, may relate to a gradual weakening of the fry's negative phototactic response as the fry ages and increases 33 in size. This argument is demonatrated in the data of Mason (1976), and indicated by White Q-915), Stuart (1953), Woodhead (1957), Roth and Geiger (1963), Dill and Northcote (1970) and Smirnov (1975).
This ontogenetic shift in photoreaponse. may depend on changes in the fish's retinal elements and photomechanical responses (Ali 1959).
Unexpectedly, the daily mean time of emergence during the diel cycle varied in a non-linear manner with daily mean temperature (Fig. 8). The latest daily mean time of emergence during the night occurred about three hours after lights-off at 9.9°C. However, daily mean time of emergence occurred progressively earlier relative to lights-off as daily mean temperature increased above or declined below 9.9°C. The latfeE occurrence of daily emergence at lower daily mean redd temperatures between 15.1 and 9.9°C could result from a' reduction in the amount of spontaneous locomotor activity of the fish with declining ambient temperature (Brett 1970) , resulting in a slower fish ascent through the gravel after dark. Bams (1969) suggested that the speed of travel of fry through the gravel is determined largely by the size of gravel interstices and by temperature. The relatively earlier occurrence of daily mean time of emergence at still lower daily mean redd temperatures between 9.9 and 4.0°C is difficult to explain. The earlier time of emergence at the lowest two temperatures results partly from the high proportion of diurnal emergence between 1000 and 1200 hours. Also, the fish may have positioned themselves closer to the gravel surface during the photophase at these two temperatures than at the higher ones, and 34 consequently had shorter distances to travel through the gravel to reach the surface during the scotophase. Visual observations of fry emerging from simulated redds at different temperatures are needed to test these suggestions.
Mean age at emergence (duration of the alevin stage) declined with increasing temperature. Appendix Figure 2 illustrates the decline in the age at emergence with increasing temperature for several salmohid species. In the majority of the studies summarized in this figure fry emerged from simulated redds. In some cases the references lacked information on age at emergence; consequently, age at emergence was estimated from published data on hatching times at different temperatures. Values for different species at a given temperature in Appendix Figure 2 are not directly comparable because of differences in experimental design among studies. The general decline in age at emergence with increasing temperature is probably a result of more rapid rates of development in fishes at higher ambient temperatures (Garside 1966, Bams 1967, Blaxter 1969a, Brett 1970, Peterson et al. 1977, Alderdice and Velsen 1978). Hence, temperature may determine the time at which pink salmon fry emerge from natural redds and migrate seaward. The seasonal timing of entry into estuaries presumably is of some importance to marine survival and growth (Koski 1975, Belford 1977).
Temperature also affected the rate of emergence negatively.
This means that the temporal synchrony of sibling fry emergence decreased with increasing temperature. A similar result was obtained 35 by T.A. Heming (unpublished data) for chinook salmon (Oncorhynchus tshawytscha) fry emerging from a simulated redd. Data for steelhead trout, Salmo gairdneri (Shapovalov 1937) also indicate lengthening of emergence period with increasing temperature. The above results are unexpected in the context of the commonly observed increase in the rate of biological processes in fishes with increasing ambient temperature (Brett 1970, Fry 1971, Hochachka and Somero 1973). The physiological correlates of emergence behavior are unknown, and an explanation of these results is difficult.
The duration of the emergence period and seaward migration period in a natural population of pink salmon reflects the duration of the adult spawning period and the temporal synchrony of fry emerging from different redds. The latter can be influenced by temperature (this study), sedimentation (Koski 1966, Mason 1976), and gravel size (Dill and Northcote 1970), among other factors. The degree of synchrony in emergence timing between redds within a population could have important influences on the predation rate on migrant fry during their seaward migration. Peterman and Gatto (1978) postulated that the functional response curve for fish predators of pink salmon fry is such that percent fry mortality per predator increases with increasing fry density up to a maximum at the lower fry densities. However, relative fry mortality declines slowly with further increases in fry density, owing to predator satiation and (or) prey handling time constraints. Neave (1953) defined an inverse 36 relationship between percent mortality and prey numbers as
"depensatory" mortality. Hence, the greater the degree of synchronization in fry emergence within and between redds, the greater the number of fry that subsequently will migrate seaward on any particular day. In turn this potentially could reduce the relative mortality rate of these fry. Moreover, the schooling behavior of seaward migrants (Hoar 1958, 1976, Bakshtanskii 1970,
Smirnov 1975) could further reduce their susceptibility to predation
(Neill and Cullen 1974, Morse 1977, Major 1978).
The tendency of pink salmon fry to emerge mainly at night is considered an anti-predator adaptation because predation likely declines at decreasingly low incident light intensities, since the visual acuity and capture efficiency of fish predators also show a corresponding decline (Brett and All 1958, Ali 1959, Brett and Groot 1963, Ginetz and Larkin 1976). Hence, the diel pattern of fry emergence in nature may result from selection from fish predators. 37
CHAPTER IV. ONTOGENY OF DIEL RHYTHMS OF SWIMMING ACTIVITY AND OF VERTICAL DISTRIBUTION
INTRODUCTION
Data presented in Chapter III indicated that emergence of pink salmon fry from gravel and their subsequent seaward migration are mainly nocturnal. In general, the majority of migrant fry enter estuaries in British Columbia from March until May (Vernon 1958, 1966, Gilhousen 1962, LeBrasseur and Parker 1964, Neave 1966), and remain in nearshore areas of estuaries and adjacent coastal waters through their first spring and part of that summer (Gilhousen 1962, Manzer and Shepard 1962, Parker 1962, LeBrasseur and Parker 1964, Ishida 1966, Neave 1966). During this coastal phase the fry school and feed during daylight hours (LeBrasseur and Parker 1964, Ishida 1966, Neave 1966, Parker and Vanstone 1966, Healey 1967, Parker 1969).
Little is known of the diel pattern of swimming activity in pink salmon fry in nature. The schooling and feeding of wild fry during the day suggest that the fry have a corresponding diurnal rhythm of swimming activity. Therefore, I hypothesize that the fry shift their nocturnal activity rhythm, characteristic of downstream migrant fry, to a diurnal rhythm in their new marine habitat. 38
The objective of this portion of the study was to test this hypothesis by examining ontogenetic changes in the diel rhythms of swimming activity and of vertical distribution in a water column in newly emerged pink salmon fry under laboratory conditions. The results are related to ontogenetic changes in the photoresponse and habitat of fishes, and also to the behavioral ecology of pink salmon fry.
MATERIALS AND METHODS
A. Fish Fry used in the current study emerged from simulated redd No. 3 (see Table 1, Fig. 2) during the night of 15-16 February 1977. One hundred and thirty eight fry emerged that night. Except for six individuals, the fish were placed in an enclosed 484-L fiberglass holding tank as they emerged. This tank had running sea water of ambient temperature, and was exposed to a LD 12:12 (600:1 lx) cycle. Lights-on and lights-off occurred at 0800 and 2000 hours, respectively. Twilight periods were not used. The fish were fed as described in Chapter II (Section B). Water temperature did not vary more than ± 1°C from experimental temperatures (Appendix Table 8).
B. Experimental tank
The experimental tank was a 190-L glass aquarium (Fig.9) with a slate bottom, and lined on three sides with white polyethylene Fig. 9. Oblique view of the experimental tank showing the observation section, delimitated by a baffle and a screen at opposite ends. Dotted lines were drawn with a grease pencil on the front viewing glass and the back wall as aids in recording behavior. See text for more details. 40 plastic. This plastic served.as a background against which the small fry could be easily seen under dim (2 lx) night illumination.
The tank had a false, perforated, Plexiglas bottom. Sea water, filtered to 5 um, entered the tank through a manifold, located at
-1 one end of the tank, at a rate of about 1 L min . A perforated sheet of Plexiglas, acting as a baffle, reduced any water current at the surface. At the other end of the tank, a standpipe, isolated from the rest of the tank by a plastic screen, regulated the water level at 42 cm. The top of the tank was covered with clear nylon mesh screening to prevent fish from escaping. The observation compartment of the tank was divided vertically into three equal sections by lines drawn at intervals of 26 cm on the front viewing glass and on the back wall, and was also divided horizontally by a line into halves of 19 cm in height each.
The tank was exposed to a LD 12:12 (430:2 lx) cycle. Twilight periods were not used. Water temperature was recorded continuously at the inlet with a thermograph. Experimental temperatures, shown in Appendix Table 8, were regulated near 9.5°C. Differences between daily maximum and minimum temperatures never exceeded 2*C, and usually were less than 1°C. Overhead illumination was incandescent. Lights-on and lights-off occurred at 0800 and 2000 hours, respectively. Salinities ranged from 25.8 to 26.9%„, and dissolved oxygen in the water always exceeded 91% of air saturation.
C. Experimental procedure
Six different groups of sibling fry were used in this 41 study. Each group, comprising six fry, was observed independently and sequentially for a 5-day period from Day 1 to Day 37 after emergence (Appendix Table 8). Group 1 was introduced in the experimental tank at 0630 hours on the night of emergence (Feb. 16). Observation of this group began two hours later and continued for five days (Days 1-5, Appendix Table 8). Group 1 had not been fed previously. However, the other five fish groups were fed in the holding tank prior to experimentation. At the end of the first 5-day observation period, group 1 was replaced by group 2, which had been held previously in the holding tank. Group 2 was observed similarly for a 5-day period (Days 7-11). Therefore, each group of six fish was replaced sequentially by another group after observation, and was observed for a 5-day period during the first 37 days after emergence. Except for group 1, all fish groups were acclimated to the experimental tank for at least 24 hours before observation. The fish were not fed during the experimental period since preliminary observations showed that feeding enhanced swimming activity.
Swimming activity and vertical distribution of the fish were recorded with the aid of a Sanyo silicon diode video camera, equipped with a wide-angle lens (1: 1.8/10), located about 1 m from the front of the tank. Manual adjustments of the lens' aperture permitted filming of the fish in the light and dark of the LD cycle. A timer activated the camera and a video tape recorder for a 10-min period every alternate hour. All observations were stored on video tapes for later analysis. 42
Swimming activity was recorded as the number of vertical
lines on the viewing glass traversed per fish per 10 min. This
value was converted to mean swimming speed by multiplying the number
of lines traversed per fish per 10 min by 26 cm. Mean swimming speed
was expressed in meters travelled per fish per 10 min, and in fish
-1 body length per second (BL s ). Vertical distribution of fish in the 38-cm water column was recorded as the tendency of the fish to swim in the upper half of the water column (i.e., top 19 cm). This tendency was estimated by noting the instantaneous (ca. 2 s) number of fish in the upper half of the water column every minute on the minute for the 10-min observation period. Therefore, a maximum score of 60 (6 fish x 10 instantaneous scans) is obtained for a 10-min observation period if all six fish are swimming in the top 19 cm of water on each of the 10 instantaneous scans of the tank. For each 10-min period, the total number of fish observed instantaneously in the upper half of the water column was divided by the maximum possible score of 60. This ratio then was multiplied by 100 to yield the "instantaneous percentage of the fish in the upper half of the water column per 10 min", which henceforth will be referred to as the index of vertical distribution. This index was considered to be a measure of the phototactic response of the fish to overhead illumination. Diel rhythms of swimming speed and of vertical distribution were obtained for each fish group by plotting, at 2-h intervals, consecutive 10-min scores of these two behavioral processes. 43
D. Estimation of rhythm parameters
1. Daily diurnal-nocturnal (D/N) ratio
The daily D/N ratio was calculated as the mean of the 10-min scores recorded during the day divided by the mean of the 10-min scores recorded during the succeeding night of a 24-h LD cycle. This ratio provides a relative index of the degree of diurnalism or nocturnalism of the recorded diel behavioral rhythm.
2. Dally mean level
Daily mean level is the grand mean of all time-series 10-min scores recorded at 2-h intervals over one 24-h LD cycle.
3. Period
Period is defined as the duration of one recurring oscillation or cycle in a rhythmic function. In the current study, a modified Enright's (1965a, b) periodogram analysis, developed by
Dorrscheidt and Beck (1975), was used to calculate the period length of diel behavioral rhythms. This periodogram analysis is described in
Appendix 1. Time-series data records, used in the periodogram analysisi contained 60 sample points (i.e., five days of observation for each fish group). 44
RESULTS
A. General Behavior of fish
In general, the fish swam continuously day and night along the length of the aquarium. Individual fish Infrequently held a stationary position in the water column for short periods of time (< 5 min). Fish were never observed to rest on the false bottom of the aquarium.
On Day 1, the fish swam about the aquarium as individuals during day and night. A schooling tendency was not apparent. Swimming fish formed small temporary aggregations of two to three individuals on Day 2. These aggregations were non-polarized. Beginning on Day 3, fish formed polarized schools of four to six individuals during the day. From Day 3 to Day 37 (end of study), schooling fish generally swam in the lower half of the water column during the day. At night the school broke up, and the fish swam as individuals near the water surface.
Aggressive behavior among fish was not observed. During the day, fish occassionally "snapped" at what appeared to be small particles or air bubbles in the water column or at the water surface. Sometimes individual fish exhibited short bouts of "fluttering" behavior (Groot 1965); that is, up and down vertical swimming along the walls of ah aquarium. This occurred too infrequently to be recorded. 45
B. Swimming activity
Figure 1Q shows the daily rhythms, of swimming activity for groups of pink salmon fry at different times after emergence. On Day 1 activity peaked shortly after lights-on, and declined during the remainder of the day. Since the fish were introduced into the aquarium only two hours before observations began, this activity peak is attributed to the incomplete acclimation of the fish to the aquarium.
Daily swimming activity peaked shortly after lights-off on Days 2 and 3.
However, a second mode is evident near lights-on on Day 3. Only the early night peak was observed on Days 4 and 5. During Days 7 to 9 this peak shifted from early evening to just before the onset of dark.
During these three days a bimodal, diurnal activity pattern emerged with peaks at the beginning and end of the day. This pattern became more pronounced during Days 10 to 13. However, the diurnal activity rhythm commonly was unimodal, but sometimes bimodal, from Day 14 to
Day 37.
Omitting Day 1, the D/N ratio declined from a value of
1.25 on Day 2 to 0.24 on Day 4 (Fig. 11). The D/N ratio was continually above 1.0 after Day 9, but it varied with fish age.
This indicates that the daily rhythm of swimming activity was diurnal from Day 9 to Day 37, but the degree of diurnalism varied during this period. The period of Day 4 to Day 15 is characterized by a gradually increasing D/N ratio that reached a maximum of 3.06 on Day 15.
The ratio declined progressively after Day 15 to 1.35 on Day 28.
Thereafter, it gradually increased until the end of the study. 46
\ /A 20
] • • 21
3 22 :w \ " . • .. / 4 23 Lj \ "\A- 5 27 7
*\ 28 8 ]vv*\ i A\_._A V—' o 9 29 / K A 10 •o 30 7 K.A\ VA- •'"-•--.A 31 M K.-.AN 13 •\A.- 33
5 14 w 34 15 ;"^"A / .' \A 16 35 / •Vx/ 36 17 >-,A\ : A-'V- 37 19 • . •. ( \
01 11 16 10 11 04 01 II 16 10 34 O* 0*
Time of day (h) Fig. 10. Daily rhythms of mean swimming speed for groups of pink salmon fry from Day . 1 to Day 37 after emergence. Each of six groups was observed independently and sequentially for five consecutive days during this period. Light and dark horizontal bars indicate day (L) and night (D), respectively. I 1 1 1 1 1 1 1 1 1 1 1 1 I I 1 I I 1 I 1 0 4 8 12 16 • 20 24 28 32 36 40
Age (days post - emerge nee )
Fig. 11. Relation between daily D/N ratio for swimming speed of pink salmon fry and age after emergence. Arrows indicate the times when one group of six fish was replaced by another group of. six fish in the experimental tank. 48
Daily mean swimming speed, expressed as mean distance -1 travelled per 10 min (Jig. 12A) and as BL s (Tig. 12B), increased significantly with fish. age. Daily mean swimming speeds varied -1 -1 between 2.10 and 8.22 m fish 10 min or between 0.097 arid -1 0.311 BL s . Swimming speed declined gradually with time for each fish group, except group 2 (Days 7-11) and group 4 (Days 19-23).
Period lengths of the daily rhythms of swimming activity for each of the six fish groups were significantly different from random "noise" (Table 4). These period lengths also were within 5% of 24.0 h,. except for group 4. Figure 13 illustrates the periodograms from which these periods were obtained. Period length did not change significantly with fish age. These period lengths (Table 4, Fig. 13) indicate that the diel rhythms of swimming activity were synchronized with the 24-h LD cycle during the first 37 days after emergence. The period of 22.5 h for group 4 suggests that the degree of synchrony of the diel activity rhythm for this group with the LD cycle was weaker compared to that for the other five groups. The significant periodic component near 12 h, a submultiple of 24 h, in the periodograms for the activity rhythms during the first 11 days after emergence results from the general bimodality of the diel rhythms during this period (Fig. 10).
Cycles of water temperature inadvertently occurred during each experiment (Table 4) despite attempts at keeping temperature constant. Water temperature affected swimming speed, as evidenced by 49
Fig. 12. Relation between daily mean swimming speed of pink salmon fry and age after emergence. Meaning of the arrows as in Fig. 11. Regression line A is Y = 2.48 + 0.12 X (rs = 0.82, P < 0.001), and regression line B is Y = 0.13 + 0.004 X (rs = 0.79, P < 0.001). 50
Table 4. Period lengths of the diel rhythms of swimming activity and of vertical distribution for six groups of fry tested at different times after emergence. Period length of the cycle of water temperature for each experiment is also given. All period lengths are significantly different (P < 0.05) from random "noise", except where indicated specifically by an asterisk-.
Period length (h)
Group Days after Swimming Vertical Temperature
No. emergence activity distribution cycle
1 1-5 23.8 23.7 30.7 2 7-11 23.7 23.7 24.7 3 13-17 23.3 23.7 22.9 4 19-23 22.5 23.7 21.8 5 27-31 23.7 23.6 25.1 6 33-37 23.7 23.7 23.4 51
Instantaneous Locomotor activity vertical distribution Days
Period (h) Fig. 13. Periodograms of rhythms of locomotor activity and of vertical distribution recorded simultaneously in LD 12:12 for each of six groups of fry from Day 1 to Day 37 after emergence. Each periodogram is based on five days of observation on each fish group. P is the major periodic component (in hours) in the time-series data record. All P values are significantly different (P < 0.05) from random "noise". 52 the positive correlation between water temperature and swimming speed scores for three of the six fish groups (Table 5). However, the period length of the temperature cycle (Table 4), recorded during each 5-day experiment, did not correlate significantly (r = 0.63) with the corresponding period length for the swimming activity rhythm of each fish group (Table 4). This indicates that temperature cycles did not synchronize the diel rhythms of swimming activity. The periods of four of the six temperature cycles were significantly different from random "noise", but only two were between 23 and 25 h.
In summary, diel rhythms of swimming activity of pink salmon fry were nocturnal during the first week after emergence. A shift from a nocturnal to a diurnal activity rhythm gradually occurred during the second week (Days 7-13). During this period the rhythms generally were bimodal, with a peak of activity early and late in the photophase (day). Thereafter, swimming activity rhythms were diurnal and usually unimodal. The synchrony between the morning peak of activity and the lights-on stimulus on the majority of days after the first week suggests that the abrupt dark to light (D-L) transition synchronized the diel rhythm of swimming activity. Periodogram analysis indicated that swimming activity rhythms were synchronized strongly with the 24-h LD cycle, but not with the daily cycles of water temperature. Daily mean swimming speed of fish increased with age and size. Table 5. Spearman rank correlation coefficients (rs) for comparisons between values of 1) water temperature and swimming speed, 2) water temperature and the index of vertical distribution, and 3) swimming speed and the index of vertical distribution. Swimming speed, vertical distribution, and water temperature were recorded simultaneously every two hours for each of six groups of fish. The fish were observed at different times after emergence.
Temperature and Temperature and Swimming speed Group Days after swimming speed vertical distribution vertical distribution No. emergence
1 1-5 0.19 NS 0.03 NS 0.18 NS 2 7-11 0.21 NS -0.03 NS -0.06 NS 3 13-17 0.40 *** -0.04 NS -0.40 ** 4 19-23 0.37 ** -0.12 NS -0.32 * 5 27-31 -0.04 NS -0.17 NS -0.28 * 6 33-37 0.36 -0.17 NS -0.65
NS, Not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001 54
C. Vertical distribution
Daily rhythms of vertical distribution for groups of pink salmon fry at different times after emergence are shown in Figure 14. During the first week after emergence the majority of the fry were observed swimming in the upper half of the water column during the night, and almost exclusively in the lower half of the water column during the day. Therefore, the fry tended to swim nearer the water surface during the night than during the day during this period. This diel pattern is referred to as a nocturnal rhythm of vertical distribution. However, the fry showed a gradually increasing tendency to swim in the upper half of the water column during the day on Days 7 to 13. This tendency was greater during the later portion of the photophase on Days 7 and 8, occurred progressively earlier on subsequent days, and reached a state on Day 13 when about half of the fry were observed instantaneously to swim in the upper half of the water column. Thereafter, the diel rhythm of vertical distribution generally remained nocturnal and unimodal, and the index values of vertical distribution during the night did not vary significantly from day to day.
The daily D/N ratio was always less than 1.0, indicating the predominance of a nocturnal rhythm of vertical distribution throughout the experimental period (Fig. 15). However, this ratio varied significantly with fish age. The lowest ratio values occurred on Days 2 to 5, reflecting the strongest degree of nocturnalism in the diel rhythm of vertical distribution. The daily D/N ratio increased 55
1 19 N—\ .-—V /\-/\/ 2 20 i X /\/ \ c ] 3 21 e o ; ] 4 r\ 22 23 ] 5 /v v\./ ] 7 27 / ' \ :
1 8 28 sA/\/ /: \ : a. n. -• 3 29 A, ./V 30
100 10 31
13 33 \AA/ \ ] 14 34 /
] 15 35
/\f"
] 14 \_ \/VA/ ] 17
•PHP 01 13 to 10 14 0* Ot oa II 16 10 14 04 Time of day (h) Fig. 14. Daily rhythms of vertical distribution in a water column for groups of pink salmon fry from Dayl to Day 37 after emergence. Each of six groups was observed independently and sequentially for five consecutive days during this period. Meaning of light and dark horizontal bars as in Fig. 10. 56
Fig. 15. Relation between daily D/N ratio for the index of vertical distribution of pink salmon fry and age after emergence. Meaning of the arrows as in Fig. 11. 57
rapidly between Day 7 and Day 13. This period corresponds to the
phase during which the fry showed a gradually Increasing tendency to
swim in the upper half of the water column during the day. The ratio
declined slightly after Day 15, but showed a weak increasing trend
from Day 27 to the end of the experimental period. Daily changes in
this ratio were relatively small from Days 13 to 37 compared to the
first two weeks of observation. The relative stability of the daily
D/N ratio during Days 13 to 37 reflects the low day-to-day variability
in the nocturnal rhythm of vertical distribution during this period.
Nocturnal rhythms of vertical distribution for each fish group had significant period lengths approximating 24.0 h (Table 4, Fig. 13). These period lengths indicate that these nocturnal rhythms were synchronized with the 24-h LD cycle during the first 37 days after emergence. Short-term fluctuations in water temperature did not affect the vertical distribution of the fish (Table 5). Further, temperature cycles did not synchronize the diel rhythms of vertical distribution. This is indicated by the weak correlation (r = -0.05) between the period of the temperature cycle in each experiment and the corresponding period of the rhythm of vertical distribution (Table 4).
In summary, the fish swam mainly in the upper half of the water column during the night and in the lower half of the water column during the day of the LD cycle from Days 1 to 7. This observation suggests that the fry exhibited a negative phototactic response 58 during this period. However, this negative phototactic response weakened between Days 7 to 13, as evidenced by the increasing proportion of the fry swimming in the upper half of the water column during light of these days. Thereafter until the end of the study, the fry showed a stable negative phototactic response, as their diel rhythms of vertical distribution remained nocturnal and relatively stable in form (Fig. 14). Periodogram analysis (Table 4, Fig. 13) indicated that the daily rhythms of vertical distribution were synchronized strongly with the 24-h LD cycle, and that daily cycles of water temperature had no apparent synchronizing effect on these rhythms. The abrupt L-D transition appeared to time the rising of the fish toward the water surface at night.
D. Relationship between swimming activity and vertical distribution
Pink salmon fry showed nocturnal rhythms of swimming activity and of vertical distribution after emergence. These rhythms indicate that the fry had a relatively negative response to high light intensities. The ontogenetic shift from a nocturnal to a diurnal rhythm of swimming activity occurred primarily 7 to 13 days after emergence. Coincident with this shift was an increasing tendency of the fry to swim in the upper half of the water column during the day. This suggests a gradual weakening of the fry's negative phototactic response during Days 7 to 13. Thereafter, diel rhythms of swimming activity were diurnal, whereas rhythms of vertical distribution remained nocturnal. This difference in phase reflects the significant negative correlations between the 10-min scores of swimming speed and the 59 10-min scores of vertical distribution on Days ,13 to 37 (Table 5).
Diel rhythms of swimming activity and of vertical distribution were strongly synchronized with the 24-h LD cycle. Corresponding periods for the rhythms of swimming activity and of vertical distribution for the same group of fish frequently were similar, but did not correlate with one another (r = -0.06). This suggests that the degree of synchrony of these two behavioral processes with the LD cycle differed in some instances (Table 4, Fig. 13). Diurnal swimming activity rhythms appeared to be synchronized with the lights-on stimulus of the LD cycle, whereas nocturnal patterns of rising toward the water surface appeared to be synchronized with the daily lights-off stimulus.
DISCUSSION
The data showed that pink salmon fry exhibited mainly nocturnal rhythms of swimming activity and of vertical distribution during the first week after emergence. However, a shift from a nocturnal to a diurnal rhythm of swimming activity occurred 7 to 13 days after emergence. Coincident with this ontogenetic shift was an increasing tendency of the fry to swim in the upper half of the water column during the day. Thereafter until the end of the study (Day 37), the fry displayed diurnal rhythms of swimming activity and nocturnal rhythms of vertical distribution. These observations support the hypothesis that pink salmon fry shift their nocturnal 60 activity rhythm to a diurnal rhythm shortly after emergence and after entering sea water.
Field and laboratory observations (see Chapter III) indicate that emergence and seaward migration of pink salmon fry are mainly nocturnal. However, the observations of Mason (1976) and mine (Table 3, Fig. 6), as well as some field observations (McCart 1967, Bakshtanskii 1970), indicate an increasing tendency of the fry of pink, coho, and sockeye salmon to emerge and migrate downstream during light as the emergence period progresses. This tendency may relate to a gradual weakening of the fry's negative phototactic response with age and (or) size, as demonstrated in the data of Mason (1976) for coho salmon. Noakes (1978) reviewed the behavioral development of salmonid fishes during the period from hatching to a few days after emergence. In general, there is a progressive development of motor patterns leading to emergence from gravel. Concurrently, the alevins show a progressive development of visual capability such that, at the time of emergence, fry are positively phototactic (change from negative phototaxis), show marked photomechanical changes in the retina, and possess full visual acuity necessary for feeding and other behaviors.
My data indicate that the progressive weakening of the negative phototactic response of pink salmon fry continues after emergence. This is evidenced by the gradual shift from a nocturnal to a diurnal rhythm of swimming activity, and by the increasing 61
tendency of the fry to swim closer to the water surface by day
following the first week after emergence. The.se laboratory findings
confirm those of Hoar and co-workers (Hoar 1958, 1976, Hoar et al. 1957) on pink salmon fry. Collectively, these observations indicate that pink salmon fry are nocturnally active, exhibit negative phototaxis, and rise toward the water surface at night during the period of downstream migration. While residing in estuaries and adjacent coastal waters, the fry school by day, gradually become day-active, and their negative phototactic response weakens somewhat. However, the fry retain their negative phototaxis, and their tendency to rise toward the water surface at night and to descend in the water column by day.
No field data are available to confirm these laboratory findings, except for the daytime schooling of pink salmon fry in marine waters
(LeBrasseur and Parker 1964, Ishida 1966, Neave 1966, Healey 1967).
Few studies have been conducted on the ontogeny of diel behavioral rhythms in vertebrates (Rensing 1965). However, recent ontogenetic studies of the locomotor activity of some fish species (Appendix Table 9) provide additional support for the hypothesis that the photobehavior and diel rhythms of locomotor activity of fishes are adapted to their environment. These studies indicate that ontogenetic changes in the phase of diel rhythms of swimming activity frequently are associated with changes in habitat. It may be argued, perhaps tautologically, that a shift in the phase of the diel activity rhythm of a fish is part of its adaptive response to changes in its habitat. Similarly, Hokanson (1977,) concluded" that fish adapt to 62 different temperature features of their environment by changing
their diel activity cycles, and by undergoing seasonal migrations.
Juvenile pink salmon are diurnal feeders in nature (see Chapter V). This fact correlates well with, the diurnal rhythm of swimming activity exhibited by fry in the current study. Since juvenile pink salmon are schooling, visual feeders, a diurnal rhythm of swimming activity is considered adaptive for feeding on small planktonic invertebrates in nearshore habitats of estuaries and adjacent coastal waters. A diurnal activity rhythm could increase the fry's prey-encounter rate during daylight hours available for feeding, and could reduce the fry's energy demands at night when little or no feeding occurs.
Daily mean.swimming speed of pink salmon fry increased gradually with age in this study. A similar observation has been reported for the larvae of several fish species (Rosenthal and Hempel 1970, Blaxter and Staines 1971, Hunter 1972, Wyatt 1972). Increases in the swimming speed of fish larvae with age in these studies were shown to increase proportionally their volume of water searched per unit time or their prey-encounter rate. This relationship between swimming speed and age of fish larvae can be explained largely by the proportional increase in swimming speed with increasing body length observed in fishes (Blaxter 1969b). However, this explanation is inadequate-to account for the observed increase in swimming speed ^1 -1 (m fish 10 min ) with age in pink salmon fry because mean swimming 63
-1 speed (BL s ), when compensated for body length., still increased with age. The rate of prey encounter potentially increases with increasing swimming speed in fishes (Rosenthal and Hempel 1970, Blaxter and Staines 1971, Hunter 1972, Ware 1975, 1978). Therefore, the increasing swimming speed of pink salmon fry with age in the laboratory may be the expression of an adaptive response to consuming increasingly larger rations in nature to meet corresponding increases in absolute energy demands as body size increases (Brett 1965) . The slight decline
in daily mean swimming speed during the course of the 5-day experiment, observed for four of the six fish groups in the current study, may have resulted from food deprivation during the experiment.
Periodogram analysis showed that the diel rhythms of swimming activity and of vertical distribution in pink salmon fry were synchronized closely with the 24-h LD cycle during the first month after emergence. The daily D-L and L-D transitions appeared to have synchronized the diel rhythms of swimming activity and of vertical distribution, respectively. Daily cycles of temperature did not exert significant synchronizing effects on either of these two types of behavioral rhythms. The daily cycle of illumination, the most predictable and consistent environmental cycle, is generally recognized as the most important synchronizer of biological rhythms for a wide spectrum of organisms in nature (Aschoff 1960, 1963, Menaker 1969, Biinning 1973, Enright 1975, Gwinner 1975). 64
CHAPTER V. TEMPORAL PATTERNS OF FEEDING BEHAVLOR
INTRODUCTION
Data presented in Chapter IV showed that pink salmon fry shift their nocturnal rhythm of swimming activity, characteristic of seaward migrants, to a diurnal rhythm shortly after emergence. This phase shift of the activity rhythm coincides with fry migration from a freshwater (river) habitat to an estuarine habitat. I argued in that chapter that a diurnal rhythm of swimming activity is adaptive to schooling and feeding during the day. This suggestion is supported by the following observations. Firstly, pink salmon fry school by day in coastal marine areas (LeBrasseur and Parker 1964, Ishida 1966, Neave 1966, Healey 1967). The fry also feed mainly during the day (Parker and Vanstone 1966, Parker 1969). Secondly, laboratory experiments have shown that pink salmon fry are visual feeders, requiring illumination of at least natural twilight intensity to feed (Brett and Ali 1958, Ali 1959, Brett and Groot 1963, Bailey et al. 1975).
The interpretation of feeding rhythms observed in nature depends on an understanding of factors that affect feeding behavior. The feeding rate of fishes, and therefore their diel feeding patterns, are determined by an interaction between external (environmental) and internal (physiological) factors (Ivlev 1961, Blaxter 1970, Curio 1976, Brett 1979, Hyatt 1979). One of these internal factors is 65
"hunger"; it may be defined operationally as the reciprocal of satiety, determined by the previous feeding schedule (or gastrointestinal stimuli) and metabolic (systemic) condition of the animal (de Ruiter 1967, Colgan 1973). Hunger has been shown to affect feeding rate and the amount of food ingested in a variety of animals (de Ruiter 1963, 1967, Ware 1972, Colgan 1973, Curio 1976, Toates and Archer 1978, Brett 1979).
The first objective of this portion of the study was to examine further the daily patterns of feeding activity and the diet of pink salmon fry in nature. This was done by collecting fish at 2-h intervals during a diel cycle in each of two marine bays. The second objective was to examine the effects of hunger on the daily pattern of feeding activity in juvenile pink salmon under laboratory conditions. Specifically, I tested two simple predictions common to many control theory models of hunger (Rozin 1964, de Ruiter 1967, Colgan 1973, Curio 1976, Toates and Archer 1978). These predictions are 1) feeding rate declines proportionally with increasing satiation during a feeding bout, and 2) the amount of food ingested increases with increasing period of food deprivation (hunger) in a compensatory manner up to a maximum before the deprived animal weakens. These tests were conducted on juvenile pink salmon, which were presented with live copepod prey at high densities, in a laboratory tank under constant temperature and constant light intensity. The latter two abiotic factors were kept constant because they affect feeding behavior in fishes (Brett and Ali 1958, Brett and Groot 1963, Blaxter 1970, Brett 1979). 66
MATERIALS AND METHODS
A. Pattern of feeding behavior in the field
Pink salmon fry were collected with a 15-m beach seine and long handle dipnet in shallow water (< 2 m depth) at 2-h intervals. Collections were made between 0715 and 2300 hours on 3 May 1976, and from 1300 hours on 20 May to 1100 hours on 21 May 1976 in Departure Bay and in Hammond Bay, Vancouver Island, British Columbia, respectively. The origin of the fish is unknown, but believed to be the Fraser River. One littoral collecting station was chosen in each of the two bays, approximately 4.0 km (shoreline distance) apart (Fig. 16). The slope of the beach at station 2 (Hammond Bay) was steeper than the one at station 1 (Departure Bay). Both stations had a cobble substrate in the upper intertidal zone and a sand-mud substrate with sea grasses and rockweeds in the lower intertidal zone. Station 2 had more large rocks than station 1. The sky was partially overcast during both collecting series. Surface water temperatures varied from about 11.0 to 13.0°C. Prey abundance was not assessed at either station.
Captured fish were lightly anesthetized with MS-222 to prevent regurgitation or defecation. Samples were preserved in 5% formalin for later analysis; Stomach contents (from the esophageal opening to the pylorus) were examined using a binocular microscope. Prey biomass and identification of prey from the fishes stomach contents were determined separately using different fish from the same sample. Prey organisms from a sample of up to 10 fish were identified usually to order. All prey from
each fish, as well as the fish with its entire alimentary tract, were 67
123° 58' 1 23° 56' W —i 1 1 1 1 r
Strait of Georgia
81
De parture Bay
Fig. 16. Map of Departure Bay and Hammond Bay, Vancouver Island, British Columbia and vicinity. Fish were collected at one littoral station in each bay. Stations are indicated by 1 and 2 in Departure Bay and in Hammond Bay, respectively. PBS denotes the Pacific Biological Station and Ck. the creeks. 68 oven-dried at 60°C for 24 hours and then weighed separately to the nearest 0.01 mg. A positive relationship existed between fish weight and prey biomass consumed (see Fig. 19). Hence, relative dry weight of prey biomass from each fish was calculated as (prey weight / fish weight) x 100. Mean relative prey biomass Cdry weight) consumed by fish at different times of the diel cycle was determined from a sample of up to 20 fish.
The daily amount of prey consumed (dry weight) by fish was estimated by summing the amounts consumed (C) during several specific time intervals within that 24-h period (Thorpe 1977). Prey consumption (C) in the interval between fish captures is calculated as
C = S2 - S-i + A where si and S2 are the mean relative dry prey biomasses in the fish's
stomach at the beginning (ti) and at the end (t2) of the interval, respectively, and A is the relative amount of prey evacuated from the stomach during that interval. If the rate of prey consumption between times of capture is linear, then A will equal the amount of prey evacuated from the stomach containing the mean dry prey biomass 0.5 (sj + S2) at time t^ over the interval (t2 - t^). At the end of
that interval a relative prey biomass sr would remain in the stomach. Hence,
A = 0.5 Csi + s2) - sr 69
Gastric evacuation, rate in salmonids is an exponential decay function of time, and thus is. proportional to the quantity of food present in the stomach (Fange and Grove 1979). Therefore, the relative amount of food remaining in the stomach (s^) after digestion has proceeded during the interval (t£ ~ t].) is
_k s^ = n0. 5r . (sif - S2) \ e (tz2 ~ t'l1 ) '
The instantaneous evacuation rate coefficient (k) used in the current study was -0.152. This value is based on the rate of evacuation of natural prey from the stomach of wild pink salmon fry (ca. 50 mm). These fry were captured in Departure Bay in the early morning, and then held in a laboratory tank without food at 10°C. There the fry were sampled serially over a period of 24 hours, and the amount of food remaining in their stomach was determined (M.C. Healey, unpublished data).
B. Pattern of feeding behavior in the laboratory 1. Experimental apparatus The experimental apparatus used consisted of a circular (83-cm diam) Plexiglas tank containing about 15 cm of sea water
(volume 81.2 L) (Fig. 17). Four water inlets equally spaced along the periphery of the tank about 1 cm from its bottom supplied it with -1 water at 1 L min . Water flowed out of the tank through small holes 15 cm up its sides. Water temperature was maintained at 10-11°C. The tank was exposed to 12.5 hours of reflected incandescent lighting 70
Came ra
Tygon tubing
Outlet
Side view is cr Inlet
83 cr
Top view Inlet
Fig. 17. Top and side views of the 83-cm diameter circular Plexiglas tank used in laboratory feeding experiments. Locations of water inlets and outlets are indicated by arrows. The bottom of the tank is divided into 10-cm2 grids. Prey were introduced into the center of the tank via a Tygon tube, and fish were filmed with a video camera located directly above the tank. 71
(550 lx) alternating with 11.5 hours of darkness (<;0.01 lx). Llghts-on and ligtits-off occurred at 0730 and 2000 hours, respectively. A Sanyo silicon diode video camera, equipped with a wide-angle lens (1: 1.8/10), provided a top view from its location directly above the tank (Fig. 17). Both sides and bottom of the tank were white, and 2 the bottom was divided into 10 cm grids with black lines (Fig. 17). The entire apparatus was enclosed by polyethylene plastic lined with white cardboard.
2. Feeding experiments
Two feeding experiments were conducted to test the effects of hunger on the daily pattern of feeding rates. The first used a group of five fish (mean fork length, 8.3 cm; mean dry weight, 946 mg) deprived of food for 24 hours prior to testing (group A); the second was conducted with a group of five fish (8.0 cm, 782 mg) deprived for 72 hours (group B). Hence, group B was considered to have a higher level of hunger compared to group A at the time the experiments were initiated.
Experimental prey consisted of three species of marine copepod (Appendix Table 10) that occur in the diet of wild juvenile pink salmon. Calanus plumchrus was the largest and most abundant species in samples. The copepods were obtained from vertical Scor net hauls in the Strait of Georgia. They were transported in large plastic containers to the laboratory, and held in darkness in a tray (2.3 x 0.17 x 0.25 m) with running sea water (10-12°C). Only live, healthy prey (< 5 days post-capture) were presented to the fish. A mean dry weight of 0.471 mg 72
was used to calculate prey biomass (dry weight) consumed by the fish
(see below) .
Each fish was given about 100 prey daily for three consecutive days prior to food deprivation while in holding tanks. Fish were acclimated to the experimental tank for 24 hours prior to testing.
Since wild pink salmon feed little at night (see Fig. 18 and Appendix
Fig. 4), the laboratory feeding sessions were restricted to the experimental light period. During both experiments a group of about
1,000 prey (in 50 mL of sea water) was introduced into the center of the tank via a tube (Fig. 17) at 0800 hours (hour 0) on the day of testing. The behavior of the fish was filmed for 10 min following the onset of feeding. On each of the following 11 hours, 500 additional prey were added to the tank in a similar manner on the hour followed by 10 min of filming. Resulting feeding behavior and swimming activity were recorded on video tape. Prey density was always high, exceeding that generally found in natural waters sampled by conventional methods
(J.C. Mason, personal communication). Hence, the feeding rate of the fish was not limited by prey availability nor by incident light intensity.
The video tapes were analyzed later. The frequency of
"snapping" at prey (discerned by a short, accelerated movement of the fish toward a prey, terminated by a rapid closing motion of the mouth at the prey) and distance travelled (number of 10-cm grid squares traversed) were quantified for each individual fish for every 10-min period during the 12-h feeding session. For each 10-min observation 73
period, a mean snapping frequency and a mean distance travelled per fish were calculated by summing these two criteria for the five fish.
Individual fish were identified by stopping a film frame at the onset of each 10-min observation period, and plotting with a marking pen the position of each of the five fish on a piece of acetate sheet fixed on the 32-cm screen of the video monitor. Through multiple playbacks of the video tape each fish was observed separately. This procedure was repeated for every observation period. Hence, distance travelled and concurrent snapping frequency for each fish were quantified for each 10-min period.
3. Prey-capture success and prey biomass consumed
To estimate the nmmber of prey consumed by the fish per unit time, an estimate of prey-capture success was required. This was obtained by placing each of 10 pink salmon (mean fork length, 8.2 cm; mean wet weight, 6.0 g) into separate 190-L aquaria. These aquaria had white sides and similar light intensities and water temperature as in the above circular experimental tank. The fish previously had been fed live calanoid copepods. After 24 hours of food deprivation and at least 7 hours of acclimation to the aquarium, each fish was presented with about 200 of the live calanoid copepods. The number of snaps subsequently directed at prey by the fish was recorded.
After about 50 snaps had been recorded, the fish was removed from
the aquarium, killed, and the copepods in its stomach counted.
Percentage capture success for each fish was calculated as (number of prey in stomach / number of observed snaps) x 100. Mean percentage of 74
prey-capture success was 92.4% (Appendix Table 11). Such a high
capture success is not uncommon for fishes (Olla et al. 1970, Blaxter
and Staines 1971, Nyberg 1971, Hunter 1972, Ware 1972, Confer and
Blades 1975, Godin 1978).
Prey biomass consumed during each hour of the 12-h session was estimated by multiplying by six the mean of the observed number of snaps per fish for the first 10 min of the hour to obtain an estimated number of snaps per fish for that hour. This value then was multiplied by 0.924 (mean capture success) and by 0.471 mg (mean dry prey weight) to obtain an estimate of prey biomass (dry weight) consumed per fish per hour. Such estimates were calculated in this manner for each hour between 0900 and 2000 hours, but not for the hour 0800-0900 for the following reason. Feeding rates declined non-linearly with time (see Fig. 21). Accordingly, for the latter hour the natural logarithms of the mean of the observed number of snaps per fish per 10 min at 0800 hours and at 0900 hours were averaged. The resulting geometric mean number of snaps per fish per min then was multiplied by 50 min to obtain an estimated number of snaps per fish between 0810-0900 hours. This latter value was added to the observed mean number of snaps for the 10 min at 0800 hours. This total then was multiplied by 0.924 and 0.471 mg to obtain an estimated prey biomass (dry weight) consumed per fish between 0800-0900 hours. Daily prey biomass consumed per fish was estimated by summing hourly estimates of prey biomass consumed per fish over the 12-h feeding session. Feeding responses of the fish were expressed as a satiation 75 curve with, cumulative number of snaps (biomass) plotted against time from the onset of feeding.
4. Gastric evacuation rate
The evacuation rate of calanoid copepods from the stomach of juvenile pink salmon also was estimated for comparison with the rates of fish predation. Eighty fish (mean total length, 10.3 cm; mean dry weight, 1.68 mg) were acclimated to a circular fiberglass tank (484 L) for several weeks. They were fed Oregon Moist Pellets (OMP) during this period and only live copepods during the three days before the experiment. The fish were exposed to a LD 12:12 (600:0.1 lx) cycle and ambient seawater temperatures during the acclimation period. The fish then were starved for 48 hours to ensure empty stomachs and high feeding motivation. At 0830 hours on the day of the experiment water flow was interrupted temporarily, and a group of about 16,000 calanoid copepods was introduced into the tank. The fish were allowed to feed spontaneously on the prey for 20 min, when their snapping rates declined to less than 1.0 per min (satiation). Later analysis revealed that their stomachs were full and distended at this time. At the end of this 20-min feeding period, 10 fish were removed, anesthetized with MS-222, and preserved in 10% formalin. The remaining 70 fish were transferred to a similar tank containing no prey. Samples of 10 fish were removed from this tank and preserved at irregular intervals during the following 48 hours as digestion proceeded. Mean (± SD) water temperature during this period was 11.4 (± 0.37)°C. Prey were removed later from the stomach of individual fish, 76 oven-dried at 60°C for 24 hours, and weighed. Relative prey biomass (dry weight) in the stomach of individual fish was expressed as a percentage of dry fish body weight.
RESULTS
A. Pattern of feeding behavior in the field
Significant diel variations (Fig. 18) were observed in relative dry prey biomass in the stomachs of Departure Bay fish (H =69.2, Kruskal-Wallis one-way ANOVA) and of Hammond Bay fish (H = 22.3).. Mean relative prey biomass in the stomachs of Departure Bay fish increased rapidly during the morning (0700-1200 hours) and early evening (1700-2100 hours), but declined slightly, although not significantly, during the afternoon (1300-1600 hours). Mean relative prey biomass declined slightly during darkness. The diel pattern of prey consumption differed for Hammond Bay fish, where mean relative prey biomass fluctuated throughout daylight hours. Mean biomass declined gradually as the night progressed. Maximum mean biomasses of 6.23 and 3.23% dry body weight were observed for Departure Bay and Hammond Bay fish, respectively. Both these daily maxima occurred near or at dusk. Of 564 fish examined from both bays, only one had an empty stomach.
Despite their smaller size, Departure Bay fish consumed a significantly greater absolute and relative dry prey biomass (t-test) 77
Departure Bay May 3. 1976
D 5.0-4 o o 4.0 J 1 i ; • /'
2.0 J CO i >. o JO 1 1 >» —i 1 r 0400 0800 1600 2000 •o
4> Hammond Bay May 20-21. 1976 i •~o o o >. T3 i\T/
—i 1— 0800 1200
Time of day (PDST)
Fig. 18. Variation in mean (± 95% confidence limits) relative dry prey biomass in the stomachs of pink salmon fry collected in Departure Bay and Hammond Bay, British Columbia with time of day and state of the tidal cycle. Means with dissimilar symbols above them are significantly different from each other (P < 0.05, Mann-Whitney U-test). All samples consisted of 20 fish, except where indicated near the mean. Open, solid and hatched horizontal bars represent hours of daylight, darkness and twilight, respectively. Times of high (+) and low (+) tides are also indicated. 78
than Hammond Bay fish (labia 6, Fig, 19). Based on mean relative prey biomasses in the stomachs of the fish collected during the two diel series, the daily amount of dry prey consumed by Departure Bay and
Hammond Bay fish was estimated to be 10.46 and 5.57% of their dry body weight, respectively (Table 6).
The diets of the fish collected in the two bays are shown in
Table 7. Fish collected in Departure Bay fed primarily and almost equally on harpacticoid copepods and on copepod nauplii, followed by barnacle (Cirripedia) larvae. These three prey categories comprised
93.1% of the fish diet. Hammond Bay fish fed mostly on harpacticoid copepods, barnacle larvae, and copepod nauplii in that order. These three prey categories comprised 86.2% of the diet in Hammond Bay.
Hammond Bay fish consumed relatively more harpacticoid copepods, barnacle larvae, amphipods, and bryozoa (cyphonautes) larvae but fewer copepod nauplii than Departure Bay fish. About 38% of the diet of
Departure Bay fish and 51% of the diet of Hammond Bay fish were classified as epibenthic in origin. The mean (± SD) number of prey found in the stomachs of Departure Bay and Hammond Bay fish are
335.6 (± 362.5) and 158.7 (± 177.1), respectively.
Tidal variations occurred in the relative abundance of some prey categories in the stomachs of fish captured in both bays (Fig. 20).
The relative abundance of barnaele larvae in the fish stomachs was greater near the times of high tide than of low tide in both bays.
Conversely, the relative abundance of harpacticoid copepods consumed Table 6. Absolute and relative prey biomasses (dry weight) and estimated daily food ration in the stomachs of juvenile pink salmon collected during diel sampling series in Departure Bay and Hammond Bay, British Columbia. All values, except daily rations, are means (± SD).
Fish Dry prey biomasses
1 Total length Dry weight . in fish stomachs Daily dry food ration Location N (mm) (mg) (mg) (% dry body weight) (% dry body weight)
Departure Bay 170 42.1 77.3 3.07 3.97 10.46 (2.1) (15.8) (1.31) (1.50)
Hammond Bay 229 44.0 98.4 1.75 1.77 5.57 (3.6) (30.4) (1.59) (1.35)
1 Calculated using Thorpe's (1977) method and a stomach evacuation rate coefficient (k) equal to -0.152 40 80 120 160 200 240 280 Dry fish weight (mg)
Fig. 19. Relation between dry prey biomass in the stomach of wild, individual pink salmon and dry body weight of the fish.. Regression line A is Y = 0.18 +0.04 X (r = 0.45, P < 0.001) and line B is Y = -0.15 + 0.02 X (r = 0.39, P < 0.001). 81
Table 7. Diets of juvenile pink salmon in Departure Bay and Hammond Bay, British Columbia on 3 May and 20-21 May 1977, respectively.
Departure Bay (N = 78) Hammond Bay (N - 87)
Occurrence Prey Occurrence Prey
Prey category (%) (No.) (%) (%) (No.) (%)
Copepoda
Calanoida 100.0 1459 5.57 70.1 534 3.87
Cyclopoida 6.4 5 0.02 8.1 7 0.05
Harpacticoida 92.3 9896 37.81 89.7 6495 47.03
Mou;5t.rtlloida 1.3 1 < 0.01 4.6 6 0.04
Nauplii larvae a 51.3 10050 38.39 37.9 1524 11.03
Amphipoda 20.5 32 0.12 56.3 421 3.05
Cirripedia - larvae 76.9 4430 16.92 78.2 3885 28.13
Cladocera 9.0 8 0.03 0.0 0 0.00
Cumacea 10.3 8 0.03 17.2 27 0.20
Decapoda - larvae ^ 23.1 29 0.11 18.4 38 0.28
Isopoda 5.2 4 0.02 .7.8.7 36 0.26
Ostracoda 5.1 6 0.02 2.3 2 0.01
Insecta
Diptera 18.0 19 0.07 26.4 40 0.29
Other 2.6 2 0.01 1.2 1 0.01
Mollusca
Gastropoda - larvae 44.9 215 0.82 10.3 13 0.09
Bivalvia - larvae 0.0 0 0.00 5.8 12 0.09
Polychaeta 0.0 0 0.00 2.3 2 0.01
Bryozoa - larvae 1.3 1 < 0.01 19.5 723 5.23
Teleostel - larvae 3.9 5 0.02 0.0 0 O.CO
Invertebrate eggs 2.6 2 0.01 20.7 30 0.22
Unidentified c 5.1 4 0.02 29.9 15 0.1.1
Total 26176 100.00 13811 100.00
^ All copepod nauplii pooled under the same category Includes zona and megalops c Number of unidentified prey include only Intact prey, whereas percentage occurrence of unidentified prey includes fish stomachs contcinint; both i;«.tao.C and fragmented, digested prey 82
• • Copepod nauplii » * Barnacle larvae • • Harpacticoid o o Calanoid . . Others
Departure Bay
May 3. 1976
Hammond Bay
May 20- 21 . 1976
l , , , ,— | i | i , 1 1 1200 1600 2000 2400 0400 0S00 1200 Time of day (PDST) Fig. 20. Variations in the relative abundance of prey found in juvenile pink salmon stomachs in Departure Bay and Hammond Bay, British Columbia with time of day and state of the tidal cycle. Prey (categories) shown were the most common ones consumed by fish. Other prey did not.show consistent tidal or daily variations. Nf and Np denote the number of fish stomachs examined for each time of day and the total number of prey found in fish stomachs, respectively. Relative abundance of a prey category in fish stomachs at each time of day is given as a percentage of the total number of prey (Np) found in the stomach of all fish (Nf) at that time. Horizontal bars and arrows are as in Fig. 18. 83 was greater near low tides in Departure Bay and, particularly, in
Hammond Bay. A similar pattern was observed for copepod nauplii in
Departure Bay fish, only, and for amphipods in Hammond Bay fish only
(not shown in Fig. 20). Departure Bay fish also had more calanoid copepods in their stomachs near the times of high tide than of low tide.
No other prey categories showed consistent tidal variations in relative abundance in the fish stomachs.
In summary, pink salmon fry showed diurnal rhythms of feeding activity in both Departure Bay and Hammond Bay. On the average, the stomachs of Departure Bay fish contained more prey (by weight) than those of the larger Hammond Bay fish. Fish from both bays fed mostly on harpacticoid copepods, copepod nauplii, and barnacle larvae, but in different proportions. Tidal variations occurred in the relative abundance of these prey and also of calanoid copepods in the stomachs of the fish.
B. Pattern of feeding behavior in the laboratory Mean snapping rate varied during the 12-h feeding session for both fish groups (Fig. 21). Mean snapping rate during the initial feeding bout (hour 0) was greater for group A fish (Fig. 21A) than for group B fish (Fig. 21B). Snapping declined rapidly during subsequent hours to a relatively constant level for both groups. Mean snapping rates between hours 1 and 11 did not vary significantly with time (Kruskal-Wallis one-way ANOVA). However, the geometric mean rate of snapping during this latter time period was significantly greater A B 90 -i
804 r70
70- •60 SI in '>*Z V) \ - 60 c 60 r-50
01 m i CL o c \ 50 r70 in 504 r-40 E E w Ol TJ CL AO 60 Ol 404 -30 CL 01 c CL /t-l-l- in E CL u> o ID 30 J, 50 c 304 C jU I i_ 01 E CL in 20 4 LAO w i in CL 10 c c n) to Ol 10J 104 -H-»-H
OJ 0J i i—i—i—i—i—i—i—i—i—i—i—i i—i—i—^—i—i—i—i—i—i—i—i—i 0 2 A 6 8 10 12 0 2 4 6 8 10 12
Time (h) Time (h)
Fig. 21. Mean (± 95% confidence limits) number of snaps directed at prey per fish per 10 min (solid line) and mean (± 95% confidence limits) swimming speed of fish (broken line) since the onset of feeding at hour 0 (0800 hours) for one group of five fish previously deprived of food for 24 h (A) and for another group of five fish deprived for 72 h (B) . Fish fed on live copepods that were available throughout the 12-h feeding session. 85
-1 -1 (t-test) for group B (8.38 snaps fish 10 min ) than for group A
-1 -1
(6.04 snaps fish 10 min ). Swimming speeds did not vary significantly with time for either groups (Kruskal-Wallis one-way ANOVA), and were not correlated with mean snapping rates (Spearman rank correlation). Group A fish had a significantly greater (t-test) geometric mean swimming speed for the 12-h period than did group B fish. In only 12 of 120 10-min periods during both 12-h sessions did swimming speed of individual fish correlate significantly with snapping frequency (Spearman rank correlation).
The evacuation rate of calanoid copepods from the stomach of juvenile pink salmon in the laboratory decays exponentially with time, with 50% evacuation occurring about 9 hours after the initial feeding bout (Fig. 22). Since fish feeding is continuous during the 12-h sessions (Fig. 21), only the evacuation rate during the first hour after feeding is considered relevant for comparison with the mean snapping rate between 0900 and 1900 hours (hours 1-11) in relation to that observed at 0800 hours. Table 8 shows that the relative mean snapping rate between 0900 and 1900 hours for group A better approximates the stomach evacuation rate for copepods estimated from the regression line. However, the feeding rate of group B fish better approximates the evacuation rate calculated directly from the data.
The stomachs of group B fish contained a greater absolute
and relative dry prey biomass at the end of the 12-h session compared
to group A fish (Table 9). The daily relative prey biomass consumed 86
9.0 0 8.0 10
- 20
30
40
so
60
70
60 JZ 90 o • 100 CD I I I I L- _1 I I 1 I l_ _l I I ' I 0 2 4 6 8 10 12 14 16 18 20 22 24 ? •o >» *> T3 03 O 3 JQ O ca >% v- > T3 c u C >_ u CL CL4)
1 _J I I 1 I L_ 0 2 4 6i 8' 1'0 12 1L4_ 16 18 20 22 24 36 46
Time (h) Fig. 22. Rate of decline in the relative dry prey biomass (G.M. ± 95% confidence limits) in the stomachs of juvenile pink salmon at 11.4 (± 0.37) °C. The upper curve is drawn from the lower regression line of logarithmic transformed data. Broken lines in the lower panel denote 95% confidence limits on Y. Table 8. Comparison of 1) the relative evacuation rate of calanoid copepods from the stomachs of juvenile.pink salmon during the first hour after "satiation" with 2) the mean prey-snapping rate of the fish between 0900 - 1900 hours (hours 1-11), expressed as a percentage of the mean prey-snapping rate at 0800 hours (hour 0). The regression equation for stomach evacuation rate is
— Y = 9.375 e 0- 104 t^wner e y is the dry weight of prey remaining in the fish's stomach, and t is the time (hours) after the initial feeding period.
Relative stomach evacuation rate Deprivation Relative mean prey-snapping during the first hour after feeding Fish period rate between hours 1-11 Regression line Raw data S3
group (h) (%) (% h"1) (% h"1)
24 9.5 9.9 13.3 72 16.3 Table 9. Absolute and relative mean dry prey biomasses in the stomachs of juvenile pink salmon after a 12-h feeding session, and estimated mean ration consumed by the two fish groups during this period.
Mean prey biomasses in fish Estimated mean ration consumed
Deprivation stomachs after 12 h of feeding per fish during 12 h of feeding
Fish period Dry weight % dry fish Number Dry weight % dry fish
group (h) (mg) body weight of prey (mg) body weight
A 24 105.9 11.5 545.1 256.7 27.1
B 72 139.7 17.9 760.2 357.9 45.8 89 per fish, was estimated at 27.1 and 45.8% dry fish, body weight for group A and group B fish., respectively (Table 9).
Figure 23 shows the cumulative prey consumed per fish for both, groups of fish over the 12-h feeding session. A power function provided the best fit to the data for both satiation curves. Group A fish fed faster than group B fish during the first 10 min. After one hour prey consumption was nearly identical. Thereafter, feeding rates declined for both groups but remained relatively constant for the remainder of the feeding session, resulting in a linear increase in prey consumed. On the average, group B fish fed faster than group A fish from hours 2 to 12 and consumed more prey over the entire 12-h session (see also Table 9).
During the 10-min feeding period at 0800 hours (hour 0), group B fish snapped more than group A fish for the first 5 min (Fig. 24). Thereafter, the opposite was observed. These patterns resulted in group A fish snapping more over the 10-min feeding period than group B fish.
In summary, juvenile pink salmon fed continuously under laboratory conditions. However, feeding rates were high initially (hour 0), and then gradually declined with increasing satiation until fish stomachs were full. Thereafter (hours 1 to 11), feeding rates were constant and approximated the evacuation rate of prey from the fish's stomach. Increased hunger level resulted in greater average O 0 -f 1 1 1 1—:—i 1 1 1 1 1 1 1 0 12 3 4 5 6 7 89 10 11 12
Time (h)
Fig. 23. Cumulative number of prey (N) consumed per fish for group A and group B fish during a 12-h feeding session. Regression, line A is N = 140.2 t °-54 and line B is N = 151.2 t °-6l+. 70 (a0 c (n 60
0800 h B 50 H
OJ ^ A — " ^ A - " E 3 40 H C
C «J 4) 30
E 20 H > (5 2 10 H
o
o -J
i r i 1 r n 1 1 1 ! 1 1 1 1 1 1 1 1 o 1 2 4 5 6 7 8 9 10 Time (min) Fig. 24. Cumulative mean number of snaps (N) directed at prey per fish for group A and group B fish during the 10-min feeding period at 0800 hours ( hour 0). Regression line A is N = 12.29 t °-71+ and regression line B is N = 61.18 t / (2.40 + t). 92 feeding rate and daily ration.
DISCUSSION
From differences in stomach contents of fish captured from the same location at different times of the day, one can infer that changes in the rate of feeding activity occurred over time, resulting in feeding periodicity (Eggers 1977, Jenkins and Green 1977). Pink salmon fry in Departure Bay and Hammond Bay fed primarily during daylight hours with little or no feeding at night. Hence, these fish exhibited diurnal rhythms of feeding activity. Such a rhythm was expected based on the fry's visual threshold for feeding determined in the laboratory (Brett and Ali 1958, Ali 1959, Brett and Groot 1963, Bailey et al. 1975). Diurnal feeding rhythms have been observed previously for juvenile pink salmon in other coastal areas (Parker and Vanstone 1966, Parker 1969, T. Kron, unpublished data; see Appendix Fig. 4). Therefore, my field data confirm these observations.
Pink salmon fry appeared to feed continuously during daylight hours in nature. That is, periods of feeding activity were not interspersed with distinct periods of fasting during the day. However, the diel feeding rhythms of the fry (Fig. 18) suggest that feeding rates during the day were neither linear nor constant (Eggers 1977) . The diurnal feeding rhythms reported by Parker and Vanstone (1966) and Parker (1969) for juvenile pink salmon (Appendix Fig. 4B and 4C) 93
confirm my observation of continuous daytime feeding. Conversely, T. Kron's data (Appendix Fig. 4A) suggest that the fry were not feeding continuously during the day as feeding apparently ceased during the afternoon, when the tide was rising.
Tidal variations occurred in the relative abundance of some prey types in the stomachs of Departure Bay and Hammond Bay fish (Fig. 20). These tidal rhythms in dietary items probably reflect tidal variations in prey availability. Changes in prey selectivity by the fish in relation to the tidal cycle also could have contributed to the observed tidal rhythms. However, this remains untested.
My data and those of other investigators (Parker and Vanstone 1966, Parker 1969, T. Kron, unpublished data) emphasize the variability in the diel timing of feeding activity and in the feeding rates of juvenile pink salmon in coastal marine waters. This variability is the result of many factors. Among these factors, the temporal and spatial variations in prey availability and factors affecting the feeding tendency ("motivation") of the fish are probably of great importance.
Salmonid fishes typically display variability and flexibility in their daily patterns of feeding activity. Brook trout, Salvelinus fontinalis (Hoar 1942) and chum salmon, Oncorhynchus keta
(M.C. Healey, unpublished data) are diurnal feeders. Juveniles of the lake-dwelling, planktivorous sockeye salmon are crepuscular feeders; 94 that is, they feed most actively at dawn and (or) dusk (Narver 1970, McDonald 1973, Eggers 1975, Doble and Eggers 1978). Adult kokanee salmon (0_. nerka) have been observed to feed on plankton during the day arid on benthos at night in a lake (Northcote and Lorz 1966). Several streams-dwelling salmonid species have been reported to feed actively during the day arid (or) night, depending on season, prey availability, and prey type, among other factors. These species are rainbow trout (Jenkins 1969, Adron et al. 1973, Landless 1976, Bisson 1978), brown trout, Salmo trutta (Jenkins 1969, Elliott 1970, 1973), Atlantic salmon (Hoar 1942, Kalleberg 1958, Hirata 1973a), and coho salmon (Mundie 1971). A major conclusion that can be drawn from these studies is that salmonid fishes are opportunistic and generalized predators. In general, fishes in north temperate communities are regarded as generalized feeders; that is, they typically include a wide range of prey types in their diets and possess few-morphological specializations in their feeding apparatus (Hyatt 1979).
The smaller Departure Bay fish consumed more food on the average than the larger Hammond Bay fish (Table 6). The opposite is expected based on the positive relationship between stomach capacity and body weight in fishes (Brett 1971, Elliott 1975a, b). My observation suggests that food availability was lower in Hammond Bay, assuming that all other limiting factors were equal in both bays. Fish in these bays fed on similar planktonic and epibenthic prey, but in differing proportions.
Between 38 and 51% of their diets comprised epibenthic prey. Gerke and Kaczynski (1972) and Kaczynski et ;al. (1973) also reported on the 95 occurrence of epibenthic prey in the diet of pink salmon fry in littoral areas of Puget Sound, Washington. Chum salmon fry also feed heavily on epibenthic invertebrates in nearshore marine habitats CFeller and Kaczynski 1975, Sibert et al. 1977, Healey 1979).
The laboratory data showed that both gastrointestinal and metabolic (systemic) factors affected the feeding rate and daily ration of juvenile pink salmon. My findings agree with two predictions of control theory models of hunger (Rozin 1964, de Ruiter 1967, Colgan 1973, Curio 1976, Toates and Archer 1978) tested in the current study.
Pink salmon fed continuously under laboratory conditions of high prey density, constant light intensity and constant temperature. However, fish feeding rates varied during the 12-h feeding session (Fig. 21). Feeding rates were initially high (hour 0), and then gradually declined with increasing satiation until fish stomachs were full. Declining feeding rate with increasing satiation during a feeding session is predicted by control theory models of hunger, and is common in fishes (e.g., Tugendhat 1960, Beukema 1968, Magnuson 1969, Reed 1971, Ware 1972, Colgan 1973). Mean snapping rates of pink salmon between hours 1 and 11 were constant. The mean feeding rate during this period, relative to that at hour 0, approximated the evacuation rate of prey from the fish's stomach. This finding suggests that after the initial filling of their stomachs with food at high prey densities, juvenile pink salmon keep their stomachs full by 96 feeding at a rate that balances gastric evacuation rate. Examination of fish alimentary tracts at the end of the 12-h feeding session revealed that the stomachs (including the esophagus) were full and distended with prey.
Control theory models of feeding behavior (e.g., Colgan 1973, Toates and Archer 1978) suggest that empty space (deficit) in the gastrointestinal tract results in neural signals being transmitted to the brain from the tract (perhaps the esophagus, Smith and Gibbs 1976). These signals would promote feeding behavior until the esophagus and stomach are filled again. However, Colgan (1973) and Toates and Archer (1978) emphasized that the control of food intake in animals is not simply homeostatic as suggested above. Nevertheless, I conclude that juvenile pink salmon feed at a relatively low hunger threshold. That is, less than 15% of stomach contents need be evacuated for spontaneous feeding to resume or occur.
Goldfish, Carassius auratus (Rozin and Mayer 1964), bluegill sunfish, Lepomis macrochirus (Windell 1966), juvenile pike, Esox lucius (Lillelund 1957, cited in Beukema 1968), and anchovy, Engraulis mordax, (Leong and O'Connell 1969, Hunter 1972) appear to be continuous feeders, like juvenile pink salmon, which maintain their stomachs full. Conversely, other salmonid species (e.g., rainbow trout, brown trout, sockeye salmon) are discontinuous feeders, and apparently feed at a higher hunger threshold than the one reported for pink salmon in the current study (Chaston 1969, Narver 1970, Brett 1971, Adron et al. 1973, Elliott 1973, 1975b, McDonald 1973, Landless 1976, Doble and Eggers 1978, 97
Grove et al. 1978). The fish, in the latter studies behaved somewhat like mammals, which are discontinuous "meal" eaters (de Ruiter 1963, 1967, Rozin 1964, Wiepkema 1971a, b, Curio 1976). Hence, the reported high growth rates of juveniles pink salmon in nature, commonly of 4 to 5% body weight per day (LeBrasseur and Parker 1964, Parker and LeBrasseur 1974, Phillips and Barraclough 1978), are explained in part by their ability to feed continuously at a rate balancing gastric evacuation rate during available feeding time. Therefore, the significance of a lower hunger threshold in pink salmon, compared to other salmonid species, may relate to their relatively higher growth rates during early marine life (Ricker 1976) and in the laboratory (Brett 1974) .
In the current study, increased hunger level resulted in greater average feeding rate and daily ration (Table 9, Fig. 23). This finding agrees with the second tested prediction of control theory models of hunger. The difference in feeding rate and daily ration between the two fish groups used is attributed to their differing food deprivation schedule (hunger level), and not to faster digestion in group B fish as a result of their longer deprivation period or greater ration for the following reason. Some studies have shown that neither food deprivation nor ration size markedly affect the instantaneous rate of gastric evacuation in fishes (Brett and Higgs 1970, Tyler 1970, Beamish 1972, Elliott 1972)• A compensatory increase in feeding rate and ration consumed with increasing duration of the deprivation period has been demonstrated for other fishes (Tugendhat 1960, Ivlev 1961, Rozin and Mayer 1964, Ishiwata 1968a, b, Magnuson 1969). Fish in most of these 98 studies were given relatively short feeding periods.(< one hour). On the one hand, the above observations support the hypothesis that a metabolic (systemic) factor controls in part food intake in fishes, including pink salmon. On the other hand, an influence of prior deprivation on food intake was not observed for the goldfish (Rozin and Mayer 1964), the stickleback, Gasterosteus aculeatus (Beukema 1968), and pumpkinseed sunfish, Lepomis gibbbsus (Colgan 1973). Except for Colgan's study, the feeding periods in these studies were equal to or greater than one hour. From available data, Colgan (1973) argued that the initial feeding rate and amount of food consumed during short periods reflect the length of the prior deprivation period. Paradoxically, if food were available for long periods, fishes consumed amounts independent of deprivation.
Group A fish consumed more food than group B fish during the initial 10-min feeding period (hour 0, Fig. 23 and 24). The opposite would be expected if increasing duration of food deprivation (hunger level) were to increase feeding rate. However, group B fish fed faster than group A fish during the first 5 min of this period. This is attributed to the greater hunger level of group B fish. The reversed pattern was observed during the second half of the 10-min period, when the feeding rate of group B fish declined faster than that of group A fish. The difference in feeding rate.between the groups during the latter 5-min period can be explained in part by the greater level of satiation of group B fish, which could result from their higher feeding rate during the first 5 min, and their relatively smaller size and hence smaller 99 stomach capacity compared to group A fish.
It is unlikely that wild pink salmon fry experience frequent deprivation periods longer than overnight; most dietary studies, including the current one, revealed few fish with empty stomachs in coastal waters. A major exception is the study of Simenstad et al. (1977) who reported 25% empty stomachs among captured pink salmon fry. However, the frequency of empty stomachs increases markedly as larger juvenile pink salmon leave coastal waters for the open ocean; this suggests that the fish experience deprivation periods longer than 12-24 hours (LeBrasseur 1965, Andrievskaya 1968, 1970). Nevertheless, if pink salmon experience food shortages during their juvenile coastal phase, the deprivation period should affect subsequent feeding rates and food consumed, as demonstrated in this study.
No significant correlation was found between mean swimming speed and mean snapping rate, recorded over the 12-h feeding session, for either of the two fish groups. This may be explained by the small tank size and high prey densities used. Under these circumstances, search time would be minimized, permitting the fish to feed rapidly and continously without modifying their swimming speed. A positive relationship between feeding rate and swimming speed in the laboratory has been reported for the air-breathing fish, Ophiocephalus striatus (Pandian and Vivekanandan 1976) and larval anchovy (Hunter and Thomas 1974). Other laboratory studies have indicated changes in fish swimming speed in response to changes in the density of their prey 100
(Ivlev 1961, Hagnuson 1969, Rosenthal and Hempel 1970, Blaxter and Staines 1971, Wyatt 1972). Therefore, It would be instructive to see if juvenile pink salmon maximize their net rate of energy intake by adjusting their swimming speed to ambient prey densities in the manner pEedicted by Ware (1978).
The major conclusions of this portion of the study are as follows. Firstly, pink salmon fry exhibited diurnal rhythms of feeding activity in the littoral zone of marine bays. Hence, the diurnal rhythms of swimming activity exhibited by fry under laboratory conditions (Chapter IV) are considered adaptive for feeding during the day. Diurnal activity, as opposed to nocturnal activity, would increase potentially the prey-encounter rates of the fry and, therefore, their feeding rates during times of the day when incident light intensity is above their visual threshold for feeding. Secondly, the current study and others (Appendix Fig. 4) showed that the timing of daytime peaks of feeding activity in juvenile pink salmon varied between study sites. These observations suggest flexibility in the diurnal rhythms of feeding activity and probably of swimming activity of pink salmon in coastal marine waters. Such flexibility has been demonstrated for other salmonid species, and permits opportunistic exploitation of prey whenever encountered, as suggested by Curio (1976) for predators in general.
Thirdly, juvenile pink salmon fed continuously under laboratory conditions in the current study. They kept their stomachs full by 101 feeding at a rate that balanced gastric evacuation rate. These observations, led to the conclusion that juvenile pink salmon feed at a low hunger threshold, unlike other salmonid species.. I suggest that such a feeding strategy relates to the rapid growth, of juvenile pink salmon in coastal marine waters. Finally, the hunger state of the fish affected their feeding rates and daily ration in ways predicted by control theory models of feeding behavior. Therefore, the laboratory findings suggest that the hunger state of wild fish determines in part their feeding rates and consequently their diel feeding patterns. However, this suggestion remains untested. Toates and Archer (1978) argued that under laboratory conditions, feeding rate likely is optimal and closely determined by metabolic factors. In comparison, prey availability and other ecological factors under natural conditions might be dominant over metabolic factors in the short-term control of feeding behavior. 102
CHAPTER VI. ANNUAL CHANGES IN THE DIEL PATTERNS OF SWIMMING
ACTIVITY, AGGRESSION, AND VERTICAL DISTRIBUTION
INTRODUCTION
In the previous chapters it was shown that pink salmon fry are nocturnally active at emergence and during seaward migration (Chapter III). The fry shift from night activity to day activity shortly after emergence, when they normally reside in estuaries and adjacent coastal waters (Chapter IV). During this juvenile coastal phase, the fry feed mainly during the day (Chapter V).
The duration of the juvenile coastal phase of pink salmon is variable. In British Columbia and southeastern Alaska, juvenile pinks of about 6-15 cm generally migrate offshore into deeper coastal waters from early June until late August (Manzer and Shepard 1962, LeBrasseur and Parker 1964, Martin 1966, Neave 1966). The fish migrate from these coastal waters into the open ocean in late summer and in the fall. The residency period of pink salmon in the open ocean is referred to as the pelagic ocean phase, and is about 10 months in duration (Parker 1962). Oceanic migratory routes of Pacific salmon have been reviewed by Neave (1964) and Royce et al. (1968). These authors concluded that pink salmon from British Columbia and southeastern Alaska undergo a long-distance, counter-clockwise, circular migration in the Gulf of Alaska. This migration is active, directional, and accurately timed on a seasonal basis. 103
Rates of oceanic travel of pink salmon vary considerably (Appendix Table 12). It has been postulated that Pacific salmon must migrate continuously to account for their rate of oceanic travel (Royce et al. 1968, Stasko 1971). The continuous swimming activity of pink salmon fry was shown previously to be affected by an artificial LD cycle (Chapter IV). Therefore, the daily illumination cycle, through its influence on the daily rhythm of swimming speed, may be an important environmental factor determining the seasonal timing of the oceanic migration of pink salmon. However, little is known about the diel activity patterns of pink salmon in the ocean (Neave 1964).
The objective of this portion of the study was to describe annual changes in the diel patterns of swimming activity, aggression, and vertical distribution of juvenile pink salmon exposed to laboratory-simulated annual changes in photoperiod and ambient temperature of 50°N latitude. The study period began in April, when wild fry normally reside in estuaries and adjacent coastal waters, and ended in April of the following year, when wild maturing fish begin their migration from the ocean toward the coast. Questions examined were 1) do the behavioral rhythms of pink salmon fry recorded previously (Chapter IV) during the period of their natural juvenile coastal phase persist during the period of their pelagic ocean phase?, 2) if the fish exhibit behavioral rhythms during the latter phase, do these rhythms vary annually?, and 3) do the parameters (phase, activity time, level) of the rhythms vary annually according 104
to the Aschoff-Wever model? The first five predictions of this model listed in Appendix Table 13 were tested. The findings are related to the migratory behavior of pink salmon, and to annual changes in the locomotor activity rhythms of other species.
MATERIALS AND METHODS
A. Fish
Fish used in this study emerged from a simulated redd similar to the one described in Figure 1. The fish were held in enclosed 484-L fiberglass tanks after emergence. About 50 fish were introduced initially in each of six holding tanks. These tanks had running sea water, and were exposed to simulated natural LD cycles for 50°N latitude. The LD cycles were regulated automatically by a timer. Twilight periods were not used. Times of lights-on and lights-off were changed manually every 10 days to correspond approximately to local times of natural sunrise and sunset, respectively, of Nanaimo, British Columbia (49°13'N, 123°57'W). Hence, artificial LD cycles ranged annually from LD 16:8 to LD 8:16. Light intensities during the photophase and scotophase of all LD cycles were 600 and 1 lx, respectively. Water temperature followed the natural temperature cycle of nearby Departure Bay. Temperature in the holding tanks was always within ± 1°C of that in the experimental tank. Fish were fed as described in Chapter II (Section B). 105
B. Experimental tank The experimental tank was a rectangular, 1.91-cm plywood tank with a front viewing glass, and a false, perforated Plexiglas bottom. The tank was painted light green on the inside. This tank is similar in design to but larger than the one illustrated in Figure 9.
Sea water, filtered to 5 um, entered the tank through a manifold,
-1 located at one end of the tank, at. a rate of about 2 L min . A perforated sheet of Plexiglas, acting as a baffle, reduced any water current at the surface. At the other end of the tank, a standpipe, isolated from the rest of the tank by a plastic screen, regulated the water level at 50 cm. The top of the tank was covered with clear nylon mesh screening to prevent fish from escaping. The baffle and the plastic screen, located at ppposite ends of the tank, delimitated the observation compartment of the tank to 2.0 x 0.6 x 0.46 m high. The observation compartment was divided vertically into four equal sections by lines drawn at intervals of 50 cm on the front viewing glass and on the back wall, and was also divided horizontally by a line into halves of 23 cm in height each.
Experimental LD cycles and water temperatures are given in Appendix Table 14 and Appendix Figure 5. A significant correlation (r =0.65, P < 0.01) existed between mean temperature and photophase. Overhead incandescent bulbs provided illumination of 600 lx during the day and 2 lx at night throughout the experimental period. Water temperature was recorded continuously at the inlet with a thermograph. Differences between daily maximum and minimum temperatures were usually less than 1°C. Mean (± SD) water salinity was 26.9 (± 1.6)%. and 106 mean dissolved oxygen in the water was 81.1 (± 16.9)% of air saturation during the entire experimental period.
C. Experimental procedure The experimental period lasted 12 months (8 April 1975 until 28 April 1976) . Twenty four different groups of six fish each were used during this period. Except for three groups, each group was observed for seven consecutive days (Appendix Table 14). For each experiment, six fish were removed randomly from the holding tanks and placed into the experimental tank. The fish were acclimated to the latter tank for two days before observations began. The LD cycle and water temperature regime of the experimental tank were similar to those of the holding tanks at the same time of the year. While still in the holding tanks, the fish to be used in a particular experiment had been exposed previously for at least one week to the LD cycle and water temperature regime that they would experience in that experiment. Fish were not fed during the period of observation.
Swimming activity, aggressive behavior, and vertical distribution of the fish were recorded with the aid of a Sanyo silicon diode video camera, equipped with a wide-angle lens (1: 1.8/10), located about 2 m from the front of the tank. Manual adjustments of the lens' aperture permitted filming of the fish in the light and dark of the LD cycle. A timer activated the camera and a video tape recorder for a 10-min period every alternate hour. All observations were stored on video tapes for later analysis. 107
Methods used to calculate, fish swimming speed and index of vertical distribution (for the top 23 cm of the water column) for each 10-min observation period were described previously in Chapter IV. Recorded aggressive behavior were "nipping" and "attack" behavior. Nipping was defined as the closing of the mandibles of one fish on the body of another. Attack was defined as an accelerated movement by one fish toward and to within one body length of another fish. Frequencies of observed nipping and attack behavior were summed over the six fish for each 10-min period. The mean rate of aggression then was expressed as the number of aggressive acts (nipping + attack) per fish per 10 min. Diel rhythms of swimming speed, rate of aggression, and vertical distribution were obtained for each fish group by plotting, at 2-h intervals, consecutive 10-min scores of these three behavioral processes.
D. Estimation of rhythm parameters 1. Diurnal-nocturnal (D/N) ratio
The D/N ratio was calculated as the mean of the 10-min scores recorded during all the photophases (days) of the experiment divided by the mean of the 10-min scores recorded during all the , scotophases (nights) of the experiment. This ratio provides a relative index of the degree of diurnalism or nocturnalism of the recorded diel behavioral rhythm.
2. Mean level
Mean level is the geometric mean of all time-series 10-min scores recorded during the experiment (usually seven days long). 108
The geometric means of swimmiLng activity and of aggression are
based on log1(j transformed scores, whereas the mean of the index of vertical distribution is based on arcsin transformed scores.
3. Mean activity time (a)
The locomotor activity rhythm of many organisms can be divided clearly into two physiological phases; an activity phase (during which the animal is active) and a rest (sleeping) phase. Pink salmon swim continuously day and night in the laboratory. Consequently, it is impossible to distinguish clearly active and rest phases from a time-series record of their behavior. Alternatively, a daily "activity time" was defined arbitrarily as the number of hours in a 24-h day during which the score for the recorded behavior was equal to or greater than its daily mean score. Since one 10-min observation period was made every two hours, daily activity time in hours was calculated by multiplying by 2 h the number of 10-min scores per day that were equal to or greater than the daily mean 10-min score. Mean activity time (a) was obtained by averaging daily activity time for all the days of the experiment. Hence, a as defined here does not relate directly to any particular physiological state of the fish. However, periods when scores of a certain behavior are above their daily mean may correspond to a specific physiological state different from another state, characterized by periods when scores are below their daily mean. My definition of a resembles conceptually the one used by Eriksson (1978b) for the perch (Perca fluviatilis). 109 Oi ) 4. Mean phase-angl• e difference max
Phase-angle difference is defined here as the difference in time between the time of the daily maximum behavioral score (regarded as the phase of the biological oscillation) and the time of the daily lights-on stimulus (phase of the physical oscillation) for swimming activity and aggression. The time of the daily lights-off stimulus was regarded as the phase of the physical oscillation for vertical distribution. Mean phase-angle difference is the average of all the daily phase-angle differences for the entire experiment. A positive sign was assigned to a phase-angle difference if the daily maximum score preceded in time lights-on (or lights-off in the case of the index of vertical distribution). A negative sign was given if the maximum score followed the lights-on stimulus (or the lights-off stimulus as in the above case). Animals that exhibit distinct activity and rest phases generally have clearly defined onset and (or) end of activity and rest phases (Aschoff 1960, 1963). It was impossible to use the onset, middle, or end of a clearly defined activity phase to relate to the external LD cycle for pink salmon. Alternatively, the daily maximum score of the recorded behavior was chosen as the daily phase of the biological rhythm for the following reason. The daily peak is commonly the most distinct, although variable, feature of the diel behavioral rhythms of pink salmon recorded in this study.
5. Period
Period was defined previously in Chapter IV. A modified Enright's (1965a, b) periodogram analysis, developed by Dorrscheidt and Beck (1975), was used to calculate the period length of diel 110 behavioral rhythms. This, periodogram analysis is described in
Appendix:1... Time-series data records, used in the periodogram analysis, contained 48 to 84 data points, (i.e., four to seven days of observation for each. fish, group).
RESULTS
A. General behavior of fish
The fish swam continuously day and night along the length of the tank. Individual fish infrequently held a stationary position in the water column for short periods (< 1 min). Fish were never observed to rest on the false bottom of the tank. They formed polarized schools during the day. However, the school broke up at the beginning of the scotophase, and the fish usually swam about as individuals during the night. Fish tended to swim closer to the water surface during the night than during the day. Aggressive behavior was observed mainly during the day. This behavior was exhibited in response to one fish approaching too closely another fish in the same school. Prolonged chasing bouts did not occur. After participating in an aggressive encounter, the aggressor and the recipient rejoined the school.
B. Swimming activity
Annual changes in the average diel pattern of swimming activity for groups of fish are illustrated in Figure 25. The fish swam continuously day and night, but generally swam faster during Ill
.-+--.-»--^_^. Apr 2fl-Mey 3
Now 24-Dae I
Dec 16-23
L4- •r-r Dec 29-Jan 5
Jon* 13-30 Jan 13-20 •—•—+—+— Jun* 28-July 4
-rfr"+ Jen 2S-Feb 1
July 12-19
Feb IB-IS
t T Juty 30- Aug 6 M- March 3- 10 /H-H-V^.
-t-f^'\ March 17-24 Aug 30-27 .^./S-l-K-M . Apr 3-10 Sapt 1-6 ..,,("K|--(-t-t-|-»-t-(- N-j-t- Sept 17-24 _tv^t^ Apr 21-28 Oct 23-30 -4—»—•—*—.——t—.—t—*—•—
Time of day (h) Fig. 25. Average daily pattern of mean swimming speed redorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines are 95% confidence limits about the mean. Light and dark horizontal bars are as in Fig. 10. 112 the day than during the night. The diel patterns were generally diurnal and unimodal from early April until late September. One exception is group 8, which did not exhibit a distinct rhythm in early August. The mean daily mode of swimming activity occurred late in the day in early April, but with time, shifted progressively to a phase position just, after lights-on in mid-June. The daily mode remained at this phase until late September. Except for group 15 in late November, diel rhythms were not apparent from the fall equinox to early January. Thereafter, diurnal rhythms were observed. Diel rhythmicity was more distinct during the second spring than during the preceding winter.
The D/N'ratio for swimming activity was above 1.0 throughout the year (Fig. 26). This indicates a predominance of diurnal activity. Three time periods (groups 2, 13 and 16) were exceptions to this generalization; the D/N ratio was just less than 1.0 for these three periods. In general, the D/N ratio was higher in the first spring and early summer than during the rest of the year. Figure 27 shows a significant linear relationship between the D/N ratio and the duration of photophase. This indicates that, on the average, the degree of diurnalism of the diel patterns of swimming activity was greater at long photophases near the summer solstice than at shorter photophases near the winter solstice.
The mean level of swimming speed showed two different seasonal trends depending on how it was measured (Fig. 28A, B). When expressed 113
3.0-i Swimming speed
2.0
1.0
J 0.0 ——r —f——"T"——i n——i —r—*~n — n
25-i Aggression
20-
15- O 10 •*-(0» 5
0J r-M- i r .i i i 1 i 1
1.0-i Vertical distribution
0.8 A
0.6
0.4 H
0.2-^
J • l.l 0.0 1 i Ai 'i Mr. i J; i r i A'S'O'N'D'i i o r J F'M'A' 1975 1976
Time of year (months)
Fig. 26. Annual changes in the D/N'ratio of swimming speed, aggression, and vertical distribution calculated for different groups of fish at different times of the year. Fig. 27. Relation between the D/N ratio for swimming speed (m fish-1 10 min-1) and the duration of the experimental photophase. The regression line is log Y = -0.19 + 0.03 X (r = 0.59, P < 0.01). 10 115
40 •D CD 0) 30 Q. (0 20 5? -= ._ » 10 E - 0 i > (0 0.3 B C CO (0 0.2 0) \ _i CD 0.1
0.0 J
0.35 c 0.30^ a> .2 1 1 •*-> CO o CO CO 0.15- \ c o> co 0.10 rs o> 0) (0 \ 0.05
o J 0.00 1 T— T (0 r— —t—T— r CD 15-1 D T3 c c E 10- o c (0 0) \
J 5 0 r A ' M ' J ' J ' A " S ' O ' N ' D 1 J ' F ' M ' A 1975 1976
Time of year (months)
Fig. 28. Annual changes in the geometric mean level of swimming speed (A, B), rate of aggression (C), and index of vertical distribution (D) calculated for different groups of fish at different times of the year. 116 as absolute mean distance travelled per 1Q min (Fig. 28A) , the mean level gradually increased during the first spring, and remained at a relatively stable level during the summer. Mean level then increased steadily throughout the fall, winter and into the second spring. This seasonal course of absolute mean swimming speed could have resulted in part from the ability of the fish to swim Increasingly longer distances per unit time as: they grew in length during the 12-month experimental period (Appendix Table 14). This ability is borne, out by Figure 29, which shows a ppsitive correlation between absolute mean swimming speed and mean fish length. When body length is compensated for by
-1 expressing swimming speed as BL s , the mean level of swimming speed generally declined steadily during the first spring and summer (Fig. 28B). This decline was followed by a gradual increase in the mean level during -1 the fall; mean level reached a relatively stable value of 0.15 BL s in early winter. This stable level was maintained until the following
-1 spring. Mean swimming speed never exceeded 0.30 BL s during the experimental period.
A complex pattern emerged when absolute swimming speed was related to the duration of photophase (Fig. 30A) and to the mean temperature of individual experiments (Fig. 31A). In the first instance, mean swimming speed increased slightly as photophase increased from 13 to 16 hours (regression line A in Fig. 30A). Yet, it also increased significantly with photophase decreasing from 16 hours in summer to 8 hours in winter (line B). After the winter solstice, swimming speed continued to increase with increasing photophase length as spring approached (line C). Despite this complex relationship with the Fig. 29. Relation between geometric mean swimming speed and body length for different groups of fish. The regression line is Y = 1.80 + 0.87 X (r = 0.88, P < 0.001). 2 \ io - Q. E
0 ->
? I 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 g 8 9 10 11 12 13 14 15 16 E i
0.1 J •
i 1 1 1 1 1 1 1 r~ 1 1 1 i 1 1 1 1 8 9 10 11 12 13 14 15 16
Photophase (h)
Fig. 30. A. Relation between geometric mean absolute swimming speed and duration of experimental photophase. Regression line A (•) is Y = -19.98 + 2.11 X (r = 0.79, P > 0.05), line B (•) is Y = 22.92 - 0.70 X (r = -0.67, P < 0.05), and line C (A) is Y = 3.78 + 2.25 X (r = 0.91, P < 0.01). B. Relation between geometric mean relative swimming speed and duration of photophase. The regression line is Y = 0.09 + 0.01 X- (r = 0.41, P < 0.05). 119
Fig. 31 A. Relation between geometric mean absolute swimming speed and temperature. Regression line A (•) is Y = A.90 + 0.60 X (r = 0.36, P > 0.05), line B (•) is Y = 40.87 - 2.31 X (r = -0.81, P < 0.01), and line C (A) is Y = -3.46 + 3.33 X (r = 0.94, P < 0.02). B. Relation between geometric mean relative swimming speed and temperature. The regression line is Y = 0.14 + 0.003 X (r = 0.09, P > 0.05). 120.
-1 duration of photophase, mean swimming speed (BL s ) increased linearly with, increasing duration of photophase (Fig. 30B) . Mean swimming speed
-1 (BL s ) still correlated significantly with the duration of photophase (r = 0.47), when temperature was held constant statistically, y 1.2
-1. -1 In the second instance, mean swimming speed (m fish 10 min ) increased linearly with 1) mean temperature increasing from:8.9°C in early spring to 13.2°C in mid-summer (line A in Fig. 31A), and 2) with mean temperature increasing from 9.1°C in mid-winter to .11.1°C in early spring (line C). Conversely, mean swimming speed also increased as temperature declined from 12.5°C in mid-summer to 8.2°C in mid-winter (line B). This complex and conflicting pattern resembles the previously described relationship between swimming speed and photophase (Fig. 30A). However, in this instance, mean swimming speed
-1
(BL s ) did not correlate significantly with mean temperature (Fig. 3IB), as it did with the duration of photophase. Yet, for 10 of the 24 fish groups used, individual 10-min scores of fish swimming speed were correlated significantly with water temperature with zero time lag (Table 10).
The period length of the observed diel patterns of swimming activity (Fig. 25) was within ± 5% of 24.0 h for 20 out of the 24 groups of fish (Table 11). All these 20 periods, except one (group 11), were significantly different from random "noise". No seasonal trend in period length was evident. Only 3 of the 24 periods calculated were not significant. The data suggest that diel rhythms of fish swimming Table 10. Spearman rank correlation coefficients (r^) for comparisons between values of 1) water temperature and swimming speed, 2) water temperature and rate of aggression, and 3) water temperature and the Index of vertical distribution. Swimming speed, aggression, vertical distribution, and temperature were recorded simultaneously every two hours for groupn of fish tested during four-' to seven-day experiments at different times of the year.
Temperature and Temperature and Temperature and
Group Dates of experiment swimming speed rate of a ggression vertical distribution
number (1975-76) r P j> r P s s
1 Apr. 8-15 -0.21 NS NR NR NR NR 2 Apr. 28 - May 5 -0.21 NS NR NR NR NR 3 May 15-22 -0.11 NS NR NR NR t!R 4 Jun. 3-7 0.38 *A 0.18 NS NR NR 5 Jun. 13-20 0.32 AA 0.01 NS NR NR 6 Jun. 28 - Jul. 4 0.12 NS -0.07 NS -0.12 NS 7 Jul. 12-19 0.33 AA -0.14 NS -0.42 AAA 8 Jul. 30 - Aug. 6 0.14 NS 0.01 NS -0.10 NS 9 Aug. 10-14 0.36 ** 0.26 NS -0.21 NS 10 Aug. 20-27 0.25 * -0.01 NS 0.02 NS 11 Sep. 1-8 0.38 *** 0.18 NS 0.40 AA 12 Sep. 17-24 -0.14 NS -0.11 NS 0.12 NS 13 Oct. 23-30 0.22 * 0.34 AAA -0.26 * 14 Nov. 7-14 0.27 * 0.18 NS -0.50 AAA 15 Nov. 24 - Dec. 1 -0.17 NS 0.06 NS 0.32 AA 16 Dec. 16-23 0.08 NS 0.02 NS -0.13 NS 17 Dec. 29 - Jan. 5 -0.15 NS 0.07 NS -0.11 NS 18 Jan. 13-20 -0.05 NS 0.23 * -0.03 NS 19 Jan. 25 - Feb. 1 -0.21 NS. -0.12 NS 0.08 NS 20 Feb. 18-25 0.07 NS 0.01 NS -0.28 A 21 Mar. 3-10 0.08 NS . 0.06 NS -0.34 A* 22 Mar. 17-24 0.21 NS 0.01 NS 0.02 NS 23 Apr. 3-10 0.49 ***• -0.05 NS -0.52 AAA 24 Apr. 21-28 0.23 A -0.14 NS -0.06 NS
NR, Not recorded; NS, Not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001 Table 11. Period lengths of rhythms of swimming activity, aggression, and vertical distribution for different groups of fish tested at different times of the year. Period length of the cycle of water temperature in each experiment is also given. All period lengths are significantly different (? < 0.05) from random "noise", unless indicated otherwise by an asterisk.
Period length (h) Group Dates of experiment Swimming Vertical Water number (1975-76) activity Aggression distribution temperature
1 Apr. 8-15 24.4 NR MR 37 4 2 Apr; 28 - May 5 24.1 NR NR 25 1 3 May 15-22 24.1 NR NR 38 1 4 Jun. 3-7 23.6- 22.0 NR 32 3 5 Jun. 13-20 23.8 23.7 NR 24 5 6 Jun. 28 - Jul. 4 23.7 24.5 24.1 21 * 7 7 Jul. 12-19 23.8 22.8 23.7 23 5 8 Jul. 30 - Aug. 6 29.1 ft 23.3 24 7 9 Aug. 10-14 23.6 23.29.43 25.6" 23 1 10 Aug. 20-27 23.7 23.8 * 38 7* * 20.7 * * Sep. 1-8 23.7 38 11 23.0 23.2 7 12 Sep. 17-24 23.7 24.0 23.8 24 * * * * 7 * Oct. 23-30 38 13 31.7 20.6 28.1 7 * * Nov. 7-14 23.6 22 14 23.3 16.0 9 15 Nov. 24 - Dec. 1 23.6 23.6 22.1 25 * * 4 * 22.2 37 16 Dec. 16-23 22.5 20.2 1 * Dec. 29 - Jan. 5 23.8 24.9 ft 23 17 7 24.4 18 Jan. 13-20 * 23.6 24.1 24 6 17.6 19 Jan. 25 - Feb. 1 24.4 23.8 23.8 24 A 20 Feb. 18-25 23.8 23.7 23.6 38 4 A 21 Mar. 3-10 23.8 23.7 23.8 24 71 22 Mar. 17-24 23.8 24.1 23.8 23 A 23 Apr. 3-10 23.8 23.7 23.8 24 8 A 24 Apr. 21-28 23.8 23.7 28.7 27 1 A 8
NR, Not recorded 123 activity were synchronized with the 24-h LD cycle during most of the year. Occasional periods of desynchronization (groups 8, 13, 16 and 18) were indicated by period lengths markedly different from 24.0 h (Table 11). These desynchronized patterns occurred in early August, late October, late November, and late January (Fig. 25). However, two (groups 8 and 16) of these four groups exhibited significant rhythmicity, but with periods different from 24.0 h. Hence, only the other two groups (13 and 18) exhibited both desynchronized and arrhythmic swimming activity. Swimming speed correlated with short-term temperature fluctuations in only one (group 13) of the four desynchronized groups (Table 10). Furthermore, the period of the activity rhythm was not correlated significantly (r = 0.20) with the period of the corresponding temperature cycle during the experiment (Table 11).
Mean daily activity time (a) did not correlate significantly (r = 0.11) with photophase (Fig. 32A). Mean activity times ranged from 8.9 to 13.7 hours daily, and were generally greater near the times of the solstices. The activity times were shorter than expected (if they were directly dependent on photophase) during spring, summer and early fall (when photophases > 11 hours), and were longer than expected during late fall and winter (when photophases < 11 hours).
Mean phase-angle difference CP correlated negatively (r = -0.57) with the duration of photophase (Fig. 33A). This means 124
16 Swimming activity 14 12
10
8
6 4
2
0 "T 1 1 1 1 1 1 1 1 1 1 1 1 f 16 Aggression B 14
12
6 > 4 u < 2H
0 ~t 1 1 T 1 T-
16 Vertical distribution 14
12H
10
8
6 4
2H
0 l—i—i—i—i—l—i—i—l—i—i—i—(—i—I—" 0 2 4 6 8 10 12 14 16
Photophase-Scotophase (h) Fig. 32. Mean daily activity time (a) for swimming activity (A) and for aggression (B) in relation to the duration of experimental photophase, and a for vertical distribution (C) in relation to the duration of experimental scotophase. 125
+ 8-, Swimming activity
+ 4
0 «— LON
-4
-8H
-12 O c -16 CD i 1 1 1 1 1 1 r
& + 8 Aggression
+ 4
o> 0 c J — LON CO - 4 I a w -8J ca r ~i r T 1 1 1 sz a. + 4-, Vertical distribution c c (0 0 '— L OFF v S -4
- 8
-12-
-16- I 1 1 1 1 1 1 i 1 8 9 10 11 12 13 14 15 16
Photophase (h) Fig. 33. Variations in mean phase-angle difference for swimming activity (A), aggression (B), and vertical distribution (C) with the duration of photophase. Regression line A is Y,= 6.77 - 0.86 X (r - -0.57, P < 0.001, one-tailed), line B is Y =0.18 - 0.21 X (r = -0.24, P > 0.05), and line C is Y = 13.71 - 0.48 X (r =0.41, P < 0.05, one-tailed). LON and LOFF,denote the daily times of lights-on and lights-off, respectively. 126 that the daily mode of swimming activity tended to occur, on the average, progressively later after the lights-on stimulus as photophase increased.
Neither ¥ (r =0.25) nor a (r =0.26) of the diel pattern of swimming activity correlated significantly with its corresponding period length for each fish group.
In summary, juvenile pink salmon.exhibited generally diurnal, unimodal rhythms of swimming activity that were synchronized with the
24-h LD cycle during most of the year. Temperature cycles did not synchronize these rhythms. On the average, fish swam increasingly faster during the day relative to their swimming speed at night (i.e., increasing diurnalism) with increasing duration of photophase. Further,
-1 mean swimming speed (BL s ) also increased with increasing duration of photophase, but did not vary significantly with mean water temperature. However, swimming speed of some fish groups was affected by short-term fluctuations in temperature. The mean daily mode of swimming activity occurred progressively later after lights-on as photophase increased in duration. Mean daily activity time was greater than expected at short photophases (< 11 hours), but less than expected at longer photophases (> 11 hours). No significant relationship existed between
the corresponding period length, a and ^max for individual diel patterns of swimming activity. 127
C. Aggressive behavior Aggressive behavior was first recorded in early June, when fish averaged 9.5 cm in length and 4.6 g in weight (Appendix Table 14). Aggression was observed too infrequently to be recorded before this time. In general, the diel rhythm of aggression was diurnal throughout the experimental period, except in late October (Fig. 34). The diurnal rhythms were usually bimodal from early June until late November. The major peak occurred within two hours after lights-on, and a second equivalent or minor peak occurred 5-8 hours after lights-on or just before lights-off. From mid-December until early March, diurnal rhythms were generally unimodal, with the mean daily mode occurring within two hours after lights-on. Diel rhythmicity was not pronounced in late March and in April.
The D/N ratio for aggressive behavior was greater than 1.0 throughout the experimental period with the exception of late October (Fig. 26). This indicates that the daily rhythms were predominantly diurnal. The D/N. ratio was greater in late spring (3-7 June), early summer (28 June - 4 July) and late winter (18 Feb. - 10 March) than during the rest of the year. Hence, the ratio increased with photophase increasing from 8 to 10.5 hours, but declined with photophase increasing further to 13 hours (Fig. 35). Thereafter, no significant variation in the D/N ratio occurred with changes in the duration of photophase.
The mean level or mean rate of aggression increased exponentially from early June until late August (Fig. 28C) . Thereafter, 128
Juna 3-7 _:_:_:^-i-UUl_;>-U. N—4^ Jun. 13-20 Nov 24-Dae 1 \- \1
Jun« 28- July i
July 12-19 Dae 29-Jan 5 -j-J-k. July 30-Aug 6
-iA Jan 25 • Feb 1
E • z ~N.__4 J
S«pt 1-8 Mirci. 3-10
-» '
March 17- 24 '.] i-
Apr 3-10
Oct 23-30 Apr 21-28 -(—I-' I ^4 . —t — 4—4. I-.-4-.
Time of day ih) Fig. 34. Average daily pattern of the mean rate of aggression recorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines and horizontal bars are as in Fig. 25. Fig. 35. Relation between the D/N ratio for mean rate of aggression and the duration of the experimental photophase. The number besides each value denotes the month of the year when the observation was made. 130
it declined to almost zero in early November, but increased sporadically
during late fall and early winter. Mean level was low again in late winter and early in the second spring. Apart from the major seasonal peak, in mid-summer, the mean rate of aggression was relatively low -1 -1 (< 0.1 acts fish 10 min ) throughout the experimental period. The mean rate of aggression was related positively to mean experimental temperature (Fig. 36), but did not vary significantly with changes in the duration of photophase (Fig. 37). Although mean temperature affected the mean level of aggression, a significant correlation between short-term fluctuations in water temperature and 10-min scores of aggression was observed for only 2 of the.21 groups of fish used (Table 10). The mean level of aggression was not related to mean fish length (r = -0.22). Therefore, annual variations in this level cannot be attributed to a crowding factor in the experimental tank.
The period length of the observed diel rhythms of aggressive behavior was within ± 5% of 24.0 h for 17 of the 21 groups used (Table 11). All these 17 periods, except one (group 5), were significantly different from random "noise". No seasonal trend in period length was evident. Only three periods calculated were not significant. The data suggest that the diel rhythms of aggressive behavior were synchronized with the 24-h LD cycle most of the year. However, sporadic periods of desynchronization, indicated by period lengths markedly different from 24.0 h, occurred in June (groups 4, 5), in early August (group 8), in late October (group 13), and in mid-December (group 16). The period of the rhythm of aggressive behavior for any 0.4
o
i i 1 1—;—i——n 1 i 1 1 1 1 1 8 9 10 11 12 13 14
Mean temperature (°C)
Fig. 36. Variation in the geometric mean rate of aggression for different groups of fish with mean experimental temperature. The regression line is Y = -0.14 + 0.02 X (r = 0.39, P < 0.05, one-tailed). Fig. 37. Relation between geometric mean rate of aggression for different groups of fish and the duration of the experimental photophase. The regression line is Y = -0.008 + 0.005 X (r - 0.18, P > 0.05). 133
group of fish, was not significantly correlated (r = 0.01) with the
period of the corresponding temperature cycle during the experiment.
No significant correlation (r =0.19) existed between a for aggression and the duration of photophase (Fig- 32). Mean daily activity times ranged from 3.0 to 9.7 hours daily, but were always less than expected if they were directly dependent on photophase.
Mean phase-angle difference correlated negatively, but not significantly (r = -0.24), with the duration of photophase (Fig. 33).
This means that the daily mode of aggressive behavior tended to occur, on the average, progressively later after the lights-on stimulus as the duration of photophase increased. Neither ¥ (r = 0.13) nor max s
a (rg =0.34) of the diel pattern of aggressive behavior correlated significantly with its corresponding period length for each fish group.
In summary, juvenile pink salmon exhibited generally diurnal rhythms of aggressive behavior throughout the experimental period. The diurnal rhythms were synchronized with the 24-h LD cycle during most of this period, with the exception of four short periods of time. Temperature cycles did not synchronize these rhythms. The diurnal rhythms tended to be bimodal from the first spring until mid-fall, but were generally unimodal in late fall and winter. Little aggression was observed at night. The degree of diurnalism of the rhythms, as measured by the D/N 134
ratio, varied in a non-linear manner with annual changes in the duration of photophase. The mean rate of aggression was relatively low and stable throughout the experimental period, with the exception of a major peak in mid-summer. Mean rate of aggression correlated positively, although weakly so, with mean water temperature on a seasonal basis, but did not correlate with the duration of photophase. With increasing duration of photophase, the mean daily mode of aggression tended to occur progressively later after lights-on. Mean daily activity times were not related to the duration of photophase, and were consistently less than expected. Mean daily activity time and
\iax n0t corre-'-ate with the corresponding period length of individual diel patterns of aggressive behavior.
D. Vertical distribution
Vertical distribution was only recorded for a 10-month period, starting on 28 June 1975 (Fig. 38). This 10-month period can be divided into two periods, based on the variability of the diel rhythm of vertical distribution among fish groups. On the one hand, the period from 28 June to 1 December shows more variability than the remainder of the 10-month experimental period. The diel rhythms of vertical distribution were mainly nocturnal from late June until mid-July. That is, greater instantaneous percentages of the fish swimming in the upper half of the water column were recorded during the night than during the day. However, the fish also exhibited a weak tendency to swim in the upper half of the water column during the day, but the majority swam 135
Jun* 28- July 4 Dac 14-23
Dec 29-Jan 5 H+u-iv'i-H- s July 12-19 o
c S
July 30-Aug 6 Jan 25- Fab 1 -f-»-.:
Aug 10-14
—T—t—+—
Aug 20-27 Feb 18-25
/I Sapt 1-8
-h-K March 3-10
Sept 17-24 . . » 4 4
Marih 17-24
Oct 23-30 ^-\- Apr 3-10
-J T-
Nov 24- Dec 1 Apr 21-28
\it—t—+ -
II 14
Time of day th) Fig. 38. Average daily pattern of the mean index of vertical- distribution in a water column recorded for different groups of pink salmon at different times of the year under simulated seasonal LD cycles and natural temperatures. Each mean value is based on seven data points obtained on seven consecutive days, except where otherwise indicated near the mean. Vertical lines and horizontal bars are as in Fig. 25. 136
in the lower half. The nocturnal rhythm of vertical distribution persisted during August, but the fish displayed only a weak tendency to swim in the upper half of the water column during the night. With the exception of groups 12 and 13 (late September and October) that displayed nocturnal rhythms, it is difficult to determine, based on visual inspection of Figure 38, whether the diel patterns of vertical distribution during early September and November are mainly diurnal or nocturnal.
On the other hand, diel rhythms were mainly nocturnal and unimodal from mid-December until early April. During this period, the tendency of the fish to swim in the upper half of the water column during the night was relatively high; this tendency was almost nil during the day, when the fish swam in the lower half. However, the diel pattern appeared arrhythmic in late April. The mean index of vertical distribution never exceeded 50% at any time of day or night throughout the experimental period. This means that at any instantaneous point in time less than 50% of the fish, on the average, were observed swimming in the upper half of the water column.
The D/N ratio was always less than 1.0 throughout the experimental period (Fig. 26). This indicates a predominance of nocturnal rhythms. In general, the D/N ratio was relatively high at photophases less than 10 hours in duration (Fig. 39). A marked decline in the ratio occurred at photophases between 8.5 and 10.0 hours in duration. Conversely, the ratio increased gradually with photophase 0.001 J i 1 1 r-i 1———T 1 1 1 1 1 1 1 1 1 1 1 8 9 10 ti 12 13 14 15 16
Photophase (h)
Fig. 39. Relation between the D/N ratio for mean index of vertical distribution and the duration of the experimental photophase. The number besides each value denotes the month of the year when the observation was made. 138 increasing from 10 to 14 hours. This latter trend indicates a progressively declining degree of nocturnalism in the daily rhythm of vertical distribution. The D/N ratio declined once again with further increases in the duration of photophase during late spring and summer.
No consistent seasonal trend was observed in the mean level or mean index of vertical distribution (Fig. 28D). Hence, no significant correlation existed between the mean level and the duration of the scotophase (Fig. 40). The mean index increased sporadically during the first spring, summer, and fall (Fig. 28D). Only in winter were mean index values consistently high. This period coincides with a period of a relatively high degree of nocturnalism in the diel patterns of vertical distribution (Fig. 26).
The period length of the observed diel rhythms of vertical distribution (Fig. 38) was within ± 5% of 24.0 h for 12 out of the 19 fish groups (Table 11). Only two (groups 11 and 17) of these 12 periods were not significantly different from random "noise". Seven of the 19 periods calculated were not significant. Therefore, the majority of the fish groups exhibited a nocturnal pattern of vertical distribution that was synchronized with the 24-h LD cycle. However, 36.8% (7 out of 19) of the groups were arrhythmic, as indicated by their non-significant periods. Five (71.4%) of these seven arrhythmic patterns occurred between August and late November; this period is characterized by a relatively high inter-group variability in the diel patterns of vertical distribution. The other two arrhythmic patterns 15 -i
8 9 10 11 12 13 14 15 16
Scotophase (h)
Fig. 40. Relation between geometric mean index of vertical distribution for different groups of fish and duration of the experimental scotophase. The regression line is Y = 3.04 + 0.18 X (r = 0.12, P > 0.05). 140
occurred in December and early January, within the time period
characterized by relatively stable nocturnal patterns. The pattern in
late April (group 24) had a significant period of 28.7 h (Table 11)
and, by definition, was not synchronized with the LD cycle. The index
of vertical distribution was correlated significantly with short-term
temperature fluctuations in 8 of the 19 groups used (Table 10). Yet,
the period of the rhythm of vertical distribution was not correlated
significantly (r = -0.38) with the period of the corresponding
temperature cycle during the experiment (Table 11).
Mean activity times ranged from 3.5 to 13.4 hours daily (Fig. 32C). The majority of a values were less than expected if they were directly dependent on the scotophase. In general, mean daily activity times were generally greater at long scotophases than at shorter scotophases (Fig. 32C). However, the relationship between a and the duration of scotophase was not significant (r = 0.38).
J Mean phase-angle difference, with respect to the lights-off stimulus, correlated positively with the. duration of photophase (Fig. 33C). This means that the daily mode of the index of vertical distribution tended to occur, on the average, progressively later after lights-off as the duration of scotophase increased (or as the corresponding photophase decreased). Neither ¥ (r = -0.31) nor nicix s a. (r =0.21) of the diel pattern of vertical distribution correlated significantly with its corresponding period length for each fish group. 141
In summary, juvenile pink salmon exhibited nocturnal rhythms of vertical distribution throughout the year.. That is, on the average, the fish always had a greater tendency to swim in the upper half of the water column during the night than during the day. However, nocturnal rhythms of vertical distribution were more variable among fish groups during the first spring, summer and fall (period 1) than in winter (period 2), when they were stable. A greater proportion of the rhythms were desynchronized during time period 1 than during the winter. Hence,
nocturnal rhythms were more strongly synchronized with the 24-h LD cycle in winter, when nights are longest, than during any other season. Temperature cycles did not synchronize the rhythms. However, short-term fluctuation in temperature affected fish vertical distribution in 42% of the fish groups used (Table 10). In these groups, the fish had a greater tendency to swim in the upper half of the water column when temperature declined, usually at night.
No seasonal trend was observed in the mean level or mean index of vertical distribution, which was low (< 13%) throughout the experimental period. The mean level was higher in winter than during the other seasons. As scotophase increased in duration, the mean nocturnal mode of vertical distribution occurred progressively later after lights-off. Mean daily activity times tended to increase with increasing duration of scotophase. However, this tendency was not significant. Mean daily activity time and mean phase-angle difference did not correlate significantly with the corresponding period length 142
of individual nocturnal rhythms of vertical distribution.
E. Relationship between swimming activity, aggression, and vertical distribution
In general, juvenile pink salmon exhibited diurnal patterns of swimming activity and aggressive behavior, and nocturnal patterns of vertical distribution throughout most of the experimental period. The inter-group variability in the form of these behavioral rhythms was greater during the first part of the experimental period (early April until early December) than during the latter part (mid-December until late April). Because rhythms of swimming activity and of aggression were diurnal, a positive correlation was found between scores of swimming speed and corresponding scores of aggression in 86% of the fish groups (Table 12). Hence, high rates of aggression were frequently associated with high swimming speeds. This relationship is exemplified in Figure 41. However, mean levels of aggression and of swimming speed did not correlate significantly on a seasonal basis (Table 13).
Diel rhythms of vertical distribution were nocturnal. A significant negative correlation between scores of vertical distribution and corresponding scores of swimming speed was found in 53% of the fish groups (Table 12). Hence, the fish occasionally swam in the upper half of the water column when they were relatively inactive, as exemplified in Figure 41. A negative correlation between scores of vertical distribution and corresponding scores of aggression was found in only 32% of the fish groups (Table 12). Therefore, on the average, the diel Table 12. Spearman rank correlation coefficients (rg) for comparisons between values of 1) fish swimming speed and rate of aggression, 2) swimming speed and the index of vertical distribution, and 3) rate of aggression and the index of vertical distribution. Swimming speed, aggression, and vertical distribution were recorded simultaneously every two hours for groups of fish tested during four- to seven-day experiments at different times of the year.
Swimming speed and Swimming speed and Rate of aggression and
Group Dates of experiments rate of aggression vertical distribution vertical distribution
number (1975-76) r P r P r P
4 Jun. 3-7 0.52 AAA NR NR NR NR 5 Jun. 13-20 0.38 AAA NR NR NR NR 6 Jun. 28 - Jul. 4 0.53 AAA -0.02 NS 0.07 NS 7 Jul. 12-19 0.18 NS -0.45 AAA 0.33 AA 8 Jul. 30 - Aug. 6 0.44 AAA -0.04 NS 0.17 NS 9 Aug. 10-14 0.81 AAA -0.05 NS 0.10 MS 10 Aug. 20-27 0.43 AAA 0.14 NS 0.06 NS 11 Sep. 1-8 0.40 AAA 0.53 AAA -0.10 NS 12 Sep. 17-24 0.75 AAA -0.38 AAA -0.35 AA 13 Oct. 23-30 0.31 AA 0.16 NS -0.06 NS 14 Nov. 7-14 0.31 AA 0.04 NS -0.05 NS 15 Nov. 24 - Dec. 1 0.64 AAA -0.34 AA -0.14 NS 16 Dec. 16-23 0.13 NS -0.25 A -0.08 NS 17 Dec. 29 - Jan. 5 0.46 AAA -0.13 NS -0.21 NS 18 Jan. 13-20 0.31 AA -0.19 NS -0.18 NS 19 Jan. 25 - Feb. 1 0.26 A -0.48 AAA -0.48 AAA 20 Feb. 18-25 0.50 AAA -0.66 AAA -0.56 AAA 21 Mar. 3-10 0.64 AAA -0.54 AAA -0.76 AAA 22 Mar. 17-24 0.24 A -0.27 A -0.33 AA 23 Apr. 3-10 0.06 NS -0.73 AAA -0.17 NS 24 Apr. 21-28 0.31 AA 0.01 NS -0.04 NS
NR, Not recorded; NS, Not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001 Sept 17-24
1? 03 11 19 03 i: 19 03 11 19 03 n 19 03 11 19 03 II 19 03 II 19 4 9 14 19 24 29 34 39
Time of day (h) Period (h)
Fig. 41. Daily rhythms of swimming speed (A), rate of aggression (B), and index of vertical distribution (C) recorded simultaneously for seven consecutive days in LD 13:11 and 11.9 °C for pink salmon group 12. Periodograms show the major periodic component (P) of each rhythmic function. All P values are significantly different (P < 0.05) from random "noise". Table 13. Spearman rank correlation coefficients (r ) for comparisons between pairs of the three types of behavioral rhythm recorded. Coefficients are given for the level, period, mean daily activity time (a), and mean phase-angle difference.(¥ ) of the rhythm. All correlations are non-significant, IQcLX except where indicated specifically by asterisks.
Level Period a Swimming activity and aggression -0.20 0.36 -0.39 0.64 Swimming activity and vertical distribution 0.44* 0.35 -0.09 -0.17 Aggression and vertical distribution -0.25 0.10 -0.27 -0.26 *, P < 0.05 (one-tailed); **, P < 0.001 146 rhythm of aggressive behavior corresponded more closely to the diel rhythm of swimming activity, whereas the rhythm of vertical distribution corresponded more closely ta the activity rhythm than to the rhythm of aggression. -1 The mean level of swimming speed (BL s ) correlated positively with the duration of photophase, but was independent of seasonal changes in mean temperature. Conversely, the mean level of aggression correlated weakly with mean temperature, but not with the duration of photophase. Neither photophase nor mean temperature affected the mean level of vertical distribution. On a seasonal basis, only the mean levels of swimming activity and of vertical distribution were correlated, significantly with each other (Table 13). Diel rhythms of swimming activity, aggression, and vertical distribution were synchronized with the 24-h LD cycle for most of the experimental period, except for a few short periods of desynchronization (Table 11). No seasonal trend in period length was observed for any of the three types of behavioral rhythm, and no significant correlation existed between the periods of these rhythms (Table 13). Proportionally more diel rhythms of swimming activity and of aggression were synchronized with the LD cycle than rhythms of vertical distribution. Further, about 37% of the diel patterns of vertical distribution were arrhythmic. Only fish groups 13 and 16 were desynchronized in all three types of behavioral rhythm. Temperature cycles apparently did not synchronize the behavioral rhythms.. Yet, short-term fluctuations 147 in temperature affected the scores of swimming activity and of vertical distribution for several groups of fish (Table IQ). The above observations indicate that the fish synchronized their diel patterns of swimming activity, aggression,.and vertical distribution to differing degrees with the LD cycle. The data also indicate that the lights-on stimulus was the part of the LD cycle that synchronized the diurnal rhythms of swimming activity and of aggression, whereas the lights-off stimulus appeared to synchronize the nocturnal rhythms of vertical distribution. Mean daily activity times for all three types of behavioral rhythms were independent of the duration of photophase or scotophase (Fig. 32). There was a weak tendency for a values of swimming activity and of aggression to be greater near the times of the solstices than near the times of the equinoxes. The a values for swimming activity and aggression were correlated negatively (Table 13). The mean daily mode of the rhythms of swimming activity and aggression occurred progressively later in time with respect to lights-on as the duration of photophasv v e increased. Hence, ¥ma x for these two JtypeV s of behavioral rhythm correlated positively with each other (Table 13). Similarly, the mean daily mode for the rhythm of vertical distribution occurred progressively later in time with respect to lights-off as the duration of scotophase increased. Therefore, ¥ for the rhythms of swimming activity and aggression are expected to be, on the average, most negative near the summer solstice, whereas V for the rhythms of max J vertical distribution are expected to be most negative near the winter 148 solstice. No correlation, existed between the corresponding a, ¥ , max and period length for any of the three types of behavioral rhythms recorded for any fish group. DISCUSSION The results showed that juvenile pink salmon exhibited diel rhythms of behavior under laboratory conditions during the periods corresponding to their juvenile coastal phase and pelagic ocean phase. In general, the fish displayed diurnal rhythms of swimming activity and of aggression, and nocturnal rhythms of vertical distribution during most of the experimental period. Annual changes occurred in many parameters of these behavioral rhythms. However, these changes did not support the Aschoff-Wever one-oscillator model of circadian rhythms (Appendix Table 13). The phase position of the daily peaks of the behavioral rhythms suggests that the D-L transition synchronized swimming activity and aggression, whereas the L-D transition synchronized vertical distribution. It is commonly recognized that the natural daily cycle of illumination is the major synchronizer of biological rhythms (Aschoff 1960, 1963, Menaker 1969, Biinning 1973, Enright 1975, Gwinner 1975). However, which portion(s) of the daily illumination cycle is (are) the effective synchronizer(s) is not well understood. Twilight periods appear to be stronger synchronizers of locomotor activity in some species than rectangular LD cycles (Gwinner 1975, Kavaliers 1978a). 149 Pink salmon were inadvertently subjected to temperature cycles in the current study. In the presence of a LD cycle, these temperature cycles did not synchronize the rhythms of swimming activity, aggression, and vertical distribution in pink salmon. However, short-term fluctuations in temperature affected swimming activity and vertical distribution in some fish groups. Temperature has been shown to affect the form and timing of swimming activity rhythms in other fish species (Northcote 1962, Byrne 1968, Dill 1970, 011a and Studholme 1972, 1978, Richkus 1974). Temperature synchronization of locomotor activity has been demonstrated in ectothermic animals such as insects (Sweeney and Hastings 1960, Wilkins 1965), crustaceans (Palmer 1974), amphibians (Adler 1969), and reptiles (Evans 1966, Hoffmann 1968, 1969, 1970, Gourley 1972, Mangelsdorff. and Hauty 1972). In most of these studies, the LD cycle was the dominant, synchronizer, when the animal was exposed simultaneously to both LD and temperature cycles. Temperature synchronization of fish swimming activity has received little attention. The few data available are conflicting. On the one hand, Kavaliers (1978a) did not observe synchronization of swimming activity with a diel temperature cycle of about 4°C amplitude in lake chub (Couesius plumbeus) under constant darkness. Similarly, Gibson (1971) subjected blennies (Blennius pholis) to temperature cycles of 10°C amplitude for several days, and found no synchronizing effect of the temperature cycle on swimming activity. On the other hand, MacAfee (1971) reported that stream-dwelling brook trout fry, subjected to constant light, synchronized their swimming activity to a diel temperature cycle of 7°C amplitude. She also found that temperature was the dominant 150 synchronizer, when both LD and temperature cycles were presented simultaneously to the fish. Annual changes occurred in the recorded parameters of the behavioral rhythms of pink salmon. Mean daily activity times (a) of swimming activity, aggression, and vertical distribution were independent of the duration of photophase or scotophase ranging from 8 to 16 hours. This is not in accordance with the prediction of the Aschoff-Wever model (Appendix Table 13). A lack of correlation between a and the duration of day or night has also been reported for the field mouse, Apodemus flavicollis (Erkinaro 1970), the flying squirrel, Pteromys volans (Hokkanen et al. 1977), and the ground squirrel, Ammospermbphilus leucurus (Kenagy 1978). Activity time of the field mouse was correlated with scotophase between 3 and 10 hours in duration, but did not change significantly with scotophase increasing further from 11 to 21 hours. On the one hand, Eriksson . (1978b) reported a positive correlation between a of perch and daylength at 66°N; however, a weak sigmoid relationship was indicated. Similarly, Miiller (1976, 1978c) found that the period of daily activity for individual and schooled whitefish (Coregonus lavaretus) was related closely to daylength at 66°N, except for a period of arrhythmicity in mid-summer. The data of Kavaliers (1978a) indicate a positive correlation between a and photophase for the lake chub at 53°N. On the other hand, a sigmoid relationship between a and the duration of day or night has been observed in birds and mammals (Aschoff et al. 1970, Daan and Aschoff 1975), 151 plants. (Muller-Haeckel 1974), and one fish, species (Figala and Miiller 1972) at Arctic and subarctic latitudes. In these latter studies, a corresponded closely to the duration of day or night ranging from about 9 to 15 hours. However, a was longer than the duration of day or night when the latter was shorter than 9 hours, and a was shorter than the duration of day or night when the latter exceeded 15 hours. Mean activity time of the barb (Barbus partipentazona) varied only between 10 and 12 hours with daylength ranging from 8 to 16 hours; a did not correspond closely to daylength. within this range (Figala and Muller 1972). The sigmoid relationship can be explained by a two-oscillator model, but not by the Aschoff-Wever one-oscillator model (Daan and Aschoff 1975). Therefore, the above studies indicate considerable inter-specific variability in the relationship between a and the duration of day or night, and that this relationship does not agree with the one predicted by the Aschoff-Wever model in many species. Activity time for pink salmon was measured by an unconventional method, and consequently it may not represent a true physiological state. However, if a is assumed to reflect a distinct physiological state in the current study, then it can be concluded that a of the behavioral rhythms of pink salmon varied independently of seasonal changes in the duration of day or night. r Mean phase-angle difference CPmax) ° mean daily mode of the pattern of swimming activity and of aggression occurred progressively 152 later in time relative to lights-on as the duration of photophase increased. Similarly, the mean daily mode of the pattern of vertical distribution occurred progressively later in time relative to lights-off as the duration of scotophase increased. Further, no correlation was observed between the period, a, and Y for any of the r ' ' max J three types of behavioral rhythm. These observations do not agree with the predictions of the Aschoff-Wever model (Appendix Table 13). However, a direct comparison of my results with the theoretical predictions of this model is difficult because the phase point QH ) used in the current study is-different from the one in the max model Qimr d.,). Support for the model's prediction of the seasonal changes in ¥ ^ has been obtained for only a few species of mammals (Aschoff 1969, Aschoff et al. 1970, Erkinaro 1970, Daan and Aschoff 1975) and birds (Aschoff 1969). In some of these studies, maximum and minimum f ' ' - onset occurred in summer and winter, respectively, for nocturnal species (Aschoff et al. 1970, Erkinaro 1970, Daan and Aschoff 1975), and in winter and summer, respectively, for diurnal species (Aschoff 1969, Aschoff et al. 1970, Daan and Aschoff 1975). Hence, species that conformed to the model's prediction of seasonal changes in * ^ also showed a consistent seasonal course in f which the Aschof f-Wever onset model does not accomodate. In general, the seasonal trend in V for • max pinr k salmon agree° s with the above seasonal trend in fonse t. for the locomotor activity of species that followed the model's prediction for \iid* Similarly, maximum and minimum leading phases C^onget) also occurred in winter and summer, respectively, for diurnal fish 153 (Bertmar and Miiller 1970, Figala. and Miiller 1972, Miiller 1978c) and mammals (Zwahlen 1975); the reverse relationship has been observed for nocturnal mammals (Hokkanen et al. 1977, Subbaraj and Chandrashekaran 1977) and lobster, Paniilirus argus (Kanciruk and r Herrnkind 1973). A very different seasonal pattern of ^onset f° locomotor activity was reported for lake chub (Kavaliers 1978a) and kangaroo rats (Kenagy 1976). These species showed maximum * nget near the equinoxes and minimum ¥ near the solstices; this pattern onset r reflects the seasonal course of twilight duration. Hence, the ¥ma x measured for the three type• s of behavioral rhythm in the current study behaved more like the ¥ _ than the onset ^mid °^ t^e a^ove species in response to seasonal changes in the LD ratio. Therefore, it is tempting to make an analogy between the ^ of pink salmon and the ¥ _ of other species that exhibit a distinct onset activity onset and a. However, a major theoretical problem arises in the a priori choice of a reference point for a rhythm (e.g., onset, midpoint, maximum, or end), which is most important in phase control and synchronization (Aschoff 1960). In previous studies on birds and mammals, the midpoint of the activity period (a) and the midpoint of the photophase or scotophase were chosen as the true corresponding phases of the two oscillations. These phases were chosen because they were less variable than other phase points, and were considered independent of the level of the biological oscillation, which varies with light intensity (Aschoff 1960, Aschoff et al. 1970). Consequently, ¥mr .d, was used in the formulation of the Aschoff-Wever model. However, many exceptions to this model's prediction of the seasonal course of 154 ^mid nave Deen reported (Aschoff et al. 1970, Erkinaro 1972, Muller-Haeckel 1974, Voute et al. 1974, Daan and Aschoff 1975). Hence, predictions of the Aschoff-Wever model concerning the seasonal courses of ^ ^ and a are supported only to a limited extent by a few species (Aschoff 1969, Erkinaro 1970, Aschoff et al. 1970, Daan and Aschoff 1975, Gwinner 1975). This suggests that either 1) too many factors not accomodated in the model are involved and thus obscure general tendencies or 2) some of the basic assumptions of the model (e.g., usefullness of the midpoint of activity as a phase reference point) are wrong (Daan and Aschoff 1975, Gwinner 1975). -1 Mean level of swimming speed (BL s ) was correlated positively with the duration of photophase, but was independent of seasonal changes in mean temperature. Data on swimming speed of pink salmon in the ocean are inadequate to correlate with daylength or water temperature. 011a and Studholme (1972) also reported a positive correlation between the swimming speed of bluefish (Pomatomus saltatrix) and photoperiod, with temperature held constant. Johnson and Groot (1963) observed a similar relationship for wild sockeye salmon smolts. These positive correlations between fish swimming speed and daylength or photophase could have resulted from photoperiodic (seasonal) changes in hormonal levels and in their effects on behavior (Fontaine 1975, Woodhead 1975, Hoar 1976, Simpson 1978). Water temperature is an important environmental factor controlling, limiting, and directing^various functions in fishes 155 (Fry 1967, 197.1, Brett 1970). The influence of temperature on the spontaneous locomotor activity of fish: has received relatively little attention (Brett 1970). However, one general pattern that has emerged is the increase in spontaneous activity in response to temperatures deviating from the acclimation temperature of the fish (Brett 1970, Fry 1971). Positive correlations between seasonal changes in mean level of swimming activity and mean temperature have been reported for several fish species (Andreasson 1969, Andreasson and Miiller 1969, Miiller 1969b, 1976, 1978a, c, Eriksson and Kubicek 1972, Eriksson 1975, 1978b, Penaz 1975, Staples 1978). Conversely, the mean swimming speed of pink salmon was independent of seasonal variations in mean water temperature, to which they had been acclimated previously in holding tanks. This observation suggests the operation of some form of behavioral homeostatic mechanism within the thermal range used in the current study. Temperature compensation or thermal homeostasis has been reported previously for various rate-dependent functions in ectothermic animals (Bullock 1955). For fishes, Fry and Hart (1948) observed that the maximum sustained cruising speed of goldfish was relatively independent of acclimation temperatures between 20 and 30°C. Respiratory or metabolic rate has been shown to be relatively independent of acclimation temperatures within certain ranges in the carp, Cyprinus carpio (Meuwis and Heuts 1957) , pumpkinseed sunfish (Roberts 1964), bluegill sunfish (Burns 1975), and yellow bullhead, Ictalurus riatalis (Morris 1965). Similarly, one of the major properties of circadian rhythms is the relative temperature independence of their spontaneous periods (Schwassman 1971a, Bunning 1973, Palmer 1976), and thus of the 156 biological clock. Annual changes.occurred in the form of the diel rhythms of swimming activity, aggression, and vertical distribution recorded in the laboratory. Little is known about behavioral rhythms in pink salmon migrating in the open ocean. Therefore, the significance of the observed annual changes in daily behavioral rhythms is difficult to assess. Juvenile pink salmon swam continuously day and night throughout the experimental period. Such behavior had been postulated previously by Royce et al. (1968) and Stasko (1971) to account for the fish's rate of oceanic travel. In general, the fish exhibited diurnal rhythms of swimming activity during the spring and early summer under laboratory conditions. This observation confirms the findings of Chapter IV, which showed that pink salmon fry were day-active during the period of their natural juvenile coastal phase. The significance of a diurnal activity rhythm was discussed previously in Chapters IV and V. Daily patterns of swimming activity were weakly rhythmic during the summer (Fig. 25), and some of the fish groups used were aperiodic (Table 11). This period coincides with the natural offshore migration of juvenile pink salmon. These data suggest that the swimming activity of the fish during this offshore migration might not be related strongly to the daily illumination cycle, and thus might not show distinct diel rhythmicity. A lack of diel rhythmicity has been shown in the swimming activity of adult pink-salmon (Stasko et al. 1973) and of the juveniles and adults of other salmonid species (Groot et al. 1975, Ichihara et al..1975, Stasko 1975, Stasko and Rommel 1977, Fried et al. 1978, LaBar et al. 1978, McCleave 1978, Tytler et al. 1978), 157 when migrating through estuaries and coastal marine waters. These fish appeared to use tidal currents to facilitate their horizontal movements. Movements with tidal currents may be more energetically efficient than other movement strategies in areas where strong tides occur (Greer Walker et al. 1978, Weihs 1978). Perhaps juvenile pink salmon use tidal ebb currents to facilitate their offshore seaward migration, as suggested by Martin (1966) and Healey (1967), and (or) use seaward salinity gradients, which vary with the tidal cycle, as orientation cues during this migration (Mclnerney 1964). Weak diel rhythms of swimming activity were commonly observed during the fall and early winter, a period when pink salmon are scattered widely in the open ocean (Royce et al. 1968). However, the daily rhythms were generally synchronized with the LD cycle during this period. Swimming activity rhythms were more distinct and more strongly synchronized with the LD cycle from February until late April. During the latter part of this period, maturing pink salmon begin to migrate northward in the eastern Gulf of Alaska in the general direction of their natal rivers (Royce et al. 1968). Several species of insects. (Miiller 1973b), fish (Andreasson 1969, Miiller 1969a, b, c, d, 1970a, b, 1973a, b, 1978a, b, Eriksson 1972, 1973, 1975, 1978a, Kroneld 1974, 1976, Pena'z 1975), and mammals (Erkinaro 1961, 1969a, b, Herman 1977) exhibit seasonal phase' inversions of their diel rhythm of locomotor activity at latitudes usually greater than 60°N. Conversely, seasonal phase inversion of the daily activity rhythm of pink salmon did not occur in the current study . 158 Temperate-zone fishes generally do not exhibit seasonal phase inversions CSiegmmnd 1969, Figala and Muller 1972, Penaz 1975, Eriksson 1978b, Kavaliers 1978a, Miiller 1978b), other than ontogenetic ones. However, exceptions are known. Richardson and McCleave (1974) tested Atlantic salmon parr in LD 12:12 at different times of the year at a temperate latitude. They observed that most of the fish were diurnal in summer, nocturnal in winter, and crepuscular in the spring. Staples (1978) observed that adult upland bully shifted from day-activity in winter to night-activity in summer under natural conditions. Chaston (1969) reported that brown trout were nocturnally active all year, except in summer when they tended to exhibit more day-activity. Juvenile pink salmon exhibited diurnal rhythms of aggressive behavior throughout most of the experimental study period (Fig. 34). The diurnal patterns were generally either unimodal or bimodal, with peak aggression occurring shortly after lights-on and shortly before lights-off. Diel variations in aggressive behavior have been reported for the territorial juveniles of several stream-dwelling salmonid species (Hoar 1953, Stringer and Hoar 1955, Newman 1956, Edmundson et al. 1968, Mason 1969). These studies indicated that levels of aggression were lower at night than during the day. In general, salmonids appear to exhibit aggressive behavior, associated with territoriality, most commonly as juveniles during periods of stream residency and as adults during the spawning season (Newman 1960). Little aggression apparently occurs during migration and pelagic life, when the fish are usually schooling. There are no 159 reports of aggression in wild pink, salmon. This may relate to their schooling behavior as fry, while residing in inshore marine waters. Little is known about the schooling behavior and other social behavior of pink salmon in the open sea (Lshida 1966, Royce et al. 1968). In the laboratory, aggressive behavior appeared to have been displayed in response to one fish being too close to another fish in the school. Therefore, aggression may. have served as a spacing mechanism in this instance, maintaining a minimum distance between individual members of a snhool. However, this aggression may be a laboratory artifact resulting from the limited space available to the fish in the experimental tank. Regardless of its apparent function under laboratory conditions, the function of.aggressive behavior in wild pink salmon remains unclear. Juvenile pink salmon displayed nocturnal rhythms of vertical distribution throughout most of the experimental period (Fig. 38). In the Gulf of Alaska, pink salmon appear to migrate predominantly in the upper 35 m of water (Manzer and LeBrasseur 1959, Neave 1960, Manzer 1964, Royce et al. 1968). However, there is little data on their diel patterns of vertical distribution. Based on gillnet catches, pink salmon swim closer to the water surface during the night than during the day in the open ocean (Manzer 1964, Machidori 1966). Hence, my laboratory findings on diel vertical distribution confirm available field data. Other salmon species apparently also rise toward the water surface at night and descend to greater depths during daylight in the ocean (Neave 1960, Manzer 1964, 160 Ishida 1966,, Machidori 1966) • Many other marine fish show nocturnal patterns of vertical migration (Woodhead 1965, 1966, Blaxter 1970, Brunei 1973, Roe 1974,. Teach. 1978). Annual changes occurred in the form of the daily rhythm of vertical distribution. On the one hand, pink salmon tended to swim near the bottom of the tank.during August and early September, and consequently did not exhibit strong diel vertical migrations. Further, the majority of fish groups used during this period, which coincides with the onset of offshore migration of wild pink salmon, were arrhythmic in their vertical distribution. Some of the fish groups displayed concurrently weak diel rhythmicity in their swimming activity. On the other hand, the fish exhibited generally strong nocturnal vertical migrations, synchronized with the LD cycle, during winter and their second spring. During their second spring, the fish are widely scattered in the Gulf of Alaska (Royce et al.' 1968). The significance of .these annual changes in the daily rhythm of vertical distribution is unclear. Seasonal variations in the form of the diel rhythm of vertical migration has been reported for cod, Gadus morhua (Brunei 1965, 1972, 1973), whitefish, Coregonus lavacetus (Eriksson and Kubicek 1972), perch (Eriksson 1978b) and brown trout (Chaston 1969). In these studies, diel vertical migrations appeared to have resulted from a response of the fish to diel changes in underwater light intensity, occurring mainly during twilight periods. Vertical migration may also be a trophic response to prey availability or an avoidance response to potential 161 predators (Woodhead 1966, Blaxter .1970). In the current study, diel vertical migration appears... to be a photic response of the fish to changes in light intensity.associated with, the L-D and D-L transitions. Many adaptive significances have been postulated for diel vertical migration in fishes and other animals, but few have been shown conclusively to operate in nature (McLaren 1963, Woodhead 1965, 1966, Blaxter 1970, Roe 1974, Swift 1976, Eggers 1975). The putative adaptive values are: 1) escape from potential predators, 2) facilitation of horizontal migration, 3) facilitation of navigation, 4) facilitation of the exploitation of prey patches near the surface, 5) demographic effects, 6) greater efficiency of energy utilization, 7) reduction of interspecific food competition, and 8) combinations of the above. The major conclusions of this portion of the study are as follows. Firstly, juvenile pink salmon swam continuously day and night throughout the experimental period. This behavior has been postulated necessary to account for the rate of oceanic travel of Pacific salmon (Royce et al. 1968, Stasko 1971). Pink salmon exhibited diurnal rhythms of swimming activity and of aggression, and nocturnal rhythms of vertical distribution in the laboratory during periods corresponding to their juvenile coastal phase and pelagic ocean phase. With the exception of the nocturnal vertical migration, these behavioral rhythms of juvenile fish remain unconfirmed by field data. As for a wide spectrum of organisms, the daily LD cycle was. the major synchronizer of the daily behavioral rhythms in pink salmon. The rhythms were synchronized with the LD cycle throughout most of the experimental period, 162 except during the summer. In general, daily patterns of swimming activity and of vertical distribution were weakly rhythmic and occasionally desynchronized in summer. This period coincides with the natural offshore migration of juvenile pink salmon. Such weak diel rhythms may relate to a natural tendency of wild fish to use tidal ebb currents to facilitate their offshore movements, as suggested by Martin (1966) and Healey 0-967). Secondly, some parameters of the behavioral rhythms of pink salmon were not entirely dependent on environmental factors such as photoperiod and temperature. Mean swimming speed was independent of acclimation temperature, and activity time for the three types of rhythm recorded was independent of photophase. However, mean swimming speed, the degree of diurnalism of the swimming activity rhythm, and the timing of the daily peak of the behavioral rhythms were affected by the duration of the day. At temperate latitudes, photoperiod (daylength) is considered the most predictable indicator of seasons (Lofts 1970). Therefore, photoperiod might be an important proximate factor that pink salmon use to time their oceanic migration on a seasonal basis, assuming that such seasonal timing results in part from adjustments in swimming speed and in timing of daily rhythms. Finally, annual changes occurred in activity time (a) and phase Cfmax) °f the behavioral rhythms recorded. However, these changes did not conform to those predicted by the Aschoff-Wever model, which is based on data obtained mainly from birds and mammals. This lack of conformity between observed and predicted seasonal changes in rhythm parameters might result from 1) differences in the rhythm parameters 163 (a, ¥) recorded for pink salmon and those used in the model, 2) fundamental differences between the physiological mechanisms underlying the behavioral rhythms of pink salmon and those underlying the locomotor activity rhythms of higher vertebrates, and 3) limitations of the model in its explanation of seasonal changes in rhythms in organisms differing widely in phylogenetic origin. 164 CHAPTER VII. TIMING OF THE DIEL SWIMMING ACTIVITY RHYTHM OF INDIVIDUAL FISH INTRODUCTION In the preceding chapters artificial LD cycles synchronized the diel rhythms of gravel emergence, swimming activity, feeding activity, vertical distribution, and aggressive behavior in juvenile pink salmon. Under natural conditions biological rhythms are normally synchronized by environmental cycles, such as the LD cycle, which follow the period of the solar cycle. However, in the absence of environmental time cues biological rhythms commonly persist, but have autonomous periods that deviate from 24 hours. Under these conditions such a rhythm is said to free-run, and thus is called a circadian rhythm (Aschoff 1960). The circadian period is thought to reflect the period of an endogenous, autonomous, biochemical oscillator or biological clock within the organism (Aschoff 1960, 1967, Hastings 1970, Biinning 1973, Pittendrigh 1974, Gwinner 1975). Since the biological clock is simply an hypothetical construct, its properties can be investigated only indirectly by recording one or more overt biological rhythms or indicator processes, The major assumption is that such processes are overt expcessions of an underlying endogenous timing mechanism to which they are coupled (Aschoff 1960, Biinning 1973). A free-running rhythm persisting for at least five to 165 seven cycles and showing a spontaneous period (T) between<20 and 28 hours under."constant" laboratory conditions have been commonly considered major criteria for the existence of an endogenous rhythmicity .(Aschoff 1960, 1963, 1967, Bunning 1973, Enright1975, Gwinner 1975). Synchronization of a biological rhythm by an environmental cycle can result from direct effects of the environmental cycle on the organism's behavior (i.e., masking, Aschoff 1960, 1963, Hoffmann 1969, Gwinner 1975) or from the synchronization (control of phase and frequency) of an endogenous rhythm by the environmental cycle (i.e., entrainment, Aschoff 1960, 1963, Gwinner 1975). If the environmental cycle entrains an endogenous circadian rhythm, then it is referred to as a Zeitgeber (Aschoff 1960, 1963, Gwinner 1975). Endogenous circadian rhythms are ubiquitous, and have been documented in organisms ranging from unicellular algae to man (Bunning 1973). However, evidence for an endogenous clock in fishes is relatively weak and equivocal (see Appendix Tables 15 and 16). In the majority of studies conducted on fishes, rhythms of swimming activity and of other functions persisted for only a few days in constant light or darkness before fading (damping) out. There is no evidence for the existence of an endogenous clock in pink salmon. The objective of this portion of the study was to test the Zeitgeber influence of an artificial LD cycle on the diel rhythm of swimming activity in individual pink salmon. Swimming activity was 166 chosen because 1) of its importance in migration and 2) of the relative ease in obtaining lengthy records of it in individual fish. Before the above objective could be addressed, the existence of an endogenous, free-running activity rhythm in individual pink salmon had to be established. This was done by subjecting fish to constant illumination of various intensities and to constant temperature, and recording concurrently their swimming activity. This was essentially a test of the generalized Circadian Rule (Aschoff 1960), which predicts that diurnal organisms should 1) increase their total amount of activity, 2) increase the frequency (1/T) of their free-running rhythm, and 3) increase the activity (a)/rest (p) ratio of their free-running rhythm as the intensity of constant illumination increases. The reverse relationships are predicted for nocturnal organisms. The first two of these three predictions were tested on individual pink salmon. Aschoff (1960) listed three types of experiment that one can conduct to test whether an environmental cycle is effective as a Zeitgeber. These are 1) catching the free-running rhythm, 2) phase-shifting of the endogenous rhythm, and 3) driving the endogenous rhythm with different external frequencies. The first two types of experiment were conducted on individual pink salmon to test the effectiveness of an artificial LD cycle as a Zeitgeber of swimming activity. 167 MATERIALS AND METHODS A. Fish and holding conditions Fish used in this study were held in enclosed 484-L fiberglass tanks after emergence. About 50 fish were introduced initially ,in each of three holding tanks. These tanks were supplied with running sea water of ambient temperature, and were exposed to a LD 12:12 .cycle without twilights. Times of lights-on and lights-off were 0800 and 2000 hours, respectively. Light intensity during the scotophase (D) was about 0.1 lx for all three holding tanks. However, light intensity during the photophase (L) was different for each tank; being 6, 60, and 600. lx for the three holding tanks, respectively. Water temperature ranged from 9 to 13°C during the experimental period, but was usually near 10°C. Fish were fed two to four times daily in the manner described in Chapter II (Section B). B. Description of the activity channels and activity monitors The "spontaneous" (unforced) swimming activity of individual fish was recorded continuously in an activity channel with the aid of photoelectric activity monitors described previously by Godin et al. (1978). The activity channel and monitors are described again below. 1. Activity channels Each of four identical, circular channels was constructed of two concentric pieces of white ABS plastic (3 mm thick) tightly sealed to a concentric Plexiglas bottom painted flat black (Fig. 42). The width of the channel is 15.5 cm and its inner and outer diameters are 49.5 and 49.5 cm I1 W-Mll Side View < : f \ IBWWWWIL* Fig. 42. Diagram illustrating the top and side views of an activity channel. Broken lines indicate transparent pieces of Plexiglas. Only the inlet and outlet tubes of the charcoal filter are shown, b, surface and bottom Plexiglas bridges; f, filter; 1, fresnel lens; Is, light source; p, phototransistor; pp, darkened Plexiglas plate with clear slit; r, aluminum light reflector; w, clear Plexiglas window. 169 80.0 cm, respectively. Transparent Plexiglas windows (6.4 mm thick, 2.5 cm wide) are mounted and sealed along the height of the inner and outer walls on opposite sides of the channel, such that all four windows are aligned with each other. Directly aligned with these windows on opposite sides of the channel are two Plexiglas bridges fixed to the bottom across the widthof the channel. These bridges are 2.54 cm high and 2.7 cm wide at the top, with their long sides tapered at 45° angles. Above each bottom bridge and submerged about 0.5 cm is a rectangular piece of clear Plexiglas (2.7 cm wide, 15.5 cm long, 0.9 cm thick) with black sides (Fig. 42).. These surface and bottom pieces of Plexiglas limit the effective height of the infrared light beams (see below) in each channel to about 12.5 cm, but do not interfere with the swimming speed of test fish in the channel. The bridges were necessary because fish interrupting the light beams near the water surface and channel bottom did not always trigger the monitoring system. Each channel is filled with sea water to a depth of 15.0 cm (total volume of about 47.1 L). This water is continuously air-lifted from the channel, via an opaque plastic pipe (1 cm O.D.) mounted on the inside of the iraner wall of the channel, to an activated charcoal filter, and returned by gravity to the opposite side of the channel through -1 Tygon tubing. Water exchange rates vary between 0.25-0^.35 L min . This water filtering system did not create any measurable water currents in the channel, and did not appear to influence the locomotor activity of test fish. The activity channels are isolated from each other in separate large refrigeration chambers. The temperature of the water in 170 the channel is thermostatically controlled to within ± 0.2°C. The channels are covered with nylon net to prevent fish from escaping. Diffused, Incident illumination is.provided by an incandescent light bulb, covered with an opaque glass: globe, located directly above each activity channel. A timer regulates a LD cycle without twilight periods. 2. Activity monitors Two light sources and two phototransistors, each separately housed.in an aluminum cylinder fixed on an aluminum bar, are aligned with the four windows of each'channel (Fig. 42). The light sources are juxtaposed to the outer wall of the channel, whereas the phototransistors are located within the central space of the channel. Each light source consists of a 60-W red, ceramic, incandescent light bmlb operated at 120 V AC and 0.5 A and housed at one end of an aluminum cylinder (15.2 cm I.D., 31.1 cm long). Heat generated by this bulb is conducted by the aluminum cylinder to the surrounding chilled air. No measurable change in water temperature due to the light source occurred in the channel after several days of operation. A fresnel lens (14.9 cm diam, 15.0 cm focal length) and Kodak No. 87A filter are mounted at the other end of the cylinder. The cylinder is sealed at this end with a circular, darkened Plexiglas plate having a clear 2,5-cm wide slit along its diameter. This vertical slit is aligned with the corresponding vertical window in the outer wall of the channel (Fig. 42). Therefore, light rays are emitted from the cylinder only through the clear slit on its frontal, acrylic plate. The filter transmits light of wave lengths greater than 870 nm. The fresnel lens collimates light 171 rays throughout the water column in the channel. These light rays are in turn focused on a Philips-OCP 71 germanium phototransistor, with peak spectral sensitivity at 1550 nm (range 500-2000 nm), by a similar fresnel lens mounted inside another cylinder juxtaposed to the inner wall of the channel opposite the light source (Fig- 42). Hence, these two matched pairs of light source and phototransistor result in two infrared light beams (2.5 cm wide, 12.5 cm high) being emitted across the width of the channel on opposite sides. The phototransistor circuitry is shown in Appendix Figure 6 and its supporting electronics in Appendix Figure;7. The components of Appendix Figure 7 are duplicated for each of the eight phototransistors (two per channel). When the infrared light beam illuminating the phototransistor is interrupted by a fish a change in voltage level occurs at the collector of the phototransistor. This signal is amplified by a CA 3000 amplifier, and fed through an emitter follower 2N 697 transistor to a triggering circuit (Appendix Figure 8), which is basically a multivibrator. The trigger level of this multivibrator is set such that a minimum threshold pulse of 200 mV is required to trigger it. The multivibrator may be operated in either a monostable or a bistable mode. The bistable mode of operation requires ^wo light sources and two phototransistors for each triggering circuit, whereas only one matched pair of light source and phototransistor is required in the monostable mode (Fig. 42). In the monostable mode, a time-delayed trigger (manually set at 5-6 s, range 2-20 s), instead of a second matched pair of light source and phototransistor, is used to reset the 172 multivibrator. Following a light beam interruption by the fish in this mode, the multivibrator triggers only once after the preset time interval (5-6 s) expires, even if the fish, traverses, the light, beam more than once during this interval. In the bistable mode, the time-delayed trigger is non-functional. Eence, for the multivibrator to trigger, the fish has to interrupt sequentially both light beams in the channel (and thus swim the equivalent of one channel revolution) in the bistable mode, whereas it only has to interrupt one of the two light beams in the monostable mode. In either case, triggering of the multivibrator simultaneously activates an event recorder and a bank of cumulative digital counters (Appendix.Figure 7). The latter were photographed every 30 min, and activity estimates were obtained by subtracting numbers of one interval from the previous one. During experimentation, the triggering circuit was operated in the monostable mode for reasons discussed in Appendix 2. 3. Evaluation of activity monitors The general behavior of the fish in the activity channel and the quantitative evaluation of the effectiveness of the activity monitors in recording their swimming activity are reviewed in Appendix 2. Data on the swimming behavior of individual fish in the circular channel indicate that the design permitted the fish to swim freely in circuits without physical obstructions impeding or water currents directing their movements. Swimming speeds of individual fish were not influenced by infrared light beams emitted across the activity channel. The photoelectric monitoring system was found to be reliable, equally sensitive 173 in detecting the locomotor movements of small (4.3-5.8 cm) and of lar; (13.0-19.0 cm) fish, and provided data that closely approximated \on average to within 94.0-102.6%) the total momentary locomotor activity of individual fish. C. Experimental procedures In all experiments described below the fish were removed from the holding tanks and placed singly in separate activity channels. They were acclimated to the channel for two days under LD 12:12 before the onset of activity recording. The experimental regime of illumination varied among experiments. Seawater temperature was thermostatically maintained at 10 ± 0.5°C. No diel temperature cycle occurred in the activity channels. Dissolved oxygen in the channels was always greater than 90% of air saturation. Fish were not fed while in the experimental activity channel. Their swimming activity was recorded continuously, and was expressed as the cumulative number of electronic counts or triggers per 30 min. 1. Experiment 1, Circadian Rule This experiment was conducted during the period of 7 May to 6 December 1977. Three different treatments of light intensity were used. They are 6, 60, and 600 lx; these light intensities are identical to the ones used during the photophase of the holding tanks. Different fish were used for each of the three treatments. During the 2-day acclimation period the fish were exposed to a LD 12:12 cycle. Light intensity during D = 0.2 lx and during L = 6, 60, or 600 lx, depending 174 on the subsequent constant light treatment (LL) the fish were subjected to. After the acclimation period, swimming activity of the fish was recorded in LD 12:12 for 24 hours, and then in LL of a specific intensity for 10 consecutive days. The three constant light treatments were conducted in random order. The number and size of fish tested in each LL treatment are listed in Table 14. Two of 40 fish examined had one opaque eye by the end of the experiment, and consequently were excluded from analysis. All other fish appeared healthy. 2. Experiments, Catching the free-running rhythm This experiment was conducted during the period of 19 February to 6 March 1978. Four fish were examined. Mean (± SD) fish length and wet weight were 19.8 (± 0.7) cm and 52.7 (± 8.0) g, respectively. The experiment lasted 15 days. After a 2-day acclimation period to the channel under LD 12:12 (600:0.2 lx), the swimming activity of individual fish was recorded continuously for five days (Days 1-5) under LD 12:12 (600:0.2 lx), followed by five days (Days 6-10) in LL (600 lx), and finally by another five days (Days 11-15) in LD 12:12 (600:0.2 lx). 3. Experiment 3, Phase-shifting of the endogenous rhythm This experiment was conducted during the period of 11 December 1977 to 14 February 1978. Fourteen fish were examined. Mean (± SD) fish length and wet weight were 17.0 (± 1.4) cm and 34.1 (± 10.2) g, respectively. The experiment lasted 11.5 days. After a 2-day acclimation period to the channel under LD 12:12 (600:0.2 lx), the swimming activity of individual fish was recorded continuously for five days under LD 12:12 (600:0.2 lx). Times of lights-on and lights-off were 0800 and 2000 175 Table 14. Number, mean (± SD) total length, and mean (± SD) wet weight of juvenile pink salmon tested under three different intensities of constant light in the Circadian Rule experiment. Light intensity Number of Total length Wet weight (lx) fish (cm) (g) 6 13 15.5 ± 1.2 26.8 ± 7.6 60 11 12.7 ± 1.5 13.9 ± 4.9 600 14 8.5 ± 1.9 4.0 ± 3.0 176 hours, respectively. On Day 5 the LD cycle was phase-shifted by 12 hours by lengthening the scotophase (D) by 12 hours, thereby delaying lights-on by 12 hours. This resulted in new times of lights-on and lights-off at 2000 and 0800 hours, respectively.. After the phase shift of the LD cycle, swimming activity of the fish was recorded for another six days (Days 6-12) under the new phase-shifted LD 12:12 (600:0.2 lx). D. Statistical procedures In all three experiments the period lengths of the rhythms of swimming activity recorded in LD and LL were estimated using the periodogram analysis described in Appendix,1. In Experiment 1 the mean level of activity or mean swimming speed, exhibited by individual fish during the course of the experiment in LL, was expressed as mean cumulative number of counts per 30 min. Non-parametric tests were used in Experiments 2 and 3 to determine the significance of differences between treatments, whereas parametric tests were used exclusively in Experiment 1. RESULTS A. Experiment 1, Circadian Rule Based on the diel distribution of their swimming activity, 68.4% of all fish were day-active for the one day in LD before being subjected to LL. However, 38.5 and 45.5% of the fish were nocturnally active at 6 and 60 lx, respectively (Fig. 43A). Under constant light conditions (LL) of this experiment, the pattern of swimming activity 177 100 n 14 •*- o> P > 80 H H 13 0 11 o 60 H n (D c n 0 40 H kO . 0 Q. 20 H 0 J c 80 -t a o o ^ 60 H 0 c *J w 4 0 c B 0 u (0 •o 20 H w o 0 0 a. a. 0 J 100 n Type 1 • Type J u 2 Q 4 0 80 H Type Type o- 0 60 C 40 H 0 u u 20 a. o J 60 600 Light intensity (lux) Fig. 43. Differences in the percentage of the fish that were day-active in LD 12:12 for one day (A) and in the proportion of calculated periods that are significantly different from random "noise" (B) with different intensity of constant illumination. Also the proportion of the four types of activity pattern observed under the three constant light intensities are given (C). The numbers above the bars are the numbers of fish examined under each light intensity. 178 of individual fish, generally lost synchrony with the previous LD cycle, and free-ran with a period deviating from 24 hours in most instances. Free-running periods (T) ranged from 20.0 to 37.7 h for the 10 days in LL (Table 15), but 86.8% of these periods were between 20 and 28 hours. The free-running period tended to decrease, but not significantly so (r = -0.19, P > 0.05), with increasing constant light intensity (Fig. 44A), whereas mean fish, swimming speed increased significantly (r = 0.64) with increasing constant light intensity (Fig. 44B). No significant correlation (r =0.06) was found between x and the corresponding mean swimming speed (activity level) of individual fish. Only 50% of all periods recorded (n = 38) in LL for 10 days were significantly different from random "noise" (Table 15). The proportion of significant periods declined from 61.5% at 6 lx to 42.9% at 600 lx (Fig. 43B). The negative relationship between T and constant light intensity remained non-significant (r = -0.36, P > 0.05) when only the significant periods were considered. Four types of activity pattern were recognized in LL. Type 1 pattern (Fig. 45A) is a free-running rhythm that has a significant periodicity during both the first half (Days 1-5) and the second half (Days 6-10) of the 10-day experiment. Generally, this type of pattern had a significant period when periodogram analysis was performed on the pooled data for the 10 days.(Table 15). Type 2 pattern (Fig. 45B) is a free-running rhythm that lacks significant periodicity during the first half of the experiment, but possesses a significant periodicity 179 Table 15. Period lengths of free-running swimming activity rhythms of individual pink salmon recorded in constant illumination (IL) of different intensities for a period of 10 days. Period lengths of the rhythms recorded during the first five days (Days 1-5) in LL and the second five days (Days 6-10) in LL are presented for each fish. The diel activity pattern of each fish was determined in LD 12:12 for one day before being subjected to LL. Light intensity Fish Activity Period length (h) umber pattern Days 1-10 Days 1-5 Days 6-10 1 D 24.2 b 18.8 24.3 d 2 D 29.1 b 15.3 25.8 d 3 D 23.9 26.8 20.2 * * * 4 N 24.6 * 23.9 * * 24.2 * 5 D 35.6 * 38.7 * 35.2 6 N 29.7 23.5 23.5 * * 7 N 24.7 * 24.2 25.7 8 N 20.0 * * 20.5 * 19.5 9 D 23.4 26.3 23.6 10 D * * 36.9 37.7 37.1 11 D 26.5 27.2 17.4 * * 12 N 24.1 * 23.2 22.5 13 D 26.5 * 28.2 26.9 * * 14 D 23.4 * 24.0 * 23.4 15 D 27.3 24.1 * 27.4 16 N 23.8 * 23.6 * 24.7 * 17 N 24.1 * * 23.8 36.7 18 N 36.4 33.7 * 35.6 19 N 24.0 * • 23.9 * 24.5 20 D 31.3 31.9 25.9 21 D 24.4 * 24.0 * * 25.9 22 N 23.6 24.2 29.9 23 D 24.7 26.1 * 22.4 24 D 26.3 17.4 21.6 25 D 25.7 23.0 36.8 * 26 D 26.3 24.6 * 23.1 27 D 26.2 26.2 21.2 * 28 D 26.4 * 26.7 23.6 29 D 24.7 b 24.3 32.6 * d 30 D 27.7 36.3 30.8 * * b 31 D 24.1 * 23.1 * d 23.3 32 D 23.5 * 22.8 * 24.2 33 N 24.4 24.1 * 23.1 34 D 24.8 26.2 * 21.9 35 D * 20.0 * 22.1 22.0 36 N * 24.0 * 23.2 * 24.0 37 D 20.9 * 20.7 22.8 a * 38 D 25.0 24.3 * c 24.2 *f Period significantly different (P < 0.05) from random "noise" a. Days 1-8; b, Days 1-9; c. Days 6-8; d, Days 6-9; D, diurnal; N, nocturnal 180 40 A 36 32 "D O 28 H 24 20 150 -| B 130 - £ o co no \ O (A 90 4-1 oca c c 3 o 70 E o E c ro 50 - 5 10 - i 6 60 600 Light intensity (lux) Fig. 44. Variations in the free-running period (A) and in the mean swimming speed (B) with intensity of constant illumination (log scale). Open and closed circles, in panel A denote non-significant and significant periods, respectively. Regression line A is Y = 27.23 - 0.85 log^o x (r = -0.19, P > 0.05) and line B is Y = 33.41 + 18.10 log10 X (r =0.64, P < 0.001, one-tailed). 125 -j J 100 -j ii Mi y * I IN ii j 1 J to j V L I' IJ 4 9 U 19 24 29 34 39 225 200 v ;j C SO0 I Mi J !'i i ii'ii ',; ii II l! '1 Iii I Hi; 1 i/ 4 9 14 I? 34 39 34 3V ii liijiJ < u 0.0 +——i T 4 9 14 19 24 2? 34 39 •5C1 D 100 o.o t^'-i • -•,—.—~i—.—,—, 08 20 08 20 08 20 09 20 08 20 08 20 08 20 08 20 08 20 08 20 08 20 08 4 9 14 1? 24 29 J>4 39 Time of day :h> Fig. 45. Examples of the four types of free-running activity pattern of juvenile pink salmon recorded in LL. Type 1 (A), Type 2 (B), Type 3 (C) and Type 4 (D) patterns are shown, and the periodogram of each pattern is given. The fish were subjected to a LD 12:12 cycle during the first day. Thereafter, open and shaded horizontal bars indicate the subjective day and night, respectively, in LL. 182 during the second half. In 76.9% .of instances., the Type 2 pattern did not have a significant period, when all 10 days of data were considered (Table 15). The lack of periodicity of the Type 2 pattern during the first five days in LL suggests that constant light affected the fish negatively. However, this negative effect apparently weakened, and the fish exhibited periodic activity during the second half of the experiment. Type 3 pattern (Fig. 45C) is a free-running pattern that has a significant periodicity during the first five days in LL, but the rhythm fades out and the pattern becomes arrhythmic during the second 5-day period of the experiment. Contrary to Type 2 patterns, 80% of Type 3 patterns showed a significant periodicity when the 10 days of data were analyzed (Table 15). Type 4 pattern (Fig. 45D) shows arrhythmic, random activity during both halves of the 10-day experiment. Generally, such random activity resulted in a non-significant period when periodogram analysis was applied to the ent ire 10—day data record (Table 15). Table 16 shows the relative proportion of the four types of activity pattern observed in LL. Type 1 and Type 2 patterns predominate. Type 2 and Type 4 patterns combined indicate that 57.9% of fish examined exhibited arrhythmic activity at least during the first five days in LL. This suggests that constant light conditions disturbed, at least temporarily, the "normal" diel activity pattern of these fish. Further, the different types of free-running activity pattern did not occur with equal frequency under the three constant light intensity treatments (Fig. 43C). Type 1 patterns occurred less frequently as constant light intensity (LL) increased from 6 to 600 lx. This means that the 183 Table 16. Percentage of the total number (n = .38) of individual pink salmon exhibiting four different types of free-running activity pattern in constant light condition (LL) for 10 days. Activity pattern Percentage Type 1 28.95 Type 2 34.21 Type 3 13.16 Type 4 23.68 184 proportion of fish exhibiting a completely periodic, free-running rhythm without fade-out in LL declined as light intensity increased. Conversely, the proportion of Type 2 patterns increased with increasing constant light intensity. This indicates that the negative influence of constant light on the diel activity rhythm during the first five days in LL increased as light intensity increased. Type 3 and Type 4 patterns occurred less frequently at LL (600 lx) than at either LL (6 lx) or LL (60 lx). In summary, the majority of free-running periods (T) recorded in LL were between 20 and 28 hours. However, T did not vary significantly with constant light intensity, whereas fish swimming speed correlated positively with constant light intensity. Not all fish were diurnally active in LD. Only half of the fish examined showed a significant T when their swimming activity was recorded in LL for 10 days. Activity patterns of the fish in LL were variable; four different types of pattern were recognized. The higher light intensities appeared to have a negative effect on the expression of significant periodic activity by the fish, at least during the first five days in LL. B. Experiment 2, Catching the free-running rhythm The four fish examined were mainly day-active during the initial five days in LD. Their activity rhythms had significant periodicities.(Table 17). With the exception of fish 4, the diel activity rhythms of the other fish were synchronized with the 24-h LD cycle, as evidenced by their periods approximating 24 hours. When subjected to constant light (LL) for the next five days, fish 1 and 2 displayed 185 Table 17. Period lengths of the swirnming activity rhythms of individual pink salmon recorded for five days in LD 12:12 (600:0.2 lx), followed by five days in LL (600 lx), and finally for another five days in LD 12:12 (600:0.2 lx) at 10°C. All periods are significantly different (P < 0.05) from random "noise", unless indicated otherwise by an asterisk. Fish Period (h) number Days 1-5 (LD) Days 6-10 (LL) Days 11-15 (LD) 1 24.0 25.3 24.0 2 24.0 24.2 29.8 3 24.3 22.0* 24.0 4 27.2 21.8* 23.9 Mean = 24.9 23.3 25.4 SEM = 0.8 0.9 1.5 186 free-running activity rhythms with, significant periods greater than the period of their respective rhythm recorded previously in LD. The other two fish (3 and 4) displayed random activity in LL. Upon returning to LD for another five days, fish 1 and 3 resynchronized their diurnal activity rhythms with the 24-h LD cycle, as evidenced by their significant periods of 24.0 h, whereas, fish 2 and 4 were arrhythmic and did not resynchronize their activity patterns with the LD cycle. The latter two fish did not show either distinct diurnal or nocturnal activity during Days 11-15 in LD. No significant difference in the mean period of the rhythms was found between the three light treatments 2 (Xr = 0.38, P > 0.05, Friedman's two-way ANOVA). Two examples of the activity patterns recorded during the 15-day experiment are given in Figure 46. In conclusion, some of the data suggest that the LD cycle acted as a Zeitgeber of swimming activity in juvenile pink salmon. These data showed that swimming activity rhythms of fish 1 and 3 resynchronized with the LD cycle after free-running in LL for five days. C. Experiment 3, Phase-shifting of the endogenous rhythm Seventy one percent of the fish examined were diurnally active during Days 1-5 in LD (Table 18). All fish exhibited diel rhythms with significant periodicities. With the exception of fish 7, the diel activity rhythms of the other fish were closely synchronized with the 24-h LD cycle. After the LD cycle was phase-delayed by 12 hours, three fish (3, 11 and 12) became arrhythmic, but the other fish retained their • ays 1-3 Days 6-10 Days 11-15 1- 19 14 II If I* It 34 39 '•ft* o* to o« io ai io o* 10 01 10 0* JO 01 10 0* 10 01 01 10 01 10 01 10 01 .10 0* 10 Time of day (h) Fig. 46. Two examples of the activity pattern of pink salmon recorded during the 15-day "Catch the free-running rhythm" experiment. For Days 1-5 and Days 11-15 light and dark horizontal bars indicate the day and night phases of the LD 12:12 cycle, respectively. For Days 6-10 open and shaded bars indicate the subjective day and night, respectively, in LL. Periodograms are given for each of the 5-day sections of the data record. 188 Table 18. Period lengths of swimming activity rhythms in individual pink salmon recorded in LD 12:12 (600:0.2 lx) and 10°C for five days (Days 1-5), and then in a 12-h phase-delayed LD 12:12 (600:0.2 lx) cycle for six and a half subsequent days (Days 6-12) . The predominant phase of the activity rhythm is indicated in parentheses. Fish Period (h) number Days 1-5 Days 6-12 1 24.9 (D) 23.1 (D) 2 25.2 (D) 24.0 (D) 3 24.0 (N) 30.9*(N) 4 23.9 (D) 23.7 (D) 5 24.3 (DN) 27.0 (DN) 6 24.3 (D) 26.1 (D) 7 22.1 (D) 24.0 (D) 8 23.8 (D) 24.1 (D) 9 23.9 (D) 23.5 (D) 10 23.9 (D) 25.7 (D) 11 23.5 (DN) 36.2*(DN) 12 23.9 (N) 25.7*(D) 13 23.7 (D) 24.1 (D) 14 24.2 (D) 24.0 (D) Mean = 24.0 25.9 SEM = 0.2 1.0 *, Period not significantly different (P > ;fitj$) from random "noise" D, diurnal; N, nocturnal; DN, no distinct D or N pattern 189 rhythmic swimming activity when subjected to the new phase-shifted LD cycle (Table 18). About 73% of the fish exhibiting significant periodic activity resynchronized their activity rhythms with the 24-h LD cycle. In general, the fish achieved their previous stable activity patterns and resynchronized their activity rhythms to the phase-shifted LD cycle within one transient cycle. An exception to this generality was fish 10, which achieved resynchronization only after three transient cycles. Two typical examples of the response of the fish's activity rhythm to a 12-h phase delay in the LD cycle are given in Figure 47. The null hypothesis that the mean of all periods recorded in LD during Days 1-5 is equal to the mean of all periods recorded under the phase-shifted LD cycle during Days 6-12 was tested. The hypothesis could not be rejected (T = 22, P > 0.05, Wilcoxon matched-pairs signed-ranks test). Therefore, I conclude that, on the average, pink salmon resynchronized their diel rhythms of swimming activity with the phase-delayed LD cycle within one transient cycle. DISCUSSION The data showed that about 87% of the fish exhibited free-running activity rhythms in LL with periods ranging from 20 to 28 hours. However, only half of the fish examined showed significant free-running periods (T). Tau did not vary significantly with constant light intensity, whereas mean fish swimming speed correlated positively Days 1-5 Days 6-12 08 20 00 20 08 20 08 20 08 20 08 20 08 20 08 20 08 20 08 20. 08 20 08 20 Time of day (h) Fig. 47. Two typical examples of the diel swimming activity pattern of pink salmon recorded for five days prior to a phase shift in the LD 12:12 cycle, and for another six days after the phase shift. Periodograms are given for each of these two sections of the data record. Light and dark horizontal bars are as in Fig. 10. 191 with, constant light intensity. Activity patterns of the fish in LL were variable, and four different types of pattern were identified. Half of the fish, tested resynchronized their.activity rhythms to the LD cycle after free-running in LL for five days. After a 12-h phase delay in the LD cycle, 73% of the fish that subsequently exhibited significant periodic activity resynchronized their activity rhythms with the phase-delayed LD cycle. This resynchronization was achieved generally within one transient cycle after the phase shift. Only half of the fish, examined in LL displayed periodicities in swimming activity significant from random "noise". These periodicities support the hypothesis that the swimming activity rhythm of pink salmon is generated by an endogenous circadian system. Since juvenile Pacific salmon (genus Oncorhynchus) do not show endogenous retinomotor rhythms in LL or DD (Ali 1959), the free-running activity rhythms of pink salmon in LL could not have resulted from a circadian rhythm in retinal sensitivity to light. Evidence for endogenous circadian rhythms in fishes is relatively weak (see Appendix Tables 15 and 16). Circadian swimming activity rhythms in fish are not always readily evident because they display a relatively rapid extinction in constant conditions. Consequently, Schwassmann (1971a) concluded that the swimming activity rhythm of fish is not a good indicator of the endogenous oscillator or clock. Unequivocal circadian swimming activity rhythms, persisting for at least 15 days in constant conditions, have been reported only for the swell shark, Cephaloscyllium veritriosum 192 (Nelson and.Johnson 1970), the labrid fish, Labroidesquadtolineatus (Casimir 1971), and the lake chub (Kavaliers 1978b). The most convincing evidence for a circadian swimming activity rhythm in salmonids was obtained for juvenile sockeye salmon by Byrne (1968). About 49 and 86% of his fish exhibited persistent, free-running rhythms for about three to six days before fading out in DD or LL, respectively, Apparent free-running activity rhythms occur in brown trout (Miiller 1969b) and brook trout (Eriksson 1972) during the Arctic summer, when the sun is continually above the horizon and the Zeitgeber strength of the illumination cycle is weakened consequently. Evidence for the existence of an endogenous clock in fishes is also provided from orientation behavior. Some fish species, including salmonids, have been shown experimentally to use sun-compass orientation as one mechanism of direction finding (Hasler et al. 1958, Braemer 1960, Hasler and Schwassmann 1960, Groot 1965, Hasler 1966, 1971, Schwassmann 1971a, Waterman 1972). Orientation in a compass direction using the sun as a reference point requires a biological clock to compensate for the daily (azimuthal) and seasonal (latitudinal) changes in the sun's position in the sky (Braemer 1960, Hasler 1966, 1971, Waterman 1972). Johnson and Groot (1963) and Groot (1965) reported that sockeye salmon smolts, held under laboratory conditions with a view of the sky, exhibit seasonal changes in their preferred pointing direction as the migration season progresses. These changes in orientation corresponded to the directions of the appropriate migration routes through the lake system that wild smolts must follow to reach the ocean. 193 The data of Johnson and Groot suggest the existence.of an innate endogenous clock that measures time of day and of season. Healey (1967) reported data on the orientation of pink salmon fry in the laboratory that support circumstantially the hypothesis of sun-compass orientation, and thus of an endogenous clock. The other half of the fish examined in Experiment 1 displayed random activity during the 10 days in LL. Arrhythmicity and extinction of a rhythm in constant conditions does not necessarily provide evidence for or against an endogenous circadian system (Aschoff 1960, Bunning 1973). Such an event could result from 1) damping of underlying oscillator(s), 2) desynchronization of a population of oscillators that share the control of the biological rhythm, 3) uncoupling of the biological rhythm from the driving circadian system, and (or) 4) stochastic processes that may be disturbing to the organism under constant conditions (Aschoff 1960, Bumning 1973, Gwinner 1975). The relative frequencies of the four types of activity pattern. observed under the three constant light intensity treatments suggest that the higher light intensities (60 and 600 lx) had a relatively strong inhibitory effect on the expression of significant circadian activity by the fish, at least during the first five days in LL. Type 2 pattern was the most common of the four types of activity pattern observed in LL. A similar type of activity pattern, wherein damping of the rhythm occurred during the initial two or three days in LL but later re-established itself, has been indicated for bluefish (011a and Studholme 1972) and Atlantic salmon (Ali 1964). 194 The period (r) of the activity pattern recorded in LL was not a significant monotonic function of constant light intensity, as predicted by the Circadian Rule (Aschoff I960). Tau declined only slightly with increasing constant light intensity. Since the fish examined in Experiment 1 were not all diurnally active in LD, this weak negative relationship may have resulted from the opposing tendencies of diurnal and nocturnal fish to shorten and lengthen, respectively, their T as constant light intensity increased (Table 15). Such tendencies have been documented for many organisms (Aschoff 1960, 1963, Hoffmann 1965, Biinning 1973, Gwinner 1975). The level of activity or mean swimming speed of the fish increased significantly with increasing constant light intensity. The Circadian Rule predicts this relationship for diurnal animals. However, the positive correlation between T and the corresponding activity level for individual fish, also predicted by the Circadian Rule, was not observed. Therefore, the above positive relationship between activity level and light intensity may simply correspond to a positive photokinetic response of the fish, rather than to the light-sensitive level of the Aschoff-Wever endogenous oscillator (Aschoff 1969, Gwinner 1975). The high inter-individual variability in the form of the activity patterns in LL suggests that a greater number of fish need to be tested for a difference in mean period length to appear between light treatments. Perhaps a significant relationship in the predicted direction between x and light intensity would emerge if each fish was tested under each light intensity. Hence, it is concluded tentatively that the data on T do not support the Circadian Rule, and that the demonstration of a circadian 195 activity rhythm in pink salmon under LL is difficult. This difficulty relates to the high inter-individual variability in the form of the observed activity rhythms. Therefore, the existence of an endogenous, circadian activity rhythm in pink salmon is suggested, but remains uncertain. The Circadian Rule has been tested in a few cases for fishes. Omly four studies present sufficient data to test the predictions of this rule. The results are conflicting. In accordance with the Circadian Rule, Schwassmann (1971b) observed that T of the free-running rhythm of electric organ discharge exhibited by nocturnal gymnotid fishes (genus Gymnorhamphichthys and Hypopomus) increased with increasing constant light intensity. Conversely, diurnal minnows (Leucaspius delineatus) shortened the x of their free-running activity rhythm with increasing intensity of constant illumination (Siegmund and Wolff 1973). Diurnally active sockeye salmon lengthened the x of their free-running activity rhythm with increasing constant light intensity (Byrne 1968); this is opposite to the predicted trend. Lake chub do not strictly obey the Circadian Rule (Kavaliers 1978a). The x of diurnal chub lengthened as -3 0-2 total spectral energy increased from 10 to about 10 uw cm , but x shortened with further increases in total spectral energy. On the one hand, many other exceptions to the Circadian Rule have been found recently in arthropods, birds, and mammals (Hoffmann 1965, Lohmann 1967, Gwinner 1975, Erkert and Kracht 1978). On the other hand, specific predictions of the Rule have been verified in experiments with a bird species and man (Aschoff et al. 1971). 196 The Aschoff-Wever one-oscillator model was formulated originally to explain the Circadian Rule. This, model is based mainly on data from perch-hopping activity of birds and running-wheel activity of mammals. Kavaliers (1978a, b) pointed out that the relationships of perch-hopping and running-wheel.activities to natural activity are unclear, and that these types of motor activity are not directly equivalent to swimming activity of fish, which is an integral part of many behaviors. Hence, the Circadian Rule may have limited application to fishes. Gwinner (1975) also expressed doubts on the general validity of the Circadian Rule. He stated that the Aschoff-Wever model assumes that locomotor activity is controlled by a single oscillator. This assumption is contradicted by many empirical data suggesting that more than one oscillator controls locomotor activity in animals (Gwinner 1975, Pittendrigh and Daan 1976), including fishes (Eriksson 1973, 1975, Kavaliers 1978a). If a free-running rhythm in constant conditions assumes the period of the environmental cycle to which it is exposed, then this cycle is considered an effective Zeitgeber entraining the biological rhythm (Gwinner 1975). Two of four fish in Experiment 2 re-established their activity rhythm with a significant period of 24.0 h after being desynchronized previously for five days in LL. This finding only provides weak evidence for the role of the LD cycle as a Zeitgeber in the entrainment of swimming activity. If the environmental cycle is an effective Zeitgeber, then a shift of its phase should result in a corresponding phase shift of the endogenous rhythm and in the period of the biological rhythm being equal to that of the Zeitgeber 197 (Aschoff i960, Gwinner 1975). Evidence for entrainment and for an endogenous rhythm is provided when the steady-state phase position and the resynchronized period of the biological rhythm are attained slowly through a series of transient cycles. Conversely, if the phase shift occurs rapidly through one or two transients, then masking rather than entrainment of the biological rhythm by the environmental cycle is suggested (Aschoff 1960, 1963, 1967, Hoffmann 1969, Bunning 1973, Gwinner 1975). In response to a 12-h phase shift of the LD cycle, the majority of pink salmon in Experiment 3 resynchronized their swimming activity rhythms with periods not significantly different, on the average, from the periods of rhythms displayed in LD before the phase shift. This meets one of the criteria of entrainment. However, the phase shift of the swimming activity rhythm, which occurred through only one transient cycle in the majority of instances, suggests masking of the activity rhythm by the LD cycle. In many organisms, environmental cycles may exert direct influence (masking) on the biological rhythm rather than entraining an underlying endogenous oscillation (Aschoff 1960, Hoffmann 1969, Gwinner 1975). Far this reason Hoffmann (1969) emphasized that, in all three types of experiment designed to test a Zeitgeber (Aschoff 1960), it is often necessary to establish the free-running period (T) of the endogenous rhythm before and after exposing it to the environmental cycle to be tested. This was not done in Experiments 2 and 3, and consequently the evidence for the effectiveness of the LD cycle as a Zeitgeber entraining swimming activity is further weakened. 198 . The majority of studies testing. the effectiveness of the LD cycle as a Zeitgeber of fish swimming activity indicate exogenous comtrol (masking) of the activity rhythm, through the direct action of the LD cycle on activity. This observation may relate to the equivocal evidence for circadian activity rhythms in fishes. On the one hand, phase shifts of the LD cycle resulted in a corresponding phase shift of the activity rhythm through only one transient cycle in Atlantic salmon (Dill 1970, Varanelli and McCleave 1974), rainbow trout (Dill 1970), American shad, Alosa sapidissima (Katz 1978), and American eel, Anguilla rostrata (Bohun and Winn 1966). Bluefish required only one or two transient cycles to re-establish their previous phase relationships with, the LD cycle after a 7-h phase delay of this cycle (011a and Studholme 1972). These authors interpreted this observation as.evidence for an endogenous rhythm controlling activity. On the other hand, resynchronization of swimming activity rhythms to a phase-shifted LD cycle required about three transient cycles in the American eel (Edel 1976) and six transient cycles in brown trout (Eriksson 1975). These latter two studies suggest entrainment of an endogenous circadian oscillator by the LD cycle. The differing results of Bohun and Winn (1966) and Edel.(1976), working on the American eel, might reflect differences in the physiological state of the fish and (or) differences in the LD intensity ratio between the two studies. It is known that the strength of the Zeitgeber (e.g., LD ratio of illumination intensity) can affect the speed of re-entrainment (Aschoff I960, Wever 1966, Biinning 1973, Aschoff et al. 1975). 199 •In.summary, some of my data suggest the existence of a circadian activity rhythm in pink salmon, but this remains uncertain owing to the difficulty in demonstrating such a rhythm in this species under constant light conditions. The LD cycle did not appear to entrain a circadian activity rhythm, but rather controlled directly the activity rhythm (i.e., masking). Hence, it is concluded that the swimming activity rhythm of juvenile pink salmon does not possess a strong endogenous component, and that this rhythm seems to be synchronized with the LD cycle through the direct masking action of this cycle, rather than through entrainment of an endogenous circadian system. ZOO- CHAPTER VIII. GENERAL DISCUSSION AND CONCLUSIONS Pink salmon undergo several major habitat changes during their invariable 2-year life cycle. These habitat changes result from a series of migrations that are accurately timed on a seasonal basis. The seasonal timing of these migrations may depend in part on an annual biological clock. Much evidence suggest the existence of an endogenous, circannual clock in some migratory passerine birds (Berthold 1974, Klein 1974, Gwinner 1975, 1977); the circannual clock has an autonomous period that deviates slightly from one year. This clock controls, independently of environmental time cues, the seasonal timing and amount of migratory activity and also seasonally-appropriate directional changes of these bird species. In nature the circannual clock is probably synchronized with environmental factors, especially photoperiod (Gwinner 1973, 1975, 1977). Further, annual rhythms or seasonally-timed events may result from the interactions between daily rhythms and annual changes in environmental factors (Gwinner 1973, Menaker 1974). Hence, knowledge of daily behavioral rhythms in pink salmon might contribute to a better understanding of the seasonal timing of its migrations. Since little was known previously about daily rhythms in pink salmon, the major objective of this study was to investigate, in a seasonal context, daily behavioral rhythms in juvenile pink salmon, and their timing mechanisms. The major findings of this study are as follows. Juvenile pink salmon were rhythmic in their behavior throughout the experimental period corresponding to the time of emergence of wild fry until the onset 201 of migration of maturing fish from the open ocean toward the coast. They generally exhibited, under laboratory conditions, nocturnal rhythms of gravel emergence activity and of vertical distribution and diurnal rhythms of swimming activity and of aggressive behavior. Wild pink salmon fry had diurnal rhythms of feeding activity. Diel temperature cycles did not appear to synchronize, these behavioral rhythms. Conversely, the daily LD cycle was the major environmental cycle synchronizing the behavior of pink salmon. However, different parts of the LD cycle synchronized different types of behavior. Based on the diel phasing of behavior, the D-L transition synchronized swimming activity, aggression and feeding activity, whereas the L-D transition synchronized the rhythms of emergence activity and of vertical distribution. The LD cycle synchronized the diel rhythm of swimming activity through a direct masking action on activity. Daily patterns of some of the above behaviors varied annually in response to simulated seasonal changes in the LD cycle and temperature. Annual changes in these rhythms did not follow the seasonal course predicted by current theory of entrainment of biological rhythms to seasonal changes in the daily illumination cycle. Many of the diel behavioral rhythms recorded in the laboratory were ecologically appropriate, and corroborated the available data on the behavioral ecology of pink salmon. Some of the ontogenetic and annual changes in these behavioral rhythms roughly coincided with changes in natural habitats experienced by wild pink salmon during their migrations. The existence of an endogenous, circadian activity rhythm in pink salmon was suggested by some of the data, but remains uncertain. 202 Diel patterns of the various behaviors of juvenile pink salmon described in the current study likely reflect visual adaptation of the fish to the daily illumination cycle (Kavanau and Ramos 1975) . Pink salmon has a duplex retina that is adapted to both scotopic and photopic visions (Ali 1959). Behavioral rhythms observed in nature are determined, not only by the properties of the organism's visual system, but also by prey availability and predator pressure (Kavanau and Ramos 1975). Therefore, the behavioral rhythms determined in the laboratory may differ from the ones exhibited by wild fish. Consequently, conclusions about the adaptive significance of diel behavioral rhythms in the laboratory can only be speculative. Emergence from the simulated gravel redd was mainly nocturnal below 13°C. This pattern corresponds well to the nocturnal seaward migration of wild pink salmon fry (e.g., Neave 1955, Roppel 1956, Ali and Hoar 1959, Vernon 1966, Bakshtanskii 1970, Smirnev 1975). On a diel basis emergence timing was synchronized by the onset of night, but was modified by temperature. I regard the nocturnal emergence and subsequent nocturnal seaward migration of pink salmon fry as anti-predator adaptations because their susceptibility to predation may decline at low incident light intensities. Generally, 90% of sibling fry emerged from the same redd within two weeks. Intra-redd synchrony of emergence (duration of emergence period) varied inversely with temperature. The temporal synchronization of emergence between and within redds in a population could affect the 203 subsequent mortality rate, due to predation, of newly-emerged fry migrating seaward. The greater the synchrony of fry emergence, the greater the population of fry that will subsequently migrate seaward on a particular night. Such a relationship could potentially reduce the relative mortality rate of the fry, through the mechanism of depensatory mortality (Neave 1953, Peterman and Gatto 1978). Further, the seasonal timing of emergence, controlled in great part by temperature, determines the seasonal timing of fry entry into estuaries. This might influence their marine growth and survival (Koski 1975, Belford 1977). Pink salmon fry exhibited mainly nocturnal patterns of swimming activity and of vertical distribution during the first week after emergence. This indicates that the fry still responded negatively to light after emergence. However, a shift from a nocturnal to a diurnal rhythm of swimming activity occurred 7 to 13 days after emergence. Fry began to school strongly during the day a few days before this shift. Coincident with this shift in activity rhythm was an increasing tendency of the fry to swim near the water surface during the day. This suggests a weakening of the fry's negative phototactic response during this period. Thereafter (until Day 37), the fish usually displayed diurnal rhythms of swimming activity and nocturnal rhythms of vertical distribution. This period corresponds to their juvenile coastal phase. Collectively, my findings and those of Hoar and co-workers (Hoar 1951, 1956, 1958, 1976, Hoar et al. 1957) indicate that pink fry are nocturnally active, exhibit negative phototaxis, and rise toward the water surface at night during the period of seaward, riverine migration. At the time that the fry are 204 in estuaries and adjacent coastal waters, they school by day, progressively become day-active, and their negative phototactic response weakens somewhat. However, the fish retain their negative phototaxis and their tendency to rise toward the water surface at night and to descend by day. Dill (1970) also reported an ontogenetic shift from nocturnal to diurnal activity in rainbow trout and Atlantic salmon. This shift coincided with the period of wild fry emergence from gravel into the water column of the stream, a.juvenile feeding habitat. Byrne. (1971) observed a similar shift in the activity rhythm of sockeye salmon fry shortly after emergence when wild fry normally migrate to a nursery lake. Pink salmon fry in Departure Bay and Hammond Bay, British Columbia fed on both planktonic and epibenthic prey during daylight hours with little or no feeding at night. Diurnal feeding patterns have also been reported for juvenile pink salmon in other coastal marine localities (Appendix Fig. 4). However, nothing appears to be known about their feeding rhythms in the open ocean. These diurnal feeding patterns correspond well with the diurnal rhythms of swimming activity exhibited by fry in the current study. Therefore, the ontogenetic phase shift in the diel rhythm of fry swimming activity is adaptive for schooling and feeding on small invertebrate prey in marine waters during the day. The above feeding patterns revealed variability in the diel timing of feeding and in the feeding rates of the fish in coastal waters. This variability may result in great part from temporal and spatial 2Q5 variations in prey availability. Feeding appeared to be continuous throughout daylight hours, as expected of energy maximizers CSchoener 1969, 1971, Rapport and.Turner 1975), but feeding rates were not constant. Variability and flexibility in daily patterns of feeding activity is common among salmonid fishes (e.g., Hoar 1942, Northcote and Lorz 1966, Jenkins 1969, Mundie 1971, Eggers 1975, Bisson 1978). Therefore, salmonids may be characterized as opportunistic and generalized predators. Laboratory findings indicated that both gastrointestinal and metabolic (systemic) factors affect fish feeding rate in manners predicted by current control theory models of hunger (Rozin 1964, de Ruiter 1967, Colgan 1973, Curio 1976, Toates and Archer 1978). Under idealized laboratory conditions, juvenile pink salmon fed continuously on live prey. During a 12-h session their feeding rate was initially high, but declined until their stomachs were full. Thereafter, the fish kept their stomachs full by feeding continuously at a rate that approximated the rate of evacuation of prey from the stomach. I conclude that juvenile pink salmon feed at a relatively low hunger threshold; that is, less than 15% of their stomachs needs to be evacuated for spontaneous feeding to occur or resume. Conversely, other salmonid species (e.g., rainbow trout, brown trout, sockeye salmon) are discontinuous feeders, and apparently feed at higher hunger thresholds (Chaston 1969, Narver 1970, Brett1971, Adron et al. 1973, Elliott 1973, 1975b, McDonald 1973, Landless 1976, Doble and Eggers 1978, Grove et al. 1978). Hence, the observed high growth rates of juvenile pink salmon in nature (LeBrasseur and Parker 1964, Parker and 2Q6 LeBrasseur 1974, Phillips and Barraclough 1978) are explained in part by their ability to feed continuously at a rate balancing gastric evacuation rate during available feeding time. However, Toates and Archer (1978) argued that in nature prey availability and other ecological factors dominate over metabolic factors in the short-term regulation of feeding behavior. Little is known about'diel behavioral rhythms in pink salmon in the open ocean. This makes the interpretation of laboratory data difficult. I suggested earlier that the daily illumination cycle, through its influence on the daily rhythm of swimming speed, might be an important environmental factor determining the seasonal timing of the oceanic migration of pink salmon. This hypothesis is supported by some of my data. Mean swimming speed of the fish, the degree of diurnalism of their swimming activity rhythm, and the timing of the daily peak of their behavioral rhythms were affected by daylength. Therefore, photoperiod may be one proximate factor that pink salmon use to time their oceanic migration on a seasonal basis, assuming that such seasonal timing results in part from adjustments in swimming speed and in timing of daily rhythms. Annual changes occurred in the diel patterns of the three types of behavior examined in response to seasonal photoperiodic and temperature changes. These diel rhythms remained synchronized with the 24-h LD cycle throughout most of the year with the exception of some short periods of desynchronization. Thus, the LD cycle was the major synchronizer of fish behavior in the current study, as for many rhythmic functions in a wide spectrum of organisms (Aschoff 1960, 1963, Menaker 1969, 207 Buhning 1973, Daan and Aschoff 1975, Enright 1975, Gwinner 1975). Temperature cycles did not have a significant synchronizing effect on rhythmic behavior of the fish.. Diel temperature variations in the open ocean are likely small, and consequently would not present strong, dependable time cues for synchronizing diel behavioral rhythms in the presence of the more predictable daily illumination cycle. Juvenile fish exhibited diurnal rhythms of swimming activity and of aggression and nocturnal rhythms of vertical distribution under simulated photoperiodic and temperature conditions of spring and early summer of their first year. This period corresponds to their inshore, juvenile residency in nature. Coincident with their offshore migration in summer, the fish occasionally exhibited weak diel rhythmicity or arrhythmicity in their swimming activity and vertical distribution in the laboratory. Such weak rhythms may relate to a natural tendency of wild fish to use tidal ebb currents to facilitate their offshore migration, as suggested by Martin 0966) and Healey (1967)• Adult pink salmon (Stasko et al. 1973) and adults and juveniles of other salmonid species (Groot et al. 1975, Stasko 1975, Stasko and Rommel 1977, Fried et al. 1978, LaBar et al. 1978, McCleave 1978, Tytler et al. 1978) lack diel rhythmicity in their swimming activity while migrating through estuarine and coastal marine waters. They apparently use tidal currents to facilitate their horizontal movements. During the fall and winter, when pink salmon are widely scattered in the North Pacific ocean (Royce et al- 1968), laboratory fish exhibited weak diurnal rhythms of swimming activity and distinct nocturnal rhythms 208 of vertical distribution synchronizedwith, the LD cycle. It has been postulated that Pacific salmon must migrate.continuously to account for their rate of oceanic travel (Royce et al. 1968, Stasko 1971). My laboratory data indicate that pink salmon swim continuously, with swimming speeds being slightly greater during the day than at night, during the period when they would be migrating in the open ocean. Nocturnal rhythms of vertical distribution observed in the laboratory correspond to the known tendency of pink salmon to migrate closer to the water surface at night than during the day, when they migrate at greater depths in the ocean (Manzer 1964, Machidori 1966). The adaptive significance of vertical migration is generally ambiguous, and many ideas have been suggested (McLaren 1963, Woodhead 1965, 1966, Blaxter 1970, Roe 1974, Swift 1976, Eggers 1975). The significance of diurnal rhythms of aggression exhibited by laboratory fish is uncertain because of the lack of documentation of such behavior in pink salmon in nature. Hence, diel patterns of swimming activity and of vertical distribution of juvenile pink salmon in the laboratory corroborate available data on their behavioral ecology in marine waters. Seasonal variations in activity time (a) and mean phase-angle difference QVma x ) of the diel behavioral rhythmJ s of the fish did not obey the predictions of the Aschoff-Wever one-oscillator model. This may have resulted in part from differences in rhythm parameters (a, ¥) recorded in the current study and those recorded for birds and mammals, for which this model was formulated. Recent studies have shown that data from many species, particularly birds and mammals, do not support predictions of the Aschoff-Wever model concerning the seasonal course 209 of a and ¥ .„ (e.g., Aschoff et al. 1970, Erkinaro 1970, 1972, Muller-Haeckel 1974, Voute et al. 1974, Daan and Aschoff 1975, Kenagy 1976, 1978, Kavaliers 1978a). . This suggests that either 1) too many factors not accomodated in the model are involved and thus obscure general tendencies or 2) some of the basic assumptions of the model are wrong (Daan and Aschoff 1975, Gwinner 1975). Some of the data reported herein suggest the existence of a circadian activity rhythm, and thus a daily biological clock, in pink salmon. However, this remains uncertain, owing to the difficulty of demonstrating such a rhythm in this species under constant light conditions. Further, the free-running period (x) of pink salmon was not a monotonic function of constant light intensity, as predicted by the Circadian Rule (Aschoff 1960). Other exceptions to this Rule, which is based on data from birds and mammals, have recently been found in fishes (Byrne 1968, Kavaliers 1978a) and in arthropods, birds and mammals (Hoffmann 1965, Lohmann 1967, Gwinner 1975, Erkert and Kracht 1978). Gwinner (1975) recently questioned the general validity of the Circadian Rule. The evidence for an endogenous biological clock in fishes in general is largely equivocal (Schwassmann 1971a, Appendix Tables 15 and 16). The strongest evidence has been obtained for specialized functions such as melanophore movement (Kavaliers and Abbott 1977), electric organ discharge (Schwassmann 1971b), mucus secretion (Casimir 1971), feeding activity and reaction time to light pulses (Eriksson 1978a, Eriksson and Veen 1979). In general circadian activity rhythms in fishes are not 210 readily evident because they display a relatively rapid extinction in constant laboratory conditions. Consequently, Schwassmann (1971a) concluded that the swimming activity rhythm of fish is not a good indicator of the circadian clock. Perhaps other rhythmic functions of pink salmon would have provided stronger evidence for the existence of such a clock. The fish examined in the current study exhibited considerable inter-individual variability in the type of activity pattern displayed in constant conditions. This is•in agreement with the general flexibility of swimming activity rhythms in fishes compared to higher vertebrates (Eriksson 1978a, Miiller 1978b), and might also reflect the lack of social stimulation experienced by isolated individual pink salmon, which are schooling fish in nature. Similarly, many other marine organisms display considerable variability in their rhythms of locomotor activity, typically show a lack of precision in the timing of these rhythms, and their rhythms are characterized by high background noise and multiple frequencies (Menaker 1976, Rawson and DeCoursey 1976). Menaker (1976) emphasized that the variability in the rhythms of marine organisms may be a function of the heterogeneity and instability of the marine environment, and of the choice of locomotor activity as an indicator process. . The ecological need or adaptation of an endogenous timing mechanism that has been demonstrated in the laboratory is seldom evident (Enright 1970). Migrating pink salmon require a daily, endogenous 211 timing mechanism if they use sun-compass orientation as one mechanism of direction finding in the open ocean (Braemer 1960, Hasler 1966, 1971). There is no experimental evidence for the operation of such a mechanism in pink salmon, although it is suggested by the observations of Healey (1967). Much of the oceanic migration of pink salmon takes place during a period when the sea is stormy and the sky is frequently overcast (Royce et al. 1968, Favorite et al. 1976). Such climatic conditions would limit the use of sun-compass orientation. However, a clock could function in maintaining synchrony among various physiological functions within the fish, and in time memory in the context of seasonal timing of migrations under these climatic conditions. Based on my laboratory data, I conclude tentatively that the swimming activity rhythm of pink salmon does not possess a strong endogenous component. Further, this rhythm seems to be synchronized by the LD cycle through the direct masking action of this cycle on activity, rather than through entrainment of an endogenous circadian system. The existence of a daily biological clock in pink salmon is suggested, but remains uncertain. The variability observed in the diel rhythms of swimming activity, feeding activity, and vertical distribution of fish in this study might be adaptive to inhabiting an unstable marine environment, and may permit the high growth rates observed in wild juvenile pink salmon. Hence, the existence of a strong endogenous component to their behavioral rhythms might not give the fish any advantage in the exploitation of marine resources. Flexibility in the daily rhythm of swimming activity is common to many fish species inhabiting unstable 212 Arctic and subarctic freshwater habitats (Eriksson 1978a, Miiller 1978b). Larkin (1956) hypothesized that freshwater fishes in general evolved flexible behaviors, in response to environmental heterogeneity and instability, which permit opportunistic exploitation of resources. It is commonly accepted that generalists are favored in fluctuating environments, like estuaries and oceans at temperate latitudes, while specialists are favored in stable ones (Schoener 1969). Although this laboratory study generated much new information on daily behavioral rhythms in pink salmon and speculations on their adaptive significance, it did not improve greatly our current poor understanding of the seasonal timing of pink salmon migrations. Perhaps only the finding that swimming speed, the degree of diurnalism of the swimming activity rhythm, and the timing of the daily peak of the behavioral rhythms are affected by daylength is significant in that regard. Since the swimming activity rhythm of pink salmon does not possess a strong endogenous component, it is doubtful that the seasonal timing of its migrations results from interactions between a circadian clock(s) and seasonal changes in environmental factors, as suggested for some higher vertebrates (Gwinner 1973, Menaker 1974). Because of methodological difficulties, Gwinner (1977) recently expressed doubts that laboratory studies of migratory restlessness activity in caged birds would ever allow rigorous testing of the endogenous timing hypothesis of bird migration. This hypothesis states that an endogenous circannual clock within the bird times the onset and end of its seasonal migration,- and also determines distance travelled 213 and directional changes. Gwinner (1977) concluded that field experiments on banded birds may be a more promising approach to testing specific predictions of this hypothesis. Similarly, the role of environmental factors in the daily and seasonal timing of fish migrations can be studied, perhaps more fruitfully, in the field using current biotelemetry techniques (Mitson 1978). However, drawbacks of this approach include high operating expenses and the uncertainty in assessing which environmental factors affect the behavior of wild fish. 214 REFERENCES Adler, K. K. 1969. Environmental control of locomotor activity in a salamander (Plethodon glutinosus). Diss. Abstr., 30 (1): 882-B. Adron, J. W., P. T. Grant, and C. B. Cowey. 1973. A system for the learning capacity of rainbow trout and its application to the study of food preferences and behaviour. J. Fish Biol., 5: 625-636. Ali, M. A. 1959. The ocular structure, retinomotor and photobehavioral responses of juvenile Pacific salmon. Can. J. Zool., 37: 965-996. . 1961. Histophysiological studies on the juvenile Atlantic salmon (Salmo salar) retina. II. Responses to light intensities, wavelengths, temperatures, and continuous light or dark. Can. J. Zool., 39: 511-526. . 1964. Diurnal rhythm in the rates of oxygen consumption, locomotor and feeding activity of yearling Atlantic salmon (Salmo salar) under various light conditions. Proc. Indian Acad. Sci. , 60 (Sect. B): 249-263. , and W. S. Hoar. 1959. Retinal responses of pink salmon associated with its downstream migration. Nature, 184: 106- 107. Alderdice, D. F., and F. P. J. Velsen. 1978. Relation between temperature and incubation time for eggs of chinook salmon (Oncorhynchus tshawytscha). J. Fish. Res. Board Can., 35: 69-75. Andreasson, S. 1969. Locomotor activity patterns of Cottus poecilopus Heckel and C. gobio L. (Pisces). Oikos, 20: 78-94. , and K. Miiller. 1969. Persistence of the diel activity phase shift in Cottus poecilopus Heckel (Pisces) after transfer from 66° to 55° northern latitude. Oikos, 20: 171-174. Andrew, F. J., and G. H. Geen. 1960. Sockeye and pink salmon production in relation to proposed dams in the Fraser River system. Intl. Pacific Salmon Fish. Comm. Bull., 11: 259 p. Andrievskaya, L. D. 1968. Feeding of Pacific salmon fry in the sea. Proc. Pacific Res. Inst. Mar. Fish. Oceanogr., 64: 73-80. (Fish. Res. Board Cari. Transl. Ser. No. 1423, 1970). . 1970. Feeding of Pacific salmon juveniles in the Sea of Okhotsk. Proc. Pacific Sci. Res. Inst. Mar. Fish. Oceanogr., 78: 105-115. (Fish. Res. Board Can. Transl. Ser. No. 2441). 215 Arey, L. B., and G. H. Mundt. 1941. A persistent diurnal rhythm in visual cones. Anat. Rec. (Suppl.), 79: 5-11. Aschoff, J. 1960. Exogenous and endogenous components in circadian rhythms. Cold. Spring Harbor Symp. Quant. Biol., 25: 11-28. . 1963. Comparative physiology: Diurnal rhythms. Annu. Rev. Physiol., 25: 581-600. . 1965. Circadian clocks. Proc. Feldafing Summer School, North-Holland Publ. Co., Amsterdam. • . 1967. Circadian rhythms in birds. Proc. 14th Intl. Ornithol. Congr., Oxford, pp. 81-105. " • . 1969. Phasenlage der Tagesperiodik in Abhangigkeit von Jahreszeit und Breitengrad. Oecologia (Berl.), 3: 125-165. , U. Gerecke, A. Kureck, H. Pohl, P. Rieger, U. von St. Paul, and R. Wever. 1971. Interdependent parameters of circadian activity rhythms in birds and man. In: Biochronometry. Edited by M. Menaker. Proc. Symposium, Friday Harbor, Washington. Natl. Acad. Sci., Washington, D.C. pp. 3-29. , E. Gwinner, A. Kureck, and K. Miiller. 1970. Diel rhythms of chaffinches Fringilla coelebs L., tree shrews Tupaia glis L. and hamsters Mesocricetus auratus L. as a function of season at the Arctic Circle. Oikos, Suppl. 13: 91-100. , K. Hoffmann, H. Pohl, and R. Wever. 1975. Re-entrainment of circadian rhythms after phase shifts of the Zeitgeber. Chronobiologia, 2: 23-78. Baggerman, B. 1972. Photoperiodic responses in the stickleback and their control by a daily rhythm of photosensitivity. Gen. Comp. Endocrinol., Suppl. 3: 466-476. Bailey, J. E., J. J. Pella, and S. G. Taylor. 1976. Production of fry and adults of the 1972 brood of pink salmon, Oncorhynchus gorbuscha, from gravel incubators and natural spawning at Auke Creek, Alaska. U.S. Fish Wildl. Serv. Fish. Bull., 74: 961-971. , S. D. Rice, J. J. Pella, and S. G. Taylor. 1979. Effects of density of pink salmon (Oncorhynchus gorbuscha) eggs on water chemistry, fry characteristics, and fry survival in gravel incubators. Unpublished manuscript. , B. L. Wing, and C. R. Mattson. 1975. Zooplankton abundance and feeding habits of pink salmon, Oncorhynchus gorbuscha, and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area. U.S. Fish Wildl. Serv. Fish. Bull., 73: 846-861. 216 Bakshtanskii, E. L. 1970. Downstream migrations of pink and red salmon . and causes of their delay in streams of the Kola Peninsula. In: "Fish Culture in Nature Waters", Tr. Vses. Nauchno-Issled Inst. Morsk. Rybn. Khoz. Okeanogr., 74: 129-143. (Fish. Res. Board Can. Transl. Ser. No. 1765, 1971). Bams, R. A. 1967. A review of the literature on the effects of changes in temperature regime on developing sockeye salmon eggs and alevins. Fish. Res. Board Can. Manuscript Rept. No. 949: 14-22. : . 1969. Adaptations of sockeye salmon associated with incubation in stream gravels. In: Symposium on salmon and trout in streams. H. R. MacMillan Lectures in Fisheries. Edited by T. G. Northcote. University of British Columbia, Vancouver, pp. 71-87. , and K. S. Simpson. 1977. Substrate incubators workshop - 1976. Report on current state-of-the-art. Can. Fish. Mar. Serv. Res; Dev. Tech. Rep. 689: 67 -p. Beamish, F. W. H. 1972. Ration size and digestion in largemouth bass, Micropterus salmoides Lacepede. Can. J. Zool., 50: 153-164. Beatty, D. D. 1966. A study of the succession of visual pigments in Pacific salmon (Oncorhynchus). Can. J. Zool., 44: 429-455. Beitinger, T. L. 1975. Diel activity rhythms and thermoregulatory behavior of bluegill in response to unnatural photoperiods. Biol. Bull., 149: 96-108. Belford, D. 1977. Simulation of coho smolt predation on pink and chum fry. The importance of relative size and growth rate. M.Sc. Thesis, University of British Columbia, Vancouver, B.C. Berthold, P. 1974. Circannual rhythms in birds with different migratory habits. In: Circannual clocks. Annual biological rhythms. Edited by E. T. Pengelley. Academic Press, Inc., New York, pp. 55-94. Bertmar, G., und K. Miiller. 1970. Die Tagesperiodik geblendeter und geruchseliminierter Bachsaiblinge Salvelinus fontinalis Mitchill (Pisces). Oikos, Suppl. 13: 13.9-142. Beukema, J. J. 1968. Predation by the three-spined stickleback (Gasterosteus aculeatus L.): The influence of hunger and experience. Behaviour, 30: 1-126. Bieniarz, K. 1974. Circadian rhythm in the content of neurosecretory material "Gomori-positive" in the cells of nucleus preopticus of the rainbow trout (Salmo irideus Gibbons). J. interdiscipl. Cycle Res., 5: 115-117. 217 Bisson, P. A. 1978. Diel food selection by two sizes of rainbow trout (Salmo gairdneri) in an experimental stream. J. Fish.. Res. Board Can., 35: 971-975. Blackett, R. F. 1974. Preliminary evaluation of pink (Oncorhynchus gorbuscha) and sockeye (0_. nerka) salmon incubation and rearing in gravel incubators and throughs. Alaska Dept. Fish Game, Tech. Data Rept. No. 17: 32 p. Blaxter, J. H. S. 1969a. Development: eggs and larvae. In:' Fish physiology. Vol. 3. Edited by W. S. Hoar and D. J. Randall. Academic Press, Inc., New York. pp. 178-252. • . 1969b. Swimming speeds of fish. In: FAO Fish. Rept. No. 62, Vol. 2: 69-100. . . 1970. 2. Light. 2.3 Animals. 2.32 Fishes. In: Marine ecology. Vol. 1 (1). Edited by 0. Kinne. Wiley-Interscience, Toronto, pp. 213-320. , and M. Staines. 1971. Food searching potential in marine fish larvae. In: Fourth European Marine Biology Symposium. Edited by D. J. Crisp. Cambridge University Press, Cambridge, pp. 467-485. Bohun, S., and H. E. Winn. 1966. Locomotor activity of the American eel (Anguilla rostrata). Chesapeake Sci., 7: 137-147. Braemer, W. 1960. A critical review of the sun-azimuth hypothesis. Gold Spring Harbor Symp. Quant. Biol., 25: 413-427. Brett, J, R. 1957. The sense organs: the eye.; In: The physiology of fishes. Vol. 2. Edited by M. E. Brown. Academic Press, Inc., New York. pp. 121-154. . 1965. The relation of size to rate of oxygen consumption and sustained swimming speed of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can., 22: 1491-1501. . 1970. 3. Temperature. 3.3. Animals. 3.32. Fishes. In: Marine ecology. Vol. T. Edited by 0. Kinne. Wiley- Interscience, Toronto, pp. 515-560. . 1971. Satiation time, appetite and maximum food intake of the sockeye salmon Oncorhynchus nerka. J. Fish. Res. Board Can., 28: 409-415. . 1974. Tank experiments on the culture of pan-size sockeye (Oncorhynchus nerka) and pink salmon (0_. gorbuscha) using environmental control. Aquaculture, 4: 341-352. . 1979. Environmental factors and growth. In: Fish physiology. Vol. 8. Edited by W. S. Hoar, D'. J. Randall and J. R. Brett. Academic Press, Inc., New York. pp. 599-675. 218 Brett, J. R., and M. A. Ali. 1958. Some observations on the structure and photomechanical responses of the Pacific salmon retina. J. Fish. Res. Board Can., 15: 815-829. • , and C. Groot. 1963. Some aspects of olfactory and visual responses in Pacific salmon. J. Fish. Res. Board Can., 20: 287-303. , and D. A. Higgs. 1970. Effect of temperature on the rate of gastric digestion in fingerling sockeye salmon, Oncorhynchus nerka. J. Fish. Res. Board Can., 27: 1767-1779. Brunei, P. 1965. Food as a factor or indicator of vertical migrations of cod in the western Gulf of St. Lawrence. ICNAF Spec. Publ. Intl. Comm. N.W. Atlantic Fish., 6: 439-448. . 1972. The Gaspe cod ecosystem in the Gulf of St. Lawrence. III. The daily and seasonal vertical migrations of cod (Gadus morhua) in 1960-62. Naturaliste can., 99: 287-357. - . 1973. Les migrations verticales de la morue: resume' critique des connaissances et des me'thodes d'etude. Trav. Pecheries Quebec, No. 36: 1-24. Bullock, T. H. 1955. Compensation for temperature in the metabolism and activity of poikilotherms. Biol. Rev., 30: 311-342. Biinning, E. 1973. The physiological clock. Springer-Verlag, New York. Revised Third Edition. 258 p. Burns, J. R. 1975. Seasonal changes in the respiration of pumpkinseed, Lepomis gibbosus, correlated with temperature, daylength, and stage of reproductive development. Physiol. Zool., 48: 142- 149. Byrne, J. E. 1968. The effects of photoperiod and temperature on the daily pattern of locomotor activity in juvenile sockeye salmon, Oncorhynchus nerka (Walbaum). Ph.D. Thesis, University of British Columbia, Vancouver, B.C. . 1971. Photoperiodic activity changes in juvenile sockeye salmon (Oncorhynchus nerka). Can. J. Zool., 49: 1155-1158. Casimir, M. J. 1971. Zur Morphologie, Histochemie, Tagesperiodik und Biologie der Operculardriise bei Labriden und Scariden (Pisces) . Marine Biology, 8: 126-146. Chan, K. K.-S. 1976. A photosensitive daily rhythm in the female medaka, Oryzias latipes. Can. J. Zool., 54: 852-856. Chaston, I. 1969. Seasonal activity and feeding pattern of brown trout (Salmo trutta) in a Dartmoor stream in relation to availability of food. J. Fish. Res. Bo.ard Can. , 26: 2165-2171. 219 Chovnick, A. (ed.). I960. Biological clocks. Cold Spring Harbor Symp. Quant. Biol., Vol. 25. Coburn, A., and P. McCart. 1967. A hatchery release tank for pink salmon fry with notes on behavior of the fry in the tank after release. J. Fish. Res. Board Can., 24: 77-85. Colgan, P. 1973. Motivational analysis of fish feeding. Behaviour, 45: 38-66. . 1975. Self-selection of photoperiod as a technique for studying endogenous rhythms in fish, J. interdiscipl. Cycle Res., 6: 203-211. Confer, J. L., and P. I. Blades. 1975. Omnivorous zooplankton and planktivorous fish. Limnol. Oceanogr., 20: 571-579. Curio, E. 1976. The ethology of predation. Zoophysiology and Ecology. Vol. 7. Springer-Verlag, New York. 250 p. Daan, S., and J. Aschoff. 1975. Circadian rhythms of locomotor activity in captive birds and mammals: Their variations with season and latitude. Oecologia (Berl.), 18: 269-316. Darnell, R. M., and R. R. Meierotto. 1965. Diurnal periodicity in the black bullhead, Ictalurus melas (Rafinesque). Trans. Am. Fish. Soc., 94: 1-7. Davis, R. E. 1962. Daily rhythm in the reaction of fish to light. Science, 137: 430-432. , and J. Bardach. 1965. Time-co-ordinated prefeeding activity in fish. Anim. Behav., 13: 154-162. de Ruiter, L. 1963. The physiology of vertebrate feeding behaviour:- towards a synthesis of the ethological and physiological approaches to problems of behaviour. Z. Tierpsychol., 20: 498-516. . 1967. Feeding behavior of vertebrates in the natural environment. In: Handbook of Physiology, Sect. 6. Alimentary canal, Voli 1. Control of food and water intake. Am. Physiol. Soc, Washington, D.C. pp. 97-116. Dill, L. M., and T. G. Northcote. 1970. Effects of gravel size, egg depth, and egg density on intragravel movement and emergence of coho salmon (Oncorhynchus kisutch) alevins. J. Fish. Res. Board Can., 27: 1191-1199. Dill, P. A. 1970. On the development of diel activity rhythms in Atlantic salmon and rainbow trout. Ph.D. Thesis, University of Western Ontario, London, Ontario. 220 Dill, P. A. 1979a. Temperature influence on time in gravel and condition at emergence for Atlantic salmon, rainbow trout and speckled trout. Unpublished manuscript. . 1979b. Changes in length and weight of pink salmon (Oncorhynchus gorbuscha Walbaum) during the eleutheroembryo stage of development. Unpublished manuscript. Doble, B. D., and D. M. Eggers. 1978. Diel feeding chronology, rate of gastric evacuation, daily ration, and prey selectivity in Lake Washington juvenile sockeye salmon (Oncorhynchus nerka). Trans. Am. Fish. Soc, 107: 36-45. Dorrscheidt, G. J., and L. Beck. 1975. Advanced methods for evaluating characteristic parameters ( T, a , p ) of circadian rhythms. J. Math. Biol., 2:-107-121. Edel, R. K. 1976. Activity rhythms of maturing American eels (Anguilla rostrata). Marine Biology, 36: 283-289. Edmundson, E., F. E. Everest, and D. W. Chapman. 1968. Permanence of station in juvenile chinook salmon and steelhead trout. J. Fish. Res. Board Can., 25: 1453-1464. Eggers, D. M. 1975. A synthesis of the feeding behavior and growth of juvenile sockeye salmon in the limnetic environment. Ph.D. Thesis, University of Washington, Seattle, Washington. • . 1977. Factors in interpreting data obtained by diel sampling of fish stomachs. J. Fish. Res. Board Can. , 34:> 290-294. Elliott, J. M. 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. . 1973. The food of brown trout and rainbow trout (Salmo trutta and S_. gairdneri) in relation to the abundance of drifting invertebrates in a mountain stream. Oecologia (Berl.), 12: 329-347. . 1975a. Weight of food and time required to satiate brown trout, Salmo trutta L. Freshwat. Biol., 5: 51-64. . 1975b. Number of meals in a day, maximum weight of food consumed in a day and maximum rate of feeding for brown trout, Salmo trutta L. Freshwat. Biol., 5: 287-303. Enright, J. T. 1965a. The search for rhythmicity in biological time-series. J. theor. Biol.,.8: 426-468. 221 Enright, J. T. 1965b. Accurate geophysical rhythms and frequency analysis. In: Circadian clocks. Edited by J. Aschoff. North-Holland Publ. Co.,.Amsterdam, pp. 31-42. . 1970. Ecological aspects of endogenous rhythmicity. Ann. Rev. Ecol. Syst., 1: 221V-238. • . 1975. Orientation in time: endogenous clocks. In: Marine ecology. Vol. 2 (2). Edited by 0. Kinne. Wiley and Sons Ltd., London, pp. 917-944. Eriksson, L.-O. 1972. Die Jahresperiodik augen- und pinealorganloser Bachsaiblinge Salvelinus fontinalis Mitchill. Aquilo, Ser. Zool., 13: 8-12. . 1973. Spring inversion of the diel rhythm of locomotor activity in young sea-going brown trout, Salmo trutta trutta L., and atlantic salmon, Salmo salar L. Aquilo, Ser. Zool., 14: 68-79. . 1975. Diel and annual locomotor activity rhythms in some fresh water fish species with special reference to the seasonal inversion in salmonids. Ph.D. Thesis, University of Umea, Umea, Sweden. . 1978a. Nocturnalism versus diurnalism - dualism within fish individuals. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York., pp. 69-89. . 1978b. A laboratory study of diel and annual activity rhythms and vertical distribution in the perch, Perca fluviatilis, at the Arctic circle. Env. Biol. Fish., 3: 301-307. , och F. Kubicek. 1972. Experimentella undersokningar av sikens vertikalvandringar. Fauna Flora, Uppsala, 67: 224-228. , and T. van Veen. 1979. Circadian rhythms and phototactic behaviour in the brown bullhead, Ictalurus nebulosus (Teleostei). Evidence for an endogenous rhythm in feeding -, locomotor - and reaction-time behaviour. Unpublished manuscript. Erkert, H. G., and S. Kracht. 1978. Evidence for ecological adaptation of circadian systems. Circadian activity rhythms of neo-tropical bats and their re-entrainment after phase-shifts of the Zeitgeber— LD. Oecologia (Berl.), 32: 71-78. Erkinaro, E. 1961. The seasonal change of the activity of Microtus agrestis. Oikos, 12: 157-163. . 1969a. Der Phasenwechsel der lokomotorischen Aktivitat bei Microtus agrestis (L.), M. arvalis (Pall.) und M. oeconomus (Pall.). Aquilo, Ser. Zool., 8: 1-31. 222 Erkinaro, E. 1969b. Free-running circadian rhythm in wood mouse (Apodemus flavicollis Melch.) under natural light-dark cycle. Experientia, 25: 649. . 1970. Wirkung von Tageslange und Dammerung auf die Phasenlage der 24 h - Periodik der Waldmaus Apodemus flavicollis Melch. im Naturtag, Oikos, Suppl., 13: 101-107. •, . 1972. Seasonal changes in the phase position of circadian activity rhythms of some voles, and their endogenous component. Aquilo, Ser. Zool., 13: 87-91. Evans, K. J. 1966. Responses of the locomotor activity rhythms of lizards to simultaneous light and temperature cycles. Comp. Biochem. Physiol., 19A: 91-103. Fange, R., and D. Grove. 1979. , Digestion. In: Fish physiology. Vol. 8. Edited by W. S. Hoar, D. J. Randall and J. R. Brett. Academic Press, Inc., New York. pp. 161-260. Favorite, F., A. J. Dodimead, and K. Nasu. 1976. Oceanography of the subarctic Pacific region, 1960-71. Intl. North Pacific Fish. Comm. Bull. No. 33: 187 p. Feller, R. J., and V. W. Kaczynski. 1975. Size selective predation by juvenile chum salmon (Oncorhynchus keta) on epibenthic prey in Puget Sound. J. Fish. Res. Board Can., 32: 1419-1429. Figala, J., und K. Miiller. 1972, Jahresperiodik und Phasenlage eines tropischen Fischen (Barbus partipentazona) am Polarkreis. Aquilo, Ser. Zool., 13: 13-20. Fontaine, M. 1975. Physiological mechanisms in the migration of marine and amphihaline fish. Adv. mar. Biol., 13: 241-355. Fried, S. M., J. D. McCleave, and G. W. LaBar. 1973. Seaward migration of hatchery-reared Atlantic salmon, Salmo salar, smolts in the Penobscot River estuary, Maine: riverine movements. J. Fish. Res. Board Can., 35: 76-87. Fry, F. E. J. 1967. Responses of vertebrate poikilotherms to temperature. In: Thermobiology. Edited by A. H. Rose. Academic Press, Inc., New York. pp. 376-409. . 1971. The effects of environmental factors on the physiology of fish. In: Fish physiology. Vol. 6. Edited by W. S. Hoar and D. J. Randall. Academic Press, Inc., New York. pp. 1-98. • , and J. S. Hart. 1948. Cruising speed of goldfish in relation to water temperature. J. Fish; Res. Board Can., 7: 169-175. Garside, E. T. 1966. Developmental rate and vertebral number in salmonids. J. Fish. Res. Board Can., 23: 1537-1551. 223 Gebhards, S. V. 1961. Emergence and mortality of chihook salmon fry in a natural redd. ,Prog. Fish-Cult., 23: 91.' Geiger, W., und H. Roth. 1962. Beobachtungen an kiinstlichen Forellenlaichgruben. Schweiz. Z. Hydrol., 25: 76-89. Gerke, R. J., and V. W. Kaczynski. 1972. Food of juvenile pink and chum salmon in Puget Sound, Washington. Wash. Dept. Fish., Tech. Rept. No. 10: 27 p. Gibson, R. N. 1965. Rhythmic activity in littoral fish. Nature, 207: 544-545. . 1967. Experiments on the tidal rhythm of Blennius pholis. J. mar. biol. Ass. U.K., 47: 97-111. ' . 1971. Factors affecting the rhythmic activity of Blennius pholis L. (Teleostei). Anim. Behav., 19: 336-343. . 1973. Tidal and circadian activity rhythms in juvenile plaice, Pleuronectes platessa. Marine Biology, 22: 379-386. . 1976. Comparative studies on the rhythms of juvenile flatfish. In: Biological rhythms in the marine environment. Edited by P. J. DeCoursey. University of South Carolina Press, Columb ia, South Carolina. pp» 199—213. , J. H. S. Blaxter, and S. J. de Groot. 1978. Developmental changes in the activity rhythms of the plaice (Pleuronectes platessa L.). In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 169-186. Gilhousen, P. 1962. Marine factors affecting survival of Fraser River pink salmon. In: Symposium on pink salmon. H. R. MacMillan Lectures in Fisheries. Edited by N. J. Wilimovsky. University of British Columbia, Vancouver, pp. 105-111. Ginetz, R. M. 1976. Fulton River upwelling gravel incubator for sockeye salmon. Can. Dept. Environm., Fish. Mar. Serv. Tech. Rep. Ser. No. PAC/T - 76-10: 45 p. , and P. A. Larkin. 1976. Factors affecting rainbow trout (Salmo gairdneri) predation on migrant fry of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can., 33: 19-24. Godin, J.-G. J. 1978. . Behavior of juvenile pink salmon (Oncorhynchus gorbuscha Walbaum) toward novel prey: influence of ontogeny and experience. Env. Biol. Fish., 3: 261-266. , C. Groot, and M. C. Armstrong. 1978. Description and evaluation of a photoelectric system for monitoring the locomotor activity of a pelagic fish. Can. J. Zool., 56: 1412-1419. 224 Gourley, E. V. 1972. Circadian activity rhythm of the gopher tortoise (Gopherus polyphemus). Anim. Behav., 20: 13-20. Greer Walker, M., F. R. Harden Jones, and G. P. Arnold. 1978. The movements of plaice (Pleuroriectes platessa L.) tracked in.the open sea. J. Cons. int. Explor. Mer, 38: 58-86. Groot, C. 1965. On the orientation of young sockeye salmon (Oncorhynchus nerka) during their seaward migration out of lakes. Behaviour, Suppl. 14: 198 p. , K. Simpson, I. Todd, P. D. Murray, and G. A. Buxton. 1975. Movements of sockeye salmon (Oncorhynchus nerka) in the Skeena River estuary as revealed by ultrasonic tracking. J. Fish. Res. Board Can., 32: 233-242. Grove, D. J., L. G. Loizides, and J. Nott. 1978. Satiation amount, frequency of feeding and gastric emptying rate in Salmo gairdneri. J. Fish Biol., 12: 507-516. Gwinner, E. 1973. Circannual rhythms in birds: Their interaction with circadian rhythms and environmental photoperiod. J. Reprod. Fert,, Suppl. 19: 51-65. . 1975. Circadian and circannual rhythms'in birds. In: Avian biology. Vol. 5. Edited by D. S. Farner and J. R. King. Academic Press, Inc., New York. pp. 221-285. . 1977. Circannual rhythms in bird migration, Ann. Rev. Ecol. Syst., 8: 381-405. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can. Bull. 180: 740 p. Hartt, A. C. 1966. Migrations of salmon in the North Pacific Ocean and Bering Sea as determined by seining and tagging, 1959-1960. Intl. North Pacific Fish. Comm. Bull. No. 19: 141 p. Hasler, A. D. 1966. Underwater guideposts: Homing of salmon. University of Wisconsin Press, Madison. 155 p. • . 1971. Orientation and fish migration. In: Fish physiology. Vol. 6. Edited by W. S. Hoar and D. J. Randall. Academic Press, Inc., New York. pp. 429-510. , R. M. Horrall, W. J. Wisby, and W. Braemer. 1958. Sun-orientation and homing in fishes. Limnol. Oceanogr., 3: 353-361. Hasler, A. D., and H. 0. Schwassmann. 1960. Sun orientation of fish at different latitudes. .Cold Spring Harbor Symp. Quant. Biol., 25: 429-441. 225 Hastings, J. W. 1970. Cellular-biochemical clock hypothesis. In: The biological clock. Two views. Edited by F. A. Brown, Jr., J. W. Hastings, and J. D. Palmer. Academic Press, Inc., New York. pp. 63-91. Healey, M. C. 1967. Orientation of pink salmon (Oncorhynchus gorbuscha) during early marine migration from Bella Coola River system. J. Fish. Res. Board Can., 24: 2321-2338. . 1979. Detritus and juvenile salmon production in the Nanaimo Estuary: I. Production and feeding rates of juvenile chum salmon (Oncorhynchus keta). J. Fish. Res. Board Can., 36: 488-496. Heard, W. R. 1964. Phototactic behaviour.of emerging sockeye salmon fry. Anim. Behav., 12: 382-389. Herman, T. B. 1977. Activity patterns and movements of subarctic voles. Oikos, 29: 434-444. Hirata, H. 1973a. Diurnal rhythm of metabolic and activity rates in juvenile Atlantic salmon, Salmo salar Linnaeus. Mem. Fac. Fish. Kagoshima University, 22: 73-93. . 1973b. Studies on the daily rhythm of feeding activity in fish, especially in goldfish, Carassius auratus (Linne). Mem. Fac. Fish. Kagoshima University, 22: 1-48. Hoar, W. S. 1942. Diurnal variations in feeding activity of young salmon and trout. J. Fish. Res. Board Can., 6: 90-101. . 1951. The behaviour of chum, pink and coho salmon in relation to their seaward migration. J. Fish. Res. Board Can., 8: 241-263. 1953. Control and timing of fish migration. Biol. Rev., 28: 437-452. 1956. The behaviour of migrating pink and chum salmon fry. J. Fish. Res. Board Can., 13: 309-325. 1958. The evolution of migratory behaviour among juvenile salmon of the genus Oncorhynchus. J. Fish. Res. Board Can., 15: 391-428. 1965. The endocrine system as a chemical link between the organism and its environment. Trans. Royal Soc. Can., 3 (Ser. 4, Sect. 3): 175-200. 1976. Smolt transformation: evolution, behavior, and physiology. J. Fish. Res. Board 'Can., 33: 1234-1252. 226 Hoar, W. S., M. H. A. Keenleyside, and R. G. Gopdall. 1957. Reactions of juvenile Pacific salmon to light. J. Fish. Res. Board Can., 14: 815-830. Hochachka, P. W., and G. N. Somero. 1973. Strategies of biochemical adaptation. W. B. Saunders Co., Toronto. 358 p. Hoffmann, K. 1965. Overt circadian frequencies and circadian rule. In: Circadian clocks. Edited by J. Aschoff. North-Holland Publ. Co., Amsterdam, pp. 87-94. . 1968. Synchronisation der circadianen Aktivitatsperiodik• von Eidechsen durch Temperaturcyclen verschiedener Amplitude. Z. vergl. Physiol., 58: 225-228. . 1969. Die relativ Wirksamkeit von Zeitgebern. Oecologia (Berl.), 3: 184-206. . 1970. Zur Synchronisation biologischer Rhythmen. Verh. dtsch. zool. Ges., 64: 266-273. Hokanson, K. E. F. 1977. Temperature requirements of some percids and adaptations to the seasonal temperature cycle. J. Fish. Res. Board Can., 34: 1524-1550. Hokkanen, H. , T. To'rmala, and H. Vuorinen. 1977. Seasonal changes in the circadian activity of Pteromys volans L. in central Finland. Ann> Zool. Fennica, 14: 94-97. Huet, M. 1972. Textbook of fish culture. Breeding and cultivation of fish. Fishing News (Books) Ltd., Surrey. 436 p. Hunter, J. G. 1959. Survival and production of pink and chum salmon in a coastal stream. J. Fish. Res. Board Can., 16: 835-886. Hunter, J. R. 1972. Swimming and, feeding behavior of larval anchovy, Engraulis mordax. U.S. Fish Wildl. Serv. Fish. Bull., 70: 821-838. , and G. L. Thomas. 1974. Effect of prey distribution and density on the searching and feeding behaviour of larval anchovy Engraulis mordax Girard. In: The early life history of fish. Edited by J. H. S. Blaxter. Springer-Verlag, New York. pp. 559-574. Hyatt, K. D. 1979. Feeding strategy. In: Fish physiology. Vol. 8. Edited by W. S. Hoar, D. J. Randall and J. R. Brett. Academic Press, Inc., New York. pp. 71-119. Ichihara, T., T. Yonemori, and H. Asai. 1975. Swimming behavior of a chum salmon, Oncorhynchus keta, on the southern migration off Etorofu Island, the southern Kurile Islands. Bull. Far Seas Fish. Res. Lab., 13: 63-77. (Transl. by M. Miki). 227 Ishida, T. 1966. Salmon of the North Pacific ocean - Part III. A review of the life history of North Pacific salmon. 2. Pink salmon in the Far East. Intl. North Pacific Fish. Comm. Bull. No. 18: 29-39. Ishiwata, N. 1968a. Ecological studies on the feeding of fishes. III. Degree of hunger and satiation amount. Bull. Jap. Soc. Sci. Fish., 34: 604-607. (Fish. Res. Board Can. Transl. Ser. No.. 1397, 1970) . . 1968b. Ecological studies on the feeding of fish. IV. Satiation curve. Bull. Jap. Soc. Sci. Fish., 34: 691-694. Ivlev, V. S. 1961. Experimental ecology of the feeding of fishes. Yale University Press, New Haven, Conn, 302 p. Jenkins, B. W., and J. M. Green. 1977. A critique of field methodology of determining fish feeding periodicity. Env. Biol. Fish., 1: 209-214. Jenkins, T. M., Jr. 1969. Night feeding of brown and rainbow trout in an experimental stream channel. J. Fish. Res. Board Can., 26: 3275-3278. John, K. R., and D. M. Gring. 1968. Retinomotor rhythms in the bluegill, Lepomis macrochirus. J. Fish. Res. Board Can., 25: 373-381. , and M. Haut. 1964. Retinomotor cycles and correlated behavior in the teleost.Astyanax mexicanus (Fillipi). J. Fish. Res. Board Can., 21: 591-595. • , and L. H. Kaminester. 1969. Further studies on retinomotor rhythms in the teleost Astyanax mexicanus. Physiol. Zool., 42: 60-70. •' , M. Segall, and L. Zawatzky. 1967. Retinomotor rhythms in the goldfish, Carassius auratus. Biol. Bull., 132: 200-210. Johnson, W. E., and C. Groot. 1963. Observations on the migration of young sockeye salmon (Oncorhynchus nerka) through a large, complex lake system. J. Fish. Res. Board Can., 20: 919-938. Kaczynski, V. W., R. J. Feller, J. Clayton, and R, J. Gerke. 1973. Trophic analysis of juvenile pink and chum salmon (Oncorhynchus gorbuscha and 0_. keta) in Puget Sound. J. Fish. Res. Board Can., 30: 1003-1008. Kalleberg, H. 1958. Observations in a stream tank of territoriality and competition in juvenile salmon and trout (Salmo salar L. and S^. trutta L.). Rep. "Inst. Freshwater Res., Drottningholm, 39: 55-98. 228 Kanciruk, P., and W. F. Herrnkind. 1973. Preliminary investigations of the daily and seasonal locomotor activity rhythms of the spiny lobster, Panulirus argus. Mar. Behav. Physiol., 1: 351-359. Kavaliers, M. 1978a. The role of photoperiod and twilight in the control of locomotory rhythms in the lake chub, Couesius plumbeus. Ph.D. Thesis, University of Alberta, Edmonton, Alberta. . -1978b. Seasonal changes in the circadian period of the lake chub Couesius plumbeus. Can. J. Zool., 56: 2591-2596. , and F. S. Abbott. 1977. Rhythmic colour change of the killifish, Fundulus heteroclitus. Can. J. Zool., 55: 553-561. Kavanau, J. L., and J. Ramos. 1975. Influences of light on activity, and phasing of carnivores. Am. Natur., 109: 391-418. Katz, H. M. 1978. Circadian rhythms in juvenile American shad, Alosa sapidissima. J. Fish Biol., 12: 609-614. Kenagy, G. J. 1976. The periodicity of daily activity and its seasonal changes in free-ranging and, captive Kangaroo rats. Oecologia (Berl.), 24: 105-140. . 1978. Seasonality of endogenous circadian rhythms in a diurnal rodent Ammospermophilus leucurus and a nocturnal rodent Dipodomys merriami. J. comp. Physiol., 128: 21-36. Kirkwood, J. B. 1962. Inshore-marine ,and freshwater life history.of the pink salmon, Oncorhynchus gorbuscha (Walbaum), and the chum salmon, Oncorhynchus keta (Walbaum), in Prince William Sound, Alaska. Ph.D. Thesis, University of Louisville, Louisville, Kentucky. Kleerekoper, H., G. Taylor, and R. Wilton. 1961. Diurnal periodicity in the activity of Petromyzon marinus and the effects of chemical stimulation. Trans. Am. Fish. Soc, 90: 73-78. Klein, H. 1974. The adaptional value of internal annual clocks in birds. In: Circannual clocks. Annual biological rhythms. Edited by E. T. Pengelley. Academic Press, Inc., New York, pp. 347-391. Kobayashi, T., and S. Harada. 1966. Ecological observations on the salmon of Nishibetsu River. II. Movements, growth and feeding habits of pink salmon fry, Oncorhynchus gorbuscha Walbaum, during seaward migration. Sci. Rep. Hokkaido Salmon Hatchery, 20: 1-10. (Fish. Res. Board Can. Transl. Ser. No. 785, 1966). Koski, K. V. 1966. The survival of coho salmon .(Oncorhynchus kisutch) from egg deposition to emergence in three Oregon coastal streams. M.Sc Thesis, Oregon State University, Corvalliss Oregon. 229 Koski, K. V. 1975. The survival and fitness of two stocks of chum salmon (Oncorhynchus keta) from egg deposition to emergence in a controlled-stream environment at Big Beef Creek. Ph.D. Thesis, University of Washington, Seattle,, Washington. Kroneld, R. 1974: Autumn and spring phase shifts in the burbot, Lota lota L. (Pisces, Gadidae), at the Arctic circle. Mem. Soc. Fauna Flora Fenn., 50: 125-136. . 1976. Phase shift of swimming activity in the burbot Lota lota L. (Pisces, Gadidae) at the Arctic circle. Physiol. Zool., 49: 49-55. Kruuk, H. 1963. Diurnal periodicity in the activity of the common sole Solea vulgaris Quensel. Neth. J. Sea Res., 2: 1-28. LaBar, G. W., J. D. McCleave, and S. M, Fried. 1978. Seaward migration of hatchery-reared Atlantic salmon (Salmo salar) smolts in the Penobscot River estuary: open-water movements. J. Cons. int. Explor. Mer, 38: 257-269. Landless, P. J. 1976, Demand-feeding behaviour of rainbow trout. Aquaculture, 7: 11-25. Larkin, P. A. 1956. Interspecific competition and population control in freshwater fish.. J. Fish, Res. Board Can., 13: 327-342. LeBrasseur, R. J. 1965. Data record. Stomach contents of salmonids caught in the Northeastern Pacific Ocean - 1958. Circular (Statistical Series) No. 15, Fish. Res.. Board Can., Pacific Biological Station, Nanaimo, British Columbia. , and R. R. Parker. 1964. Growth rate of Central British Columbia pink salmon. J. Fish. Res. Board Can., 21: 1101- 1127. Leong, R. J., and C. P. O'Connell. 1969. A laboratory study of particulate, and filter feeding of the northern anchovy (Engraulis mordax). J. Fish. Res. Board Can., 26: 557-582. Lichtenheld, R. W. 1966. Effect of light, temperature and gamma radiation on the locomotor activity of juvenile steelhead trout (Salmo gairdneri). Ph.D. Thesis, University of Washington, Seattle, Washington. Lillelund, K. 1957. Versuche zur Bestimmung der Verdauungsgeschwindigkeit bei der Brut vom Hecht (Esox lucius L.), Kleiner Marane (Coregonus albula L.) und Groszer Marane (Coregonus holsatus (Thienemann)). Kurze Mit. Inst. Fisch. Biol. Universitat Hamburg, 7: 29-39. 230 Livingston, R. T. 1971. Circadian rhythms in the respiration of eight species of cardinal fishes (Pisces: Apogonidae): comparative analysis and adaptive significance. Marine Biology, 9: 253-266. Lofts, B. 1970. Animal photoperiodism. Institute of Biol. Studies in Biology, No, 25. Edward Arnold Ltd., London. 64 p. Lohmann, M. 1967. Phase-dependent changes of circadian frequency after light steps. Nature, 213: 196-197. MacAfee, N. 1971. Effect of light and temperature cycles on diel activity in speckled trout (Salvelinus fontinalis (Mitchill)). B.Sc. (Honors) Thesis, St. Francis Xavier University, Antigonish, Nova Scotia. Machidori, S. 1966. Vertical distribution of salmon (Genus Oncorhynchus) in the North-western Pacific. , Bull. Hokkaido Reg. Fish, Res. Lab. , 31: 11--17. MacKinnon, D., and J. R. Brett. 1955. Some observations on the movement of Pacific salmon fry through a small impounded water basin. J. Fish. Res. Board Can., 12: 362-368. Magnuson, J. J. 1969. Swimming activity of the scombrid fish Euthynnus affinis as related to search for food. In: FAO Fish. Rept. No. 62, Vol. 2: 439-451. Major, P. F. 1978. Predator-prey interactions in two schooling fishes, Caranx ignobilis and Stolephorus purpureus. Anim. Behav., 26: 760-777. , Mangelsdorff, A. D., and G. T. Hauty. 1972. Effects of temperature and illumination on the locomotor activity and internal temperature of the turtle (Terrapene Carolina Carolina). J. interdiscipl. Cycle Res., 2: 167-177, Manzerj J. I,. 1964. Preliminary observations on the vertical distribution of Pacific salmon (Genus Oncorhynchus) in the Gulf of Alaska. J. Fish. Res. Board Can., 21: 891-903. , and R. J. LeBrasseur. 1959. Further observations on the vertical distribution of salmon in the northeast Pacific. Fish., Res. Board Can. Manuscript Rept. No. 689. , and M. P. Shepard. 1962. Marine survival, distribution and migration of pink salmon (Oncorhynchus gorbuscha) off the British Columbia coast.. In: Symposium on pink salmon. H. R. MacMillan Lectures in Fisheries. Edited by N. J. Wilimovsky. University of British Columbia, Vancouver, pp. 113-122. 231 Martin, J. W; 1966. Early sea life of pink salmon. Proc. 1966 Northeast Pacific Pink Salmon Workshop, Alaska Dept. Fish and Game Infor. Leaflet, 87: 111-125. Mason, J. C. 1969. Hypoxial stress prior to emergence and competition among coho salmon fry. J. Fish. Res. Board Can., 26: 63-91. . 1976. Some features of coho salmon, Oncorhynchus kisutch, fry emerging from simulated'redds, and concurrent changes in photobehavior. U.S. Fish Wildl. Serv. Fish. Bull., 74: 167-175. McCart, P. 1967. Behaviour and ecology of sockeye salmon fry in the Babine River. J. Fish. Res. Board Can., 24: 375-428. McCleave, J. D. 1978. Rhythmic aspects of estuarine migration of hatchery-reared Atlantic salmon (Salmo salar) smolts. J. Fish Biol., 12: 559-570. McDonald, J. G; 1960. : The behaviour of Pacific salmon fry during their downstream migration to freshwater and saltwater nursery areas. J. Fish. Res. Board Can., 17: 665-685. ' . 1973. Diel vertical movements and feeding habits of underyearling sockeye salmon (Oncorhynchus nerka), at Babine Lake, B.C. Fish. Res. Board Can. Tech. Rept. No. 378: 55 p. Mclnerney, J. E. 1964. Salinity preference: an orientation mechanism in salmon migration. J. Fish. Res, Board Can., 21: 995-1018. McKenzie, J. A., 1960. The spontaneous activity of the speckled trout (Salvelinus fontinalis). M.A. Thesis, University of Toronto, Toronto, Ontario. McLaren, I. A. 1963. Effects of temperature on growth of zooplankton and the adaptive value of vertical migration. J. Fish. Res. Board Can., 20: 685-727. Meier, A. H. 1970. Thyroxin phases the circadian fattening response to. prolactin. Proc. Soc. Exp. Biol. Med., 133: 1113-1116. . 1975. Chronophysiology of prolactin in lower vertebrates. Amer. Zool., 15: 905-916. , T. N. Trobec, M. M. Joseph, and T. M. John. 1971. Temporal synergism of prolactin and adrenal steroids in the regulation of fat stores. Proc. Soc. Exp. Biol. Med., 137: 408-415. Menaker, M. 1969. Biological clocks. Bioscience, 19: .681-689. (ed.). 1971. Bipchronometry. Proc. Symp., Friday Harbor, Wash., Natl. Acad. Sci., Washington, D.C. 662 p. 232 Menaker, M. 1974. Circannual rhythms in circadian perspective. In: Circannual clocks. Annual biological rhythms. Edited by E. T. Pengelley. Academic Press, Inc., New York., pp. 507-520. • . 1976. Overview: A "terrestrial" biologist looks at biorhythms in the marine environment. In: Biological rhythms in the marine environment. Edited by P. J. DeCoursey. University of South Carolina Press, Columbia, South Carolina, pp. 273-278. Meuwis, A. L., and M. J. Heuts. 1957. Temperature dependence of breathing rate in carp. Biol..Bull., 112: 97-107. Mitson, R. B. 1978. A review of biotelemetry techniques using acoustic tags. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 269-283. Morris, R. W. 1965. Thermal acclimation of metabolism of the yellow bullhead, Ictalurus natalis (LeSueur). Physiol., Zool., 38: 219-227. Morse, D. H. 1977. Feeding behavior and predator avoidance in heterospecific groups. Bioscience, 27: 332-339. Miiller, K. 1969a. Untersuchungen zur Tagesperiodik von Wasserorganismen am Polarkreis. Umschau, 69: 18-19. • . 1969b. Die Tages- und Jahresperiodik der Bachforelle (Salmo trutta L.) am Polarkreis. Aquilo, Ser. Zool., 8: 50-62. 1969c. Jahreszeitliche Wechsel der 24 h-Periodik bei der Bachforelle (Salmo trutta L.) am Polarkreis. Oikos, 20: 166-170. 1969d. Nachweis circannualer Periodik bei der Quappe (Lota lota L.) (Pisces: Gadidae). Experientia, 25: 1268-1269. 1970a. Phasenwechsel der lokomotorischen Aktivitat bei der Quappe Lota lota L. Oikos, Suppl. 13: 122-129. 1970b. Die Tages- und Jahresperiodik der Buntflossenkoppe Cottus poecilopus Heckel am Polarkreis. Oikos, Suppl. 13: 108-121. 1973a. Seasonal phase shift and the duration of activity time in the burbot, Lota lota (L.) (Pisces, Gadidae). J. comp. Physiol., 84: 357-360.; 1973b. Circadian rhythms of locomotor activity in aquatic organisms in the subarctic summer. Aquilo, Ser. Zool., 14: 1-18. 233 Mu'ller, K. 1976. Zur Chronobiologie der Marane (Coregonus lavaretus L.) am Polarkreis. [Chronobiological studies on whitefish (Coregonus lavaretus L.) at the arctic circle]. Arch. Fisch. Wiss., 27: 121-132.. (Can. Fish. Mar. Serv. Transl. Ser. No. 4005, 1977). . 1978a. Locomotor activity of fish and environmental oscillations. J-n: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 1-19. . 1978b. The flexibility of the circadian system of fish at different latitudes. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic:Press, Inc., New York. pp. 91-104. .. . 1978c. Locomotor activity in whitefish - shoals (Coregonus lavaretus). In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp.. 225-233. Muller-Haeckel, A. 1974. Jahresperiodik der Blattbewegungen zweiter Oxalis-arten im Naturtag der subarktischen Region. Oecologia (Berl.), 15: 329-336. Mundie, J. H. 1971. The diel drift of Chironomidae in an artificial stream and its relation to the diet of coho salmon fry, Oncorhynchus kisutch (Walbaum). Can. Entomol., 103: 289-297. Narver, D. W. 1970. Diel vertical movements and feeding of underyearling sockeye salmon and the limnetic zooplankton in Babine Lake, British Columbia. J. Fish. Res. Board Can., 27: 281-316. Neave, F. 1953. Principles, affecting the size of pink and chum salmon populations in British Columbia. J. Fish. Res. Board Can., 9: 450-491. - . 1955. Notes on the seaward migration of pink and chum salmon fry. J. Fish. Res. Board Can., 12: 369-374. . 1960. Observations on the vertical distribution of salmon in the Northeast Pacific, 1960. Fish. Res. Board Can. Manuscript Rept. No. 701. . 1964. Ocean migration of Pacific salmon. J. Fish. Res. Board Can., 21: 1227-1244. ' . 1966. Salmon of the North Pacific Ocean - Part III. A review of the life history of North Pacific salmon. 5. Pink salmon in British Columbia. Intl. North Pacific Fish. Comm. Bull. No. 18: 71-79. Neill, S. R. St. J., and J. M. Cullen. 1974. Experiments on whether schooling by their prey affects' the hunting behaviour of cephalopods and fish predators. J. Zool., Lond., 172: 549- 569. 234 Nelson, D. R., and R. H. Johnson. 1970. Diel activity rhythms in the nocturnal, bottom dwelling sharks Heterodontus francisci and Cephaloscyllium ventriosum. Copeia, .1970 (4): 732-739. Newman, M. A. 1956. Social behaviour and interspecific competition in two trout species. Physiol. Zool., 29: 64-81. . 1960. A comparative study of the residential behavior of juvenile salmonids. Ph.D. Thesis, University of British Columbia, Vancouver, B.C. Nicol, J. A. C. 1963. Some aspects of photoreception and vision in fishes. Adv. mar. Biol., 1: 171-208. Noakes, D. L. G. 1978. Ontogeny of behavior in fishes: a survey and suggestions. In: The development.of behavior. Comparative and evolutionary aspects. Edited by G. M. Burghart and M. Bekoff. Garland Publ., Inc., New York. pp. 103-123. Northcote, T. G. 1962. Migratory behaviour of juvenile rainbow trout (Salmo gairdneri) in outlet and inlet streams of Loon Lake, British Columbia. J. Fish. Res. Board Can., 19: 201-270. ' , and H. W. Lorz. 1966. Seasonal and diel changes in food of adult kokanee (Oncorhynchus nerka) in Nicola Lake, British Columbia. J. Fish. Res. Board Can., 23: 1259-1263. Nyberg, D. W. 1971. Prey-capture in the largemouth bass. Am. Midland Naturalist, 86: 128-144. O'Connor, J. M. 1972. Tidal activity rhythm in the hogchoker Trinectes maculatus (Bloch and Schneider). J. exp. mar. Biol. Ecol., 9: 173-177. 011a, B. L., H. M. Katz, and A. L. Studholme. 1970. Prey capture and feeding motivation in the bluefish, Pomatomus saltatrix. Copeia, 1970 (2): .360-362. 011a, B. L., and W. W. Marchioni. 1968. Rhythmic movements of cones in the retina of bluefish, Pomatomus saltatrix, held in constant darkness. Biol. Bull., 135: 530-536. 011a, B. L., and A. L. Studholme. 1972. Daily and seasonal rhythms of activity in the bluefish (Pomatomus saltatrix). In: Behavior of marine animals. Vol. 2. Edited by H. E. Winn and B. L. 011a. Plenum Press, Inc., New York. pp. 303-326. . 1978. Comparative aspects of the activity rhythms of tautog, Tautoga onitis, bluefish, Pomatomus saltatrix, and Atlantic mackerel, Scomber scombrus, as related to their life habits. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 131-151. 235 Palmer, J. D. 1974. Biological clocks in marine organisms. Wiley-Interscience, New York. 192 p. : . 1976. An introduction to biological rhythms. Academic Press, Inc., New York. 375 p. Pandian, T. J.,.and E. Vivekanandan. 1976. Effects of feeding and starvation on growth and swimming activity in an obligatory air-breathing fish. Hydrobiologia, 49: 33-:39. Parker, R. R. 1962. A concept of the dynamics of pink salmon populations. In: Symposium on pink salmon. H. R. MacMillan Lectures in Fisheries. Edited by N. J. Wilimovsky. University of British Columbia, Vancouver, pp. 203-211. . 1969. Foods and feeding of juvenile pink salmon in Central British Columbia waters. I. 1966 diel series. Fish. Res. Board Can. Manuscript Rept. No. 1017: 67 p. . , and R. J. LeBrasseur. 1974. Ecology of early sea life of pink and chum juveniles. Proc. 1974.Northeast Pacific Pink and Chum Salmon Workshop, Richmond, B.C. pp. 161-184. , and W. E. Vanstone. 1966. Changes in chemical composition of Central British Columbia pink salmon during early sea life. J. Fish. Res. Board Can., 23: 1353-1384. Penaz, M. 1975. Die lokomotorische Aktivitat larvaler und juveniler Elritzen (Phoxinus phoxinus). Zool. Listy, 24: 263-276. Peterman, R. M., and M. Gatto. 1978. Estimation of functional responses of predators on juvenile salmon. J. Fish. Res. Board Can., 35: 797-808. Peterson, R. H., H. C. E. Spinney, and A. Sreedharan. 1977. Development of Atlantic salmon (Salmo salar L.) eggs and alevins under varied temperature regimes. J. Fish. Res. Board Can., 34: 31-43. Phillips, A. C, and W. E. Barraclough. 1978. Early marine growth of juvenile Pacific salmon in the Strait of Georgia and Saanich Inlet, British Columbia. Can. Fish, Mar. Serv. Tech. Rept. No. 830: 19 p. Pittendrigh, C. S. 1974.. Circadian oscillators in cells and the circadian organization of multicellular systems. In: The neurosciences. Third study program. Edited by F. A. Schmitt. M.I.T. Press, Cambridge, Mass. pp. 437-458. , and S. Daan. 1976. A functional analysis of circadian pacemakers in nocturnal rodents. V. Pacemaker structure: A clock for all seasons. J. comp. Physiol., 106: 333-355. 236 Poston, H. A. 1978. Neuroendocrine mediation of photoperiod and other environmental influences on physiological responses in salmonids: a review. U.S. Fish Wildl. Serv. Tech. Paper No. 96: 14 p. Pritchard, A. L. 1944. Physical characteristics and.behaviour of pink salmon fry at McClinton Creek, B.C. J. Fish. Res. Board Can., 6: 217-227. Rapport, D. J., and J. E. Turner. 1975. Feeding rates and population growth. Ecology, 56: 942-949. Rawson, K. S., and P. J. DeCoursey. 1976. A comparison of the rhythms of mice and crabs from intertidal and terrestrial habitats. In: Biological rhythms in the marine environment. Edited by P. J. DeCoursey. University of South Carolina Press, Columbia, South Carolina, pp. 33-52. Reed, R. C. 1971. An experimental study of prey selection and regulatory capacity of bluegill sunfish (Lepomis macrochirus). M.A. Thesis, University of California, Santa Barbara, Cal. Rensing, L. 1965. Circadian rhythms in the course of ontogeny. In: Circadian clocks. Edited by J. Aschoff. North-Holland Publ. Co., Amsterdam, pp. 399-405. Reynolds, W. W,, and M. E. Casterlin. 1976. Locomotor activity rhythms in the bluegill sunfish, Lepomis macrochirus. Am. Midland Naturalist, 96: 221-225. " . 1978. Ontogenetic change in preferred temperature and diel activity of the yellow bullhead, Ictalurus natalis. Comp. Biochem. Physiol., 59A: 409-411. Richardson, N. E., and J. D. McCleave. 1974. Locomotor activity rhythms of juvenile Atlantic salmon (Salmo salar) in various light conditions. Biol.. Bull., 147: 422-432. Richkus, W. A. 1974. Factors influencing the seasonal and daily patterns of alewife (Alosa pseudoharengus) migration in a Rhode Island river. J. Fish. Res. Board Can.,.31: 1485-1497. Ricker, W. E. 1976. Review of the rate of growth and mortality of Pacific salmon in salt water, and noncatch mortality caused by fishing. J. Fish. Res. Board Can., 33: 1483-1524. Roberts, J. L. 1964. Metabolic responses of fresh-water sunfish to seasonal photoperiods and temperatures. Helgolander wiss. Meeresunters, 9: 459-473. v Roberts, N. J; 1963. An analysis of the daily activity pattern of the catfish, Ictalurus nebulosus LeSueur. M.Sc. Thesis, University of Maryland, College Park, Maryland. 237 Roe, H. S. J. 1974. Observations on the diurnal vertical migrations of an oceanic animal community. Marine Biology, 28: 99-113. Roppel, A. Y. 1956. Peak hours of pink (Oncorhynchus gorbuscha), and chum (0_. keta) salmon fry migration from Old Tom Creek, S.E., Alaska, 1952. Copeia, 1956 (2): 110-111. Rosenthal, H., and G. Hempel., 1970. Experimental studies in feeding and food requirements of herring larvae (Clupea harengus). In: Marine food chains. Edited by J. H. Steele. Oliver and Boyd, Edinburgh, pp. 344-364. Roth, H., und W. Geiger. 1963. Experimentelle Untersuchungen (iber das Verhalten der Bachforellenbrut in der Laichgrube. Schweiz. Z. Hydrobiol., 25: 202-218. Royce, W. F., L. S. Smith, and A. C. Hartt. 1968. Models of oceanic migrations of Pacific salmon and comments on guidance mechanisms. U.S. Fish Wildl. Serv. Fish. Bull., 66: 441-462. Rozin, P. N. 1964. Comparative biology of feeding patterns and mechanisms. Federation Proc., 23: 60-65. • , and J. Mayer. 1964. Some factors.influencing short-term intake in the goldfish. Am. J. Physiol., 206: 1430-1436. Schneider, H. 1964. Physiologische und Morphologische Untersuchungen zur Bioakustik der Tigerfische (Pisces, Theraponidae). Z. vergl. Physiol., 47: 493-558. Schoeneri, T. W. 1969. Optimal size and specialization in constant and fluctuating environments: An energy-time approach. Brookhaven Symp. Biol., 22: 103-114. . 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst., 2: 369-404. Schwassmann, H. 0. 1971a. Biological rhythms. In: Fish physiology. Vol. 6. Edited by W. S. Hoar and D. J. Randall. Academic Press, Inc., New York. pp. 371-428. . 1971b. Circadian activity patterns in gymnotid electric fish. In: Biochronometry. Edited by M. Menaker. Proc. Symp., Friday Harbor, Wash., Natl. Acad. Sci., Washington, D.C. pp. 186-199. Shapovalov, L. 1937. Experiments in hatching steelhead eggs in gravel. Calif. Fish Game, 23: 208-214. • , and W. Berrian. 1940. An experiment in hatching silver salmon (Oncorhynchus kisutch) eggs in gravel. Trans. Am. Fish. Soc, 69: 135-140. 238 Shelton, J. M. 1955. The hatching of chinook salmon eggs under simulated stream conditions. Prog. Fish-Cult., 17: 20-35. Sibert, J., T. J. Brown, M. C. Healey, B. A. Kask, and R. J. Naiman. 1977. Detritus-based food webs: exploitation by juvenile chum salmon (Oncorhynchus keta). Science, 196: 649-650. Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Co., Toronto. 312 p. Siegmund, R. 1969. Lokomotorische Aktivitat und Ruheverhalten bei einheimischen SuBwasserfischen (Pisces; Percidae, Cyprinidae) [Locomotory activity and resting behaviour of native freshwater fish (Pisces; Percidae, Cyprinidae)]. Biol. Zentralbl., 88: 295-312. (Can. Fish. Mar. Serv. Transl. Ser. No. 4139, 1977). , and D. L.. Wolff. 1972a. Influence of acoustical stimulation on circadian rhythms of locomotor activity with two species of fish. Forma et Functio, 5: 33-44. . 1972b. Die Aktivitatsperiodik von Fischen (Leucaspius delineatus und Carassius carassius) unter Beriicksichtigung der extraretinalen Lichtwahrnehmung. Forma et Functio, 5: 273-298. . 1973. Circadian-Rhythmik und Gruppenverhalten bei Leucaspius delineatus (Pisces, Cyprinidae). Experientia, 29: 54-58. Simenstad, C. A., B. S. Miller, J. N. Cross, K. L. Fresh, S. N. Steinfort, and J. C. Fegley. 1977. Nearshore fish and macroinvertebrate assemblages along the Strait of Juan de Fuca including food habits of nearshore fish. NOAA Tech. Memorandum ERL MESA - 20: 144 p. Simpson, T. H. 1978. An interpretation of some endocrine rhythms in fish. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 55-68. Smirnov, A. I. 1975. The biology, reproduction and development of the Pacific salmon. Izd. Mosk. Univ., 335 p. (Can. Fish. Mar. Serv. Transl. Ser. No. 3861, 1976). , and M. S. Kamyshnaya. 1965. The biology of young pink salmon in relation to certain questions of their propagation and acclimatization. Zool. Zhurn., 44: 1813-1824. (Fish.. Res. Board Can. Transl. Ser. No. 782, 1966). Smith, G. P., and J. Gibbs. 1976. What the gut tells the brain about feeding behavior. Life Science Research Project, 2: 129-139. Sokal, R, R., and F. J. Rohlf. 19,69. Biometry, The principles' and practice of statistics in-biological research. W. H. Freeman and Co., San Francisco. '776 p. 239 Solem, J. 0. 1973. Diel rhythm of fry and young of Lota lota (Pisces). Oikos, 24: 325-327. Spencer, W. P. 1939. Diurnal activity rhythms in fresh-water fishes. Ohio J. Sci., 39: 119-132. Staples, D. J. 1978. The significance of locomotor and feeding rhythms in the regulation of population biomass and production of the upland bully Philypnodon breviceps Stokell. In: Rhythmic activity of fishes. Edited by J. E. Thorpe. Academic Press, Inc., New York. pp. 243-252. Stasko, A. B.' 1971. Review of field studies on fish orientation. Ann. N.Y. Acad. Sci., 188: 12-29. . 1975. Progress of migrating Atlantic salmon (Salmo salar) along an estuary, observed by ultrasonic tracking. J. Fish Biol., 7: 329-338. , R. M. Horrall, A. D. Hasler, and D, Stasko. 1973. Coastal movements of mature Fraser River pink salmon (Oncorhynchus gorbuscha) as revealed by ultrasonic tracking. J. Fish. Res. Board Can., 30: 1309-1316. Stasko, A. B., and S. A. Rommel, Jr. 1977. Ultrasonic tracking of Atlantic salmon and eels. Rapp. P.-v.- Reun. Cons. int. Explor. Mer, 170: 36-40.' Steel, R. G. D., and'J. H. Torrie. 1960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill Co., Inc., Toronto. 481 p. Stickney, A. P. 1972. The locomotor activity of juvenile herring (Clupea harengus harengus L.) in response to changes in illumination. Ecology, 53: 438-445. Stringer, G. E., and W. S. Hoar. 1955. Aggressive behavior of underyearling Kamloops trout. Can. J. Zool., 33: 148-160. Stuart, T. A. 1953. Spawning migration, reproduction and young stages of loch trout (S. trutta L.). Scottish Home Dept. Fresh Water Salmon Fish. Res., 5: 1-39. Subbaraj, R., and M. K. Chandrashekaran. 1977. "Rigid" internal timing in the circadian rhythm of flight activity in a tropical bat. Oecologia (Berl.), 29: 341-348. Sweeney, B., and J. W. Hastings. 1960. Effects of temperature upon diurnal rhythms. Cold Spring Harbor Symp. Quant. Biol., 25: 87-104. Swift, M. C. 1976. Energetics of vertical migration in Chaoborus trivittatus larvae. Ecology, 57:.900-914. 240 Tesch, F.-W. 1978. Horizontal and vertical swimming of eels during the spawning migration at the edge of the continental shelf. In: Animal migration, navigation and homing. Edited by K. Schmidt-Koenig and W. T. Keeton. Springer-Verlag, New York, pp. 378-391. Thines, G. 1970. Periodic variation in the photic behaviour of the tench (Tinea tinea L.) in self-regulating conditions. J. interdiscipl. Cycle Res., 1: 363-377. , F. Wolff, C. Boucquey, et M. Softie". 1966. Etude comparative de l'activite'du poisson cavernicole Anoptichtys antrobius Alvarez, et de son ancestre epige'Astyanax mexicanus (Fildppi). Ann. Soc. Roy. Zool. Belg., 96: 61-116. Thorpe, J. E. 1977. Daily ration of adult perch, Perca fluviatilis L. during summer in Loch Leven, Scotland. J. Fish Biol., 11: 55-68. ... (ed.). 1978. Rhythmic activity of fishes. Academic Press, Inc. , New York. 312 p. Toates, F. M., and J. Archer. 1978. A comparative review of motivational systems using classical control theory. Anim. Behav., 26: 368-380. Tomita, T. 1971. Vision: electrophysiology of the retina. In: Fish physiology. Vol. 5. Edited by W. S. Hoar and D. J. Randall. Academic Press, Inc., New York. pp. 33-57. Tugendhat, B. 1960. The normal feeding behaviour of the three-spined stickleback (Gasterosteus aculeatus). Behaviour, 15: 284-318. Tyler, A. V. 1970. Rates of gastric emptying in young cod. J. Fish. Res. Board Can., 27: .1177-1189. Tytler, P., J. E. Thorpe, and W. M. Shearer. 1978. Ultrasonic tracking of the movements of Atlantic salmon smolts (Salmo salar L.) in the estuaries of two Scottish rivers. J. Fish Biol., 12: 575-586. Varanelli, C. C., and J. D. McCleave. 1974. Locomotor activity of Atlantic salmon parr (Salmo salar L.) in various light conditions and in weak magnetic fields. Anim. Behav., 22: 178-186. Vernon, E. H. 1958. An examination of factors affecting the abundance of pink salmon in the Fraser River. Intl. Pacific Salmon Fish. Comm. Progr. Rept.' No. 5: 49 p. • . 1966. Enumeration of migrant pink salmon fry in the Fraser River estuary. Intl. Pacific Salmon Fish. Comm. Bull. No. 14: 83 p. 241 Vernon, E. H., A. S. Hourston, and G. A. Holland. 1964. The migration and exploitation of pink salmon runs in and adjacent to the Fraser River convention area in 1959, Intl. Pacific Salmon Fish. Comm. Bull. No. 15: 296 p. Voute, A. M., J. W. Sluiter, and M. P. Grimm. 1974. The influence of the natural light-dark cycle on the activity rhythm of pond bats (Myotis dasycneme Boie, 1825) during summer. Oecologia (Berl.), 17: 221-243. Wagner, H.-J. 1975. Quantitative changes of synaptic ribbons in the cone pedicles of Nannacara: light dependent or governed by a circadian rhythm. In: Vision in fishes; new approaches in research. Edited by M. A. Ali. Plenum Press, Inc., New York, pp.'679-686. , and M. A. Ali. 1977. Cone synaptic ribbons and retinomotor changes in the brook troutSalvelinus fontinalis (Salmonidae, Teleostei), under various experimental conditions. Can. J. Zool., 55: 1684-1691. Ware, D. M. 1972. Predation by rainbow trout (Salmo gairdneri): The influence of hunger, prey density, and prey size. J. Fish. Res. Board Can., 29: 1193-1201. . 1975. Growth, metabolism, and optimal swimming speed of a pelagic fish. J. Fish. Res. Board Can., 32: 33-41. . 1978. Bioenergetics of pelagic fish: theoretical change in swimming speed and ration with body size. J. Fish. Res. Board Can., 35: 220-228. Waterman, T. H. 1972. Visual direction finding by fishes. In: Animal orientation and navigation. Edited by S. R. Galler, K. Schmidt-Koenig, G. J; Jacobs, and R. E. Belleville. NASA, Washington, D.C. • pp. 437-456. Wehrmann, L. 1968. Messungen der Bewegungsaktivitat des Aales (Anguilla anguilla L.) unter Verwendung einer Markierung mit Magnetmarken. Dipl.-Arb. mathem. - naturwiss. Fakultatj. Universifeat Hamburg, Hamburg. Weihs, D. 1978. Tidal stream transport as an efficient method for migration. J. Cons. int. Explor. Mer, 38: 92-99. Wever, R. 1966. The duration of re-entrainment of circadian rhythms after phase shifts of the Zeitgeber. A theoretical investigation. J. theor. Biol., 13: 187-201. White, G. M. 1915. The behaviour of brook trout embryos from the time of hatching to the absorption of the yolk sac. J. Anim. Behav., 5: 44-60. 242 Wiepkema, P. R. 1971a. Positive feedbacks at work during feeding. Behaviour, 39: 266-273. . 1971b. Behavioural factors in the regulation of food intake. Proc. Nutrition Soc, 30: 142-149. Wilkins, M. B. 1965. The influence of temperature and temperature changes on biological clocks.. In: Circadian clocks. Edited by J. Aschoff. North-Holland Publ., Co., Amsterdam, pp. 146-163. Williams, J. A., and E. Naylor. 1978. A procedure for the assessment of significance of rhythmicity in time-series data. Intl. J. Chronobiol., 5: 435-444. Wilson, G. 1974. Pilot size trials of chum salmon, layer-planted, gravel incubation boxes utilizing upwelling flow. In: Proc. 1974 Northeast Pacific Pink and Chum Salmon Workshop. Richmond, B.C. pp. 12-20. Windell, J. T. 1966. Rate of digestion in the bluegill sunfish. Invest. Indiana Lakes Streams, 7: 185-214. Woodhead, A. D. 1975. Endocrine physiology of fish migration. Ann. Rev. Oceanogr. Mar. Biol., .13: 287-382. Woodhead, P. M. J. - 1957. Reactions of salmonid larvae to light. J. Exp. Biol., 34: 402-416. • . 1965. Effects of light upon behaviour and distribution of demersal fishes of the North Atlantic. ICNAF Spec Publ. No. 6: .267-287. - . 1966. The behaviour of fish in relation to light in the sea, Ann. Rev. Oceanogr. Mar. Biol., 4: 337-403. Wyatt, T. 1972. Some effects of food density on the growth and behaviour of plaice larvae. Marine Biology, 14: 210-216. Young, J. Z. 1935. The photoreceptors of lampreys. II. The function of the pineal complex. J. Exp, Biol., 12: 254-270. Zwahlen, R. 1975. Die lokomotorische Aktivitat des Eichhbrnchens (Sciurus vulgaris). Oecologia (Berl.), 22: 79-98. 243 APPENDIX 1. PERIODOGRAM ANALYSIS Periodogram analysis is a form of analysis of variance in which the relative variation of mean values in a time series about their grand mean is partitioned into contributions at equally spaced intervals within a predetermined range of period values. Periodogram analysis is based essentially on the truism that if one divides a time-series record containing purely periodic activity of period p^ into sections having the length of p^, then all sections will be identical and thus equal to the grand mean period p^. However, if the record is divided into sections not equal to p^, then individual sections will differ from one another and from their mean. Hence, differences between predetermined sections of the time series and their grand mean can be used as an indicator of the relative similarity between the true period of the data record and predetermined trial periods (sections). This similarity is expressed as the periodogram amplitude q(p). This amplitude is systematically evaluated for each trial period (p) examined, and is calculated by the equation of Dorrscheidt and Beck (1975). q(p) = 244 where = sample or data point recorded at a fixed interval (e.g., at 2-h intervals in this study). j = a specific data point (a) in the total time-series record of L data points. p = integer trial period for which q(p) is to be estimated, n = L/p, number of repetitions of the reference activity (section) with trial period p in the data record of length L. i = integer column number in the Buys-Ballot table or form estimate of the data (Enright 1965a). k = integer row number in the Buys-Ballot table or form estimate of the data (Enright 1965a). 1 np a = — ^ a^ , total mean activity n? "3=1 n m. — \ a. + (k- l)p , profile or mean activity for l n the itn column in the Buys-Ballot table generated with trial period p. The mean of all is m and is identical to the total mean a. Hence, we have 0 < q(p) < 1, where q(p) = 1 corresponds to purely periodic activity of period p. The periodogram consists of a plot of q(p) values agamnst a predetermined range of trial periods (p), as exemplified in Appendix Figure 3. It has been assumed in the past that the maximum estimated amplitude q(p) of the periodogram represents the best estimator of the period of a primary oscillation in the time-series record, provided that one exists (Enright 1965a, b). 245 However, peaks in the periodogram resulting from periodic components in the time-series must be distinguished from those peaks resulting from random fluctuations. To accomplish this, Dorrscheidt and Beck (1975) proposed a statistical procedure delimitating an upper statistical limit for these random fluctuations in the time series. Based on Gaussian distributed activity, they defimed Fs (fi , f2) (p is an integer) P(n - 1) + Fs (fx , f2) P - 1 where qg is the periodogram amplitude (q) expected with a probability of S%, and Fg the corresponding fractile of the F-dastribution with the degrees of freedom f^ = p and f2 = np. The upper confidence limit qg is computed independently for each trial period p examined. For the given analysis parameters (n, p), it can be stated that below any calculated qg value a corresponding q(p) is expected with a probability of S%. In this study S% = 95%. Hence, q(p) values in the periodogram that are equal to or greater than the calculated 95% confidence limit (see Appendix Fig. 3) are considered significantly different from random fluctuations ("noise"), and have a high probability of representing the true period of the rhythmicity, if one exists, in the time-series record. The limitations of the periodogram and the inherent difficulties in its interpretation have been discussed previously by 246 Enright (1965a, b). One major difficulty is the interpretation of multiple, significant peaks in a periodogram. Secondary peaks may be due to a periodic component that is either a multiple or a submultiple of the true period of the oscillation, which may be represented by a major peak in the periodogram (Enright 1965a). The resolution or sharpness of periodogram peaks increases with increasing length (L) of the time-series record (Enright 1965a, Dorrscheidt and Beck 1975, Williams and Naylor 1978). In this study three criteria were used to determine the major periodic component from a periodogram having several peaks. These are in the following order of prioriy 1) only. significant peaks were considered, 2) significant peaks that were closer to p = 24.0 h were considered over those which were further away from p = 24.0 h in instances when the fish were exposed to a 24-h LD cycle, and 3) if several significant peaks were similarly close to p = 24.0 h, then the one that was associated with the greatest difference between q(p) and q was chosen as the one most likely s representing the true period of the time-series data. In this study, trial periods (p) were tested at intervals of 0.1 h within the range of 4 to 39 hours. 247 APPENDIX 2. EVALUATION OF THE ACTIVITY MONITORS In the experiments described below, the fish experienced either a normal or a reversed LD 12:12 (600:6 lx) and a water temperature of about 10°C. Fish were acclimated to these conditions in holding tanks for at least two weeks before testing, and were acclimated to the channels for two days prior to visually and electronically recording their swimming activity. All evaluation tests at 6 and 600 lx were done between 0730 - 1930 hours under normal hours, D ^ hours) and reversed LD cycles, (JVonse t^ = 0730 onset = 1930 respectively. Fish were not fed because the presence of food modifies their swimming activity. A. •.Swimming behavior Visual observations revealed that individual salmon of various sizes typically swam in circuits in the channel. They exhibited low frequencies of 180° body turns and spent low proportions of time in a stationary position (Appendix Table 17). Some fish showed a strong directional (clockwise vs. counterclockwise) preference in their swimming behavior, while others exhibited weaker directional preferences (Appendix Table 17). Most fish appeared to spend more time swimming nearer the outer wall than the inner wall of the channel. No apparent differences were observed in these aspects of the swimming behavior of individual fish under different incident light intensities. These observations suggest that the design of the experimental channel permitted juvenile pink salmon to swim freely in circuits without physical obstructions, impeding or water currents directing their 248 movements. B. Influence of the infrared light beam on swimming activity There is much biochemical and electrophysiological evidence indicating that teleost fish do not perceive light of wave lengths greater than 700 nm (Brett 1957, Nicol 1963, Beatty 1966, Tomita 1971). However, to test the possibility that juvenile pink salmon might perceive and react to the two infrared light beams emitted across the channel, the following experiment was conducted. The swimming speed of individual fish, expressed as the number of channel revolutions per 10 min, was recorded for successive 10-min periods before, during, and after the operation of the light sources. This sequence of observation was repeated several times for each fish over a period of five to six days. The null hypothesis, that the swimming speed of individual fish was.equal during the three treatment periods (i.e., lights off, lights on, lights off), was tested. The experiment was replicated for several fish of various sizes at a low (6 lx) and a high (600 lx) incident light intensity. Appendix Table 18 shows that the null hypothesis could only be rejected at P < 0.05 for 1 out of 23 fish tested; the lone rejection occurred at the low light intensity. Further, fish were never observed to remain stationary in the infrared light beam for more than 1-2 s, nor to orient toward the source of the emitted light. Therefore, I conclude that the swimming speeds of juvenile pink salmon were not influenced by the two infrared light beams. However, the data should not be construed to indicate that pink salmon do not perceive infrared 249 light simply because their swimming behavior was unaltered during transmission of infrared light across the activity channels. Nevertheless, the data do not contradict published evidence indicating that the retina of teleost fish does not respond to infrared radiation. C. Sensitivity and accuracy of the activity monitors For the evaluation of the sensitivity and accuracy of the activity monitors and for all other experimentation, the triggering circuits were operated in the monostable mode rather than in the bistable mode. Theoretically, the probability of this circuit not closing when a fish interrupts one light beam in the monostable mode is less than when the fish has to interrupt both light beams in the bistable mode. Because of this possibility and also because the fish continuously swim in circuits in the channels with low frequencies of body turning and low proportions of time spent in a stationary position (Appendix Table 17), I think it justifiable to operate the triggering circuits in the monostable mode (only one infrared beam per channel), and retain the other matched pair of light source and phototransistor as a replacement in the event of electronic failures during operation. In evaluating the performance of the activity monitors, visual observations were made, for 10-min periods, of the number of interruptions of one infrared beam per channel by individual fish of various sizes. Concurrently, the number of electronic (monitor) counts or triggers was noted for the same observation period. These observations were made at incident light intensities of 6 and 600 lx. An estimate of triggering accuracy for each activity monitor during 250 each 10-min period of operation was calculated as the number of electronic counts expressed as a percentage of the concurrent number of visual counts or light beam interruptions. These percentage values of triggering accuracy were normalized by common logarithmic transformation. Since, in the majority of cases, no significant differences (P > 0.05) were observed between mean estimates obtained from the locomotor activity of individual fish within each of two distinct size-classes and between incident light intensities, the data were pooled under two classes of fish length for each of the four activity monitors. Final mean estimates of triggering accuracy for each monitor were obtained by summing the estimate of triggering accuracy for each 10-min observation period on all fish within each size-class and dividing by the total number of 10-min periods of observation conducted on fish of that size-class. Appendix Table 19 shows that the difference between the total sum of paired visual and electronic counts is significant (P < 0.05) in 3 out of 8 cases, namely for both the small and large size-classes of fish tested with Monitor A, and for the large size-class of fish tested with Monitor D. Closer inspection of the data for these latter three cases reveals that this difference, although significant, is only large in one case (small size-class of fish tested with Monitor A), and in the other two cases this difference was just significant. Therefore, in general, the data in Appendix Table 19 indicate a close correspondence between the total sums of paired visual and electronic counts for each activity monitor during ii number of 10-min observation periods. The mean triggering accuracy for the four activity monitors ranged from 94.0 251 to 102.6% (Appendix Table 19). The difference between the mean percentages of triggering accuracy obtained from small and large fish was only significant (P < 0.05, t-test) for Monitor D. Five out of the eight mean percentages underestimated the visual counts of locomotor activity or the number of light beam interruptions. Generally, this was due to the fish breaking the very top or bottom of the vertical light beam which occasionally resulted in the circuit not closing. Overestimation of the visual counts was generally due to the activation of electronic relays controlling the thermostat of each refrigeration chamber, which occasionally resulted in the simultaneous triggering of one or more of the activity monitors. Despite these shortcomings, the data in Appendix Table 19 indicate that the electronic data output from each of the four activity monitors is representative, within acceptable limits (94.0-102.6%), of the total momentary locomotor activity of individual pink salmon. Further, the data indicate little difference in the sensitivity of the activity monitors in detecting interruptions of the infrared light beams by juvenile pink salmon of lengths ranging from 4.3 to 19.0 cm. Hence, based on the behavior of the fish in the channel (Appendix Table 17) and the performance of the activity monitors (Appendix Table 19), I can state with confidence that, on the average, one electronic trigger or count from any one of the activity monitors is equivalent to one revolution of the activity channel by a swimming pink salmon whose length is within the above range. 252 Appendix Table 1. Total length and wet weight of individual adult female pink salmon captured on their spawning grounds in the Eve River, British Columbia during two brood years. Redd number is the simulated redd into which newly hatched sibling alevins, originating from eggs of a particular female, were introduced. The water temperature regime of each redd is shown in Table 1. Brood Redd Total length Wet weight year number (cm) (kg) 1976 1 52.5 1.25 1977 2 61.0 2.10 1976 3 57.0 1.45 1977 4 59.0 2.10 1976 5 51.0 1.25 1977 6 61.0 2.00 Appendix Table 2. Mean (95% confidence limits) total dry weight, dry body weight, and dry yolk weight of pink salmon alevins sampled about five days after median hatching time. Mean values are based on samples of 20 sibling alevins taken from each of six different groups incubated separately under similar temperature regimes from fertilization until hatching. Siblings from each of these six groups were introduced separately into six simulated redd. Redd Total dry weight Dry body weight Dry yolk weight number (mg) (mg) (mg) (% total dry weight) 1 56.03 9.25 46.71 83.41 (55.03 - 57.04) (8.93 - 9.58) (45.66 - 47.78) (82.76 - 84.10) 2 69.20 13.63 55.50 80.27 (67.97 - 70.45) (13.16 - 14.12) (54.33 - 56.70) (79.53 - 80.96) 3 67.24 10.72 56.50 84.07 (66.04 - 68.45) (10.34 - 11.10) (55.31 - 57.73) (83.46 - 84.71) 4 68.66 14.85 53.70 78.28 (67.44 - 69.90) (14.33 - 15.38) (52.56 - 54.86) (77.56 - 79.03) 5 60.92 10.23 50.67 83.20 (59.84 - 62.02) (9.88 - 10.60) (49.60 - 51.76) (82.51 - 83.86) 6 71.58 14.37 57.16 79.96 (70.31 - 72.87) (13.87 - 14.89) (55.95 - 58.40) (79.21 - 80.64) 254 Appendix Table 3. Efficiency of the plastic sheet and tube arrangement in preventing newly emerged fry from re-entering the gravel of a simulated redd at two temperatures. Twenty newly emerged pink salmon fry were placed in the water column above the plastic sheet in a redd containing no other fry. The fry were left undisturbed over one LD 12:12 cycle. At the end of this period the number of fry remaining in the water column above the plastic and the fry collected in the outflow container were counted. Temperature Date No. of fry Recaptures (°C) (1977) introduced (No.) (%) 5.0 Mar. 8 20 20 100.0 II 9 20 20 100.0 II 12 20 20 100.0 II 13 20 20 100.0 10.0 Jan. 27 20 20 100.0 II 28 20 18 90.0 II 29 20 20 100.0 Mean =98.6 SD = 3.8 255 Appendix Table 4. Incident light intensities (incandescent) during the light (L) and dark (D) phases of a LD 12:12 cycle at the water surface (ca. 8 cm above the gravel) and at various depths in gravel without water. Light intensities were measured with a Model IL 600 photometer coupled to a Model IL photomultiplier (International Light Inc.). Incident Incident light intensity Depth of gravel light intensity at the gravel surface (cm) (lx) (%) Water surface (L) 4.734 X 102 * (D) 1.074 X 101 * 0 (L) 3.551 X 102 * 100.00 (D) 9.340 X 10° * 100.00 2 (L) 2.789 X 102 78.54 (D) 3.623 X 10"1 3.88 5 (L) 1.599 X 101 4.50 (D) 9.575 X IO"2 1.03 7 (L) 1.380 X 0.39 (D) 2.450 X IO-2 0.26 9 (L) 2.518 X IO"1 0.07 (D) 7.700 X IO"3 0.08 11 (L) 1.018 X IO"2 0.0029 (D) 2.350 X IO"3 0.0250 13 (L) 8.877 X IO"3 0.0025 (D) 2.270 X IO"3 0.0243 3 15 (L) 8.670 X IO" 0.0024 (D) 2.220 X IO"3 0.0241 3 17 (L) 3.140 X IO" 0.0009 (D) 2.023 X IO"3 0.0217 3 20 (L) 2.150 X IO" 0.0006 (D) 1.170 X IO"3 0.0125 *, Measured with a Model 210 photometer (Photovolt Corp., New York) 256 Appendix Table 5. Estimated fish survival in gravel from hatching to emergence, and cumulative thermal units (C° - days) experienced by alevins from median hatching time until 50% emergence for each of six simulated redds. Redd Mean temperature Survival Thermal units number (°C) (%) (C° - days) 1 3.4 84.2 447 2 5.0 99.1 556 3 7.9 94.3 588 4 9.9 100.2 605 5 12.3 99.1 642 6 15.0 99.3 515 Appendix Table 6. Geometric mean (95% confidence limits) fork length, mean total dry weight, and weighted mean dry yolk weight of pink salmon fry emerging from a simulated redd at different mean temperatures. Data for fork length, dry fish weight, and absolute dry yolk weight were transformed by common logarithms, whereas data for yolk weight (%) were transformed by the arcsin. Mean temperature Fork length (mm) Total dry weight (mg) Dry yolk weight (N = 60) (°C) N Mean N Mean (mg) (% total dry weight) 3.4 232 32.26 172 31.57 0.00 0.00 (32. 17 - 32.26) (31.12 - 32.02) 5.0 326 35.00 326 51.56 1.64 3.39 (34. 92 - 35.08) (51.03 - 52.09) (1.42 - 1.89) (3.30 - 4.13) 7.9 185 33.22 125 39.76 1.30 3.33 (33. 12 - 33.33) (39.10 - 40.43) (0.99 - 1.72) (2.50 - 4.26) 9.9 322 34.38 322 49.51 5.07 10.63 (34. 30 - 34.46) (49.00 - 50.03) (4.34 - 5.93) (10.17 - 11.10) 12.3 327 31.38 327 31.07 2.05 7.04 (31. 30 - 31.45) (30.76 - 31.39) (1.65 - 2.54) (5.60 - 8.62) 15.0 475 32.29 475 54.42 17.92 33.90 (32. 23 - 32.35) (53.97 - 54.88) (16.79 - 19.13) (33.37 - 34.43) 258 Appendix Table 7. Mean fork lengths and mean dry weights of sibling pink salmon fry emerging from a simulated redd at different mean temperatures during the light and dark phases of the diel LD cycle on each of eight different days. All samples contained 20 fish. Mean Mean fork length Mean dry weight temperature Light Dark Light Dark (°C) (mm) (mm) (mg) (mg) 3.4 32.35 31,88 32.35 30.48 5.0 35.17 35.41 52.43 52.68 34.97 35.00 48.95 50.14 35.05 35.08 52.63 52.53 35.04 35.13 51.59 51.43 9.9 34.26 34.52 50.35 50.93 15.0 32.09 32.27 58.88 58.05 31.97 32.32 57.77 57.68 U = 29 a U = 32 3 df = 7 df = 7 P > 0.05 P > 0.05 Mann-Whitney U-test Appendix Table 8. Experiments conducted on six groups of six fish each at different times after emergence from gravel. Water temperature, total lengths and wet weights of fish are given for each experiment. Dates of Mean ± SD Mean ± SD Mean ± SD Group experiments Days after temperature total length wet weight iiumber (1977) emergence (°C) (mm) (mg) Ul 1 Feb. 16-21 1-5 9.6 ±0.7 36.2 ± 0.2 253 ± 5 2 Feb. 22-27 7-11 9.5 ± 0.4 36.0 ± 1.3 246 ± 16 ,3 Feb. 28 - Mar. 5 13-17 9.6 ± 0.5 38.5 ± 0.5 301 ± 20 4 Mar. 6-11 19-23 9.4 ± 0.6 36.4 ± 0.7 362 ± 33 5 Mar. 14-19 27-31 9.5 ± 0.4 42.7 ± 0.9 435 ± 34 6 Mar. 20-25 33-37 9.6 ± 0.4 44.0 ± 1.7 456 ± 47 260 Appendix Table 9. Summary of ontogenetic 01 soason.il chunks in the diel pattern of locomotor activity of name fish species recorded mainly in the laboratory. Changes in activity pattern an rolaLod to changes in habitat that the fish would normally experience under natural conditions at the time of study. Directions of habitat change s coinciding with phase shifts in the activity pattern are indicated. Time of Natural habitat Pattern Direction of year at time of study of activity habitat change Nocturnal Byrne (1971) Sockeye salmon 0-10 days Spring Stream (Oncorhvnchus nerka) Diurnal 10 d - 1 yr Spring - Spring Lake Rainbow trout 15-20 d (post-hatch) Winter - Spring Stream gravel Nocturnal (Salmo gairdneri) to emergence > 5 d after emergence Spring Stream water column Diurnal Nocturnal Atlantic salmon 20-25 d to emergence Winter - Spring Stream gravel (Saloo salar) > 5 d after emergence Spring Stream water column Diurnal Reynolds and Yellow bullhead Juvenile Stream and lake Crepuscular Casterlin (1978) (Ictalurus natalis) Adult Stream and lake Nocturnal 1 Darnell and Black, bullhead Juvenile Jun., Aug.-Sep. Schooling in open water Meierotto (1965) (Ictalunjs tnelas) of stream by day Adult Jun., Aug.-Sep. Hiding ITI vegetation of stream by day 1 Solera (1973) Burbot Fry May Littoral zone of lake Diurnal (Lota lota) Jul.-Aug. Deeper, offshore waters Nocturnal of lake 1 Staples (1978) Upland bully 0+ yr Mar.-Jul. Pelagic zone of lake Diurnal (Philypnodoa breviceps) Mar.-May Benthic zone of lake Notturncl (with diel onshore- o.ffshore migration) 011a and Tautog Mult Jan.-Nov. Inshore marine zone Diurnal (Tautoga onitls) Studholme (1978) Adult Dec. Offshore marine zone Nocturnal Gibson et al. (1978) Plaice Larvae Pelagic marine zone Nocturnal (Pleuronectes platessa) vertical migration Demersal intertidal zone Nocturnal circatidal rhythm Offshore Nocturnal circadian rhythm 1 Field study; S, Stream; L, Lake; C, Cravel; OW, Open water; MV, Marginal vegetation; IN, Inshore; HIT, Intertidal; OFF, OffBhore P, Pelagic; B, Benthic Appendix Table 10. Percent frequency of three species of calanoid copepod presented collectively as prey to experimental fish. Length and weight measurements for each prey species are based on samples of 100 prey. Frequency Mean total length Mean dry weight Prey species (%) (mm) (mg) OS Calanus plumchrus 76.2 4.30 0.577 ^ Calanus marshallae 15.7 2.90 0.169 Pseudocalanus minutus 8.1 1.39 0.023 Grand mean = 0.471 262 Appendix Table 11. Percentage capture success of ten juvenile pink salmon preying on live calanoid copepods. Attempts at No. of prey in Capture success Fish. prey capture the fish's stomach (%) 1 48 43 89.6 2 50 50 100.0 3 52 46 88.5 4 49 46 93.9 5 53 48 90.6 6 47 45 95.7 7 53 52 98.8 8 50 44 88.0 9 50 43 86.0 10 49 46 93.9 Mean = 92.4 SD = 4.9 Appendix Table 12. Estimated minimum swimming speeds of North American pink salmon in different locations and at different times of the year during their migrations in marine waters. Life history Direction Source Fish length Swimming speed - 1 phase Time of year Location of movement of data (cm) km day * BL s" Author(s) Juvenile coastal Late Aug. Chatham Strait NW Seining 15 5.6 - 7.4 0.37 - 0.50 Martin (1966) Mid Sep. Chatham Strait NW Seining 18 18.5 - 22.2 1.04 - 1.24 Martin (1966) Pelagic ocean Jul.-Sep. Gulf of Alaska NNW Tagging 10 - 25 18.5 - 22.2 0.86 - 2.14 Royce et al. (1968) Oct.-Dec. Gulf of Alaska S Tagging 20 - 30 18.5 + 0.71 - 1.08 + Royce et al. (1968) Jan.-Feb. Gulf of Alaska E Tagging 25 - 40 18.5 + 0.54 - 0.86 + Royce et al. (1968) K3 ON Mar.-Aug. Gulf of Alaska N Tagging 35 - 60 18.5 - 83.4 0.36 - 2.76 Royce et al. (1968) Jun.-Jul. North Pacific N Tagging ca. 44 54.3 1.43 Hartt (1966) 1 May-Aug. Gulf of Alaska ? Tagging ? 46.3 - 55.6 Neave (1966) Adult coastal Jun.-Jul. Bering Sea and N Tagging ca. 44 45.0 1.18 Hartt (1966) . Gulf of Alaska Aug. Northern Strait SE Tagging ca. 55 20.9 . 0.44 Vernon et al. (1964) of Georgia Aug. Central Strait SE Tagging ca. 55 11.7 0.25 Vernon et al. (1964) of Georgia Aug. Southern Strait NE Tagging ca. 55 38.9 0.82 Vernon.et al. (1964) of Georgia Aug. Central Strait NE Tagging ca. 55 26.3 0.55 Vernon et al. (1964) of Georgia Aug. Near mouth of NE Tagging ca. 55 2.8 0.06 Vernon et al. (1964) Fraser River Sep. San Juan Channel NW Sonic tracking 64 53.7 3 0.97 a Stasko et al. (1973) Active fish only 264 Appendix Table 13. Major predictions of the Aschoff-Wever model of circadian rhythms. 1. Increases in daylength result in advances of phase .j^) ln diurnal animals and in delays of phase in nocturnal animals. 2. Maximum leading (positive) phases occur in mid-summer for diurnal animals and in mid-winter for nocturnal animals; minimum leading phases occur at the equinoxes. 3. A positive correlation exists between activity time (a) and the duration of day and night for diurnal and nocturnal animals, respectively. 4. A positive correlation exists between a and advancing phase-angle differences .,) for diurnal and nocturnal animals, mid 5. A positive correlation exists between frequency (1/period) of the free-running rhythm, its a, and its phase 0^^) • 6. Increases in the duration of twilight result in advances in mmi .d, for both diurnal and nocturnal animals, 7. Diurnal animals have shorter spontaneous periods (x) in summer than in winter; the reverse is predicted for nocturnal animals. Sources: Aschoff (1960, 1967, 1969), Aschoff et al. (1970), Daan and Aschoff (1975), Gwinner (1975) Appendix Table 14. Dates and duration of experiments conducted on different groups of six fish each at different times of the year. Duration of the photophase (L) and scotophase (D) of the light-dark cycle and mean water temperature are given for each experiment. Mean length and mean weight of each group of fish are also given. Mean (± SD) Mean (+ SD) Mean (+ SD) Group Dates of experiments No. of Duration (h) temperature total length wet weight number (1975-76) days L D (°C) (cm) (g) 1 Apr. 8-15 7 13.0 11.0 12.4 + 0.3 5.2 + 0.4 0.7 + 0.2 2 Apr. 28 - May 5 7 14.0 10.0 8.9 + 0.5 7.3 + 0.4 2.2 ± 0.4 3 May 15-22 7 15.0 9.0 9.0 + 0.5 8.1 + 0.2 3.0 ± 0.2 4 Jun. 3-7 4 15.5 8.5 10.1 + 0.4 9.5 + 0.5 4.6 ± 0.6 5 Jun. 13-20 7 16.0 8.0 10.2 + 0.7 10.2 + 0.2 5.7 + 0.5 6 Jun. 28 - Jul. 4 6 16.0 8.0 12.4 i 1.6 10.6 + 0.5 6.8 ± 0.9 7 Jul. 12-19 7 16.0 8.0 12.4 + 0.4 11.7 ± 0.5 9.5 ± 1.4 8 Jul. 30 - Aug. 6 7 15.0 9.0 13.2 + 0.5 13.2 ± 0.4 14.3 ± 1.6 9 Aug. 10-14 5 14.75 9.25 12.5 + 0.6 14.5 + 0.3 19.4 ± 2.6 10 Aug. 20-27 7 14.25 9.75 12.1 + 1.1 16.3 ± 0.4 29.0 + 2.5 11 Sep. i-8 7 l-'i.O 10.0 12.3 + 1.0 17.1 + 0.5 32.9 + 2.2 12 Sep. 17-24 7 13.0 11.0 11.9 + 0.6 18.7 + 0.4 46.6 ± 4.5 13 Oct. 23-30 7 11.0 13.0 10.4 + 0.6 19.8 + 0.9 53.9 ± 7.7 14 Nov. 7-14 7 9.75 14.25 10.0 + 0.5 20.6 + 0.8 50.3 ± 5.1 15 • Nov. 24 - Dec. 1 7 3.75 15.25 9.6 + 0.4 22.0 + 0.5 71.4 ± 5.0 16 Dec. 16-23 7 8.0 16.0 9.6 + 1.4 22.6 + 0.5 85.0"± 6.3 17 Dec. 29 - Jan. 5 7 8.0 16.0 8.6 + 0.4 22.5 + 1.2 89.5 ± 17.0 13 Jan. 13-20 7 8.5 15.5 8.5 + 0.6 23.2 + 0.6 95.3 ± 11.0 19 Jan. 25 - Feb. 1 7 8.75 15.25 8.2 ± 0.4 24.9 + 0.8 120.0 ± 10.4 20 Feb. 18-25 7 10.0 14.0 9.2 + 1.3 25.5 + 1.3 125.4 ± 21.3 21 Mar. 3-10 7 10.5 13.5 9.1 + 0.6 27.8 + 0.8 166.6 ± 11.0 22 Mar. 17-24 7 11.5 12.5 9.1 + 0.4 28.5 + 0.2 176.5 ± 4.2 23 Apr. 3-10 7 12.25 11.75 11.1 2.0 30.2 + 1.1 202.1 ± 21.5 24 Apr. 21-28 7 13.25 10.75 11.1 + 1.4 31.3 + 1.1 226.9 ± 22.8 266 Appendix Table IS. Results of laboratory studies conducted on salmonU fishes under constant Illumination and usually constant temperature conditions. Results not showing a persistent rhythm in constant light (LL) or constant darkness (DU) arc excluded. Species Author(s) Swimming Activity Sockeye salmon 16 of 33 fish showed persistent rhythms for 5-6 days in DD. 30 of 35 fish Byrne (1968) (Oncorhynchus nerka) showed persistent rhythms for at least 3-4 days in LL before damping out Atlantic salmon Persistent rhythm for 5 days in DD All (1964) (Salmo salar) 2 of 9 fish showed persistent rhythms for a maximum of 5 days In DD Dill (1970) 5 of 30 fish showed evidence of rhychmicity in DD Richardson and McCleave (1974) 2 of 7 fish showed rhythmlcity for 5 days In DD before damping out Varanelll and McCleave (1974) Rainbow trout 2 of 6 fish showed a persistent rhythm for a maximum of 5 days In DD Dill (1970) (Salmo pairdneri) Steelhead trout Diurnal rhythms In LL and nocturnal rhythms in DD for 1 day Lichtenheld (1966) (Salmo gairdnerl) Brown trout Suggestion of a rhythm for I day in DD Charton (1969) (Salmo trutta) Free-running rhythms for ca. 10-18 days under natural conditions In nailer (1969b) June - July at 66*N Brook trout Suggestion of a rhythm for 1 day in DD McKenzle (1960) (Salvelinus fontinalls) Rhythms persisting for at least 1 day in LL before damping out MacAfee (1971) 2 fish showed free-running rhythm under natural conditions In Eriksson (1972) mid-summer at 66°N Retinomotor Movements Atlantic salmon Rhythm of retinal pigment thickness persisted for 1 day In DD before All (1961) (Salmo salar) damping out Brook trout Rhythmic movement of cones and pigment for 1 day in DD before damping out. Wagner and Ali (1977) (Salvellnus fontinalis) Rhythm of the relative number of synaptic ribbons in cone pedicles persisted for 3 days in LL and DD Oxygen Metabolism Atlantic salmon Rhythm In rate of oxygen uptake persisted for 1 day in DD before damping out All (1964) (Salmo salar) Suggestion of a rhythm In the rate of oxygen uptake for 4 days In DD Hlrata (1973a) Endocrine Physiology; Rainbow trout Rhythm in concentration of neurosecretory material "Goraori-positlve" in flieniarz (1974) (Salmo Irldeus) nucleus preopticus cells of the brain for 1 day in DD and LL Feeding Behavior Atlantic salmon Suggestion of a rhythm in DD All (1964) (Salmo salar) Rainbow trout 8-h rhythm in LL Adron et al. (1973) (Salmo galrdnerl) 267 Appendix Table U. Results of laboratory siuJU-s eomlm-ted on non-salmonW fishes under eonslnnt illumin.UK.it ami usually nmstant teoperature conditions. Results not showing a persistent rhythm in constant llftht (LI.) or constant darkness (DD) are excluded. Author(s) Species Swimming Activity Edel (1976) Eel Matured fish showed a persistent rhythm for 3 days In DD, but only for I day in LL before damping out (Anpullla rostrata) Wchrmann (1968, cited in Eel Free-running rhythms for at least 1 week in constant conditions Edel 1976) (Anpuilla anfiuilla) Swell shark Free-running rhythms persisting for 15 days in DD and for 18 days in LL Nelson and Johnson (1970) {Cephaloscyllium yentrlosurn) Herring Persistent rhythms for 2 days In DD and LL before damping out Stlckney (1972) (Clupea harengus harengus) Lake chub Free-running rhythms for 15-20 days in DD before damping out Kavaliers (1973b) (Couesius plumbeus) Bluefish Free-running rhythm for 5 days in LL Olla and Studholac (1972) CPomatomus saltatrix) Individual fish showed free-running rhythms in LL (1 lx) for at least Whitefish Muller (1976) (Coregonus lavaretus) 12 days before damping out. Schooled fish showed a persistent rhythm for at least 10 days in DD before damping out Cyprlnid Free-running rhythms for at least 13 days in LL (4 lx) and for at least Siegraund and Wolff (1972a, b, 1973) (Leucaspius delineatus) 6 days in DD before damping out Reynolds and C.isterlio (1976) Bluegill sunfish Evidence for a rhythm persisting for 1 day in DD (Lepoiols macrochirus) Free-running rhythms persisting for A days in LL Beitinger (1975) Characid Evidence for a free-running rhythm persisting for 7 days in DD Thines et al. (1966) C&styanax mexicanus) Labrid Free-running rhythm persisting for 32 days in LL Casimir (1971) (Labroldes quadrolineatus) Burbot Free-running rhythms persisting for about 10 days in natural conditions MGller (1969a) (Lota lota) In December at 66°N Free-running rhythms persisting for about 9 days in natural conditions Muller (1970a, 1973a, 1978a) in May - June at 66°N Goldflch Diurnal rhythms persisting for up to 7 days in LL Spencer (1939) (Carassius auratus) Drown bullhead 2 of 15 fish showed rhythmicity in DD. 9 of 12 fish showed rhythmicity Eriksson (1978a), (Ictalurus nebulosus) for ac least 20 days in LL with 15-min pulses of darkness every hour Eriksson and Veen (1979) Free-running rhythm persisting for 5 days in DD Roberts (1963) Plaice Newly-caught juveniles showed free-running circatidal rhythms for 1-3 days Gibson (1973, 1976) (Ple»ronectes platessa) in DD, These rhythms became circadian and persisted for another 5-10 days before damping out Fish previously held in LD showed circadian rhythms persisting for about Gibson (1976) 7 days in DD Adults shoved circadian rhythms of bottom activity persisting for 6 days Gibson et al. (1978) in LL and of surface activity for 6 days in DD Hogchcker Free-running circatidal rhythms persisting for 3 days in IL O'Connor (1972) (Trlnectes roaculatus) Sculpln Nocturnal rhythm persisting for 3 consecutive days under overcast skies Andreasson (1969) (Cottus goblo) in December at 66°H Sculpin Free-running rhythm under natural conditions in June - July at 66°N Muller (1970b, 1978a) (Cottu3 poecilopus) 268 Appendix Table 16 (cont'd) Swimming Activity Larvae showed briefly :i noeLurnal rhythm in LI. before damping nut Kleerekoper et al. (1961) (Petrumyzon marinus) Common sole Nocturnal rhythm persisting in DD for 2 days before damping out Kruuk (1963) (Solea solea) Shanny Free-running circatidol rhythm persisting for at least 3-5 days in LL and DD Gibson (1965, 1967) ^Blennius pholis) Tench Nocturnal rhythm persisting for 7 days In LL Thlnes (1970) (Tinea tinea) Retinomotor Movements Characid Rhythmic movement of retinal cones and pigments for up to 6 days in DD John and Haut (1964), (Astyanax mexicanus) John and Kaminester (1969) Bluefish Rhythm cf retinal cone movement persisting for 1-2 days In DD Olla and Studholme (1972) (Pomatonus saltatrix) Goldfish Rhythm of retinal cone movement persisting for 3 days In DD John et al. (1967) (Carassius auratus) Black bullhead Rhythra of retinal cone movement persisting for 4 days In DD Aiey and Mundt (1941, cited In (Ameiurus nebuiosus) John et al. 1967) Bluegill sunfish Rhythm of retinal cone and pigment movements for 1.5 days In DD John and Gring (1968) (Lepomis tnacrochirus) Cichlid Rhythm of the relative number of synaptic ribbons In cone pedicles Wagner (1975) (Sannacara anon; a la) persisting for 2 days in DD Schooling Behavior Bluefish Diurnal rhythm persisting for 5 davs In LL Olla and Studholme (1972) (Pomatoous saltatrix) Characid Rhythra of latency time for school formation persisting for i days in DD John and Haut (1964) (Astyanax mexicanus) Self-Selection of Light and Dark Pumpklnseed sunfish "17 of 26 fish showed a diel rhythm of self-selection of light, but did not Colgan (J975) (Lepomis gibbosus) select distinct days and nights Tench Fish showed a diel rhythm of activity and rest phases during self-selected Thines (1970) (Tinea tinea) day and night, respectively Melanophore Movement Brook lamprey Larvae showed a rhythm of melanophore dispersion for several days Young (1935) (Latapetra planer!) (number unspecified) in DD Kllllfish Free-running rhythm of melanophore dispersion persisting for 5-21 days Kavr.llers and Abbott (1977) (Fundulus heteroclitus) in LL and DD Electric Organ Discharge Gymnotids Free-running rhythm persisting for up to 45 days In LL Schwassraann (1971b) (Cymnorha.-3pM.chthy_s and Hypopomus sp.) Sound Production Tigerflsh Crepuscular rhythm persisting for about 8 days In LL and DD before Schneidr.r (1964) (Therapon Jarbus) damping out 269 Appendix Table 16 (cont'd) Llfiht-Shock Reaction Davis (1962) Bluegill sunfish Rhythm In the duration of reaction to a light pulse in DD is suggestive of (Lepomis macrochirus) an Internal rhythm of anticipation of lights-on and (or) feeding time Eriksson (1978a) Brown bullhead Some fish showed a free-running circadian rhythm in their reaction time to Eriksson and Veen (1979) (Ictalurus ncbulosus) light stimuli that persisted for at least 20 days in DD Feeding Behavior Hirata (1973b) Goldfish Diurnal rhythm persisting for at least 5 days in LL and DD (Carassius auratus) Eriksson and Veen (1979) Brown bullhead 2 of 3 fish showed free-running rhythms with periods of 22-24 h for about (Ictalurus nebulosus) 20 days in LL interrupted with 15-min pulses of darkness every hour Oxygen Metabolism Livingston (1971) Cardinal fish Free-running rhythm in rate of oxygen uptake persisting for 3 days in LL (Phaeoptyx conklini, Astrapogon and Apogon sp.) Mucus Secretion Labrid and Scarid sp. Free-running rhythm of mucus secretion from the opercular gland persisting Casimir (1971) for at least 32 days In LL Gonadal Development Baggerman (1972) Threespine stickleback Dally rhythm of photosensitivity indicated by the photoperiodic response (Gasterosteus aculeatus) of the gonads Chan (1976) Medaka Daily rhythm of photosensitivity indicated by the photoperiodic response (Oryzias latipes) of the ovaries of female fish Lipid Deposition Golden topminnow Evidence for a circadian rhythm of fattening sensitivity to prolactin that Meier (1970) (Fuadulus chrysotus) persisted for several days in LL Gulf killifish Evidence for a circadian rhytha of fattening sensitivity to prolactin that Meier et al. (1971, (Fundulus grandis) persisted for 4 days in LL In Meier 1975) 270 Appendix Table 17. Percentage of total time individual pink salmon were nbsorved to be stationary (S) or svimming in clockwise (C) and counterclockwise (CC) directions in a circular channel. Mean number of 160° turns of the body per 10 min is also presented for each fish. Fish tested under incident light intensities of 6 and 600 lx are indicated by D snd L, respectively. Total length of fish Total observation time % of total t.te Mean no. of 180" (cm) (min) C CC S turns per 10 min 4.3 (L) 240 41.63 58.27 0.10 1.7 4.8 (D) 240 91.70 7.50 0.80 1.7 4.9 (D) 230 68.31 30.28 1.41 1.6 4.9 (1.) 250 76.22 23.64 0.14 2.0 5.0 (L) 200 14.80 84.54 0.66 2.6 5.1 CD) 150 98.84 0.88 0.28 0.8 5.3 (D) 240 99.69 0.24 0.07 0.7 5.3 (L) 250 12.06 87.87 0.07 0.9 6.3 (L) 290 96.05 2.72 1.23 4.1 6.4 (L) 290 94.50 5.13 0.37 4.5 6.5 (L) 260 16.39 83.23 0.38 3.0 6.8 (L) 290 99.64 0.19 0.17 1.8 6.8 (D) 280 5.36 93.63 1.01 4.6 13.2 (L) 300 2.23 95.62 2.15 6.7 13.8 (D) 240 15.10 84.10 0.80 3.4 14.5 (L) 300 59.81 39.84 0.35 2.4 15.0 CL) 240 2.6!: 93.44 1.91 5.4 15.6 CL) 300 1.36 98.23 n.41 2.9 16.3 CL) 210 5.00 94.61 0.39 5.8 16.5 CL) 300 3.57 96.11 0.32 6.5 16.6 CL) 270 1.70 98.23 0.07 2.9 17.3 CD) 250 33.35 64.91 1.74 10.2 17.5 CD) 230 31.21 68.71 0.08 11.4 18.0 CD) 260 12.62 87.06 0.32 1.1 18.0 CD) 250 90.71 9.22 0.07 1.6 19.0 CD) 200 11.09 88.65 0.26 12.3 19.0 CD) 250 34.67 65.13 0.20 2.8 19.0 CD) 250 35.96 63.21 0.83 8.7 271 Appendix Table 18. Locomotor activity (Lor.nl number of channel revolutions) of individual pink salmon recorded tor 10-min periods before, during, and after the operation of the activity monitor at two incident light intensities (N is th e number cf replicates of the above sequence of observations). Incident light Total length intensity of fish Monitor operation 7 a Before During After N P (lx) (cm) *r 6 4.8 338 337 316 0.600 10 > 0.70 4.9 234 225 241 1.550 10 > 0.30 5.1 264 242 227 0.350 10 > 0.80 5.3 274 254 254 3.150 10 > 0.20 13.8 220 236 231 0.350 10 > 0.80 17.3 257 243 230 3.200 10 > 0.20 17.5 235 199 227 6.450 10 < 0.05 18.0 236 234 229 0.350 10 > 0.80 18.5 213 235 233 2.450 10 > 0.20 19.0 161 123 137 4.050 10 > 0.10 19.0 140 111 115 1.400 10 > 0.40 19.0 222 247 239 2.600 10 > 0.20 600 4.3 281 303 306 1.850 10 > 0.30 4.9 330 306 319 3.800 10 > 0.10 5.1 225 222 225 0.450 10 = 0.80 5.3 242 235 237 0.800 10 > 0.50 13.2 353 333 335 5.450 10 > 0.05 14.5 206 199 214 2.600 10 > 0.20 15.0 368 346 371 0.278 11 > 0.98 15.6 228 241 239 0.150 10 > 0.99 16.3 157 169 164 2.699 7 > 0.20 16.5 182 184 193 0.150 10 > 0.99 16.6 314 314 328 1.711 9 > 0.40 a Calculated from Friedman's two-way ANOVA Appendix Table 19. Comparisons of total sums of visual (V) and electronic (E) counts of the number of light beam interruptions by small and large pink salmon, and of estimates of mean triggering accuracy (E counts expressed as a percentage of V counts) for four identical photoelectric activity monitors. Each monitor was tested for n number of 10-min periods. Data were pooled from observations at incident light intensities of 6 and 600 lx. Monitor A Monitor B Monitor C Monitor D V E V E V E V E V E V E V E V E Total sum of counts 2244 2135° 781 753" 1127 1133 693 679 1087 1110 1001 985 1338 1315 1157 1172" Mean accuracy, %" 94.0 96.8 100.9 98.6 102.6 98.2 97.7 101.4 (91.6-96.4)c (94.2-99.5) (99.5-102.3) (96.2-101.2) (99.5-105.7) (94.8-101.6) (95.7-99.8) (99.8-103.0) 41 n 40 37 40 40 40 39 40 Total fish length, cm 5.0-5.8 14.0-19.0 4.3-5.3 14.5-19.0 4.8-5.3 13.0-19.0 4.9-5.7 13.4-18.0 Fish width, cm 0.71-0.82 2.14-3.0 0.54-0.82 2.46-3.0 0.67-0.69 2.09-3.16 0.62-0.69 2.12-3.10 •Difference between paired visual (V) and electronic (E) counts is significant (P < 0.05; Wilcoxon matched-pairs sign-ranks test (Siegel 1956)). 'Geometric mean of logarithmic transformed data. '95% confidence limits. 0) 300 340 380 420 460 500 540 580 620 660 700 740 780 820 Wavelength (nm) Appendix Fig. 1. Power spectrum of a General Electric tungsten filament lamp (CIE - 3400°K). 274 Appendix Figure 2. Relation between mean or median age at emergence (in days post-hatch) of fry of various salmonid species and mean or median incubation temperature. Legend: Oncorhynchus gorbuscha (1: Bailey et al. 1979, 2: Bailey et al. 1976, 3: Blackett 1974, 5: This study); Oncorhynchus keta (4: Koski 1975, 6: Wilson 1974); Oncorhynchus klsutch (7: Mason 1976, 8: Koski 1966, 9: Shapovalov and Berrian 1940); Oncorhynchus tshawytscha (10: T. A. Heming, unpublished data, 11: Gebhards 1961, 12: Shelton 1955); Oncorhynchus nerka (13: Ginetz 1976, 14: Blackett 1974); Salmo gairdneri (15: Dill 1979a, 16: Shapovalov 1937, 17: Dill 1979b); Salmo salar (18: Dill 1979a, 19: Dill 1979b); Salmo trutta fario (20: Geiger and Roth 1962); Salvelinus fontinalis (21: Dill 1979a). The regression line is Y = 132.36 - 7.61 X (r = -0.85, P < 0.001). Age at emergence (days) •o ^ O" 00 O KJ o ooooooooo I I I I I I I I—I—I—I—I—I—I—I—I—I 1.0 -i 0.8 J 0.6 H 3 0.4 H Q. E < 0.2 J 0.0 14 1 9 24 2 9 34 39 Period (h) Appendix Fig. 3. Periodogram of the swimming activity rhythm of one pink salmon (17.0 cm, 30.2 g) recorded for five consecutive days in LD 12:12 (600:0.2 lx) and 10°C. A, major significant periodicity at 23.9 h; B, component at 32.3 h, multiple of A; C and D, components at 15.9 and 8.0 h, respectively, sub-multiples of A; E, 95% confidence limits. 277 Parker & Vanstone (1966) Burke Channel, B.C. o a May 23-24, 1963 N = ioo o u « E 1 1 1 1 1 1 r o 0500 0900 1300 1700 2100 Parker (1969) S Fougner Bay, B.C. >> "2 ° May 27-28, 1966 2 2 1 N=30 TO o o u (0 E o Time of day (PDST) Appendix Fig. 4. Diel and tidal variations in mean number of prey or in mean (± SD) relative wet prey biomass in the stomachs of juvenile pink salmon. Collections were made in different coastal marine areas of the west coast of North America. Raw data for B and C were obtained from the original papers. The time scale is local time at each study site. Numbers near the means indicate where sample sizes differ from the marginal value (N). Meaning of horizontal bars and arrows as in Fig. 18. o o o 16 JZ h 14 a) (0 h 12 j= a o h 10 o o 14 -, L 8 o~ o o o w 13 - o 3 12- •I (111 •*-> 03 N3 v. 1 1 - oo 0) a E 10 - o> 9 - c II •ft to 8 - Mi o 2 7 - •"A'M'J'J'A'S'O'N'D'J'F'M'A1 1975 1976 Time of year (months) Appendix Fig. 5. Mean (± SD) temperature of sea water in the experimental tank and duration of experimental photophase (approximating daylength at 50°N) at different times of the year during the experimental period. Phototransistor circuit Triggering circuit is v dc 75 0 _T\. Ri -3 -1.9 v R4 CN ±6, to iOA202 CN 6 V cs CO OA202 r-r^ < R, z Appendix Fig. 6. Circuit diagram of the phototransistor, amplifier, and emitter follower transistor.. All components are duplicated for each of the eight phototransistors. All resistors have a tolerance and a wattage of ± 10% and 0.5 W, respectively. RT. = 120 kfl; R2 - 22 kfi; R3 - 100 kfi trimpot; R4 = 470 kfl; R5 = 10 kfi; R6 = 3.3 kQ; R7 = 100 kfi; R8 = R9 = 1 kfi; Ri0 = 68 kfi; Qi = 50 uF, 25 V DC; C2 = 100 yF, 20 V DC; C3 - 100 uF, 15 V DC; C4 = 50 uF, 25 V DC; C5 - 0.0047 uF, 100 V DC; Qx = OCP 71; Q2 • 2N 697. Triggering Junction Digital Power supply amera Amplifier —>- box counter i 6 V DC ci rcuit u •o t > 00 I Infrared Event Phototransistor Timer light source recorde r T 120 V AC Appendix Fig. 7. Block diagram of the phototransistor, associated electronics, power supplies, and recorders. Broken lines indicate the path of the signal generated when the fish interrupts the infrared light beam illuminating the phototransistor. -IB VDC Intarf ac« Appendix Fig. 8. Circuit diagram of the triggering circuit for the activity monitor. All components are duplicated for each of the four monitors. All resistors have a tolerance and a wattage of ± 10% • and 0.5 W, respectively. Rx = R21 = 680fi; R2 = R2o - 750Q; R3 = R19 = 180 kfi; . - R18 =12 kn; ;R5 •- R16 - R30 • 10 kn; R6 = R17 R31 = 1 kn; R7 = R15 = 82 kn; R8 = R11+ = 27 kfi; R9 = R10 - 2.7 kn; R R . U - 12 = 39 kn; R13 = R25 = 330 Q; R22 = 1000 kn; R23 = R26 = 100 kn; R21t = 100 n; R27 = 15 kn; R28 - 220 kn; R29 - 33 kn; R32 - 1.5 kft; R33 • 5 kn; Cx = C10 = 0.01 uF, 400 V DC; C2 - C5 • C6 - C9 - 50 yF, 25 V DC; C3 - C8 - 0.1 yF, 50 V DC; Ck - C7 - 0.003 yF, 400 V DC; Cn - 10 yF, 20 V DC; C12 - 0.5 yF, 400 V DC; C13 = 1.0 yF, 40 V DC; = 1000 yF, 30 V DC; C15 - 10 yF, 25 V DC;. C16 - 0.01 yF, 150 V DC; C17 = 500 yF, 30 V DC; Lj = L2 = 1 mHenry; Ti - Hammond 116 G36; Q1 > Q2 = Q3 = - 2N1 305; Q5 = 2N4 851; Q6 = MPS 6518; Q7 = 2N3 053; Q8 = 2N3 612; Dx = D2 = D3 = IN 34A; D4 = IN 4005; Ki = KHU 17D11, 24 V DC.