Torleif Eriksson Umeå 1988
Department of Animal Ecology University of Umeå
S-901 87 Umeå, Sweden AKADEMISK AVHANDLING
som med tillstånd av rektorsämbetet vid Umeå universitet för erhållande av filosofie
doktorsexamen, framlägges till granskning fredagen den 12 februari 1988, kl 10.00 i Hörsal C, LUO, (Johan Bures väg).
Examinator: prof. C. Otto, Umeå
Opponent: Dr. Kees Groot, Nanaimo, Canada
S-901 87 Umeå, Sweden Date of issue Fe b r u a r y 1988 Author Torleif Eriksson
Migratory behaviour of Baltic salmon (Salmo salar);
T itle adaptive significance of annual cycles.
A b stra c t
This thesis evaluates the adaptive significance of annual cycles on the migratory behaviour of Baltic salmon (Salmo salar L.). The studies have included field experiments as well as laboratory studies gf maturity and migratory behaviour patterns of smolts and postsmolts mainly from the Angerman river population.
Contrasting to the phenotypical elasticity in life-history traits, Baltic salmon was found to have a rather strict temporal organization of their annual behavioural patterns. Two year old smolted Baltic salmon showed drastic differences in migratory behaviour when compared in tanks containing either fresh or brackish water. Freshwater kept fish showed an annual cycle where downstream displacement in the upper water column was followed by a stationary behaviour, indicating a readaption to a freshwater life. Fish in brackish water behaved as a migratory fish throughout tne study. Baltic salmon also showed differences in maturation patterns in fresh and brackish water. Three summer old males detained in
freshwater all matured sexually the following autumn. If transferred to sea and kept in net- pens a low proportion matured, mainly previously matured males. Furthermore there was a size- dependent relationship of sexual rematuration. Many small previously matured males did not migrate, similarly small previolusly matured males were unable to respond to shifts in the environment. With larger size the alternatives of sexual maturation and high growth rate wi more related to the environmental conditions the fish experienced.
A hypothesis has been tested assuming that Baltic salmon migration is influenced by an annual time program. According to the hypothesis the migratory distance covered in the Baltic should be a result of a migratory activity sequence rather than a definite goal orientation.
Fish detained before release generally showed a shorter distance between release point and area of recapture compared to fish released at normal time of smoltrun. The migratory distance appeared to be inversly related to the period of delay before release. A seasonal difference in migratory propensity was recorded.
An instantaneous mortality curve for Baltic salmon during seaward migration and early sea-phase was estimated based on recapture data on sequentially related fish. The weekly mortality rate decreased from an initial mean of 271 during onset of migration, to 3.51 in mean during autumn. The high risk of mortality during the first part of migration indicated that strong selective forces act on a precise seasonal timing of migration.
were
Key words
Migratory behaviour, annual cycles, Baltic salmon, Salmo salar, annual time program, swimming behaviour, mortality, maturity, life history.
Language English ISBN 9 1 - 7 1 7 4 - 3 2 2 - 7 Number of pages 86
Signature
ÿtfn
to by their Roman numerals.
I. Eriksson L-0., Lundqvist H., Brännäs E. and Eriksson T 1982. Annual periodicity of activity and migration in the Baltic salmon , Salmo salar L., p. 415-430. In Müller K. ed. Coastal research in the Gulf of Bothnia. W.Junk/ the Hauge.
II. Eriksson T., Eriksson L-0. and Lundqvist H. 1987.
Adaptive flexibility in life-history tactics of mature male Baltic salmon parr in relation to body-size and environment. Am. Fish. Soc. Symp. 1: 236-243.
III. Eriksson T. 1984. Adjustments in annual cycles of swimming behaviour in juvenile Baltic salmon in fresh and brackish water. Trans. Am. Fish. Soc. 113:467-471.
IV. Eriksson T. 1988. The effect of release time on the migratory behaviour of Baltic salmon (Salmo salar):
influence of an annual time program. Manuscript submitted for publication.
V. Eriksson T. 1988. Mortality risks of Baltic salmon (Salmo salar L.) during downstream migration and early sea-phase; effects of body-size and season. Manuscript
Papers no I, II and III are reproduced by permission of the journals concerned.
Introduction
Migration generally results in a displacement in time and space. The term is frequently used, but hard to define satisfactory since migration has different meanings within different research fields (Baker 1982). I have adopted the definition by Dingle (1980); "Migration is a specialized behaviour especially evolved for the displacement in space".
This definition is rather broad but the term "migration" is restricted to events with an adaptive value.
Why do fish migrate? From an evolutionary point of view, the question can be explained by a higher fitness in migrating versus non-migrating fish. A high fitness can be reached through reduced mortality and/or increased growth rate. Ultimately this will lead to more offspring produced in response to long-term predictable patterns of
environmental changes.
In an ecological time scale it is natural to focus on proximate factors, revealing the mechanisms responsible for different behaviours. Immediate changes of environmental stimulii (e.g. daylength, temperature, food conditions) trigger and/or synchronize hormonal changes resulting in migratory behaviour (Pianka 1978).
Baker (1982) concluded that the most successful
individuals in terms of number and quality of offspring they produce (i.e. reproductive success), will be those with a path that as its lifetime track, manifest the individual’s solution to spatial and temporal problems and possibilities.
Animals living at northern latitudes have to cope with pronounced, often predictable changes in their environment.
Generally seasonal differences of the environment increase with higher latitude. Animals adjusting their use of
energetic resourses in agreement with the environmental demands should be at an advantage. Thus, adaptions to seasonal changes must have been of vital importance to animals inhabiting temperate and arctic zones (Gwinner 1981).
The anadromous Baltic salmon (Salmo salar L.) exhibits two major habitat shifts during its lifecycle. Adult salmon enter the rivers in autumn to spawn. After 1-4 years in freshwater, the majority of the juveniles leave their riverine environment and migrate to their feeding areas in the central Baltic proper. During the migration from the freshwater habitat to the sea the mortality risks are very high (Carlin 1969, Lindroth et al. 1982, Larsson 1984).
Thus, I predict that strong selective forces act on timing of juvenile migration of the Baltic salmon.
In this thesis I evaluate the adaptive significance of annual cycles on the migratory behaviour of Baltic salmon.
Two main questions are put forward: First the flexibility in life-history options of Baltic salmon is considered. The interactions between phenotypical plasticity and predisposed behaviours are discussed on an annual basis. Second, the significance of an annual time program in the migratory behaviour of Baltic salmon is discussed. The costs and benefits of migration are analysed.
The basis of the thesis is the evaluation of a number of successive experimental releases of Baltic salmon carried out in the years 1980-82. In addition, migratory behaviour in terms of locomotory activity and swimming behaviour have been studied experimentally in the laboratory.
Summary of papers
Annual periodicity of activity and migration in the Baltic salmon, Salmo salar L. (1).
One of the major environmental variables to which most living organisms have been exposed throughout their
evolution is the day-night cycle. The adaptive response of many animals to the temporal structure of the environment has been manifested in pronounced rhythmic patterns. This
paper discusses important ecological and physiological factors effecting the initiation as well as the pattern of migration in smolts of the Baltic salmon. The seasonality of annual locomotory activity patterns in parr and smolts is presented. It is concluded that photoperiod and temperature largely influence the activity patterns in juvenile Baltic salmon. The photoperiod synchronizes the basically
crepuscular circadian activity patterns, while temperature and season affect activity levels. In contrast to parr, smolts showed an activity overshot at the period of smoltrun in June at temperatures between 9-13°.
The seasonality of smoltification and migration in salmonids is summarized. Seawater adaption is an important part of smolting. Smolting fish adopt to seawater throughout spring prior the time of migration. If the smolts are
retained in freshwater beyond the period of the natural smolt run the salinity tolerance disappears. It is concluded that physiological and behavioural changes related to
smoltification, and subsequently to seaward migration are adaptive. These are basically reversible and reocurring circannual processes, mainly timed by the annual changes in the photoperiod.
Evidence for an innate circannual rhythm in salmon is discussed. An internal timing process related to condition factor and silvering was exhibited by Baltic salmon parr when kept for nearly two years under constant conditions of
light, water temperature and food. The circannual rhythm had a period length of about 10 months. A circannual timing program affecting physiological and behavioural processes related to season, is also assumed to affect the migratory behaviour of salmon. The migratory track of salmon is
hypothesised to be more a result of a temporal organization of migratory activity and not so much the result of a
definite goal orientation. Thus if a salmon, that has initiated its migration, is prevented from migrating the fish should show a reduced distance travelled.
Adaptive flexibility in life-history tactics of mature male Baltic salmon parr in relation to body-size and environment
(II).
The importance of size and sexual maturation (mature male parr vs immature parr) on behaviour, growth and survival of individually tagged two-summer-old fish Baltic salmon (Salmo salar L.) were studied in two populations from northern Sweden.
Data on growth, sexual maturity and recapture rate were gathered from fish; (1) released in the river at time of normal smolt run, (2) transferred to brackish water at time of smolting and subsequently released, and (3) transferred to brackish water at time of smolting and subsequently reared in net-pens for one summer.
Recapture rate of previously mature males released in the river were low, about five times below that of immature fish of similar size. Electro-fishing during the autumn in the release stream indicated that a majority of the two year old salmon parr remaining in the stream were sexually mature males. Recapture rates of previously mature parr released in the brackish water were indistinguishable from those of immature fish of the same size. In fish retained in net pens in brackish water, only a small fraction of all males
matured sexually after the first summer. The incidence of sexual maturation was highest in previously mature small males. Growth rate was comparatively high among previously mature males that did not remature. These males showed a higher instantaneous growth rate than immature fish.
A discrete size-dependent relationship to sexual
rematuration among previously mature males strongly affected their life-history. Many small previously matured males did not migrate, and they were unable to respond to shifts in
the environment. With larger size the alternatives of sexual maturation and high growth rate were related to the
environmental conditions the fish experienced. If fish reached the sea, the majority allocated their energetic resources to somatic growth. On the other hand, when
retained in fresh water, males showed a capacity to mature at all sizes, indicating a flexibility in the choice of life-history traits.
Adjustments in annual cycles of swimming behaviour in juvenile Baltic salmon in fresh and brackish water (III).
An experimental study in tanks containing either fresh or brackish water compared the vertical distribution and swimming direction of 2-year old smolted Baltic salmon (Salmo salar). The fish were transferred to the tanks at normal time of smoltrun i.e. at the end of May, and registrations were performed during one year.
Fish in both fresh and brackish water were distributed in the upper part of a water column, and they mowed with the current during the time May to August. Salmon kept in
brackish water continued this swimming behaviour throughout the year whereas fish maintained in freshwater changed their behaviour drastically. From September, freshwater fish were distributed close to the bottom, and they tended to swim against the current. The following spring (April-May),
freshwater kept fish again started to swim in the upper part of the column, indicating a repetetition of the smolt
behaviour.
The results correspond well with the behaviour of salmon in the nature. Apparently there is an interaction between alternative circannual programs in Baltic salmon.
The selection of a specific program is suggested to be based in part on the environmental information the fish perceive.
The effect of release time on the migratory behaviour of Baltic salmon (Salmo salar): influence of an annual time program (IV).
A field experiment testing a hypothesis that migration of Baltic salmon is influenced by an annual time program affecting the migratory pattern during the sea phase was performed. According to the hypothesis, the migratory distance covered in the Baltic should be a result of a migratory activity sequence rather than a definite goal orientation.
During three years, 1980-82, a total of about 18.000 individually tagged two year old hatchery-reared Baltic salmon smolts were released. At the time of normal smoltrun smolts of the Ångerman river stock were transferred to the sea and kept in netpens located north of the river mouth.
Each year there were two groups of delayed releases in addition to control groups released in the river and in the sea at normal time of smoltrun. These groups of fish were released 2-6 months after transfer to the sea.
A seasonal difference in migratory propensity (i.e.
readiness for migratory activity) was apparent. Nettings carried out 2-3 weeks after release showed that a
considerable fraction of the salmon released, after a delay of 4-6 months, had lost their migratory propensity. These fish stayed in the vicinity of the release area during the following winter. This conclusion was supported by a high rate of recaptures reported by local fishermen at the place of release.
Fish that were detained before release showed a
generally shorter distance between release point and area of recapture compared to fish released at normal time of
smoltrun. The migratory distance appeared to be inversely
related to the period of delay before release. During all three years, salmon released with a 4-6 months1 delay were recaptured significantly closer to the release point than fish released in June.
The results did not reject the hypothesis, that the migratory behaviour of Baltic salmon is influenced by an annual time program. This program seems to affect the
propensity for migration and thereby the distance travelled.
Mortality risks of Baltic salmon (Salmo salar L.) during downstream migration and early sea-phase; effect of body- size and season (V).
Mortality risks in Baltic salmon during early migration was estimated through a sequential release experiment. Effects of fish size and time of release on survival rate were studied.
About 9000 two year old hatchery reared smolts from the Angermanälven stock were used. The fish were either released in the river at the time of normal smoltrun or transferred to netpens in the brackish Bothnian sea, where the fish were protected and given a surplus of food. Batches of
approximately 2000 individually tagged fish were released from the netpens at three occasions yearly for three years.
The first release from the netpens took place after an acclimatization period of 10-14 days. Two delayed releases were performed each year after periods of 2 to 6 months.
A protected transfer to the sea and an acclimatization period of about two weeks prior to release increased the recapture rate by 1.6 to 2.0 times compared to fish released in the river. Furthermore, fish that were released at a delay of 2-6 months had a 2.8 to 5.0 times higher recapture rate than smolts released in the river.
I found a strong positive correlation between size of the fish and recapture rate during all experimental years.
Mortality rates peaked during the downstream migration and entry in to the sea. The weekly risk of mortality during these first two weeks was estimated to be 27.8%. Thereafter the mortality risk declined rapidly, and it was found to be 6.1% per week during the following 8-9 weeks. From mid September until the end of November the mortality rate estimated was only 3.5% per week.
The mortality curve obtained by the release experiments stresses the high mortality risk for Baltic salmon during their downstream migration and first weeks in the sea.
Baltic salmon appear to migrate at a sub-optimal size with respect to survival. Despite the gain in survival during migration, because of larger size by a prolonged freshwater residency, this is considered be outweighted by an
accelerated growth in the sea.
Discussion
Age and size at smolting and sea-ward migration, duration of the sea-phase, age and size at sexual maturation vary
considerably within and between salmon populations (II, Thorpe 1986). In addition, males frequently mature as parr prior to sea-migration. Saunders and Schom (1986) reported 32 alternative life-history traits of a single Atlantic salmon stock.
In contrast to this phenotypical plasticity in life- history traits, at least the Baltic salmon seem to have a rather strict temporal organization of their behaviour (I, II). In fact the young salmon appear to have the opportunity to choose between its life-style options only during a short critical period of each annual cycle (II,III). At this
period in spring, the decision to migrate and grow or stay and mature, is taken in relation to signals from the abiotic and biotic environment.
During other parts of the year, Baltic salmon show a predisposed rather than opportunistic or flexible behaviour (III). This could be illustrated by the migratory behaviour of young Baltic salmon. During spring the parr gradually
"prepare" for a transfer to a life in the sea. This preparation, an integrated part of the annual cycle, includes behavioural, physiological and morphological
changes (Hoar 1976, Wedemeyer et al. 1980, Lundqvist 1983).
The preparation is not fulfilled by all parr. A variable proportion of small immature individuals and of previously matured male parr stay in the river for another year. Many small male parr participate in spawning (Österdahl 1969). I have demonstrated that a transfer of smolts to the sea during the accurate period results in enhanced survival and a low incidence of sexual maturation (II, V). Contrary, smolt sized parr retained in the river reversed to a
freshwater adopted fish and matured sexually (Lundqvist and Friberg 1982, Lundqvist and Eriksson 1985).
As the risk of predation is very high during migration (V, Larsson 1984), it seems legitimate to assume that strong selective forces act on a precise seasonal timing of
migration (IV). An annual timing device is proposed to be of adaptive value in several aspects. A contemporary migration and schooling behaviour generally reduce the risk of
mortality (Bertram 1978). Furthermore, an accurate timing could result in a smolt migratory period that coincides with optimal food supply in the sea. In a field experiment it was possible to demonstrate a reduced migratory distance if the salmon were prevented to migrate during their first summer in the sea (IV). This result is in agreement with the
hypothesis that the migratory behaviour of Baltic salmon is under the influence of an annual time program affecting the propensity for migration, and thereby the distance travelled
(I,IV). Since there is a size dependent mortality acting on Baltic salmon postsmolts during the early sea-phase (II,V),
a high growth rate during the early sea-phase will be highly advantageous.
Do the results in my thesis support the hypothesis of an annual time program affecting the migratory behaviour in Baltic salmon? Although the reduction of the migratory distance obtained was less than hypothesized, there is an significant overall negative correlation between distance travelled and and duration of the delay period prior to release (IV). Furthermore the observations of locomotory activity, swimming direction and vertical distribution in juvenile fish, strongly emphasize a seasonality in the migratory behaviour in Baltic salmon (I, III). The
advantages of a reliable timing device is also stressed by the mortality curve demonstrated (V). A contemporary
migration through a synchronised initiation of migration seems to have a significant selective value.
From this study, I can not distinguish between an
internal or external nature of the annual cycle in migratory behaviour. As circannual rhythms dependend on external cues to synchronize them with the natural annual cycle, the advantages of such a rhythm is not clear (Enright 1970).
Althought, external information varies considerably within and between years. If relying entirely on external cues, a year to year variation in timing of seasonal activities would be expected due to the variability in weather conditions. Thus precise timing of seasonal events is
assumed to be the most important trait of circannual systems (Gwinner 1981).
Acknowledgements
I am deeply indepted to my supervisor and friend Prof.
Lars-Ove Eriksson for his never-failing support during these years. His enthusiasm, helpful criticism and friendship is greatly appreciated.
As a member of the fish research group at the
Department I have always had the advantage of having someone eager to discuss fish ecology or give a helping hand. My sincere thanks to all of you. I especially want to mention Dr. Hans Lundqvist who always has been interested in
discussing experimental designs, statistical treatment of data and critical reading my manuscripts. You really are a good member of the team.
I express my sincere gratitude to all colleagues and friends at the department for creating such a friendly and stimulating atmosphere during these years. I thank Prof.
Karl Müller, former head of the department and Prof.
Cristian Otto, present head of the department for their encouragement during the study. Also I would like to thank Prof. Otto for critically reading some of my manuscripts as this improved them tremendously. Many thanks are also due to Göran Andersson, Gunnar Borgström and Mona Lisa Löfgren for invaluable technical assistance, Lena Andersson for typing some of the manuscripts and to Görel Marklund for skilfull preparation of the figures.
Without the help and interest of Göran and Per
Björkland the experimental releases at the Ulvö islands had been impossible to realize. With them and their families I have found friends for life.
Finally, I am especially indebted to my wife Ann and my sons Anton and Björn for their patience and support through out the study.
This work was supported by grants from the County Council of Västernorrland, the Swedish Natural Science Research Council and the Swedish Council for Forestry and Agricultural Research.
References
Baker R.R. 1978. The evolutionary ecology of animal migration. Holmes & Meier, New York.
Baker R.R. 1982. Migration, paths through time and space.
Hodder and Stoughton.
Dingle H. 1980. Ecology and evolution of migration, p. 1- 103. In S.A. Gauthreaux, Jr. (ed.) Animal migration, orientation and navigation. Academic press, London New York.
Enright J. T. 1970. Ecological aspects of endogenous rhythmicity. Ann. Rev. Ecol. Syst. 1:221-238.
Gwinner E.B. 1981a. Circannual systems, p. 381-389. In J.
Aschoff (ed.) Handbook of behavioural neurobiology, voi.4. Plenum.
Hoar W.S. 1976. Smolt transformation; Evolution, behaviour and physiology. J. Fish. Res. Board. Can. 33:1234- 1252.
Larsson P-0. 1984. Some characteristics of the Baltic salmon, Salmo salar L., population. Ph.D. thesis, University of Stockholm, Stockholm, Sweden.
Lindroth A., P-0. Larsson and G. Bertmar. 1982. Where does the Baltic salmon go, p. 387-414. In K. Müller, (ed.) Coastal research in the Gulf of Bothnia. Dr W. Junk pubi., the Hauge.
Lundqvist H. 1983. Precocious sexual maturation and smolting in Baltic salmon (Salmo salar L.); Photoperiod
synchronization and adaptive significance of annual biological cycles. Ph.D. thesis, University of Umeå, Umeå, Sweden.
Lundqvist H. and G. Friberg. 1982. Sexual maturation versus immaturity: different tactics with adaptive values in Baltic salmon (Salmo salar L.) male smolts. Can. J.
Zool. 60:1822-1827.
Lundqvist H. and L-0. Eriksson. 1985. Annual rhythms of swimming behaviour and seawater adaption in young
salmon, Salmo salar, associated with smolting. Envir.
Biol. Fish. 14:259-267.
Österdahl L. 1969. The smolt run in a small Swedish river, p. 205-215. In T.G. Northcote (ed.) Salmon and trouts in streams. H.R. MacMillan lectures in fisheries.
Univ. of British Columbia, Vancouver, B.C.
Saunders R.L. and G.B. Schom. 1985. Importance off the variation in life-history parameters of Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci.
42:615-618.
Thorpe J.E. 1986. Age at first maturity in Atlantic salmon, Salmo salar: freshwater period influences and
conflicts with smolting, p. 7-14. In D.J. Meerburg (ed.) Salmonid age at maturity. Can. Spec. Pubi. Fish.
Aquat. Sci. 89.
Wedemeyer G.A., R.L. Saunders and W.C. Clarke. 1980.
Environmental factors affecting smoltification and early marine of anadromous salmonids. Marine Fish.
Review. 42:1-14.
