Effects of sea-ranching and family background on fitness traits in ...

10 downloads 1105 Views 141KB Size Report
fitness traits in brown trout Salmo trutta reared under ... *Institute of Freshwater Research, SE-178 93 Drottningholm, Sweden; †Department of Population ...
Journal of Applied Ecology 2003 40, 241 – 250

Effects of sea-ranching and family background on fitness traits in brown trout Salmo trutta reared under near-natural conditions Blackwell Science, Ltd

JOHAN DANNEWITZ*†, ERIK PETERSSON‡§, TORE PRESTEGAARD* and TORBJÖRN JÄRVI* *Institute of Freshwater Research, SE-178 93 Drottningholm, Sweden; †Department of Population Biology and ‡Department of Animal Ecology, Evolutionary Biology Centre, Uppsala University, Norbyvägen 18D, SE-752 36 Uppsala, Sweden; and §Laboratory of Streamwater Ecology, National Board of Fisheries, SE-810 70 Älvkarleby, Sweden

Summary 1. Many threatened populations of salmonids depend on supplemental releases of hatchery-produced fish. Laboratory studies suggest that altered selection regimes in the hatchery may result in evolutionary changes of traits connected to fitness. Such changes can have profound effects on the performance of the hatchery fish following release in the natural environment, and may also affect the genetic characteristics of locally adapted wild populations. However, surprisingly few studies have looked at the ability of hatchery fish to compete with wild conspecifics under natural conditions. 2. We studied growth, survival and life-history adoption of a wild and a multigeneration sea-ranched strain of brown trout Salmo trutta in a semi-natural stream. The fish were planted in the stream as eyed eggs and their family and strain origins were later revealed by microsatellite markers. 3. In the first experiment, in which the experimental fish originated from a full-sib mating design, there were strong family effects on both growth and survival over the first growth season. In the second experiment, in which the experimental fish originated from a half-sib mating design, there were significant male and female effects on growth parameters but not on survival over the first growth season. 4. When family and male–female effects were accounted for, there were no differences between wild and sea-ranched trout in body size and condition factor after the first growth season, or in survival up to this stage. Nor was there any difference between the groups in the proportions that metamorphosed into the migratory smolt phase at 1 year of age. 5. Synthesis and applications. Our results suggest that wild-born trout of sea-ranched origin can successfully compete with trout of wild origin under semi-natural conditions. This indicates that the impact of hatchery selection on the performance of sea-ranched fish in the wild may not be as pronounced as previously thought. It is suggested that for salmonid populations that depend on supplemental stocking, more effort should be paid to minimizing negative environmental effects during hatchery rearing. The observed differences in fitness characters between families suggest that family effects should be taken into account in stocking programmes because the amount of genetic variation maintained within populations is related to the variance in family performance. Key-words: domestication, growth, life histories, microsatellites, salmonids, survival. Journal of Applied Ecology (2003) 40, 241–250

© 2003 British Ecological Society

Correspondence: Johan Dannewitz, Institute of Freshwater Research, SE-178 93 Drottningholm, Sweden (fax + 46 8759 03 38; e-mail [email protected]).

242 J. Dannewitz et al.

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 000–000

Introduction In salmonid fish, there is an increasing use of hatchery programmes for farming or artificial propagation of endangered wild populations. Within conservation programmes, releases of hatchery fish often aim to reintroduce populations that have become extinct, or to boost small declining populations until the factors responsible for the decline have been identified (Fleming & Petersson 2001). In the latter case, native fish are usually used as broodstock to produce fish that are reared in the hatchery during a period before they are released. In sea-ranching programmes for anadromous salmon and trout in Scandinavia, the fish are commonly released at the smolt stage, which in many rivers is reached at 2 years of age (Finstad & Jonsson 2001). Recent studies suggest that artificial breeding and rearing of fish may result in an evolutionary divergence of the hatchery fish away from their wild conspecifics (Swain, Riddel & Murray 1991; Fleming & Gross 1992, 1993; Petersson & Järvi 1993; Fleming, Jonsson & Gross 1994; Petersson et al. 1996; Einum & Fleming 1997). The mechanisms underlying such changes are alterations of sexual and natural selection in the hatchery, and random genetic processes during breeding (Petersson et al. 1996). Releases of genetically altered organisms may have profound ecological and genetic effects on wild populations (Daniels et al. 2001). Escapes and intentional releases of hatchery-reared fish constitute a potential threat to locally adapted wild populations through the effects of genetic introgression (Hindar, Ryman & Utter 1991; Ryman 1991; Ryman, Utter & Laikre 1995; Fleming et al. 2000; Hansen et al. 2000). However, few studies have addressed the ecological consequences of releasing fish. Previous studies comparing the performance of wild and hatchery fish in a natural environment have been conducted using farmed fish, i.e. fish that have been kept under artificial conditions for many generations and selected for certain economically important traits. These studies suggest that the time in captivity changes life-history characteristics and behaviour in ways that negatively affect the success of the farmed fish in the wild (Einum & Fleming 1997; McGinnity et al. 1997; Fleming et al. 2000). However, the effects of unintentional hatchery selection on the performance in the wild of fish produced for supplemental releases (e.g. sea-ranched fish) are still uncertain. In addition, family effects have been neglected in most, if not all, studies comparing hatchery and wild fish, either because of limited information about parentage of experimental fish and/or because experimental fish from too few families have been used to control for this effect. Family effects can account for differences in numerous fitness traits (Wimberger 1992; Geiger et al. 1997), suggesting that this source of variation must be recognized in comparative studies of hatchery and wild fish. Artificial breeding and rearing of fish for conservation purposes may result in evolutionary changes in

several traits. For example, by providing a predatorfree environment with surplus food, the hatchery environment may select for decreased anti-predator responses and increased growth potential (Petersson & Järvi 1995, 2000; Johnsson et al. 1996; Fernö & Järvi 1998; Einum & Fleming 2001). However, it is not clear how these changes might affect the fish following release into the natural environment because the hatchery environment may induce genetically based phenotypic variation that would not be expressed in the wild (cf. Reznick & Travis 1996). We therefore undertook a large-scale field experiment in which we compared the performance of a wild and a multigeneration sea-ranched strain of brown trout Salmo trutta L. from River Dalälven in central Sweden. By using molecular markers to assess parentage of experimental fish, we were able to include, for the first time to our knowledge, family effects in the analyses of fish growth and survival.

Materials and methods      The River Dalälven in central Sweden (60°38′N and 17°26′E) harbours two strains of anadromous (= breeding in freshwater but migrating to sea to grow) brown trout, both originating from the native population that existed prior to the construction of a dam and the hydro-electric power station in Älvkarleby. The first is a strict sea-ranched strain established in 1967. Fish belonging to this strain are reared in the hatchery until 2 years of age and are then marked and released in the river. Only marked fish returning to the river have been used as parents to produce the next generation of hatchery fish, which means that this strain has been genetically isolated for about seven generations. The second strain consists of trout spawning naturally downstream from the power station. This wild strain has probably been affected by sea-ranched trout spawning in the wild and possibly also by strayers from nearby populations (Petersson et al. 1996). However, the amount of gene flow from the sea-ranched into the wild strain seems to be limited because previous laboratory studies comparing these strains have found differences in growth rate (Petersson & Järvi 1995, 2000; Johnsson et al. 1996), life-history characteristics (Petersson et al. 1996), anti-predator behaviour (Fernö & Järvi 1998), competitive ability (Petersson & Järvi 2000), reproductive behaviour (Petersson & Järvi 1997) and stress response (Lepage et al. 2000). In most of these studies, the experimental fish were reared under identical environmental conditions in the hatchery, suggesting that the differences observed had a genetic basis. In autumn 1997, four different cross-types of wild (W) and sea-ranched (S) fish were made. Ten unique crosses of W × W and nine unique crosses of S × S were made on 21–22 October. Furthermore, 10 unique

243 Fitness of wild and sea-ranched trout

Fig. 1. Schematic view of the experimental stream in Älvkarleby.

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 241–250

crosses of W × S and S × W (female parent first) were made between 11 and 25 November, making a total of 39 full-sib families. The eggs from each family were kept in separate egg trays until April 1998, when 100 eyed eggs from each family (in total 3900 eggs) were introduced into an experimental stream nearby the power station using Vibert egg incubation boxes (Barlaup & Moen 2001). The experimental stream in Älvkarleby has a length of 110 m and a total area of 345 m2 (Johnsson et al. 1999). A tube supplies the stream with river water from the dam of the hydro-electric power station. The stream consists of four pools with riffles in between (Fig. 1). Wolf traps are situated at the upper and lower ends of the stream, and all water runs through the traps. A system of stainless steel gradings leads the fish to a collection tank, the mesh being 1 × 1 mm. Thus, no trout fry could escape or invade the stream. The water flow is

adjustable and it is possible to drain the stream so that water only remains in the pools, thus making it possible to catch all the fish if necessary. Fish stocked in the stream only have access to naturally occurring food, and are exposed to natural predation from mink Mustela vison Schreber and heron Ardea cinerea L. The eggs from each female were divided and placed in two boxes (50 eggs in each). In total, 20 sites in the stream were used for burial of the incubation boxes. Four boxes (representing four families, one from each of the four different cross-types) were buried at each site except for two sites without eggs of S × S origin. This experiment is hereafter referred to as experiment 1. In autumn 1999 the experiment was repeated using a half-sib mating design with 20 wild and 20 sea-ranched parental fish (10 males and 10 females from each strain). Between 19 and 27 October, eggs of two females were fertilized with milt from two males to create five 2 × 2 breeding sets for each strain. Hence, the total number of unique crosses (full-sib families) was 20 in each strain. The eggs were kept in the hatchery as in experiment 1 with the exception that eggs from half of the crosses from each strain were divided and placed in two separate egg trays. In April 2000, 100 eggs from each egg tray were placed in Vibert egg incubation boxes and were introduced into the experimental stream at 15 sites. Thus, 10 W × W and 10 S × S crosses were represented by 200 eggs each, and 10 W × W and 10 S × S crosses were represented by 100 eggs each, making a total of 6000 stocked eggs. Four boxes (representing four different crosses, two of W × W origin and two of S × S origin) were buried at each site. This experiment is hereafter referred to as experiment 2. In September 1998 and 2000 the experimental stream was drained slowly and the trout were caught by hand-netting; 226 surviving trout were caught in experiment 1 and 303 in experiment 2, which corresponds to densities of 66 and 88 individuals 100 m−2, respectively. This is within the range of natural densities observed in the area (T. Järvi, personal observation). In both experiments, the section of the stream (run 1–3; Fig. 1) where each fish was caught was noted and length and body mass were measured to the nearest mm and 0·1 g, respectively. The trout were individually marked with Passive Integrated Transponder (PIT)-tags, tissue samples for DNA analyses were taken (approximately 4 mm2 was cut from the caudal fin) and each fish was then returned to the section where it had been caught. The traps were checked 5 days a week from hatching until the final draining of the stream in August 1999 and September 2001, when the experiments were terminated. Length and mass of fish trapped after marking in September 1998 and 2000 were measured, their PITtag number was noted and their smolt status was checked according to external smolt characteristics (Järvi, Lofthus & Sigholt 1991). Condition factor (C) was calculated using the formula: C = 100W/Lb

244 J. Dannewitz et al.

Table 1. Total number of distinct alleles and expected heterozygosity (H) at microsatellite loci examined in the wild (W) and searanched (S) trout from River Dalälven used as parents to produce experimental fish for experiments 1 and 2. *Significant deviation (P < 0·05) from expected Hardy–Weinberg genotype frequencies analysed using Markov chain methods in  3·1d (Raymond & Rousset 1995) Experiment 1

Experiment 2

Locus

No. alleles

H (W, n = 40)

H (S, n = 38)

Locus

No. alleles

H (W, n = 20)

H (S, n = 20)

Str15 Str60 Str73 Str543 Ssa85 Ssa197 SsoSL417

4 3 3 6 5 8 8

0·68 0·51 0·68* 0·72* 0·69 0·76 0·78

0·67 0·52 0·65* 0·66 0·72 0·73 0·78

Str15 Str60 Str73 One9 Ssa85 Ssa197 SsoSL417

5 3 3 6 6 8 7

0·74 0·61 0·65 0·49 0·79 0·66 0·78

0·72 0·56 0·65 0·51 0·71 0·82 0·84

where W is body mass in g, L is length in cm and b is the slope coefficient of the regression of log mass and log length on all fish within each experiment (Bolger & Connolly 1989).

 

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 000–000

In experiment 1, parental assessment of fish caught after the first growth season was performed using the seven microsatellite loci Str15, Str60, Str73 (Estoup et al. 1993), Str543 (Presa & Guyomard 1996), Ssa85, Ssa197 (O’Reilly et al. 1996) and SsoSL417 (Slettan, Olsaker & Lie 1995). DNA was extracted from frozen fin tissue using the chelex protocol described by Walsh, Metzger & Higuchi (1991). Amplification of the microsatellites was made in 20-µl volumes with approximately 100 ng of template DNA, 0·4 µ of each primer, 1·0 m dNTP, 1·5 m MgCl2, 0·5 units of DyNAzymeTM II DNA polymerase and 1 × reaction buffer. The polymerase chain reaction (PCR) amplification was initiated with a denaturation step at 94 °C for 3 min followed by 30 cycles of 30 seconds at 94 °C, 30 seconds at the primerspecific annealing temperature and 30 seconds at 72 °C, and ended with a 10-min elongation step at 72 °C. The PCR products were mixed with one volume of formamide loading buffer and heated to 95 °C and were then separated on standard 6% acrylamide sequencing gels. After electrophoresis the fragments were visualized using silverstaining. Size determination of alleles was made from comparisons with a 10-bp DNA ladder. Allele sizes were later confirmed on an ABI Prism 310 Genetic Analyser (Applied Biosystems, Foster City, CA, USA) used according to the manufacturer’s recommendations (http://www.appliedbiosystems.com). In experiment 2, we used the same loci as in experiment 1 with the exception that One9 (Scribner, Gust & Fields 1996) was used instead of Str543 to assure that all loci labelled with the same dye had non-overlapping size ranges (see below). All seven loci were co-amplified in the same 25 µl PCR reaction (multiplex PCR) using Pharmacia Ready-To-GoTM PCR beads (Amersham Pharmacia Biotech Inc., Piscataway, NJ, USA) and approximately 100 ng of template DNA. Primers were

endlabelled with fluorescent dyes so as to enable comigration of all loci in the same capillary during electrophoresis, i.e. loci labelled with the same dye had non-overlapping size ranges. A uniform signal intensity among loci was achieved by adjusting primer concentrations. The multiplex PCR amplification was initiated with a denaturation step at 94 °C for 5 min followed by 28 cycles of 30 seconds at 94 °C, 30 seconds at an annealing temperature of 58 °C and 1 min at 72 °C, and ended with a 5-min elongation step at 72 °C. Electrophoresis and size determination of alleles was made on an ABI Prism 310 Genetic Analyser. The microsatellite loci examined were moderately to highly polymorphic (Table 1). In experiment 1, the parents (in total 78) were genotyped at all seven loci. Parentage was assessed by comparing the alleles at a given locus from each fish caught in the stream with the alleles in each of the 39 parental pairs (cf. O’Reilly, Herbinger & Wright 1998). Potential parental crosses with alleles incompatible with those of a particular offspring, at one or more loci, were excluded from the set of possible parental crosses. The 223 offspring (the tissue samples of three offspring were accidentally destroyed in the field) were first genotyped at the loci Str15, Str60, Str73, Str543 and Ssa197. Those offspring that could not be unambiguously assigned to only one parental cross (65 out of 223) were also genotyped at the loci Ssa85 and SsoSL417, and it was then possible to assign 215 (96%) of the offspring to a single parental pair. Eight offspring matched two parental pairs and were omitted from further analyses. Also, 10 offspring were omitted from analyses because of possible unintended mixing of tissue samples in the field. In experiment 2, parents and surviving offspring were genotyped at all seven loci, which made it possible to assign 274 (90%) of the offspring to a single parental pair. Twenty-nine offspring matched two or more parental crosses and were omitted from further analyses. In both experiments, allele comparisons between parents and offspring were conducted using the software  (written by Will Eichert, Bodega Marine Laboratory, Bodega Bay, CA; available on http://www-bml.ucdavis.edu/whichparents.htm).

245 Fitness of wild and sea-ranched trout

  We investigated the influence of strain origin, family background and stream section on length, mass and body condition after the first growth season, survival over the first growth season and smolt age (whether or not the trout were trapped as smolts). Analyses were performed separately for the two experiments. Hybrids in experiment 1 were not included in the statistical analyses, because we were not able to control in an appropriate way for variation caused by differences between cross-types in date of stripping. Growth variables were analysed with mixed model s using type III sums of squares as implemented in the GLM module in the software package  6.0 (StatSoft Inc. 2001). Family was nested within strain in experiment 1 and male–female were nested within set, which in turn was nested within strain in experiment 2 (Lynch & Walsh 1998). Strain, stream section and set were treated as fixed effects, whereas family and male– female were treated as random effects. Variation in survival between families in experiment 1 was analysed by comparing the observed distribution of surviving offspring among families with the Poisson distribution predicted assuming identical survival rates among families. Differences in survival between the strains in experiment 1 were analysed with a one-way  using numbers of surviving offspring from parental crosses as independent samples. Survival data in experiment 2 were analysed using arcsin-transformed proportions of surviving offspring among parental crosses following the model used for growth parameters, i.e. male and female were nested within set, which in turn was nested within strain. Smolt age was treated as a binary variable because the trout were categorized as either smolts or nonsmolts (parr), and was analysed using generalized linear models (the GLZ module) in  6.0 (StatSoft Inc. 2001). Significance tests were performed using likelihood ratio tests (McCullagh & Nelder 1989). Included in the analyses were offspring that were either trapped as smolts or parr in the trap, or were caught as parr when the experiments were terminated. Because previous studies have shown that age at smoltification in salmonids is a phenotypically plastic trait, which is

dependent on individual growth and/or condition during the period leading up to smolt migration in spring (Metcalfe et al. 1989), we included length and condition factor in autumn prior to smolt migration as continuous variables in the analyses. An initial analysis was performed to study effects of family/male–female origin, length and condition factor. Differences between the strains were analysed in a model incorporating strain, stream section, length and condition. Data were checked for normality and transformed if necessary before parametric statistical analyses. The level of significance was set at 0·05. Preliminary analyses of experiment 2 revealed no differences in fitness traits due to stripping date (data not shown). Therefore, stripping date was not included as a factor in the final analyses of experiment 2.

Results Data on growth parameters, survival and frequency of smolts among the cross-types in experiment 1 and 2 can be seen in Table 2. In experiment 1, stream section had no significant effects on either length, mass or condition factor (Table 3). Nor were there differences between trout of wild and sea-ranched origin in length, mass or condition factor (Table 3). However, the effect of family background was highly significant considering all growth variables investigated (Table 3). Results obtained in experiment 2 were very similar to those obtained in experiment 1. There were no differences in length, mass or body condition between trout caught in the different stream sections, or between trout of wild and sea-ranched origin (Table 4). However, there were significant male and female effects on all growth variables investigated (Table 4). Family survival over the first growth season ranged from 0% to 25% in experiment 1 and from 0% to 15% in experiment 2. In experiment 1, the distribution of the number of surviving offspring among families deviated significantly from the Poisson distribution predicted assuming identical survival rates among families (Fig. 2). Many families had fewer offspring whereas a few families had more offspring than predicted. There was no difference in survival between trout of wild and sea-ranched origin in experiment 1 (F1,17 = 0·00, P =

Table 2. Number (n) of pure wild (W × W), pure sea-ranched (S × S) and hybrid (W × S, S × W, female parent first) trout caught in the experimental stream in September 1998 (experiment 1) and September 2000 (experiment 2), survival over the first growth season, length (mm), mass (g) and condition factor after the first growth season, and frequency of 1-year-old smolts

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 241–250

Experiment

Cross-type

n

Survival (%)

Length ± SD

Mass ± SD

Condition* ± SD

Smolt frequency (%)

1 1 1 1 2 2

W×W S×S W×S S×W W×W S×S

63 78 25 39 126 148

6·3 8·7 2·5 3·9 4·2 4·9

132 ± 15 134 ± 14 109 ± 14 113 ± 14 103 ± 15 105 ± 15

24·9 ± 9·0 25·4 ± 8·5 13·4 ± 5·7 15·4 ± 6·1 11·8 ± 5·3 12·3 ± 5·4

0·59 ± 0·03 0·58 ± 0·03 0·59 ± 0·02 0·60 ± 0·05 0·92 ± 0·05 0·91 ± 0·05

28·6 34·6 4·0 10·3 3·2 5·4

*Condition factor could not be compared between experiment 1 and 2 because different slope coefficients were used (see text).

246 J. Dannewitz et al.

Table 3.  tables for length, mass and condition factor from analyses of trout of sea-ranched and wild origin in experiment 1; d.f.N, degrees of freedom in the numerator; d.f.D, degrees of freedom in the denominator Length

Mass

Condition

Factor

d.f.N

d.f.D

F

P

d.f.D

F

P

d.f.D

F

P

Stream section Strain Family (strain)

2 1 16

121·0 28·2 121·0

2·75 0·79 3·65

0·07 0·38 < 0·001

121·0 28·4 121·0

2·96 1·00 3·61

0·06 0·33 < 0·001

121·0 32·5 121·0

0·40 0·65 2·79

0·67 0·43 < 0·001

Table 4.  tables for length, mass and condition factor from analyses of trout of sea-ranched and wild origin in experiment 2; d.f.N, degrees of freedom in the numerator; d.f.D, degrees of freedom in the denominator Length

Mass

Condition

Factor

d.f.N

d.f.D

F

P

d.f.D

F

P

d.f.D

F

P

Stream section Strain Set (strain) Male (set × strain) Female (set × strain)

2 1 8 10 10

242·0 16·3 13·2 242·0 242·0

0·72 0·47 0·70 2·68 2·14

0·49 0·50 0·69 < 0·01 < 0·05

242·0 15·9 13·0 242·0 242·0

0·52 0·29 0·74 2·95 1·98

0·60 0·60 0·65 < 0·01 < 0·05

242·0 16·1 13·0 242·0 242·0

1·11 1·01 0·37 2·77 2·03

0·33 0·33 0·92 < 0·01 < 0·05

Fig. 2. Individual family size (number of offspring that survived the first growth season) among families of wild and sea-ranched origin stocked in the experimental stream in experiment 1. The observed distribution of offspring among families (vertical bars) deviates significantly from the Poisson distribution (—) predicted assuming identical survival rates among families (Kolmogorov– Smirnov Dmax = 0·39, P < 0·01).

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 000–000

1·00). In experiment 2, there were no significant effects of strain, male or female origin on survival rates (strain origin: F1,2·4 = 0·94, P = 0·42; male: F10,10 = 1·00, P = 0·50; female: F10,10 = 0·80, P = 0·63). However, the statistical power in this analysis was low due to the three factors investigated and the few observations available. The effect of stream section on survival was not studied because dead fish could not be traced back to a particular section or were not found at all. In experiment 1, there were no associations between either family origin or body condition and probability of smolt migration at age 1 year (family: χ2 = 14·14, d.f. = 10, P = 0·17; condition factor: χ2 = 0·74, d.f. = 1, P = 0·39) whereas length in September was strongly associated with probability of smolt migration (χ2 = 14·01, d.f. = 1, P < 0·001): trout that were large in autumn

prior to smolt migration were more likely to metamorphose into the smolt stage the following spring (Fig. 3). The same pattern emerged in experiment 2 (male effect: χ2 = 12·58, d.f. = 8, P = 0·13; female effect: χ2 = 18·74, d.f. = 19, P = 0·47; effect of condition factor: χ2 = 0·39, d.f. = 1, P = 0·53; effect of length: χ2 = 6·81, d.f. = 1, P < 0·01; Fig. 3). Further analyses including strain origin and stream section were performed to study other variables likely to affect age at smoltification. However, there were no associations between either strain origin or stream section and probability of smolt migration in experiment 1 (strain origin: χ2 = 0·00, d.f. = 1, P = 0·98; stream section: χ2 = 2·12, d.f. = 2, P = 0·35) or in experiment 2 (strain origin: χ2 = 0·13, d.f. = 1, P = 0·72; stream section: χ2 = 3·22, d.f. = 2, P = 0·20) whereas the effect of length in September remained

247 Fitness of wild and sea-ranched trout

Fig. 3. Length (mean ± SD) in September of non-smolting trout (squares) and trout that metamorphosed into the migratory smolt stage the following spring (circles). Individual data points are shown next to the vertical bars. In a logistic regression model, there were significant associations between length in September and probability of smolt migration the following spring in both experiments (see text).

significant in both analyses (experiment 1: χ2 = 15·67, d.f. = 1, P < 0·001; experiment 2: χ2 = 6·11, d.f. = 1, P < 0·05).

Discussion

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 241–250

When family effects were accounted for, we found no significant differences between trout of wild and searanched origin in length, mass or condition factor after the first growth season in a semi-natural stream. Also, there were no differences between the strains in survival over the first growth season, nor in the proportion that metamorphosed into the migratory smolt phase at 1 year of age. These results suggest that wild-born trout of sea-ranched origin have the capacity to compete successfully with wild trout under natural conditions, at least during the freshwater phase that includes the intense natural selection of early life responsible for much of the total mortality within a typical cohort (Elliott 1989a,b). Our results further suggest that the impact of hatchery selection on the performance in the wild of sea-ranched salmonids may not be as pronounced as previously thought (Einum & Fleming 2001). It must be emphasized, however, that the study was conducted in a semi-natural stream, rather than a wholly natural stream system, and lacks true replication at the stream level. Also, we cannot rule out the possibility that differences might occur between the groups during subsequent life-history stages or under different environmental conditions. The present findings are not consistent with previous studies on trout from the River Dalälven, which showed that, under hatchery conditions, sea-ranched trout had a higher growth potential (Petersson & Järvi 1995, 2000; Johnsson et al. 1996) and were more prone to take risks than wild fish (Fernö & Järvi 1998). There are at least two possible explanations for this discrepancy. First, sea-ranched trout simply may not be able to utilize their growth potential fully when reared under natural conditions because food abundance in the wild

is lower and less predictable than in the hatchery. If so, we would expect to find fitness costs of the higher food demands among the sea-ranched trout, such as a reduced survival rate or a relatively low body condition (Johnsson et al. 2000). However, this was not observed. Secondly, trout of wild and sea-ranched origin may follow different reaction norms for particular traits, i.e. the relative expression of traits between the two types of fish may differ among environments (cf. Reznick & Travis 1996). Previous studies have shown that salmonids are able to change risk-taking behaviour and growth in relation to, for example, food abundance, predation risk and maturation status (Damsgård & Dill 1998; Dannewitz & Petersson 2001). Trout of sea-ranched origin are likely to differ from wild conspecifics with respect to reaction norms for certain traits because they should be able to utilize extreme conditions in the hatchery but must also be able to face completely different environmental conditions following release in the river. As pointed out by Einum & Fleming (2001), any lack of correspondence between studies performed in the hatchery and data from the wild may be attributable to this problem. The results of the present study also contradict previous studies on Atlantic salmon conducted under natural conditions. McGinnity et al. (1997) and Fleming et al. (2000) both reported high growth but low survival in farm relative to wild salmon during the first growth season. However, the farm salmon used in these studies had spent their whole life cycle in captivity, and had undergone intentional selection for rapid growth and other commercial traits for many generations. It is therefore not surprising that hatchery programmes for farming have resulted in a more pronounced divergence between the hatchery fish and their wild conspecifics than is the case for hatchery programmes associated with sea-ranching. Indeed, escapes of farmed salmonids is a serious problem in many areas because these domesticated fish may invade and interact with locally adapted wild populations (Fleming et al. 2000). It is not clear to what extent family effects may have affected results in previous studies comparing wild and hatchery fish. Our results suggest that family background may strongly affect important fitness traits like growth and survival, although it is difficult to separate genetic from environmental effects in the present design. However, as we observed no significant differences in growth parameters between fish caught in the different stream sections, it seems reasonable to assume that the differences in growth among families observed in both experiments 1 and 2, and the non-random distribution in survival among families in experiment 1, were at least partly caused by differences between families in genetic background and/or non-genetic maternal effects. The only variable that had a significant effect on whether or not the trout metamorphosed into the migratory smolt phase at 1 year of age was length in autumn prior to smolt migration. This is consistent

248 J. Dannewitz et al.

with previous studies showing that those individuals best able to compete for resources gain in growth and tend to be the first to become smolts (Metcalfe et al. 1989; Huntingford et al. 1990; Metcalfe 1991). Even though we observed no significant effect of family background on the probability of smoltification, family origin may still have indirect effects on smolt age through the effects of differential growth between families. This may, during subsequent life stages, further increase the variance in fitness between families. Thus, our findings suggest that comparative studies on salmonid fish in which the results are based on only a few families must be interpreted with caution because the presence or absence of significant treatment effects may be due to family effects not accounted for.

 

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 000–000

We have shown that trout of sea-ranched origin, which differ from wild trout in several fitness-related traits when reared under hatchery conditions, appear to perform well in competition with wild conspecifics in a near-natural environment, indicating that results obtained from laboratory studies should be used with caution to predict interactions in the wild between hatchery and wild fish. Our results further indicate that sea-ranching programmes may not result in the pronounced genetic effects of hatchery selection observed in farmed salmonids, suggesting that generalizations about the influence of hatchery selection on the fish phenotype should be avoided. Our observation of variation in fitness between families has implications for conservation biology because the amount of genetic variation maintained within populations is related to the variance in family size (Kapuscinski & Lannan 1986; Geiger et al. 1997), suggesting that family effects should be taken into account in calculations of the minimum number of parental fish used to produce fish for supplementation of recovering salmonid populations. Stocking of eyed eggs, as in our experiments, is not a typical strategy for releasing fish. Normally, the production of fish for stocking involves a period of hatchery rearing before the fish are released. There is, however, growing evidence that the period of hatchery rearing may have profound environmental effects on fish performance following release in the natural environment. Recent studies suggest that such environmental effects can be pronounced and may override genetic effects of hatchery selection (Fleming, Lamberg & Jonsson 1997; Johnsson, Höjesjö & Fleming 2001; Sundström & Johnsson 2001) and may be at least partly responsible for the low success of many stocking programmes (Jonsson, Hansen & Jonsson 1991; Finstad & Jonsson 2001; Fleming & Petersson 2001). Our observation of no detectable differences in fitness between wild-born trout of sea-ranched and wild origin, in combination with new knowledge about negative environmental effects of culture, suggests that for populations

dependent on supplemental stocking more attention should be paid to minimizing negative environmental effects during hatchery rearing, for example by decreasing the time spent in captivity or by stocking newly fertilized or eyed eggs. Such an approach, alongside the development of strategies for habitat improvements and maintenance of genetic variation, is likely to be important for a successful management of endangered salmonids and other fish populations.

Acknowledgements We thank Anna Löf and Leif Johansson for excellent technical assistance, and Carl-Gustaf Thulin, Jacob Höglund, Michael Hansen, Steve Ormerod and one anonymous referee for valuable comments on earlier versions of the manuscript. This work has been carried out with financial support from the Commission of the European Communities, Agricultural and Fisheries (FAIR) specific RTD program, CT-97-3498. It does not necessarily reflect the Commission’s view and in no way anticipates the Commission’s future policy in this area. The work was approved by the Ethical Committee of Animal Research and complies with the standards and procedures laid down by the Swedish Ministry of Agriculture (licence 34 3632.92).

References Barlaup, B.T. & Moen, V. (2001) Planting of salmonid eggs for stock enhancement – a review of the most commonly used methods. Nordic Journal of Freshwater Research, 75, 7–19. Bolger, T. & Connolly, P.L. (1989) The selection of suitable indices for the measurement and analysis of fish condition. Journal of Fish Biology, 34, 171–182. Damsgård, B. & Dill, L.M. (1998) Risk-taking behavior in weight-compensating coho salmon, Oncorhynchus kisutch. Behavioral Ecology, 9, 26 –32. Daniels, M.J., Beaumont, M.A., Johnson, P.J., Balharry, D., Macdonald, D.W. & Barratt, E. (2001) Ecology and genetics of wild-living cats in the north-east of Scotland and the implications for the conservation of the wildcat. Journal of Applied Ecology, 38, 146 –161. Dannewitz, J. & Petersson, E. (2001) Association between growth, body condition and anti-predator behaviour in maturing and immature brown trout parr. Journal of Fish Biology, 59, 1081 –1091. Einum, S. & Fleming, I.A. (1997) Genetic divergence and interactions in the wild among native, farmed and hybrid Atlantic salmon. Journal of Fish Biology, 50, 634–651. Einum, S. & Fleming, I.A. (2001) Implications of stocking: ecological interactions between wild and released salmonids. Nordic Journal of Freshwater Research, 75, 56 –70. Elliott, J.M. (1989a) Mechanisms responsible for population regulation in young migratory trout, Salmo trutta. I. The critical time for survival. Journal of Animal Ecology, 58, 987 –1001. Elliott, J.M. (1989b) The critical-period concept for juvenile survival and its relevance for population regulation in young sea-trout, Salmo trutta. Journal of Fish Biology, 35, 91– 98. Estoup, A., Presa, P., Krieg, F., Vaiman, D. & Guyomard, R. (1993) (CT)n and (GT)n microsatellites: a new class of genetic markers for Salmo trutta L. (brown trout). Heredity, 71, 488 – 496.

249 Fitness of wild and sea-ranched trout

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 241–250

Fernö, A. & Järvi, T. (1998) Domestication genetically alters the anti-predator behaviour of anadromous brown trout (Salmo trutta) – a dummy predator experiment. Nordic Journal of Freshwater Research, 74, 95 –100. Finstad, B. & Jonsson, N. (2001) Factors influencing the yield of smolt releases in Norway. Nordic Journal of Freshwater Research, 75, 37 –55. Fleming, I.A. & Gross, M.R. (1992) Reproductive behaviour of hatchery and wild coho salmon (Oncorhynchus kisutch): does it differ? Aquaculture, 103, 101–121. Fleming, I.A. & Gross, M.R. (1993) Breeding success of hatchery and wild coho salmon (Oncorhynchus kisutch) in competition. Ecological Applications, 3, 230 –245. Fleming, I.A. & Petersson, E. (2001) The ability of released, hatchery salmonids to breed and contribute to the natural productivity of wild populations. Nordic Journal of Freshwater Research, 75, 71– 98. Fleming, I.A., Hindar, K., Mjølnerød, I.B., Jonsson, B., Balstad, T. & Lamberg, A. (2000) Lifetime success and interactions of farm salmon invading a native population. Proceedings of the Royal Society of London B, 267, 1517 –1523. Fleming, I.A., Jonsson, B. & Gross, M.R. (1994) Phenotypic divergence of sea-ranched, farmed and wild salmon. Canadian Journal of Fisheries and Aquatic Sciences, 51, 2808 –2824. Fleming, I.A., Lamberg, A. & Jonsson, B. (1997) Effects of early experience on the reproductive performance of Atlantic salmon. Behavioral Ecology, 8, 470 – 480. Geiger, H.J., Smoker, W.W., Zhivotovsky, L.A. & Gharrett, A.J. (1997) Variability of family size and marine survival in pink salmon (Oncorhynchus gorbuscha) has implications for conservation biology and human use. Canadian Journal of Fisheries and Aquatic Sciences, 54, 2684 –2690. Hansen, M.M., Ruzzante, D.E., Nielsen, E.E. & Mensberg, K.-L.D. (2000) Microsatellite and mitochondrial DNA polymorphism reveals life-history dependent interbreeding between hatchery and wild brown trout (Salmo trutta L.). Molecular Ecology, 9, 583 –594. Hindar, K., Ryman, N. & Utter, F. (1991) Genetic effects of cultured fish on natural fish populations. Canadian Journal of Fisheries and Aquatic Sciences, 48, 945 – 957. Huntingford, F.A., Metcalfe, N.B., Thorpe, J.E., Graham, W.D. & Adams, C.E. (1990) Social dominance and body size in Atlantic salmon parr, Salmo salar L. Journal of Fish Biology, 36, 877 – 881. Järvi, T., Lofthus, R. & Sigholt, T. (1991) On growth and smoltification in Atlantic salmon parr – the effect of sexual maturation and competition. Nordic Journal of Freshwater Research, 66, 72 – 88. Johnsson, J.I., Höjesjö, J. & Fleming, I.A. (2001) Behavioural and heart rate responses to predation risk in wild and domesticated Atlantic salmon. Canadian Journal of Fisheries and Aquatic Sciences, 58, 788 –794. Johnsson, J.I., Jönsson, E., Petersson, E., Järvi, T. & Björnsson, B.Th (2000) Fitness-related effects of growth investment in brown trout under field and hatchery conditions. Journal of Fish Biology, 57, 326 –336. Johnsson, J.I., Petersson, E., Jönsson, E., Björnsson, B.Th & Järvi, T. (1996) Domestication and growth hormone alter antipredator behaviour and growth patterns in juvenile brown trout, Salmo trutta. Canadian Journal of Fisheries and Aquatic Sciences, 53, 1546 –1554. Johnsson, J.I., Petersson, E., Jönsson, E., Järvi, T. & Björnsson, B.Th (1999) Growth hormone-induced effects on mortality, energy status and growth: a field study on brown trout (Salmo trutta). Functional Ecology, 13, 514 – 522. Jonsson, N., Hansen, L.P. & Jonsson, B. (1991) Variation in age, size and repeat spawning of adult Atlantic salmon in relation to river discharge. Journal of Animal Ecology, 60, 937 – 947.

Kapuscinski, A.R.D. & Lannan, J.E. (1986) A conceptual genetic fitness model for fisheries management. Canadian Journal of Fisheries and Aquatic Sciences, 43, 1606–1616. Lepage, O., Överli, Ö., Petersson, E., Järvi, T. & Winberg, S. (2000) Differential stress coping in wild and domesticated sea trout. Brain, Behavior and Evolution, 56, 259–268. Lynch, M. & Walsh, B. (1998) Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland, MA. McCullagh, P. & Nelder, J.A. (1989) Generalized Linear Models. Chapman & Hall, London, UK. McGinnity, P., Stone, C., Taggart, J.B., Cooke, D., Cotter, D., Hynes, R., McCamley, C., Cross, T. & Ferguson, A. (1997) Genetic impact of escaped farmed Atlantic salmon (Salmo salar L.) on native populations: use of DNA profiling to assess freshwater performance of wild, farmed, and hybrid progeny in a natural river environment. ICES Journal of Marine Science, 54, 998 –1008. Metcalfe, N.B. (1991) Competitive ability influences seaward migration age in Atlantic salmon. Canadian Journal of Zoology, 69, 815 – 817. Metcalfe, N.B., Huntingford, F.A., Graham, W.D. & Thorpe, J.E. (1989) Early social status and the development of life-history strategies in Atlantic salmon. Proceedings of the Royal Society of London B, 236, 7 –19. O’Reilly, P.T., Hamilton, L.C., McConnell, S.K. & Wright, J.M. (1996) Rapid analysis of genetic variation in Atlantic salmon (Salmo salar) by PCR multiplexing of dinucleotide and tetranucleotide microsatellites. Canadian Journal of Fisheries and Aquatic Sciences, 53, 2292 –2298. O’Reilly, P.T., Herbinger, C. & Wright, J.M. (1998) Analysis of parentage determination in Atlantic salmon (Salmo salar) using microsatellites. Animal Genetics, 29, 363–370. Petersson, E. & Järvi, T. (1993) Differences in reproductive traits between sea-ranched and wild sea-trout (Salmo trutta) originating from a common stock. Nordic Journal of Freshwater Research, 68, 91–97. Petersson, E. & Järvi, T. (1995) Evolution of morphological traits in sea trout (Salmo trutta) parr (0+) through domestication. Nordic Journal of Freshwater Research, 70, 62–67. Petersson, E. & Järvi, T. (1997) Reproductive behaviour of sea trout (Salmo trutta) – the consequences of sea-ranching. Behaviour, 134, 1– 22. Petersson, E. & Järvi, T. (2000) Both contest and scramble competition affect the growth performance of brown trout, Salmo trutta, parr of wild and sea-ranched origins. Environmental Biology of Fishes, 59, 211 –218. Petersson, E., Järvi, T., Steffner, N.G. & Ragnarsson, B. (1996) The effect of domestication on some life history traits of sea trout and Atlantic salmon. Journal of Fish Biology, 48, 776 –791. Presa, P. & Guyomard, R. (1996) Conservation of microsatellites in three species of salmonids. Journal of Fish Biology, 49, 1326 –1329. Raymond, M. & Rousset, F. (1995) GENEPOP: population genetics software for exact tests and ecumenicism. Journal of Heredity, 86, 248–249. Reznick, D. & Travis, J. (1996) The empirical study of adaptation. Adaptation (eds M.R. Rose & G.V. Lauder), pp. 243 – 289. Academic Press, San Diego, CA. Ryman, N. (1991) Conservation genetics considerations in fishery management. Journal of Fish Biology, 39 (Supplement A), 211–224. Ryman, N., Utter, F. & Laikre, L. (1995) Protection of intraspecific biodiversity of exploited fishes. Reviews in Fish Biology and Fisheries, 5, 417 – 446. Scribner, K.T., Gust, J.R. & Fields, R.L. (1996) Isolation and characterization of novel salmon microsatellite loci: crossspecies amplification and population genetic applications. Canadian Journal of Fisheries and Aquatic Sciences, 53, 833 – 841.

250 J. Dannewitz et al.

© 2003 British Ecological Society, Journal of Applied Ecology, 40, 000–000

Slettan, A., Olsaker, I. & Lie, O. (1995) Atlantic salmon, Salmo salar, microsatellites at the SSOSL25, SSOSL85, SSOSL311, SSOSL417 loci. Animal Genetics, 26, 281–282. StatSoft Inc. (2001) STATISTICA (Data Analysis Software System), Version 6. StatSoft Inc., Tulsa, OK. Sundström, L.F. & Johnsson, J.I. (2001) Experience and social environment influence the ability of young brown trout to forage on live novel prey. Animal Behaviour, 61, 249 – 255. Swain, D.P., Riddel, B.E. & Murray, D.B. (1991) Morphological differences between hatchery and wild populations of coho salmon (Oncorhynchus kisutch): environmental versus

genetic origin. Canadian Journal of Fisheries and Aquatic Sciences, 48, 1783 –1791. Walsh, P.S., Metzger, D.A. & Higuchi, R. (1991) Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Bio Techniques, 10, 506–513. Wimberger, P.H. (1992) Plasticity of fish body shape. The effects of diet, development, family and age in two species of Geophagus (Pisces: Cichlidae). Biological Journal of the Linnean Society, 45, 197 –218. Received 21 December 2001; final copy received 16 September 2002