Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
Migration costs drive convergence of threshold traits for migratory tactics Genki Sahashi1,† and Kentaro Morita2
rspb.royalsocietypublishing.org
Research Cite this article: Sahashi G, Morita K. 2013 Migration costs drive convergence of threshold traits for migratory tactics. Proc R Soc B 280: 20132539. http://dx.doi.org/10.1098/rspb.2013.2539
Received: 27 September 2013 Accepted: 8 October 2013
Subject Areas: evolution, ecology Keywords: alternative migratory tactics, life-history traits diversity, migration distance, partial migration, salmonid fish, selection
Author for correspondence: Genki Sahashi e-mail: [email protected]
†
Present address: Division of Biosphere Science, Graduate School of Environmental Sciences, Hokkaido University, Hakodate 041-8611, Japan. Electronic supplementary material is available at http://dx.doi.org/10.1098/rspb.2013.2539 or via http://rspbroyalsocietypublishing.org.
1 Division of Marine Bioresource and Environmental Science, Graduate School of Fisheries Sciences, Hokkaido University, Hakodate 041-8611, Japan 2 Hokkaido National Fisheries Research Institute, Fisheries Research Agency, Sapporo 062-0922, Japan
Partial migration of some, but not all, members of a population is a common form of migration. We evaluated how migration costs influence which members migrate in 10 populations of two salmonid species. The migratory patterns of both species were evaluated based on the size at maturity for resident males, which is the threshold trait that determines the migratory tactics used within a population. In both species, this size was smaller in males located further from the sea, where migration costs are presumably higher. Moreover, the threshold sizes at maturity in males were correlated between both species. Our results suggest that migration costs are a significant convergent selective force on migratory tactics and life-history traits in nature.
1. Introduction Convergent evolution is the process by which similar features evolve independently in different evolutionary lineages [1]. This evolutionary phenomenon provides an ideal means for evaluating adaptation, because the same phenotype evolves repeatedly and independently under similar circumstances [2]. Migratory animals often exhibit similarities in behavioural, physiological, morphological and life-history traits, such as the migratory syndromes shared by insects, fish and birds [3]. These similarities are thought to have arisen through natural selection, which favours convergent solutions to common problems facing migratory animals [4]. To better understand migration, its evolution and its ecology, there is a need for the development of hypotheses that advance general migratory theory and have cross-species applicability [3]. Partial migration is one of the common forms of migration in which a population contains both migratory and resident individuals [5,6]. Whether alternative migratory tactics are adopted is dependent on life-history traits called ‘threshold traits’ [7]. These traits exhibit two properties: a threshold value that divides the phenotypic distribution (e.g. migrants and residents), and an underlying ‘liability’ trait that differs in a continuous variable [8]. If the value of the liability trait exceeds the threshold, an individual will choose a different migratory tactic than if the value of the trait is below the threshold [9]. The threshold value often varies among populations because genetic variability and environmental differences can affect the trait [10]. Threshold traits have been used to describe the inheritance and evolution of partial migration. A number of studies have examined ecological factors underlying the diversity of specific threshold traits in partially migratory animals, but there have been few cross-species investigations in nature. The trade-off between benefits and costs is a fundamental mechanism driving life-history divergence. Thus, the threshold traits of partially migratory animals will coevolve to maximize fitness under a given set of migratory conditions. Costs of migration include increased metabolic and energetic requirements during migration [11], decreased reproductive output [12] and higher predation risk [13]. These effects will presumably increase as the migration distance increases. For partially migratory animals, migration costs affect only the migratory individuals, so migration cost is expected to be a general factor that results in the diversity of threshold traits in partially migratory animals.
& 2013 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
(i) The threshold size at maturity in males decreases with increasing distance from the sea as an indicator of migration cost. (ii) The proportion of male individuals that mature without migrating to the sea increases with increasing distance from the sea. (iii) The divergence of threshold size at maturity in males covaries between the two species among tributaries.
2. Material and methods This study was conducted in the Kushiro river system in eastern Hokkaido, Japan. We chose one reach in each of 25 tributaries based on the presence of typical biotopes for salmonid parr and spawning. We measured a number of environmental factors in each reach (see below for details). We then chose 10 of the tributaries (figure 1) in which distance from the sea was not correlated with other environmental factors at the site (see electronic supplementary material, table S1). In these 10 tributaries, the mean length of the study reach was 81 m (range: 55–90 m; see electronic supplementary material, table S2). In the Kushiro river system, records of hatchery releases of white-spotted charr and masu salmon exist only for Lake Kussharo.
(b) Fish collection and measurement Sampling was conducted between 13 and 16 September 2010, during the spawning season of salmon and about one month before the spawning season of charr. Parr were collected using an electrofisher (200– 300 V DC, model 12B, Smith-Root, Vancouver, WA) and a dip net (30 cm width, 3 mm mesh). All parr were immediately frozen for subsequent laboratory measurements. Anadromous adults were collected in some tributaries, but were not analysed. Tributary 9 was re-sampled on 9 November 2010 as no mature parr were collected during the first sampling. Because age 1þ parr cease to grow from September to November [23], age greater than or equal to 1þ parr were included in the analyses without any size transformation. The following variables were determined for each individual after defrosting: fork length, sex, reproductive state (mature or immature, by visual inspection of gonads) and age (determined using otoliths).
(c) Measurement of environmental factors As a surrogate for migration cost, the distance from the sea was measured for each tributary using a portable GPS receiver (GPSMAP60CSx, Garmin, Olathe, KS) and 1 : 25 000 scale topographic maps (http://www.gsi.go.jp/). Previous studies have used altitude as a surrogate for migration cost, but altitude itself can be an indicator of other environmental factors such as water temperature (e.g. [24]). To avoid multicollinearity, altitude was not used as a model parameter. In addition to distance from the sea, eight environmental factors were measured in each tributary: water temperature, tributary width, tributary depth, velocity, substrate, slope, riverine fish growth and fish density (see electronic supplementary material, tables S2 and S3). The environmental factor values in the study reaches represent those of each tributary (see electronic supplementary material, appendix A). Water temperature was measured at hourly intervals in each tributary study reach using fixed data loggers (Stow-Away TidbiT, Onset Computer Corp., Bourne, MA) deployed on the stream beds from June to August 2010 and November 2010 to May 2011. Mean temperature (averaged for the whole study period) and monthly temperature (averaged for each month) were used as indices of water temperature. Tributary width was measured at 5 m intervals in each study
Proc R Soc B 280: 20132539
(a) Study area
2
rspb.royalsocietypublishing.org
Partial migration is evident in many species of salmonid fish, such as the white-spotted charr, Salvelinus leucomaenis and masu salmon, Oncorhynchus masou. All females and most males in both species live for several years in rivers, then descend to the sea to feed [14,15]. After this feeding migration, these anadromous fish become sexually mature and return to the natal river to spawn [14,15]. By contrast, some males in both species become sexually mature at the parr stage (the darkly pigmented, stream-dwelling juvenile stage of the salmonid life cycle) [14,15]. As with other salmonids, the alternative migratory tactics in both species are associated with alternative reproductive phenotypes during spawning. Anadromous males have high fertilization success by competing for females using their large body size, whereas smaller precocious mature parr achieve fertilizations by adopting a sneaking behaviour [16]. In salmonid males, life-history divergence is thought to depend on whether the current size or growth rate at the time of the life-history decision will result in the male exceeding a threshold size [17,18], such as the size at maturity for parr. Males that exceed this size become precocious mature parr, whereas those below this size become anadromous (see electronic supplementary material, figure S1a). As the threshold size at maturity in males decreases in a population, the proportion of mature male parr increases within the population (see electronic supplementary material, figure S1a,b). Thus, in salmonid populations with high migration costs, the threshold size at maturity in males will be small, resulting in more resident individuals within a population. However, quantifying the effects of migration costs on threshold size at maturity in males is difficult because other environmental factors can covary with migration costs. Namely, the threshold size at maturity in males may be shaped by other evolutionary and plastic responses to environmental gradients that correlate with migration distance. Recent studies on the partially migratory Atlantic salmon, Salmo salar, have shown that threshold size at maturity in males varies among populations, and the variation is influenced by migration costs such as altitude or migration distance [19,20]. However, these studies did not measure the effects of migration costs on threshold size at maturity in males independently, because they did not examine correlations between migration costs and other environmental factors. For example, water temperature often decreases with increasing altitude, so it can also affect the diversity of threshold size at maturity in males (e.g. [21,22]). Thus, to test the hypothesis that migration cost is a general factor in shaping the diversity of threshold traits, the effects of migration costs on the threshold traits must be quantified under a sampling design that considers multicollinearity among environmental factors. In this study, we collected data from 10 tributaries in a single river system to investigate the hypothesis that increased migration cost (i.e. distance of migration) causes a decrease in the threshold size at maturity in males (i.e. promotes residency) in two sympatric salmonids: white-spotted charr and masu salmon. The study was designed to independently quantify the effects of migration costs on the threshold sizes (see Material and methods). Our objective was to identify the evolutionary factors that promote diversity in threshold size at maturity in males of the two species. We would expect to find the following three schematic scenarios (see electronic supplementary material, figure S1c).
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
145º
140º 44º
2
42º
1
Hokkaido Island study area
3 4
1 Ginnosawa 5
7
2 Ishikaribetsu 3 Isobunnai
6 Mitnamitawa 7 Ponsetsuri 8 Oriyomappu 9 8
10
Old river channel
9 Tirishinne 10 Ooshima
10 km
Pacific Ocean
Figure 1. Location of the 10 tributaries in the Kushiro river system. The old river channel has no current because a closed water gate was constructed at the junction of the main stream and the old river channel.
reach. Tributary depth, velocity and substrate were measured at 5 m intervals with three evenly spaced transects (1/4, 1/2, 3/4). Substrate composition was sorted into six classes: sand-silt (less than 2 mm), category 1 (as categorized in electronic supplementary material, table S2); gravel (2–16 mm), category 2; pebble (17–64 mm), category 3; cobble (65–256 mm), category 4; boulder (greater than 256 mm), category 5; irregular bedrock, category 6. The mean and standard deviation of the categorical values were used as indices of substrate composition [25]. Velocity was measured at 60% of the depth from the surface to the bed using a propeller-type meter (VR-301; Kenek Co., Tokyo, Japan). Riverine fish growth was quantified using the average fork length in age 0þ parr. Fish densities were estimated in a reach of each tributary during fish collection using the two-pass removal method (model M(b), program CAPTURE; available at http://www.mbrpwrc.usgs.gov/software/index.html). The density of each species and the combined density of the two salmonids were used as indices of fish density.
(d) Data analyses To assess threshold sizes at maturity in males, we first calculated the probability of being mature at a particular combination of variables (i.e. maturity ogive) using logistic regression. The logistic regression model was described as logit ( p) ¼ a þ b1FL þ b2Age þ Ctributary, where p is the probability of maturation, a is the regression constant, bn are regression coefficients, Ctributary is a tributary-specific constant, FL is the fork length and Age is age in years. The model parameters were estimated using the maximum-likelihood method with a binomial distribution of errors. The significance of model parameters was evaluated using the likelihood-ratio test and the Akaike information criterion (AIC). The threshold size at maturity in males (L50, first) was defined as the fork length at which the probability of maturing at the first age at maturity is 50%. The ages at first maturity are age 1þ in charr [15] and age 0þ in salmon [22]. At the age of first maturity, the fork length at 50% probability of maturing (L50, first) is the same as the fork length at which the probability of becoming
3. Results A total of 1106 charr and 890 salmon were caught during the study. Mature male parr were captured in all tributaries, but no mature female parr were collected (see electronic supplementary material, table S3). The maturity rate for all fish exceeding the minimum age at maturity varied considerably among tributaries (range: 17–100% for charr 1þ and 3–93% for salmon 0þ). The logistic regression analyses identified significant effects of the tributary, fork length and age on the probability of maturing for both charr and salmon males (table 1). The probability of maturing increased with increasing fork length and age, but these relationships differed among tributaries, resulting in substantial variation in L50, first values among tributaries (see electronic supplementary material, table S3 and figure S2). Although the effect of age was insignificant for charr, the model incorporating all three variables (Ctributary, Age and FL) had the smallest AIC value for both species (DAIC . 1.5). There was a positive correlation between the L50, first values of charr and salmon (figure 2), suggesting there was spatial covariation of threshold size at maturity in males between the species. Among the nine environmental variables we measured, only distance from the sea was negatively correlated with L50, first values in both species (table 2). None of the monthly temperatures were significantly correlated with L50, first values in both species (see electronic supplementary material, table S4). Moreover, more complex models did not perform better at explaining the probability of maturing for either species (see electronic supplementary material, appendix B). Distance from the sea was not significantly correlated with the other eight environmental variables (see electronic supplementary material, table S1). This indicates that threshold size at maturity in males (L50, first) decreased with increasing distance from the sea, an indicator of migration cost (figure 3). The decrease in the threshold size at maturity in males (L50, first) was associated with a significant increase in the maturity rate for all fish exceeding the minimum age at maturity in both charr (r ¼ 20.802, p ¼ 0.005) and salmon (r ¼ 20.861, p ¼ 0.001). The maturity rate for all fish exceeding the minimum age at maturity increased as the distance from the sea increased, although the correlation with charr was insignificant (charr, r ¼ 0.265, p ¼ 0.460; salmon, r ¼ 0.663, p ¼ 0.037). The CV of L50, first differed significantly between charr and salmon (Z1,18 ¼ 2.07, p ¼ 0.038); it was more than twice as large in salmon (23%) as in charr (11%).
Proc R Soc B 280: 20132539
6 4 Numaoro 5 Hitominai
3
rspb.royalsocietypublishing.org
Lake Kussharo
mature is 50%, a feature of the probabilistic maturation reaction norm (PMRN) [26]. We calculated L50, first using regression coefficients of the logistic regression described as 2(a þ b2Age þ Ctributary)/b1, where Age is 0 and 1 for salmon and charr, respectively. We used L50, first to test for interspecific covariation and to examine intraspecific variation along environmental gradients. The proportion of male individuals that matured without migrating to the sea was quantified using the maturity rate for all fish exceeding the minimum age at maturity (i.e. charr 1þ; salmon 0þ). The coefficient of variation (CV) was used as a measure of variation in L50, first among tributaries within species. Interspecific differences in CV of L50, first were tested using the Z-test [27]. All statistical analyses were performed using R v. 2.14.0 for Windows (http://www.r-project.org/).
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
Table 1. Results of a logistic regression of fork length, age and tributary on the probability of maturing for age 1þ male charr and 0þ male salmon. ln L, log likelihood for the full model; G 2, statistic of the likelihood-ratio test. ln L
(a) white-spotted charr fork length
275.2
age tributary
G2
d.f.
36.2
1
3.5 21.4
1 9
0.084
,0.001
1.172 22.470 to 2.017
0.062 0.011
298.0
age tributary
57.9
1
10.1 181.5
1 9
0.133 2.852 27.099 to 4.328
,0.001 0.001 ,0.001
213.443
constant
160 r = 0.718 p < 0.05
120
100
80
60 120
160 170 130 140 150 white-spotted charr L50, first (mm)
180
Figure 2. Relationship between white-spotted charr L50, first and masu salmon L50, first in 10 tributaries. r : Pearson’s correlation.
4. Discussion The threshold sizes at maturity (L50, first) in male charr and salmon in a given tributary were correlated. In both species, the size decreased with increasing distance from the sea. This supports our hypothesis that increased migration costs result in a decreased threshold size at maturity in males of both species, thereby promoting residency (see electronic supplementary material, figure S1c). Distance from the sea to the spawning ground is an indicator of migration cost (energy expenditure and mortality) in anadromous fishes [28]. For example, increased migration costs are associated with a decrease in female ovarian investment and male gonadal mass (e.g. females [12]; males [29]). Male energy reserves and secondary sexual characters decrease with increasing migration cost [30]. The median natural mortality during downriver migration in Atlantic salmon is 2.3% km21 ([31] and references therein). Therefore, the fitness of anadromous individuals decreases with an increase in the distance between the spawning grounds and the sea. By contrast, the fitness of resident individuals is expected to be unrelated to distance from the sea, because
they remain their entire life in their natal stream. Bohlin et al. [32] showed that the fitness of resident brown trout, Salmo trutta, did not change with altitude, whereas the fitness of anadromous brown trout tended to decline with altitude, an indicator of migration cost in that study. Therefore, the threshold size at maturity in male salmonid parr is expected to decrease with distance from the sea, an indicator of migration cost, to promote residency. Genetic factors and plasticity can affect the diversity of lifehistory traits. The threshold size at maturity of fishes is thought to be influenced by both genetic [33] and environmental (i.e. plastic) [34] factors, but their relative contributions to the PMRN remain unclear [17]. With respect to plasticity, Jonsson et al. [35,36] noted that the threshold size at maturity in Atlantic salmon was affected by both water temperature and food quality. In addition, Morita et al. [21] showed in a transplant experiment that the threshold size at maturity in male charr varies with river width and temperature. However, it is difficult to explain why the plasticity affected the divergence of threshold size at maturity in males, which is related to migration cost. In general, some proportion of salmonids in a population mature as precocious males and they stay in their natal stream without going to the sea. For plasticity to be able to affect the threshold size at maturity in males, male parr would have to be able to estimate their distance from the sea without going to sea. On the other hand, in salmonid fishes, anadromous adults can exhibit very precise natal homing [37], and most phenotypic traits are heritable [38]. In Atlantic salmon, the threshold sizes at maturity of male parr differ genetically among geographically distinct populations [39]. Taken together with our results, this suggests that genetic factors that correspond to the variation in migration cost drive the diversity of threshold size at maturity in males. Convergent evolution of similar phenotypes in similar environments may be a general evolutionary phenomenon given its prevalence at a variety of taxonomic scales [1]. A number of recent specific studies have tested factors underlying the evolution of partial migration (e.g. table 1 of Chapman et al. [6]). Our results imply that migration costs drive convergence of the threshold traits for partial migration in two sympatric salmonid fishes. Migration costs such as energy expenditure, mortality and predation during migration are expected to be critical factors affecting the fitness of
Proc R Soc B 280: 20132539
(b) masu salmon fork length
masu salmon L50, first (mm)
p-value
213.34
constant
140
coefficient
rspb.royalsocietypublishing.org
variable
4
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
Table 2. Pearson’s correlation between environmental factors and whitespotted charr L50, first or masu salmon L50, first. Statistical significance in probability tests (p , 0.05) indicated by asterisk.
20.632*
20.646*
mean temperature width
20.068 0.151
0.337 20.165
depth velocity
0.392 0.340
0.371 0.183
mean of substrate value
0.316
20.155
s.d. of substrate value slope
L50, first (mm)
distance
150
100 r = –0.646 p < 0.05
50
20
20.265
40 60 80 distance from sea (km)
0.303
20.146
20.218
0þ FL for each species
20.218
20.053
density for each
20.508
0.385
species density of two
20.510
20.441
species
migratory animals, thus animals determine their migratory tactics to reduce the migration costs (birds [13]; mammals [40]). In addition, populations that migrate long distances optimize their life-history traits to adapt to the increased migration costs (fish [12,30]; birds [41]; invertebrates [42]). Therefore, migration costs might be one general selective factor that produces convergence in migratory tactics and life-history traits. Understanding how continuous phenotypic variation across environmental gradients is generated and maintained is critical to our understanding of adaptive evolution [43]. However, most previous studies of adaptive evolution have considered dichotomous environments [44]. This approach is useful for identifying the causes of phenotypic variation, but is problematic for understanding the incidences of underlying patterns of phenotypic variation [44]. Our results suggest that variation in L50, first covaried with distance to the sea in two species across 10 tributaries covering a range of 100 km. Parallel patterns of threshold sizes at maturity and migration costs have also been detected in Atlantic salmon and brown trout at a river system scale [19,20]. Therefore, migration costs appear to be an evolutionary force that produces clear parallel divergence of life-history traits at a microgeographical scale within river systems.
100
Figure 3. Relationship between the distance from the sea and L50, first of masu salmon (squares) and white-spotted charr (triangles). r, Pearson’s correlation. (Online version in colour.)
Precise homing of salmonid fish would maintain genetic adaptation by promoting reproductive isolation among populations [37]. Such strong homing can generate significant genetic variation among populations only a short distance apart [45]. Our current study showed that within-species variation in L50, first differed between charr and salmon, whereby the CV for L50, first in salmon was nearly twice as large as in charr. The difference in the CV of L50, first might reflect the homing accuracy of the two species, with salmon homing more accurately (and so showing more local adaptation) than charr. We demonstrated an independent effect of migration cost on threshold size after accounting for correlations among possible environmental factors. Nonetheless, the possibility of unaccounted factors affecting plastic effects on the threshold size at maturity in males cannot be ruled out. Further study is needed to evaluate this possibility by conducting common garden experiments to better understand how migration costs can affect threshold size. Acknowledgements. We thank T. Matsuishi for his support and invaluable suggestions, K. Nogami for helping with the fieldwork, K. Okuda and Y. Koganeyama for helping with fish measurements in the laboratory, K. Ohkuma for obtaining the sampling permit issued by the Governor of Hokkaido, Mr Konishi for granting permission to enter a state forest, and J. R. Bower for his critical readings of an earlier version of the manuscript. This study was conducted based on a joint research agreement between Hokkaido University and the Hokkaido National Fisheries Research Institute, Fisheries Research Agency. Finally, we are grateful to N. B. Metcalfe and an anonymous reviewer for their constructive comments on earlier versions of this article. Data accessibility. Fish and environmental data are uploaded as electronic supplementary material.
References 1. 2.
Schluter D. 2000 The ecology of adaptive radiation. Oxford, UK: Oxford University Press. Reznick DN, Bryga HA. 1996 Life-history evolution in guppies (Poecilia reticulata: Poeciliidae). V. Genetic basis of parallelism
3.
in life histories. Am. Nat. 147, 339–359. (doi:10.1086/285855) Dingle H. 2006 Animal migration: is there a common migratory syndrome? J. Ornithol. 147, 212 –220. (doi:10.1007/s10336-005-0052-2)
4.
Cresswell KA, Satterthwaite WH, Sword GA. 2011 Understanding the evolution of migration through empirical examples. In Animal migration: a synthesis (eds EJ Milner-Gulland, JM Fryxell, ARE Sinclair), pp. 7–16. Oxford, UK: Oxford University Press.
Proc R Soc B 280: 20132539
masu salmon L50, first
200
rspb.royalsocietypublishing.org
white-spotted charr L50, first
5 r = –0.632 p < 0.05
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
5.
7.
8.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43. 44.
45.
migration cost. J. Anim. Ecol. 70, 112–121. (doi:10. 1046/j.1365-2656.2001.00466.x) Hutchings JA. 2005 Life history consequences of overexploitation to population recovery in Northwest Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 62, 824–832. (doi:10.1139/F05-081) Uusi-Heikkila¨ S, Kuparinen A, Wolter C, Meinelt T, O’Toole AC, Arlinghaus R. 2011 Experimental assessment of the probabilistic maturation reaction norm: condition matters. Proc. R. Soc. B 278, 709–717. (doi:10.1098/rspb.2010.1507) Jonsson B, Jonsson N, Finstad AG. 2013 Effects of temperature and food quality on age and size at maturity in ectotherms: an experimental test with Atlantic salmon. J. Anim. Ecol. 82, 201–210. (doi:10.1111/j.1365-2656.2012.02022.x) Jonsson B, Finstad AG, Jonsson N. 2012 Winter temperature and food quality affect age at maturity: an experimental test with Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 69, 1817–1826. (doi:10.1139/f2012-108) Quinn TP. 2005 The behavior and ecology of Pacific salmon and trout. Seattle, WA: University of Washington Press. Carlson SM, Seamons TR. 2008 A review of quantitative genetic components of fitness in salmonids: implications for adaptation to future change. Evol. Appl. 1, 222 –238. (doi:10.1111/j. 1752-4571.2008.00025.x) Piche´ J, Hutchings JA, Blanchard W. 2008 Genetic variation in threshold reaction norms for alternative reproductive tactics in male Atlantic salmon, Salmo salar. Proc. R. Soc. B 275, 1571 –1575. (doi:10. 1098/rspb.2008.0251) Wall J, Douglas-Hamilton I, Vollrath F. 2006 Elephants avoid costly mountaineering. Curr. Biol. 16, R527 –R529. (doi:10.1016/j.cub.2006.06.049) Fiedler W. 2005 Ecomorphology of the external flight apparatus of blackcaps (Sylvia atricapilla) with different migration behavior. Ann. NY Acad. Sci. 1046, 253–263. (doi:10.1196/annals.1343.022) Altizer S, Davis AK. 2010 Populations of monarch butterflies with different migratory behaviors show divergence in wing morphology. Evolution 64, 1018 –1028. (doi:10.1111/j.1558-5646.2009. 00946.x) Endler JA. 1977 Geographic variation, speciation, and clines. Princeton, NJ: Princeton University Press. Torres-Dowdall J, Handelsman CA, Ruell EW, Auer SK, Reznick DN, Ghalambor CK. 2012 Fine-scale local adaptation in life histories along a continuous environmental gradient in Trinidadian guppies. Funct. Ecol. 26, 616 –627. (doi:10.1111/j.13652435.2012.01980.x) Garcia de Leaniz C et al. 2007 A critical review of adaptive genetic variation in Atlantic salmon: implications for conservation. Biol. Rev. 82, 173–211. (doi:10.1111/j.1469-185X.2006.00004.x)
6
Proc R Soc B 280: 20132539
9.
20.
parr. J. Anim. Ecol. 73, 253 –260. (doi:10.1111/j. 0021-8790.2004.00803.x) Aubin-Horth N, Bourque J-F, Daigle G, Hedger R, Dodson JJ. 2006 Longitudinal gradients in threshold sizes for alternative male life history tactics in a population of Atlantic salmon (Salmo sala). Can. J. Fish. Aquat. Sci. 63, 2067–2075. (doi:10. 1139/F06-103) Morita K, Tsuboi J, Nagasawa T. 2009 Plasticity in probabilistic reaction norms for maturation in a salmonid fish. Biol. Lett. 5, 628–631. (doi:10.1098/ rsbl.2009.0290) Morita K, Nagasawa T. 2010 Latitudinal variation in the growth and maturation of masu salmon (Oncorhynchus masou) parr. Can. J. Fish. Aquat. Sci. 67, 955– 965. (doi:10.1139/F10-028) Morita K, Morita SH, Nagasawa T. 2010 Seasonal growth patterns of wild masu salmon parr in a subarctic river. J. Natl Taiwan Museum, Spec. Publ. 14, 87 –93. Nakano S, Kitano F, Maekawa K. 1996 Potential fragmentation and loss of thermal habitats for charrs in the Japanese archipelago due to climatic warming. Freshwater Biol. 36, 711– 722. (doi:10. 1046/j.1365-2427.1996.d01-516.x) Bain MB, Finn JT, Booke HE. 1985 Quantifying stream substrate for habitat analysis studies. North Am. J. Fish. Manage. 5, 499–506. (doi:10.1577/ 1548-8659(1985)5,499:QSSFHA.2.0.CO;2) Heino M, Dieckmann U, Godø OR. 2002 Measuring probabilistic reaction norms for age and size at maturation. Evolution 56, 669 –678. (doi:10.1111/ j.0014-3820.2002.tb01378.x) Zar JH. 1996 Biostatistical analysis, 3rd edn. Upper Saddle River, NJ: Prentice Hall. Hendry AP, Bohlin T, Jonsson B, Berg OK. 2004 To sea or not to sea? Anadromy versus non-anadromy in salmonids. In Evolution illuminated: salmon and their relatives (eds AP Hendry, SC Stearns), pp. 92– 125. Oxford, UK: Oxford University Press. Jonsson B, Jonsson N. 2006 Life-history effects of migratory costs in anadromous brown trout. J. Fish Biol. 69, 860 –869. (doi:10.1111/j.1095-8649.2006. 01160.x) Kinnison MT, Unwin MJ, Quinn TP. 2003 Migratory costs and contemporary evolution of reproductive allocation in male chinook salmon. J. Evol. Biol. 16, 1257–1269. (doi:10.1046/j.1420-9101.2003. 00631.x) Thorstad EB, Whoriskey F, Uglem I, Moore A, Rikardsen AH, Finstad B. 2012 A critical life stage of the Atlantic salmon Salmo salar: behaviour and survival during the smolt and initial post-smolt migration. J. Fish Biol. 81, 500–542. (doi:10.1111/j. 1095-8649.2012.03370.x) Bohlin T, Pettersson J, Degerman E. 2001 Population density of migratory and resident brown trout (Salmo trutta) in relation to altitude: evidence for a
rspb.royalsocietypublishing.org
6.
Jonsson B, Jonsson N. 1993 Partial migration: niche shift versus sexual maturation in fishes. Rev. Fish Biol. Fish. 3, 348– 365. (doi:10.1007/BF00043384) Chapman BB, Bro¨nmark C, Nilsson J-A˚ , Hansson L-A. 2011 The ecology and evolution of partial migration. Oikos 120, 1764 –1775. (doi:10.1111/j. 1600-0706.2011.20131.x) Roff DA. 1996 The evolution of threshold traits in animals. Q. Rev. Biol. 71, 3–35. (doi:10.1086/ 419266) Falconer DS, Mackay TFC. 1996 Introduction to quantitative genetics, 4th edn. Harlow, UK: Longman. Pulido F. 2011 Evolutionary genetics of partial migration—the threshold model of migration revis(it)ed. Oikos 120, 1776 –1783. (doi:10.1111/j. 1600-0706.2011.19844.x) Hazel WN, Smock R, Johnson MD. 1990 A polygenic model for the evolution and maintenance of conditional strategies. Proc. R. Soc. Lond. B 242, 181–187. (doi:10.1098/rspb.1990.0122) Wikelski M, Tarlow EM, Raim A, Diehl RH, Larkin RP, Visser GH. 2003 Costs of migration in free-flying songbirds. Nature 423, 704. (doi:10.1038/423704a) Kinnison MT, Unwin MJ, Hendry AP, Quinn TP. 2001 Migratory costs and the evolution of egg size and number in introduced and indigenous salmon populations. Evolution 55, 1656– 1667. (doi:10. 1111/j.0014-3820.2001.tb00685.x) Ydenberg RC, Butler RW, Lank DB, Smith BD, Ireland J. 2004 Western sandpipers have altered migration tactics as peregrine falcon populations have recovered. Proc. R. Soc. Lond. B 271, 1263–1269. (doi:10.1098/rspb.2004.2713) Kato F. 1991 Life histories of masu and amago salmon (Oncorhynchus masou and Oncorhynchus rhodurus). In Pacific salmon life histories (eds C Groot, L Margolis), pp. 448 –520. Vancouver, Canada: UBC Press. Morita K, Morita SH, Yamamoto S. 2009 Effects of habitat fragmentation by damming on salmonid fishes: lessons from white-spotted charr in Japan. Ecol. Res. 24, 711–722. (doi:10.1007/s11284-008-0579-9) Watanabe M, Takamura S, Maekawa K. 2008 Effects of timing of nest entry and body size on the fertilization success of alternative male reproductive phenotypes of masu salmon (Oncorhynchus masou). Can. J. Zool. 86, 1121 –1130. (doi:10.1139/Z08-093) Hutchings JA. 2011 Old wine in new bottles: reaction norms in salmonid fishes. Heredity 106, 421–437. (doi:10.1038/hdy.2010.166) Dodson JJ, Aubin-Horth N, The´riault V, Pa´ez DJ. 2013 The evolutionary ecology of alternative migratory tactics in salmonid fishes. Biol. Rev. 88, 602–625. (doi:10.1111/brv.12019) Baum D, Laughton R, Armstrong JD, Metcalfe NB. 2004 Altitudinal variation in the relationship between growth and maturation rate in salmon
Migration costs drive convergence of threshold traits for migratory tactics Genki Sahashi1,† and Kentaro Morita2
rspb.royalsocietypublishing.org
Research Cite this article: Sahashi G, Morita K. 2013 Migration costs drive convergence of threshold traits for migratory tactics. Proc R Soc B 280: 20132539. http://dx.doi.org/10.1098/rspb.2013.2539
Received: 27 September 2013 Accepted: 8 October 2013
Subject Areas: evolution, ecology Keywords: alternative migratory tactics, life-history traits diversity, migration distance, partial migration, salmonid fish, selection
Author for correspondence: Genki Sahashi e-mail: [email protected]
†
Present address: Division of Biosphere Science, Graduate School of Environmental Sciences, Hokkaido University, Hakodate 041-8611, Japan. Electronic supplementary material is available at http://dx.doi.org/10.1098/rspb.2013.2539 or via http://rspbroyalsocietypublishing.org.
1 Division of Marine Bioresource and Environmental Science, Graduate School of Fisheries Sciences, Hokkaido University, Hakodate 041-8611, Japan 2 Hokkaido National Fisheries Research Institute, Fisheries Research Agency, Sapporo 062-0922, Japan
Partial migration of some, but not all, members of a population is a common form of migration. We evaluated how migration costs influence which members migrate in 10 populations of two salmonid species. The migratory patterns of both species were evaluated based on the size at maturity for resident males, which is the threshold trait that determines the migratory tactics used within a population. In both species, this size was smaller in males located further from the sea, where migration costs are presumably higher. Moreover, the threshold sizes at maturity in males were correlated between both species. Our results suggest that migration costs are a significant convergent selective force on migratory tactics and life-history traits in nature.
1. Introduction Convergent evolution is the process by which similar features evolve independently in different evolutionary lineages [1]. This evolutionary phenomenon provides an ideal means for evaluating adaptation, because the same phenotype evolves repeatedly and independently under similar circumstances [2]. Migratory animals often exhibit similarities in behavioural, physiological, morphological and life-history traits, such as the migratory syndromes shared by insects, fish and birds [3]. These similarities are thought to have arisen through natural selection, which favours convergent solutions to common problems facing migratory animals [4]. To better understand migration, its evolution and its ecology, there is a need for the development of hypotheses that advance general migratory theory and have cross-species applicability [3]. Partial migration is one of the common forms of migration in which a population contains both migratory and resident individuals [5,6]. Whether alternative migratory tactics are adopted is dependent on life-history traits called ‘threshold traits’ [7]. These traits exhibit two properties: a threshold value that divides the phenotypic distribution (e.g. migrants and residents), and an underlying ‘liability’ trait that differs in a continuous variable [8]. If the value of the liability trait exceeds the threshold, an individual will choose a different migratory tactic than if the value of the trait is below the threshold [9]. The threshold value often varies among populations because genetic variability and environmental differences can affect the trait [10]. Threshold traits have been used to describe the inheritance and evolution of partial migration. A number of studies have examined ecological factors underlying the diversity of specific threshold traits in partially migratory animals, but there have been few cross-species investigations in nature. The trade-off between benefits and costs is a fundamental mechanism driving life-history divergence. Thus, the threshold traits of partially migratory animals will coevolve to maximize fitness under a given set of migratory conditions. Costs of migration include increased metabolic and energetic requirements during migration [11], decreased reproductive output [12] and higher predation risk [13]. These effects will presumably increase as the migration distance increases. For partially migratory animals, migration costs affect only the migratory individuals, so migration cost is expected to be a general factor that results in the diversity of threshold traits in partially migratory animals.
& 2013 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
(i) The threshold size at maturity in males decreases with increasing distance from the sea as an indicator of migration cost. (ii) The proportion of male individuals that mature without migrating to the sea increases with increasing distance from the sea. (iii) The divergence of threshold size at maturity in males covaries between the two species among tributaries.
2. Material and methods This study was conducted in the Kushiro river system in eastern Hokkaido, Japan. We chose one reach in each of 25 tributaries based on the presence of typical biotopes for salmonid parr and spawning. We measured a number of environmental factors in each reach (see below for details). We then chose 10 of the tributaries (figure 1) in which distance from the sea was not correlated with other environmental factors at the site (see electronic supplementary material, table S1). In these 10 tributaries, the mean length of the study reach was 81 m (range: 55–90 m; see electronic supplementary material, table S2). In the Kushiro river system, records of hatchery releases of white-spotted charr and masu salmon exist only for Lake Kussharo.
(b) Fish collection and measurement Sampling was conducted between 13 and 16 September 2010, during the spawning season of salmon and about one month before the spawning season of charr. Parr were collected using an electrofisher (200– 300 V DC, model 12B, Smith-Root, Vancouver, WA) and a dip net (30 cm width, 3 mm mesh). All parr were immediately frozen for subsequent laboratory measurements. Anadromous adults were collected in some tributaries, but were not analysed. Tributary 9 was re-sampled on 9 November 2010 as no mature parr were collected during the first sampling. Because age 1þ parr cease to grow from September to November [23], age greater than or equal to 1þ parr were included in the analyses without any size transformation. The following variables were determined for each individual after defrosting: fork length, sex, reproductive state (mature or immature, by visual inspection of gonads) and age (determined using otoliths).
(c) Measurement of environmental factors As a surrogate for migration cost, the distance from the sea was measured for each tributary using a portable GPS receiver (GPSMAP60CSx, Garmin, Olathe, KS) and 1 : 25 000 scale topographic maps (http://www.gsi.go.jp/). Previous studies have used altitude as a surrogate for migration cost, but altitude itself can be an indicator of other environmental factors such as water temperature (e.g. [24]). To avoid multicollinearity, altitude was not used as a model parameter. In addition to distance from the sea, eight environmental factors were measured in each tributary: water temperature, tributary width, tributary depth, velocity, substrate, slope, riverine fish growth and fish density (see electronic supplementary material, tables S2 and S3). The environmental factor values in the study reaches represent those of each tributary (see electronic supplementary material, appendix A). Water temperature was measured at hourly intervals in each tributary study reach using fixed data loggers (Stow-Away TidbiT, Onset Computer Corp., Bourne, MA) deployed on the stream beds from June to August 2010 and November 2010 to May 2011. Mean temperature (averaged for the whole study period) and monthly temperature (averaged for each month) were used as indices of water temperature. Tributary width was measured at 5 m intervals in each study
Proc R Soc B 280: 20132539
(a) Study area
2
rspb.royalsocietypublishing.org
Partial migration is evident in many species of salmonid fish, such as the white-spotted charr, Salvelinus leucomaenis and masu salmon, Oncorhynchus masou. All females and most males in both species live for several years in rivers, then descend to the sea to feed [14,15]. After this feeding migration, these anadromous fish become sexually mature and return to the natal river to spawn [14,15]. By contrast, some males in both species become sexually mature at the parr stage (the darkly pigmented, stream-dwelling juvenile stage of the salmonid life cycle) [14,15]. As with other salmonids, the alternative migratory tactics in both species are associated with alternative reproductive phenotypes during spawning. Anadromous males have high fertilization success by competing for females using their large body size, whereas smaller precocious mature parr achieve fertilizations by adopting a sneaking behaviour [16]. In salmonid males, life-history divergence is thought to depend on whether the current size or growth rate at the time of the life-history decision will result in the male exceeding a threshold size [17,18], such as the size at maturity for parr. Males that exceed this size become precocious mature parr, whereas those below this size become anadromous (see electronic supplementary material, figure S1a). As the threshold size at maturity in males decreases in a population, the proportion of mature male parr increases within the population (see electronic supplementary material, figure S1a,b). Thus, in salmonid populations with high migration costs, the threshold size at maturity in males will be small, resulting in more resident individuals within a population. However, quantifying the effects of migration costs on threshold size at maturity in males is difficult because other environmental factors can covary with migration costs. Namely, the threshold size at maturity in males may be shaped by other evolutionary and plastic responses to environmental gradients that correlate with migration distance. Recent studies on the partially migratory Atlantic salmon, Salmo salar, have shown that threshold size at maturity in males varies among populations, and the variation is influenced by migration costs such as altitude or migration distance [19,20]. However, these studies did not measure the effects of migration costs on threshold size at maturity in males independently, because they did not examine correlations between migration costs and other environmental factors. For example, water temperature often decreases with increasing altitude, so it can also affect the diversity of threshold size at maturity in males (e.g. [21,22]). Thus, to test the hypothesis that migration cost is a general factor in shaping the diversity of threshold traits, the effects of migration costs on the threshold traits must be quantified under a sampling design that considers multicollinearity among environmental factors. In this study, we collected data from 10 tributaries in a single river system to investigate the hypothesis that increased migration cost (i.e. distance of migration) causes a decrease in the threshold size at maturity in males (i.e. promotes residency) in two sympatric salmonids: white-spotted charr and masu salmon. The study was designed to independently quantify the effects of migration costs on the threshold sizes (see Material and methods). Our objective was to identify the evolutionary factors that promote diversity in threshold size at maturity in males of the two species. We would expect to find the following three schematic scenarios (see electronic supplementary material, figure S1c).
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
145º
140º 44º
2
42º
1
Hokkaido Island study area
3 4
1 Ginnosawa 5
7
2 Ishikaribetsu 3 Isobunnai
6 Mitnamitawa 7 Ponsetsuri 8 Oriyomappu 9 8
10
Old river channel
9 Tirishinne 10 Ooshima
10 km
Pacific Ocean
Figure 1. Location of the 10 tributaries in the Kushiro river system. The old river channel has no current because a closed water gate was constructed at the junction of the main stream and the old river channel.
reach. Tributary depth, velocity and substrate were measured at 5 m intervals with three evenly spaced transects (1/4, 1/2, 3/4). Substrate composition was sorted into six classes: sand-silt (less than 2 mm), category 1 (as categorized in electronic supplementary material, table S2); gravel (2–16 mm), category 2; pebble (17–64 mm), category 3; cobble (65–256 mm), category 4; boulder (greater than 256 mm), category 5; irregular bedrock, category 6. The mean and standard deviation of the categorical values were used as indices of substrate composition [25]. Velocity was measured at 60% of the depth from the surface to the bed using a propeller-type meter (VR-301; Kenek Co., Tokyo, Japan). Riverine fish growth was quantified using the average fork length in age 0þ parr. Fish densities were estimated in a reach of each tributary during fish collection using the two-pass removal method (model M(b), program CAPTURE; available at http://www.mbrpwrc.usgs.gov/software/index.html). The density of each species and the combined density of the two salmonids were used as indices of fish density.
(d) Data analyses To assess threshold sizes at maturity in males, we first calculated the probability of being mature at a particular combination of variables (i.e. maturity ogive) using logistic regression. The logistic regression model was described as logit ( p) ¼ a þ b1FL þ b2Age þ Ctributary, where p is the probability of maturation, a is the regression constant, bn are regression coefficients, Ctributary is a tributary-specific constant, FL is the fork length and Age is age in years. The model parameters were estimated using the maximum-likelihood method with a binomial distribution of errors. The significance of model parameters was evaluated using the likelihood-ratio test and the Akaike information criterion (AIC). The threshold size at maturity in males (L50, first) was defined as the fork length at which the probability of maturing at the first age at maturity is 50%. The ages at first maturity are age 1þ in charr [15] and age 0þ in salmon [22]. At the age of first maturity, the fork length at 50% probability of maturing (L50, first) is the same as the fork length at which the probability of becoming
3. Results A total of 1106 charr and 890 salmon were caught during the study. Mature male parr were captured in all tributaries, but no mature female parr were collected (see electronic supplementary material, table S3). The maturity rate for all fish exceeding the minimum age at maturity varied considerably among tributaries (range: 17–100% for charr 1þ and 3–93% for salmon 0þ). The logistic regression analyses identified significant effects of the tributary, fork length and age on the probability of maturing for both charr and salmon males (table 1). The probability of maturing increased with increasing fork length and age, but these relationships differed among tributaries, resulting in substantial variation in L50, first values among tributaries (see electronic supplementary material, table S3 and figure S2). Although the effect of age was insignificant for charr, the model incorporating all three variables (Ctributary, Age and FL) had the smallest AIC value for both species (DAIC . 1.5). There was a positive correlation between the L50, first values of charr and salmon (figure 2), suggesting there was spatial covariation of threshold size at maturity in males between the species. Among the nine environmental variables we measured, only distance from the sea was negatively correlated with L50, first values in both species (table 2). None of the monthly temperatures were significantly correlated with L50, first values in both species (see electronic supplementary material, table S4). Moreover, more complex models did not perform better at explaining the probability of maturing for either species (see electronic supplementary material, appendix B). Distance from the sea was not significantly correlated with the other eight environmental variables (see electronic supplementary material, table S1). This indicates that threshold size at maturity in males (L50, first) decreased with increasing distance from the sea, an indicator of migration cost (figure 3). The decrease in the threshold size at maturity in males (L50, first) was associated with a significant increase in the maturity rate for all fish exceeding the minimum age at maturity in both charr (r ¼ 20.802, p ¼ 0.005) and salmon (r ¼ 20.861, p ¼ 0.001). The maturity rate for all fish exceeding the minimum age at maturity increased as the distance from the sea increased, although the correlation with charr was insignificant (charr, r ¼ 0.265, p ¼ 0.460; salmon, r ¼ 0.663, p ¼ 0.037). The CV of L50, first differed significantly between charr and salmon (Z1,18 ¼ 2.07, p ¼ 0.038); it was more than twice as large in salmon (23%) as in charr (11%).
Proc R Soc B 280: 20132539
6 4 Numaoro 5 Hitominai
3
rspb.royalsocietypublishing.org
Lake Kussharo
mature is 50%, a feature of the probabilistic maturation reaction norm (PMRN) [26]. We calculated L50, first using regression coefficients of the logistic regression described as 2(a þ b2Age þ Ctributary)/b1, where Age is 0 and 1 for salmon and charr, respectively. We used L50, first to test for interspecific covariation and to examine intraspecific variation along environmental gradients. The proportion of male individuals that matured without migrating to the sea was quantified using the maturity rate for all fish exceeding the minimum age at maturity (i.e. charr 1þ; salmon 0þ). The coefficient of variation (CV) was used as a measure of variation in L50, first among tributaries within species. Interspecific differences in CV of L50, first were tested using the Z-test [27]. All statistical analyses were performed using R v. 2.14.0 for Windows (http://www.r-project.org/).
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
Table 1. Results of a logistic regression of fork length, age and tributary on the probability of maturing for age 1þ male charr and 0þ male salmon. ln L, log likelihood for the full model; G 2, statistic of the likelihood-ratio test. ln L
(a) white-spotted charr fork length
275.2
age tributary
G2
d.f.
36.2
1
3.5 21.4
1 9
0.084
,0.001
1.172 22.470 to 2.017
0.062 0.011
298.0
age tributary
57.9
1
10.1 181.5
1 9
0.133 2.852 27.099 to 4.328
,0.001 0.001 ,0.001
213.443
constant
160 r = 0.718 p < 0.05
120
100
80
60 120
160 170 130 140 150 white-spotted charr L50, first (mm)
180
Figure 2. Relationship between white-spotted charr L50, first and masu salmon L50, first in 10 tributaries. r : Pearson’s correlation.
4. Discussion The threshold sizes at maturity (L50, first) in male charr and salmon in a given tributary were correlated. In both species, the size decreased with increasing distance from the sea. This supports our hypothesis that increased migration costs result in a decreased threshold size at maturity in males of both species, thereby promoting residency (see electronic supplementary material, figure S1c). Distance from the sea to the spawning ground is an indicator of migration cost (energy expenditure and mortality) in anadromous fishes [28]. For example, increased migration costs are associated with a decrease in female ovarian investment and male gonadal mass (e.g. females [12]; males [29]). Male energy reserves and secondary sexual characters decrease with increasing migration cost [30]. The median natural mortality during downriver migration in Atlantic salmon is 2.3% km21 ([31] and references therein). Therefore, the fitness of anadromous individuals decreases with an increase in the distance between the spawning grounds and the sea. By contrast, the fitness of resident individuals is expected to be unrelated to distance from the sea, because
they remain their entire life in their natal stream. Bohlin et al. [32] showed that the fitness of resident brown trout, Salmo trutta, did not change with altitude, whereas the fitness of anadromous brown trout tended to decline with altitude, an indicator of migration cost in that study. Therefore, the threshold size at maturity in male salmonid parr is expected to decrease with distance from the sea, an indicator of migration cost, to promote residency. Genetic factors and plasticity can affect the diversity of lifehistory traits. The threshold size at maturity of fishes is thought to be influenced by both genetic [33] and environmental (i.e. plastic) [34] factors, but their relative contributions to the PMRN remain unclear [17]. With respect to plasticity, Jonsson et al. [35,36] noted that the threshold size at maturity in Atlantic salmon was affected by both water temperature and food quality. In addition, Morita et al. [21] showed in a transplant experiment that the threshold size at maturity in male charr varies with river width and temperature. However, it is difficult to explain why the plasticity affected the divergence of threshold size at maturity in males, which is related to migration cost. In general, some proportion of salmonids in a population mature as precocious males and they stay in their natal stream without going to the sea. For plasticity to be able to affect the threshold size at maturity in males, male parr would have to be able to estimate their distance from the sea without going to sea. On the other hand, in salmonid fishes, anadromous adults can exhibit very precise natal homing [37], and most phenotypic traits are heritable [38]. In Atlantic salmon, the threshold sizes at maturity of male parr differ genetically among geographically distinct populations [39]. Taken together with our results, this suggests that genetic factors that correspond to the variation in migration cost drive the diversity of threshold size at maturity in males. Convergent evolution of similar phenotypes in similar environments may be a general evolutionary phenomenon given its prevalence at a variety of taxonomic scales [1]. A number of recent specific studies have tested factors underlying the evolution of partial migration (e.g. table 1 of Chapman et al. [6]). Our results imply that migration costs drive convergence of the threshold traits for partial migration in two sympatric salmonid fishes. Migration costs such as energy expenditure, mortality and predation during migration are expected to be critical factors affecting the fitness of
Proc R Soc B 280: 20132539
(b) masu salmon fork length
masu salmon L50, first (mm)
p-value
213.34
constant
140
coefficient
rspb.royalsocietypublishing.org
variable
4
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
Table 2. Pearson’s correlation between environmental factors and whitespotted charr L50, first or masu salmon L50, first. Statistical significance in probability tests (p , 0.05) indicated by asterisk.
20.632*
20.646*
mean temperature width
20.068 0.151
0.337 20.165
depth velocity
0.392 0.340
0.371 0.183
mean of substrate value
0.316
20.155
s.d. of substrate value slope
L50, first (mm)
distance
150
100 r = –0.646 p < 0.05
50
20
20.265
40 60 80 distance from sea (km)
0.303
20.146
20.218
0þ FL for each species
20.218
20.053
density for each
20.508
0.385
species density of two
20.510
20.441
species
migratory animals, thus animals determine their migratory tactics to reduce the migration costs (birds [13]; mammals [40]). In addition, populations that migrate long distances optimize their life-history traits to adapt to the increased migration costs (fish [12,30]; birds [41]; invertebrates [42]). Therefore, migration costs might be one general selective factor that produces convergence in migratory tactics and life-history traits. Understanding how continuous phenotypic variation across environmental gradients is generated and maintained is critical to our understanding of adaptive evolution [43]. However, most previous studies of adaptive evolution have considered dichotomous environments [44]. This approach is useful for identifying the causes of phenotypic variation, but is problematic for understanding the incidences of underlying patterns of phenotypic variation [44]. Our results suggest that variation in L50, first covaried with distance to the sea in two species across 10 tributaries covering a range of 100 km. Parallel patterns of threshold sizes at maturity and migration costs have also been detected in Atlantic salmon and brown trout at a river system scale [19,20]. Therefore, migration costs appear to be an evolutionary force that produces clear parallel divergence of life-history traits at a microgeographical scale within river systems.
100
Figure 3. Relationship between the distance from the sea and L50, first of masu salmon (squares) and white-spotted charr (triangles). r, Pearson’s correlation. (Online version in colour.)
Precise homing of salmonid fish would maintain genetic adaptation by promoting reproductive isolation among populations [37]. Such strong homing can generate significant genetic variation among populations only a short distance apart [45]. Our current study showed that within-species variation in L50, first differed between charr and salmon, whereby the CV for L50, first in salmon was nearly twice as large as in charr. The difference in the CV of L50, first might reflect the homing accuracy of the two species, with salmon homing more accurately (and so showing more local adaptation) than charr. We demonstrated an independent effect of migration cost on threshold size after accounting for correlations among possible environmental factors. Nonetheless, the possibility of unaccounted factors affecting plastic effects on the threshold size at maturity in males cannot be ruled out. Further study is needed to evaluate this possibility by conducting common garden experiments to better understand how migration costs can affect threshold size. Acknowledgements. We thank T. Matsuishi for his support and invaluable suggestions, K. Nogami for helping with the fieldwork, K. Okuda and Y. Koganeyama for helping with fish measurements in the laboratory, K. Ohkuma for obtaining the sampling permit issued by the Governor of Hokkaido, Mr Konishi for granting permission to enter a state forest, and J. R. Bower for his critical readings of an earlier version of the manuscript. This study was conducted based on a joint research agreement between Hokkaido University and the Hokkaido National Fisheries Research Institute, Fisheries Research Agency. Finally, we are grateful to N. B. Metcalfe and an anonymous reviewer for their constructive comments on earlier versions of this article. Data accessibility. Fish and environmental data are uploaded as electronic supplementary material.
References 1. 2.
Schluter D. 2000 The ecology of adaptive radiation. Oxford, UK: Oxford University Press. Reznick DN, Bryga HA. 1996 Life-history evolution in guppies (Poecilia reticulata: Poeciliidae). V. Genetic basis of parallelism
3.
in life histories. Am. Nat. 147, 339–359. (doi:10.1086/285855) Dingle H. 2006 Animal migration: is there a common migratory syndrome? J. Ornithol. 147, 212 –220. (doi:10.1007/s10336-005-0052-2)
4.
Cresswell KA, Satterthwaite WH, Sword GA. 2011 Understanding the evolution of migration through empirical examples. In Animal migration: a synthesis (eds EJ Milner-Gulland, JM Fryxell, ARE Sinclair), pp. 7–16. Oxford, UK: Oxford University Press.
Proc R Soc B 280: 20132539
masu salmon L50, first
200
rspb.royalsocietypublishing.org
white-spotted charr L50, first
5 r = –0.632 p < 0.05
Downloaded from http://rspb.royalsocietypublishing.org/ on March 29, 2015
5.
7.
8.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43. 44.
45.
migration cost. J. Anim. Ecol. 70, 112–121. (doi:10. 1046/j.1365-2656.2001.00466.x) Hutchings JA. 2005 Life history consequences of overexploitation to population recovery in Northwest Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 62, 824–832. (doi:10.1139/F05-081) Uusi-Heikkila¨ S, Kuparinen A, Wolter C, Meinelt T, O’Toole AC, Arlinghaus R. 2011 Experimental assessment of the probabilistic maturation reaction norm: condition matters. Proc. R. Soc. B 278, 709–717. (doi:10.1098/rspb.2010.1507) Jonsson B, Jonsson N, Finstad AG. 2013 Effects of temperature and food quality on age and size at maturity in ectotherms: an experimental test with Atlantic salmon. J. Anim. Ecol. 82, 201–210. (doi:10.1111/j.1365-2656.2012.02022.x) Jonsson B, Finstad AG, Jonsson N. 2012 Winter temperature and food quality affect age at maturity: an experimental test with Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 69, 1817–1826. (doi:10.1139/f2012-108) Quinn TP. 2005 The behavior and ecology of Pacific salmon and trout. Seattle, WA: University of Washington Press. Carlson SM, Seamons TR. 2008 A review of quantitative genetic components of fitness in salmonids: implications for adaptation to future change. Evol. Appl. 1, 222 –238. (doi:10.1111/j. 1752-4571.2008.00025.x) Piche´ J, Hutchings JA, Blanchard W. 2008 Genetic variation in threshold reaction norms for alternative reproductive tactics in male Atlantic salmon, Salmo salar. Proc. R. Soc. B 275, 1571 –1575. (doi:10. 1098/rspb.2008.0251) Wall J, Douglas-Hamilton I, Vollrath F. 2006 Elephants avoid costly mountaineering. Curr. Biol. 16, R527 –R529. (doi:10.1016/j.cub.2006.06.049) Fiedler W. 2005 Ecomorphology of the external flight apparatus of blackcaps (Sylvia atricapilla) with different migration behavior. Ann. NY Acad. Sci. 1046, 253–263. (doi:10.1196/annals.1343.022) Altizer S, Davis AK. 2010 Populations of monarch butterflies with different migratory behaviors show divergence in wing morphology. Evolution 64, 1018 –1028. (doi:10.1111/j.1558-5646.2009. 00946.x) Endler JA. 1977 Geographic variation, speciation, and clines. Princeton, NJ: Princeton University Press. Torres-Dowdall J, Handelsman CA, Ruell EW, Auer SK, Reznick DN, Ghalambor CK. 2012 Fine-scale local adaptation in life histories along a continuous environmental gradient in Trinidadian guppies. Funct. Ecol. 26, 616 –627. (doi:10.1111/j.13652435.2012.01980.x) Garcia de Leaniz C et al. 2007 A critical review of adaptive genetic variation in Atlantic salmon: implications for conservation. Biol. Rev. 82, 173–211. (doi:10.1111/j.1469-185X.2006.00004.x)
6
Proc R Soc B 280: 20132539
9.
20.
parr. J. Anim. Ecol. 73, 253 –260. (doi:10.1111/j. 0021-8790.2004.00803.x) Aubin-Horth N, Bourque J-F, Daigle G, Hedger R, Dodson JJ. 2006 Longitudinal gradients in threshold sizes for alternative male life history tactics in a population of Atlantic salmon (Salmo sala). Can. J. Fish. Aquat. Sci. 63, 2067–2075. (doi:10. 1139/F06-103) Morita K, Tsuboi J, Nagasawa T. 2009 Plasticity in probabilistic reaction norms for maturation in a salmonid fish. Biol. Lett. 5, 628–631. (doi:10.1098/ rsbl.2009.0290) Morita K, Nagasawa T. 2010 Latitudinal variation in the growth and maturation of masu salmon (Oncorhynchus masou) parr. Can. J. Fish. Aquat. Sci. 67, 955– 965. (doi:10.1139/F10-028) Morita K, Morita SH, Nagasawa T. 2010 Seasonal growth patterns of wild masu salmon parr in a subarctic river. J. Natl Taiwan Museum, Spec. Publ. 14, 87 –93. Nakano S, Kitano F, Maekawa K. 1996 Potential fragmentation and loss of thermal habitats for charrs in the Japanese archipelago due to climatic warming. Freshwater Biol. 36, 711– 722. (doi:10. 1046/j.1365-2427.1996.d01-516.x) Bain MB, Finn JT, Booke HE. 1985 Quantifying stream substrate for habitat analysis studies. North Am. J. Fish. Manage. 5, 499–506. (doi:10.1577/ 1548-8659(1985)5,499:QSSFHA.2.0.CO;2) Heino M, Dieckmann U, Godø OR. 2002 Measuring probabilistic reaction norms for age and size at maturation. Evolution 56, 669 –678. (doi:10.1111/ j.0014-3820.2002.tb01378.x) Zar JH. 1996 Biostatistical analysis, 3rd edn. Upper Saddle River, NJ: Prentice Hall. Hendry AP, Bohlin T, Jonsson B, Berg OK. 2004 To sea or not to sea? Anadromy versus non-anadromy in salmonids. In Evolution illuminated: salmon and their relatives (eds AP Hendry, SC Stearns), pp. 92– 125. Oxford, UK: Oxford University Press. Jonsson B, Jonsson N. 2006 Life-history effects of migratory costs in anadromous brown trout. J. Fish Biol. 69, 860 –869. (doi:10.1111/j.1095-8649.2006. 01160.x) Kinnison MT, Unwin MJ, Quinn TP. 2003 Migratory costs and contemporary evolution of reproductive allocation in male chinook salmon. J. Evol. Biol. 16, 1257–1269. (doi:10.1046/j.1420-9101.2003. 00631.x) Thorstad EB, Whoriskey F, Uglem I, Moore A, Rikardsen AH, Finstad B. 2012 A critical life stage of the Atlantic salmon Salmo salar: behaviour and survival during the smolt and initial post-smolt migration. J. Fish Biol. 81, 500–542. (doi:10.1111/j. 1095-8649.2012.03370.x) Bohlin T, Pettersson J, Degerman E. 2001 Population density of migratory and resident brown trout (Salmo trutta) in relation to altitude: evidence for a
rspb.royalsocietypublishing.org
6.
Jonsson B, Jonsson N. 1993 Partial migration: niche shift versus sexual maturation in fishes. Rev. Fish Biol. Fish. 3, 348– 365. (doi:10.1007/BF00043384) Chapman BB, Bro¨nmark C, Nilsson J-A˚ , Hansson L-A. 2011 The ecology and evolution of partial migration. Oikos 120, 1764 –1775. (doi:10.1111/j. 1600-0706.2011.20131.x) Roff DA. 1996 The evolution of threshold traits in animals. Q. Rev. Biol. 71, 3–35. (doi:10.1086/ 419266) Falconer DS, Mackay TFC. 1996 Introduction to quantitative genetics, 4th edn. Harlow, UK: Longman. Pulido F. 2011 Evolutionary genetics of partial migration—the threshold model of migration revis(it)ed. Oikos 120, 1776 –1783. (doi:10.1111/j. 1600-0706.2011.19844.x) Hazel WN, Smock R, Johnson MD. 1990 A polygenic model for the evolution and maintenance of conditional strategies. Proc. R. Soc. Lond. B 242, 181–187. (doi:10.1098/rspb.1990.0122) Wikelski M, Tarlow EM, Raim A, Diehl RH, Larkin RP, Visser GH. 2003 Costs of migration in free-flying songbirds. Nature 423, 704. (doi:10.1038/423704a) Kinnison MT, Unwin MJ, Hendry AP, Quinn TP. 2001 Migratory costs and the evolution of egg size and number in introduced and indigenous salmon populations. Evolution 55, 1656– 1667. (doi:10. 1111/j.0014-3820.2001.tb00685.x) Ydenberg RC, Butler RW, Lank DB, Smith BD, Ireland J. 2004 Western sandpipers have altered migration tactics as peregrine falcon populations have recovered. Proc. R. Soc. Lond. B 271, 1263–1269. (doi:10.1098/rspb.2004.2713) Kato F. 1991 Life histories of masu and amago salmon (Oncorhynchus masou and Oncorhynchus rhodurus). In Pacific salmon life histories (eds C Groot, L Margolis), pp. 448 –520. Vancouver, Canada: UBC Press. Morita K, Morita SH, Yamamoto S. 2009 Effects of habitat fragmentation by damming on salmonid fishes: lessons from white-spotted charr in Japan. Ecol. Res. 24, 711–722. (doi:10.1007/s11284-008-0579-9) Watanabe M, Takamura S, Maekawa K. 2008 Effects of timing of nest entry and body size on the fertilization success of alternative male reproductive phenotypes of masu salmon (Oncorhynchus masou). Can. J. Zool. 86, 1121 –1130. (doi:10.1139/Z08-093) Hutchings JA. 2011 Old wine in new bottles: reaction norms in salmonid fishes. Heredity 106, 421–437. (doi:10.1038/hdy.2010.166) Dodson JJ, Aubin-Horth N, The´riault V, Pa´ez DJ. 2013 The evolutionary ecology of alternative migratory tactics in salmonid fishes. Biol. Rev. 88, 602–625. (doi:10.1111/brv.12019) Baum D, Laughton R, Armstrong JD, Metcalfe NB. 2004 Altitudinal variation in the relationship between growth and maturation rate in salmon
