Canadian Journal of Fisheries and Aquatic Sciences
Trade-offs in the adaptation towards hatchery and natural conditions drive survival, migration, and angling vulnerability in a territorial fish in the wild
Journal: Canadian Journal of Fisheries and Aquatic Sciences
Manuscript ID cjfas-2018-0256.R1
Manuscript Type: Article
Date Submitted by the 09-Nov-2018 Author:
Complete List of Authors: Tsuboi, Jun-ichi; Fisheris Research Agency, National Research Institute of Aquaculture Kaji, Kohichi; Yamanashi Prefectural Fisheries Technology Center Baba, Shinya; Logics of Blue Arlinghaus, Robert; Humboldt University of Berlin,
Adaptation,Draft Captive breeding, Hatchery-induced selection, Stocking, Keyword: Territorial behaviour
Is the invited manuscript for consideration in a Special Not applicable (regular submission) Issue? :
https://mc06.manuscriptcentral.com/cjfas-pubs Page 1 of 47 Canadian Journal of Fisheries and Aquatic Sciences
1 Trade-offs in the adaptation towards hatchery and natural conditions drive
2 survival, migration, and angling vulnerability in a territorial fish in the wild
3
4 Jun-ichi Tsuboi, Kohichi Kaji, Shinya Baba, and Robert Arlinghaus
5
6 J. Tsuboi. * Yamanashi Prefectural Fisheries Technology Center, Kai, Yamanashi
7 400-0121, Japan
8 e-mail: [email protected]
9 K. Kaji. Yamanashi Prefectural Fisheries Technology Center, Kai, Yamanashi
10 400-0121, Japan
11 e-mail: [email protected]
12 S. Baba. Logics of Blue, Hyogo, 650-0025, Japan
13 e-mail: [email protected]
14 R. Arlinghaus. Department of Biology and Ecology of Fishes, Leibniz-Institute of
15 Freshwater Ecology and Inland Fisheries, Müggelseedamm 310, 12587 Berlin,
16 Germany; Division of Integrative Fisheries Management, Albrecht-Daniel-Thaer
17 Institute of Agricultural and Horticultural Sciences, Department for Crop and Animal
18 Sciences, Faculty of Life Sciences, Humboldt-Universität zu Berlin, Invalidenstrasse
19 42, 10115 Berlin, Germany.
20 e-mail: [email protected]
21
22 Corresponding author:
23 Jun-ichi Tsuboi
24 National Research Institute of Fisheries Science, Japan Fisheries Research and
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 2 of 47
25 Education Agency, Nikko, Tochigi 321-1661, Japan.
26 Phone: +81-288-55-0055
27 Fax: +81-288-55-0064
28 e-mail: [email protected]
29
30 *Present address: National Research Institute of Fisheries Science, Japan Fisheries
31 Research and Education Agency, Nikko, Tochigi 321-1661, Japan
32
33
34 Draft
https://mc06.manuscriptcentral.com/cjfas-pubs Page 3 of 47 Canadian Journal of Fisheries and Aquatic Sciences
35 Abstract
36 Hatchery fish to support capture fisheries need to thrive in both hatchery and natural
37 environments. We conducted joint experiments in both environments with individuals
38 stemming from multiple generations held in captivity to test the performance of
39 hatchery-reared ayu (Plecoglossus altivelis). Ayu has an annual, herbivorous, territorial
40 and amphidromous riverine fish native to Japan of high importance to recreational
41 fisheries. Hatchery fish of the 1st hatchery generation exhibited poor growth and highest
42 malformation rates relative to the 2nd and following hatchery generations. The 1st
43 generation offspring stocked into a natural stream also showed low survival and poor
44 vulnerability to angling, suggesting that maladaptation to the hatchery environment
45 explained the performance in the wild.Draft By contrast, offspring of the 7th to 9th
46 generations exhibited high growth in the hatchery environment, but when stocked into
47 the wild they also exhibited low survival, maladapted migratory behaviour, and again
48 poor vulnerability to angling. Consequently, intermediate generations held in captivity
49 were found to offer the best fisheries performance and can thus be recommended for
50 enhancements to support recreational fisheries.
51
52 Keywords Adaptation; Captive breeding; Hatchery-induced selection; Stocking;
53 Territorial behaviour
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 4 of 47
54 Introduction
55 Stocking of hatchery-reared fish has been frequently applied in fisheries
56 enhancements and conservation, particularly in freshwater environments (Cowx 1994;
57 Lorenzen et al. 2012). Stocking programs are successful if hatchery-fish survive and
58 grow in the wild, contributing to fisheries catch or maintaining threatened or declining
59 populations. However, rapid environmental change from natural to artificial conditions
60 is expected to drive rapid evolutionary change (reviewed by Bay et al. 2017). Indeed,
61 hatchery selection negatively affects fitness in the wild, particularly in salmonids (Araki
62 et al. 2007; Christie et al. 2014, 2016, 2018). The hatchery environment offers strikingly
63 different selection pressures to natural conditions (e.g., in terms of chemical conditions,
64 density of fish, access to food, presenceDraft of predators, parasite load), such that adaptation
65 to the hatchery environment can be expected to exert rapid fitness effects on genotypes,
66 gene expression and phenotype development (Araki et al. 2008; Christie et al. 2016;
67 Uusi-Heikkilä et al. 2017). Hatchery selection effects leading to trait divergence among
68 hatchery and natural populations likely accumulate over generational time, but are most
69 severe early on in the adaptation process to artificial conditions (Araki et al. 2007, 2008)
70 and may occur even within the first generation in captivity through epigenetic effects
71 (Christie et al. 2016). In particular, hatchery selection has been shown to rapidly modify
72 phenotypic variance and mean trait values involved in maturation and growth and
73 associated behavioural and physiological traits considered important for efficient
74 biomass production in aquaculture settings (Huntingford 2004; Kostow 2004). In this
75 study, we followed Teletchea and Fontaine (2014) and defined domestication as the
76 gradual adaptation of an organism to living conditions that are determined by some
77 form of human intervention. Releasing highly domesticated fishes into the wild
https://mc06.manuscriptcentral.com/cjfas-pubs Page 5 of 47 Canadian Journal of Fisheries and Aquatic Sciences
78 promises high mortality post release (Lorenzen 2006), at the possible benefit of high
79 return to the fisheries catch when large, recruited fishes are stocked that are immediately
80 targeted by recreational anglers (Mezzera and Largiadèr 2001; Baer et al. 2007;
81 Lorenzen et al. 2012).
82 Although a range of studies have shown that hatchery fish have lower fitness (e.g.,
83 survival) in the wild (Lorenzen 2000, 2006), it is less clear which phenotypic traits are
84 exactly affected and which fitness components – behaviour, growth, survival or
85 reproduction – are altered alone or in combination in hatchery selection (e.g., Brown
86 and Day 2002; Miller et al. 2004; Araki and Schmid 2010). Moreover, from an
87 anthropocentric perspective, the current body of literature suggests that some traits
88 important to fisheries may even benefitDraft from hatchery selection – notably individual
89 vulnerability to capture, and by the same token population-level catchability, which is
90 often found to be higher in domesticated phenotypes compared to wild phenotypes
91 because domesticated fish are more explorative, bolder and grow faster, in turn showing
92 greater food intake rates and vulnerability to angling (Lorenzen et al. 2012; Klefoth et al.
93 2012, 2013). Thus, although hatchery selection should lower natural fitness, it should at
94 the same time benefit vulnerability to fishing, assuming that fish that are stocked are
95 morphologically vulnerable to fishing gear (Lennox et al. 2017). However, no studies
96 have quantified the trade-offs among natural and hatchery selection in terms of natural
97 and fisheries performance over several generations in captivity when these fish are
98 stocked into the wild to support recreational fisheries catch.
99 In stock enhancements, hatchery-reared fish are released to be recaptured after
100 stocking after some period in the wild (Lorenzen et al. 2012). Hatchery fish are usually
101 more-exploratory and bolder than wild conspecifics, and as a result hatchery fish are
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 6 of 47
102 often highly vulnerable to angling gear (Klefoth et al. 2012, 2013; Härkönen et al. 2014,
103 2016; Koeck et al. in press; reviewed in Lennox et al. 2017 and Arlinghaus et al. 2017).
104 Because domestication changes all of these traits, domesticated fish of both salmonids
105 and cyprinids were indeed found to be more readily captured than less domesticated fish
106 (Mezzera and Largiadèr 2001; Klefoth et al. 2012), leaving behind individuals, which
107 are less explorative and more timid (Alós et al. 2016; Tsuboi et al. 2016; Arlinghaus et
108 al. 2017). The fact that vulnerability to angling is a function of domestication relates to
109 a range of behavioural and physiological traits that co-vary with adaption to artificial
110 environments and that render domesticated fish a good model for trait complexes
111 affecting vulnerability to angling (Klefoth et al. 2012, 2013; Arlinghaus et al. 2017).
112 According to recent reviews andDraft empirical studies, passive gear types, such as gill
113 nets or rod-and-reel, are supposed to selectively catch bold, explorative, aggressive,
114 stress resilient, proactive and generally risk taking fish (Biro and Post 2008; Arlinghaus
115 et al. 2017; Diaz Pauli and Sih 2017; Louison et al. 2017; Klefoth et al. 2017). However,
116 does selection in captivity always and in every generation change (plastically or
117 genetically) all behaviours toward more vulnerable phenotypes? Or is it also possible
118 that some key behaviours determining vulnerability, such as aggression (Sutter et al.
119 2012), might be lost over time when fish are held in captivity (Ruzzante 1994)?
120 Aggression is often a key component of angling vulnerability in top predators (Sutter et
121 al. 2012; Wilson et al. 2015). Some fishing styles used in non-predatory fishes also take
122 advantage of aggressive and territorial behaviour, such that a possible loss of aggression
123 over generational time in hatcheries might have repercussions for fisheries catch also in
124 these species. For example, in Japan ayu (Plecoglossus altivelis) constitutes a popular
125 target for anglers. The fish is an annual, amphidromous and highly territorial fish that is
https://mc06.manuscriptcentral.com/cjfas-pubs Page 7 of 47 Canadian Journal of Fisheries and Aquatic Sciences
126 herbivorous by grazing periphyton from substrates in streams. Angling of ayu uses a
127 very peculiar method that takes advantage of the territorial behaviour shown by ayu to
128 monopolize benthic algae attached to substrates (Fig. 1). In brief, a conspecific is
129 attached to a fishing line and held with a long pole in vicinity to a territorial ayu who
130 attacks the supposed intruder and is then hooked on a freely hanging treble hook (Fig.
131 1). Thus, the angling style of ayu does not use natural foraging of a baited hook as cue,
132 but takes advantage of the attack behaviour of a territorial grazer on conspecifics.
133 Because of the popularity of ayu fishing in Japan, captive breeding and release
134 programs are widely used to supplement wild populations (Iguchi 1997). Several
135 comparative studies on wild and hatchery-reared ayu have demonstrated significant
136 changes in genetic diversity, behaviour,Draft morphology, and immune function in hatchery
137 fish over time (Iguchi et al. 1999; Miwa et al. 2003; Yoshizawa 2003; Ikeda et al.
138 2005).
139 In ayu angling, it is known that larger and more aggressive individuals are more
140 vulnerable relative to smaller ones (Katano and Iguchi, 1996). Consequently, ayu
141 angling shows positively size-dependent vulnerability patterns to angling (Miura et al.
142 2012; Sato and Tsuboi 2018). Moreover, ayu individuals migrating upstream after
143 stocking were found to be more vulnerable to angling (Tsukamoto et al. 1990a, b).
144 However, territorial and (or) stronger migration tendencies are likely maladaptive in
145 highly-crowded hatchery rearing conditions (Ruzzante 1994; Brockmark and Johnsson
146 2010) and might thus be lost over generational time in captivity. Long-term
147 domesticated ayu may in response also be less vulnerable to angling after stocking, in
148 contrast to what the literature on the “domesticated phenotype” currently assumes
149 (Lorenzen et al. 2012).
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 8 of 47
150 We tested the performance of hatchery-reared ayu in both hatchery and natural
151 environments in a number of generations in captivity, ranging from the 1st to the 9th
152 generation, in a real-scale hatchery and stocking program. In the hatchery, as fitness
153 measures, the growth rate and malformation rates were monitored in the 1st up to the
154 9th generation, and subsequently after stocking the migration, survival and vulnerability
155 to angling in the wild were assessed. We used experimental exploitation in a natural
156 stream for assessing fitness in the wild and studied the trade-off among natural and
157 hatchery adaptation on natural and fisheries-driven fitness components.
158 Both evolutionary adaption and cultured experience may contribute to the
159 domestication syndrome (Fleming et al., 1997), but our study was not designed to
160 disentangle among genetic and plasticDraft effects. Instead, we aimed at more generally
161 testing the aggregate effects of domestication over several generations on phenotypic
162 expressions and angling vulnerability using ayu as a model for a non-piscivorous
163 territorial fish, and use results to inform fisheries management based on stocking.
164
165 Material and Methods
166
167 Hatchery rearing
168 Ayu were cultured in 29 independent artificial ponds in the Yamanashi Prefectural
169 Fisheries Technology Center in Japan (35°42'16"N, 138°31'26"E), where approximately
170 2 million individuals are reared per year and stocked into the natural streams in the Fuji
171 River basin (Fig.2). Each study year, we produced two cohorts of ayu, which were
172 characterized by different number of generations held in captivity. They were always
173 separated by artificial ponds. Both cohorts in all study years originally originated from
https://mc06.manuscriptcentral.com/cjfas-pubs Page 9 of 47 Canadian Journal of Fisheries and Aquatic Sciences
174 wild fish caught around the estuarine region of the Fuji River (35°06'55"N,
175 138°38'25"E, Fig. 2).
176 Between 30 September and 7 October 2009, fertilization was conducted. Firstly,
177 1,500 females of parent candidates were checked whether ovulation had occurred by
178 gently pressing fish abdomen. Subsequently, all ovulated females were used for
179 fertilization. Gametes from multiple females (5 to 7 individuals) and equal number of
180 males are combined during fertilization. A total of 121 and 107 female ayu (body wet
181 mass; 123 ± 22 g, mean ± SD, egg mass, c. 0.46 mg) were used for producing fertilized
182 eggs of females of the 4th and 8th generation in captivity, respectively (Table S1). A
183 total of 5,620,000 and 5,130,000 fertilized and cohesive eggs of the new 5th and 9th
184 generation, respectively, were attachedDraft to 125 plates (100 ×100 × 3 cm,
185 width×length×thickness) made out of vinylidene chloride fiber (1,000 denier,
186 Asahi-kasei, Co., Ltd., Tokyo), following standard hatchery practice for ayu. The
187 average ratio of eyed eggs (N of eyed eggs / N of total eggs) were 50.8 and 53.4 % in
188 the 5th and 9th generations, respectively (Table S1). 14 days after fertilization, the egg
189 plates were transferred from the holding tanks (100 × 200 × 100 cm,
190 width×length×depth) supplied with freshwater to octagon-shaped 50 m2 tanks filled
191 with artificial saltwater (70 cm in depth). The two cohorts were kept separated. The day
192 after transfer, the embryos started to hatch in the octagon-shaped tanks. Transfer from
193 freshwater to saltwater was needed because in the wild ayu migrate to the estuary
194 coastal zone just after hatching in downstream sections of the river basin in autumn and
195 then back upstream up the river in spring (Katano and Iguchi, 1996). The artificial
196 seawater contained NaCl：0.301 %，MgSO4：0.078 %，MgCl2：0.060 %，CaCl2：
197 0.018 %，KCl：0.008 %，NaHCO3：0.003 %, maintained at 15 ºC with recirculating
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 10 of 47
198 filtration (Miura et al. 2012). Each day during the holding period after hatching, larval
199 and juvenile fish were given live rotifers (Brachionus plicatilis) and a formulated ayu
200 diet composed mainly of fish meal, vitamins C, E, and eicosapentaenoic and
201 docosahexaenoic acid (“Rescue A” produced by Scientific Feed Laboratory Co., Ltd.,
202 Tokyo) to satiation using automatic feeders. Every 10 days after hatching, from each
203 hatchery pond 10 individuals were weighed as total body mass and the average mass per
204 individual was calculated.
205 Ayu were sorted by body size using 4 mm mesh 90 days after hatch to minimize
206 cannibalism. Large-size individuals were used for this study as these are usually used
207 for stocking. After sorting, fish were stocked back into the 50 m2 tanks per 70,000
208 individuals (estimated by mean bodyDraft mass as above) and acclimatized to freshwater by
209 decreasing salinity over five days. Afterwards, the fish were reared with spring water at
210 16 ºC in a flow-through system. From sorting to stocking, fish were given a formulated
211 diet composed mainly of fish meal and euphausiid (“AYU SOFT No.2” produced by
212 NOSAN Co., Ltd., Tokyo) to satiation using automatic feeders. Although two cohorts of
213 ayu were reared in different tanks throughout cultivation, environmental and dietary
214 conditions were identical among the cohorts. Flavobacterium psychrophilum, which
215 produces coldwater disease, were never detected using PCR (polymerase chain reaction)
216 targeted at the Peptidyl-prolyl cis-trans isomerase C Gene in 60 individuals randomly
217 sampled from each cohort just before stocking (Yoshiura et al. 2006). At the same time,
218 the malformation rates (lack of gill cover, lack or underdevelopment of each fin, throat
219 projection) were checked in 313 individuals per cohort via random sampling using a dip
220 net (Table 1).
221 Using the same approach as above, we cultivated among 153 to 723 thousand
https://mc06.manuscriptcentral.com/cjfas-pubs Page 11 of 47 Canadian Journal of Fisheries and Aquatic Sciences
222 individuals of other generations of ayu; 1st generation (originating from wild fish
223 caught in the study river basin in March 2010) and 6th generations in captivity in 2010,
224 2nd and 7th generations in 2011, and 3rd and 8th generations in 2012 (Table S1, Table
225 1). In this way, early and long-term domesticated cohorts were directly compared in a
226 paired fashion for their performance both in hatchery and natural conditions in each
227 year. The limitation of the design is that each generation is only used once and
228 compared within a year to one other generation. However, as we analyzed our data
229 pooling all study years, we considered the design robust to see among the generation
230 patterns in performance would they emerge.
231
232 Stocking and fishing Draft
233 Six months after hatching, paired releases were conducted for four years using two
234 cohorts (Table 1). The two cohorts were stocked into the Arakawa River (35°42'14"N
235 138°31'29"E, altitude 360 m, river width 14.84 ± 4.18 m, mean ± SD), just beside the
236 Prefectural Fisheries Technology Center, located about 95 km upstream from the river
237 mouth. One of two cohorts of ayu were clipped their adipose fins as group marking
238 (Table 2). Swimming performance of ayu has been found before to be unaffected by this
239 practice (Katano and Uchida 2006). Approximately 10,000 individuals (mean body
240 mass 11.3‒14.3g) were stocked per cohort per year (Table 3). In 2013, only 7,515 and
241 8,082 individuals of the two cohorts (of the 3rd and 8th generations) were stocked,
242 respectively, because of scale malfunctioning in some fishes at the time of sorting. Wild
243 ayu were not able to migrate to the Arakawa River due to weirs (Fig. 1). No other
244 hatchery-reared ayu were stocked into the study site. Note that the ayu has an annual
245 life-cycle and they need temperatures of at least 9 ˚C for wintering (Tachihara and
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 12 of 47
246 Kimura 1988; Sakae et al. 1996). In the study reach, water temperature drops to 2 ˚C
247 during mid-winter (Tsuboi unpublished data). Therefore, there were no wild recruits in
248 the study reach to compete with the stocked individuals.
249 Between late June and early October, angling experiments were conducted by staff of
250 the Prefectural Fisheries Technology Center in a 1 km river stretch that was the
251 designed study area (500 m up- and downstream from the stocking point, Fig. 2), using
252 a 7.2 m pole (long rod without a reel), a 1 lbs. monofilament line (1 pound = 0.453 kg),
253 without sinker, equipped with a hatchery-reared ayu fixed by a nose ring, to which a
254 barbless treble hook (gape width of 6.0 mm) was attached just after the anal fin (Fig. 1,
255 Table 2). A small cascade (2.5 m in height) prevented almost all individuals to migrate
256 upstream at the upper end of study area.Draft Experimental angling was conducted in a
257 randomly chosen site (up- or downstream from a stocking point) for 0.5 – 8 hours per
258 angling day. First, hatchery-reared ayu from the Prefectural Fisheries Technology
259 Center were used as a live-bait, subsequently the caught individuals (i.e., stocked and
260 angled ayu) were used as the bait. During the angling experiments, water temperature
261 ranged from 14.8 to 27.0 ˚C (21.3 ± 2.4 ˚C, mean ± SD), which was a suitable range for
262 ayu foraging (and fishing) based on agonistic behaviour (15.0 – 27.5 ˚C, Uchida et al.
263 1995). Angling pressure of any other recreational anglers was negligible, as riparian
264 trees covered part of river line of the study area, preventing the use of typical
265 angling-rods (c. 9.0 m in length). Between June and August, alternative sampling was
266 also undertaken twice using a casting net (mesh size: 1 cm, net diameter: 3.8 m,
267 maximum distance from fisher to fish: 6.8 m) as active gear type. Approximately 50
268 casts were conducted per sampling episode using random sampling around potential ayu
269 habitats all over the study area.
https://mc06.manuscriptcentral.com/cjfas-pubs Page 13 of 47 Canadian Journal of Fisheries and Aquatic Sciences
270
271 Data analyses
272 1) Growth and malformation rate in hatchery conditions
273 The body size at 80 days after hatching in each tank was compared among 1st, 2nd
274 and the 3rd and more generations in captivity using Mann-Whitney's U test. The
275 frequencies of malformed individuals were compared between two cohorts in each year
276 using Fisher's exact tests.
277 2) Survival, migration behaviour, and vulnerability to angling in a field experiment
278 Firstly, to evaluate the performance in the natural stream after stocking, survival,
279 migration behaviour, and vulnerability to angling were compared between two cohorts
280 each year using G-tests. We used a DraftBonferroni correction of the p values to account for
281 a total of 12 G-tests that were conducted (Demšar 2006). The number of individuals
282 caught by the casting net was used as an index of survival. The vulnerability to angling
283 was evaluated by comparing the number of individuals caught by angling and casting
284 nets, respectively, among cohorts. Migration behaviour was evaluated by the recapture
285 points where ayu were caught by casting nets (up- or downstream from the stocking
286 points) because juvenile ayu migrate upstream into the river from the coastal littoral
287 zone in their natural life cycle (Nishida 1986).
288 3) Angling catch-per-unit effort (CPUE) in the field experiment
289 Using all data collected over four years of angling experiments, a generalized linear
290 mixed model (GLMM) with a Poisson error distribution was used. This model is
291 follows as
292
2 293 f(Catch) = β0 + β1 Generation + β2 (Generation) +β3 Point + β4 Date + log (Effort) +
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 14 of 47
294 αi(j) + δj
2 295 αi(j) ~ Normal (0, 휎훼 )
2 296 δj ~ Normal (0, 휎훿 ),
297
298 where f is log function. Catch is the number of individuals caught by angling.
299 Generation is the number of generations in captivity assumed as continuous variable.
300 Square of the number of generations in captivity was used to test for non-linear effects.
301 Point is recapture point (upstream from stocking point = 1, downstream = 0). Date is the
302 number of days from 1st June (stocking date), which is normalized to a mean of 0 and a
303 standard deviation of 1. Effort is angling effort (h), which was used as offset term. αi(j) is
304 a random effect of cohort (brood lines)Draft i of data ID j; i = A; 5th, 6th, 7th, 8th, i = B; 9th,
305 i = C; 1st, 2nd, 3rd generations in captivity. Since overdispersion was recognized in a
306 preliminary data analysis, a random effect of individual data was included in the model.
307 δj is random effect of individual data ID; j=1, 2, 3, ..., N, where N is sample size (164).
308 To clarify the trade-off of the number of generations in captivity on angling CPUE, we
309 calculated generation effect which is as follows,
310
2 311 Generation effect = β1 Generation + β2 (Generation)
312
313 95% confidence intervals of the generation effects were estimated using nonparametric
314 bootstrap method (bootstrap sample size is same as data sample size; bootstrap sample
315 was generated as 5,000 times).
316 4) Vulnerability to angling in the field experiment
317 Note that CPUE data integrate both abundance and vulnerability to angling and thus
https://mc06.manuscriptcentral.com/cjfas-pubs Page 15 of 47 Canadian Journal of Fisheries and Aquatic Sciences
318 are not a clean measure of vulnerability to the gear. Instead, a generalized linear mixed
319 model (GLMM) with a binomial error distribution was used to compare the probability
320 of being caught by angling (i.e., recapture probability of an individual by angling
321 relative to an angling-independent fish availability estimate using casting nets) across
322 the number of generations in captivity assumed as continuous variable. This model is
323 follows as
324
2 325 f(Angling prob) = β0 + β1 Generation + β2 (Generation) +β3 Length +β4 Date
326 + β5 Point + αi(j)
2 327 αi(j) ~ Normal (0, 휎훼 ),
328 Draft
329 where f is logit function. Angling prob expresses the probability of being caught by
330 angling (not by casting nets). Length is total length (mm) at recapture. Both Length and
331 Date were normalized to a mean of 0 and SD of 1. αi(j) is a random effect of cohort
332 (brood lines) i of data ID j; i = A; 5th, 6th, 7th, 8th, i = B; 9th, i = C; 1st, 2nd, 3rd
333 generations in captivity. As with the angling CPUE model, the generation effects and
334 confidence interval were calculated.
335 All analyses were conducted using the R machinery (version 3.3.2). The Akaike
336 information criterion (AIC) was used to identify the best-fitting models comparing all
337 possible subsets of main effects for all GLMM analyses.
338
339 Results
340 Over four years, two cohorts of ayu were cultured each year ranging from the first to
341 the ninth generation held in captivity, except for the fourth generation where no data
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 16 of 47
342 were available (Table 1). Despite the identical culture conditions, the cohorts of the 1st
343 and 2nd generations in captivity revealed a smaller body size than those of the third and
344 more generations 80 days after hatching (Mann–Whitney U test, n = 21, p = 0.024, Fig.
345 3). The 1st generation also showed the highest malformation rates, which were
346 significantly higher than those of the 6th generation in 2011 (Fisher's exact test, p =
347 0.023, Table 1). In other study years, no significant differences in size at day 80 and
348 malformation rates was detected among cohorts (p = 0.695 in 2010, p = 0.079 in 2012, p
349 = 1.000 in 2013, Table 1).
350 After stocking hatchery-reared ayu into a natural stream, a total of 422 and 840
351 individuals were recaptured by angling (Fig. 1) and casting nets, respectively,
352 throughout the four-year experimentDraft (Tables 2, 3). Ayu recaptured by angling were
353 larger than those recaptured by casting nets (angling; 176 ± 27.1 mm, casting net; 149 ±
354 28.6 mm, mean fork length ± SD, F = 260.6, p < 0.001). In 2010, the cohorts of the 9th
355 generation in captivity showed significantly lower survival than those of the 5th
356 generation (G-test, p = 0.024, Table 3a). Similarly, the 1st generation in 2011 revealed
357 lower survival than those of the 6th generation. Moreover, the 1st generation fish were
358 less vulnerable to angling relative to the 6th generation offspring (Table 3b), indicating
359 less pronounced territorial behaviour during foraging of the 1st generation offspring
360 relative to the 6th generation offspring. Although only marginally significant, the
361 upstream migration behaviour of the 2nd and 3rd generations were more pronounced
362 than that shown by fish of the 7th and 8th generations in 2012 and 2013, respectively
363 (Table 3c). The cohort of the 3rd generation in 2013 were also more vulnerable to
364 angling relative to fish from the 8th generation.
365 Based on the AIC, both the number of generations in captivity and the square of the
https://mc06.manuscriptcentral.com/cjfas-pubs Page 17 of 47 Canadian Journal of Fisheries and Aquatic Sciences
366 number of generations in captivity were selected as best model of angling CPUE (N /
367 person hour) based on territorial behaviour throughout the four-year hatchery rearing
368 and angling experiments (Table 4). The estimated generation effect (linear and quadratic
369 effects of the number of generations in captivity) on angling CPUEs exhibited a
370 dome-shape curve where intermediate generations held in captivity (i.e., generation 5)
371 exhibited a relatively higher coefficient of angling CPUE compared to other generations
372 (Fig. 4). Accordingly, ayu cultured in the 1st and 9th generations in captivity showed
373 substantially lower fisheries potential as a stocking cohort for recreational angling:
374 while the 1st generation fish were poor performers in the wild, the 9th generation fish
375 performed even less well, thereby strongly reducing abundance through mortality and
376 hence penalizing catch rates by angling.Draft
377 We also performed an analysis explicitly modelling the vulnerability to angling
378 independent of abundance using individual-level capture data based on angling
379 compared to casting nets. Based on the AIC, the full model was identified as best model
380 (Table 5). Accordingly, larger individuals were much more vulnerable to angling than
381 smaller ones (Fig. S1). We also detected a negative impact of season on vulnerability.
382 This can be explained by the fact the vulnerability to angling in ayu is strongly size
383 dependent and fish growth very fast. Thus, a say, 180 mm fish in June is among the
384 fast-growing fish, while the same sized fish in September is a slow-growing fish. Then,
385 all else being equal (including length of fish), season exerts a negative impact on
386 vulnerability (Fig. S1). The individuals recaptured in upstream sections from the
387 stocking point were also caught by angling more frequently compared to the fishes
388 caught by the casting net (Table 5). The 3rd generation held in captivity exhibited a
389 relatively higher coefficient of vulnerability (Fig. 5). Accordingly, ayu cultured in 8th
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 18 of 47
390 and 9th generations in captivity showed substantially lower vulnerability to angling.
391
392 Discussion
393
394 Trade-offs in hatchery and natural selection
395 Our results showed that hatchery selection creates trade-offs in ayu performance in
396 the wild. We found the return to the catch was both low when fish were initially
397 adapting to the hatchery environment and when they were domesticated long term, with
398 intermediate generations performing better when judged from a fisheries perspective.
399 Thus, long domestication results in a penalty to catchability. The reason relates to
400 domestication effects both on naturalDraft fitness components (survival, migration) and the
401 pattern by which the ayu forage and are captured using an aggression mechanism, as
402 will be elaborated below.
403 We found both early generations, particularly the first generation held in captivity,
404 and late generations exhibiting lower survival in nature and reduced return to catch after
405 stocking into a natural stream relative to intermediate generations. The 1st generation in
406 captivity was challenged by the novel hatchery environment to which the fish were not
407 previously exposed. They in turn showed weak growth at the larval stages in the
408 hatchery and high malformation rates as indicator of poor performance. However, the
409 growth rate in the hatchery improved after 3rd generations held in captivity, suggesting
410 ayu responded to selection and adapted to hatchery conditions within a few generations.
411 Such pattern of rapid adaptation (“contemporary evolution”) is in agreement with a
412 large body of research in salmonids (Huntingford 2004; Huntingford and Adams 2005;
413 Waples and Hendry 2008; Christie et al. 2016).
https://mc06.manuscriptcentral.com/cjfas-pubs Page 19 of 47 Canadian Journal of Fisheries and Aquatic Sciences
414 The first generation struggled in captivity. This generation did not easily adapt to
415 forage on formulated diet, particularly just after hatching, consequently the
416 malformation rate was significantly greater in the 1st compared to the 6th generation
417 fish in the same study year, and the growth rates were substantially smaller. These
418 results have been found in other wild fish when initially adapting to hatchery conditions
419 (e.g., Eknath et al. 1993; Gjedrem 2000; Härkönen et al. 2017). Supporting our results,
420 Kaji (2016) reported the first generation of ayu had also much higher malformation rate
421 (71.2 %) than those of the 5th generation held in captivity (1.3 %), and work by
422 Kataoka and Suzuki (2007) showed that individuals of ayu that initially were slow
423 growers were much more difficult to angle than initially fast growing fishes, despite a
424 uniform size at the time of stockingDraft into the wild. A major limitation of our study is the
425 lack of replication in the number of generations in captivity, and hence we cannot
426 conclusively relate our study findings to domestication in a cause-and-effect or
427 evolutionary manner.
428
429 Negative effect of high rearing density
430 The rearing density (400-500 individuals / m2) in the hatchery setting was much
431 higher than the densities typically reported in the wild (c. 2 individuals / m2; Katano
432 2014). These dramatic differences in density (Jørgensen et al. 1993), together with the
433 control of diseases and parasites in the hatchery environment and the circumvention of
434 sexual selection (Thériault et al. 2011) are candidate factors involved in rapid hatchery
435 selection, particularly in the 1st and 2nd generation. High density and simplified
436 physical environments are known to cause loss of territorial behaviour in dominance
437 hierarchies in salmonids (Hasegawa and Yamamoto 2008, 2010; Warnock and
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 20 of 47
438 Rasmussen 2013), and thus likely affected the hatchery selection in ayu because the cost
439 of territory was likely too high when foraging on formulated diets exclusively in large
440 densities (compare also work in salmonids and pike, Esox lucius, Brockmark and
441 Johnsson 2010; Hühn et al. 2014). As a result, the generation effect on vulnerability to
442 angling started to be decreased at earlier generation (i.e., 3rd generation) relative to the
443 generation effect on angling CPUE that started to decline only by the 7th generation.
444 The 1st generation in captivity may also have suffered substantial stress induced by the
445 high density and non-natural food, and in turn their juvenile growth curve was steeply
446 depressed compared to further generations. The high rearing density is known to
447 suppress the development of the thymus, which administers disease resistance in ayu
448 (Iguchi et al. 2003; Miwa et al. 2003).Draft By contrast, substantially lower densities (20 %
449 of the usual hatchery density) has been shown to enable adaptation of the 1st generation
450 of ayu successfully, which led to the same survival rates and angling vulnerabilities
451 compared to fish from the 2nd generation in other studies (Yogo 2010; Moriyama
452 2013). Density reduction alone has been found to strongly alter the survival rate of
453 fishes when released from hatcheries into the wild (Brockmark and Johnsson 2010;
454 Larsen et al. 2016), suggesting that density and related stressors maybe among the most
455 important selective pressures in the adaptation from the wild to hatcheries.
456
457 Long-term domestication impacts on behavioural patterns
458 With increasing generation time, we found long-term domesticated ayu to become
459 less vulnerable to angling (particularly in the 7th - 9th generation). Long domestication
460 reduced the natural migration patterns upstream after stocking, suggesting difficulties to
461 adapt to natural environments because upstream migration is the natural behavioural
https://mc06.manuscriptcentral.com/cjfas-pubs Page 21 of 47 Canadian Journal of Fisheries and Aquatic Sciences
462 pattern in this species. Long term domestication is known to reduce the swimming
463 performance in salmonids (Reinbold et al. 2009; Bellinger et al. 2014), which might
464 have played a role in explaining our findings. In hatchery environments, ayu have also
465 been found to sacrifice natural territorial behaviour to achieve higher growth rate and
466 stress tolerance (Awata et al. 2011). A reduction in territorial behaviour comes along
467 with the difficulty of defending natural territories using benthic algae thriving on rocky
468 substrates, and given the way ayu altered territorial behaviour it must penalize catch
469 rates. Less territorial fishes might suffer from lower fitness, due to a lower rank in a
470 dominance hierarchy (Nakano 1995) and lower defense capacity for spawning nests
471 (Sutter et al. 2012). A decrease in fitness in the wild is well known in salmonids with
472 increasing number of generations inDraft captivity (Araki et al. 2008), which we also found
473 in ayu. Importantly, however, these patterns are not a linear function of generation time
474 because we found a dome-shaped relationship of generation time and performance in
475 nature and in relation to ayu fishing. Supporting our findings, low-inbred strains (4th to
476 6th generations in captivity) of ayu were previously found to exhibit higher survival rate
477 and greater vulnerability to angling than high-inbred strains (13th to 15th) (Yoneyama
478 et al. 1997).
479 After stocking, the 7th and 8th generations held in captivity were significantly less
480 frequently migrating upstream relative to the 2nd and 3rd generations. Upstream
481 migration behaviour is natural in the life cycle of ayu, which is characterized by
482 migration from coastal sites up the rivers in spring (Nishida 1986). The degree of
483 upstream migration is also correlated with the vulnerability to angling in ayu
484 (Tsukamoto et al. 1990a, b), which was also shown in the present work. However,
485 hatchery-reared ayu were found to more often migrate downstream relative to wild
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 22 of 47
486 individuals after stocking (Tsukamoto et al. 1990a, b), and similarly migratory
487 behaviour in salmonids have been found to be different from the behaviour of wild fish
488 (e.g., Jonsson et al. 1991; Finstad et al. 2005; Kostow 2009). Swimming performance
489 and migration are very important in the natural foraging of benthic algae, which are
490 more abundant in high velocity upstream sites with sufficient oxygen. High velocity
491 riffles are suitable habitat to form exclusive territories (Katano and Iguchi 1996), and
492 upstream sites were accordingly found to lead to higher vulnerability to angling in the
493 present work. The reduced swimming performance and altered migratory behaviour
494 upstream, the higher mortality and possibly the reduction of general aggressiveness are
495 good candidates why high levels of domestication reduced vulnerability of ayu to
496 angling based on territorial behaviour.Draft
497
498 Conclusions and management implications
499 In conclusion, both unintentional domestication selection and relaxation of natural
500 selection, due to artificially modified and high density rearing environments, are
501 occurring in hatcheries of ayu. Initial challenges to adapt to the hatchery environment
502 reduces health and in turn reduces performance after stocking, penalizing vulnerability
503 to fishing. As domestication proceeds to multiple generations, fish become well adapted
504 to the artificial environments but loose behavioural patterns (upstream migration,
505 territorial behaviour) that are key for both natural fitness and for ayu being captured
506 using the peculiar fishing method based on agonistic interactions. Therefore, our study
507 challenges the commonly expressed belief that increasing domestication selection
508 increases catchability by promoting boldness and growth (Lorenzen et al. 2012). Instead
509 our work suggests that there is a middle ground in terms of number of generation in
https://mc06.manuscriptcentral.com/cjfas-pubs Page 23 of 47 Canadian Journal of Fisheries and Aquatic Sciences
510 captivity where the vulnerability to angling, survival post stocking and thus abundance
511 and population-level catchability are maximized in ayu. At the same time, each
512 generation in captivity seem to render the fish increasingly well adapted to the artificial
513 environment, which pays large costs for natural fitness, particularly in terms of
514 reproductive fitness (Araki et al. 2007). Management using domesticated fishes thus
515 seems to be mainly providing benefits to fisheries in terms of rapid recapture following
516 put-and-take type of fisheries (Lorenzen et al. 2012), and ayu seems no exception. The
517 contribution of strongly domesticated hatchery fish to supporting wild fish conservation,
518 however, is doubtful (Araki et al. 2007; Lorenzen et al. 2012; Glover et al. 2017) and
519 should be studied in future work. As a possible future direction, gene flow from wild
520 fish could restore fitness of later generationsDraft of hatchery fish. Indeed, the hybrid ayu
521 between hatchery female and wild male has shown intermediate susceptibility to
522 bacterial coldwater disease in between hatchery and wild fish (Nagai and Sakamoto
523 2006). Before such work becomes available, we suggest to stock, where considered
524 necessary, intermediate generation-time-ayu of genotypes most similar to the wild
525 stocks, being cognizant that this type of management most likely creates and supports a
526 socially valued put-and-take type fishery.
527
528
529 Acknowledgements
530 We thank Kouji Hada, Youji Omori, Kenji Yoshino, Masayuki Miura, Akihiko
531 Ashizawa, Takumi Okazaki, Shoko Amano and the staff at Yamanashi Prefectural
532 Fisheries Technology Center for culturing fish, stocking, and angling experiment
533 throughout the study and reviewers for feedback. Funding was provided by JSPS
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 24 of 47
534 KAKENHI Grant Numbers JP21925006 (J.T.).
535 536
Draft
https://mc06.manuscriptcentral.com/cjfas-pubs Page 25 of 47 Canadian Journal of Fisheries and Aquatic Sciences
537 References
538 Alós, J., Palmer, M., Rosselló, R., and Arlinghaus, R. 2016. Fast and behavior-selective
539 exploitation of a marine fish targeted by anglers. Sci. Rep. 6: 38093, doi:
540 10.1038/srep38093.
541 Araki, H., Cooper, B., and Blouin, M.S. 2007. Genetic effects of captive breeding cause
542 a rapid, cumulative fitness decline in the wild. Science 318: 100–103. doi:
543 10.1126/science.1145621.
544 Araki, H., Berejikian, B.A., Ford, M.J., and Blouin, M.S. 2008. Fitness of
545 hatchery-reared salmonids in the wild. Evol. Appl. 1(2): 342–355. doi:
546 10.1111/j.1752-4571.2008.00026.x.
547 Araki, H., and Schmid, C. 2010. Is hatcheryDraft stocking a help or harm?: Evidence,
548 limitations and future directions in ecological and genetic surveys. Aquaculture
549 308Suppl.1: S2–S11. doi: 10.1016/j.aquaculture.2010.05.036.
550 Arlinghaus, R., Laskowski, K.L., Alós, J., Klefoth, T., Monk, C.T., Nakayama, S., and
551 Schröder, A. 2017. Passive gear-induced timidity syndrome in wild fish
552 populations and its potential ecological and managerial. Fish and Fish. 18(2):
553 360–373. doi: 10.1111/faf.12176.
554 Awata, S., Tsuruta, T., Yada, T., and Iguchi, K. 2011. Effects of suspended sediment on
555 cortisol levels in wild and cultured strains of ayu Plecoglossus altivelis.
556 Aquaculture 314: 115–121. doi: 10.1016/j.aquaculture.2011.01.024.
557 Baer, J., Blasel, K., and Diekmann, M. 2007. Benefits of repeated stocking with adult,
558 hatchery-reared brown trout, Salmo trutta, to recreational fisheries. Fish. Manag.
559 Ecol. 14(1): 51–59. doi: 10.1111/j.1365-2400.2006.00523.x.
560 Bay, R.A., Rose, N., Barrett, R., Bernatchez, L., Ghalambor, C.K., Lasky, J.R., Brem,
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 26 of 47
561 R.B., Palumbi, S.R., and Ralph, P. 2017. Predicting responses to contemporary
562 environmental change using evolutionary response architectures. Am. Nat. 189(5):
563 463–473. doi: 10.1086/691233.
564 Bellinger, K.L., Thorgaard, G.H., and Carter, P.A. 2014. Domestication is associated
565 with reduced burst swimming performance and increased body size in clonal
566 rainbow trout lines. Aquaculture 420-421: 154–159. doi:
567 10.1016/j.aquaculture.2013.10.028.
568 Biro, P.A., and Post, J.R. 2008. Rapid depletion of genotypes with fast growth and bold
569 personality traits from harvested fish populations. Proc. Natl. Acad. Sci. U.S.A.
570 105(8): 2919–2922. doi: 10.1073/pnas.0708159105.
571 Brockmark, S., and Johnsson J.I. 2010.Draft Reduced hatchery rearing density increases
572 social dominance, postrelease growth, and survival in brown trout (Salmo trutta).
573 Can. J. Fish. Aquat. Sci. 67(2): 288–295. doi: 10.1139/F09-185.
574 Brown, C., and Day, R.L. 2002. The future of stock enhancements: lessons for hatchery
575 practice from conservation biology. Fish and Fish. 3(2): 79–94. doi:
576 10.1046/j.1467-2979.2002.00077.x.
577 Christie, M.R., Ford, M.J., and Blouin, M.S. 2014. On the reproductive success of
578 early-generation hatchery fish in the wild. Evol. Appl. 7(8): 883–896.
579 doi:10.1111/eva.12183.
580 Christie, M.R., Marine, M.L., Fox, S.E., French, R.A., and Blouin, M.S. 2016. A single
581 generation of domestication heritably alerts the expression of hundreds of genes.
582 Nature Communications 7: 10676. doi: 10.1038/ncomms10676.
583 Christie, M.R., McNickle, G.G., French, R.A., and Blouin, M.S. 2018. Life history
584 variation is maintained by fitness trade-offs and negative frequency-dependent
https://mc06.manuscriptcentral.com/cjfas-pubs Page 27 of 47 Canadian Journal of Fisheries and Aquatic Sciences
585 selection. Proc. Natl. Acad. Sci. U.S.A. doi: 10.1073/pnas.1801779115.
586 Cowx, I.G. 1994. Stocking strategies. Fish. Manag. Ecol. 1(1): 15–30. doi:
587 10.1111/j.1365-2400.1970.tb00003.x.
588 Demšar, J. 2006. Statistical Comparisons of Classifiers over Multiple Data Sets. J.
589 Mach. Learn. Res. 7: 1–30. Available from
590 http://www.jmlr.org/papers/volume7/demsar06a/demsar06a.pdf [accessed 9
591 November 2018].
592 Diaz Pauli, B., and Sih, A. 2017. Behavioural responses to human-induced change:
593 Why fishing should not be ignored. Evol. Appl. 10(3): 231–240. doi:
594 10.1111/eva.12456.
595 Eknath, A.E., Tayamen, M.M., Palada-deDraft Vera, M.S., Danting, J.C., Reyes, R.A.,
596 Dionisio, E.E., Capili, J.B., Bolivar, H.L., Abella, T.A., Circa, A.V., Bentsen,
597 H.B., Gjerde, B., Gjedrem, T., and Pullin, R.S.V. 1993. Genetic improvement of
598 farmed tilapias: the growth performance of eight strains of Oreochromis niloticus
599 tested in different farm environments. Aquaculture 111(1-4): 171–188. doi:
600 10.1016/0044-8486(93)90035-W.
601 Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., and McKinley, R.S. 2005.
602 Migration of hatchery-reared Atlantic salmon and wild anadromous brown trout
603 post-smolts in a Norwegian fjord system. J. Fish Biol. 66(1): 86–96. doi:
604 10.1111/j.0022-1112.2005.00581.x.
605 Fleming, I.A., Lamberg, A., and Jonsson, B. 1997. Effects of early experience on the
606 reproductive performance of Atlantic Salmon. Behav. Ecol. 8(5): 470–480. doi:
607 10.1093/beheco/8.5.470.
608 Gjedrem, T. 2000. Genetic improvement of cold-water fish species. Aquacult. Res.
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 28 of 47
609 31(1): 25–33. doi: 10.1046/j.1365-2109.2000.00389.x.
610 Glover, K.L., Solberg, M.F., McGinnity, P., Hindar, K., Verspoor, E., Coulson, M.W.,
611 Hansen, M.M., Araki, H., Skaala, Ø., and Svåsand, T. 2017. Half a century of
612 genetic interaction between farmed and wild Atlantic salmon: Status of
613 knowledge and unanswered questions. Fish and Fish. 18(5): 890–927. doi:
614 10.1111/faf.12214.
615 Härkönen, L., Hyvärinen, P., Paappanen, J., and Vainikka, A. 2014. Explorative
616 behavior increases vulnerability to angling in hatchery-reared brown trout (Salmo
617 trutta). Can. J. Fish. Aquat. Sci. 71(12): 1900–1909. doi:
618 10.1139/cjfas-2014-0221.
619 Härkönen, L., Hyvärinen, P., Paappanen,Draft J., and Vainikka, A. 2015. Behavioural
620 variation in Eurasian perch populations with respect to relative catchability. Acta
621 Ethol. 19(1): 21–31. doi: 10.1007/s10211-015-0219-7.
622 Härkönen, L., Hyvärinen, P., Mehtätalo, L., and Vainikka, A. 2017. Growth, survival
623 and interspecific social learning in the first hatchery generation of Eurasian perch
624 (Perca fluviatilis). Aquaculture 466: 64–71. doi:
625 10.1016/j.aquaculture.2016.09.027.
626 Hasegawa K., and Yamamoto S. 2008. Effects of competitor density and physical
627 habitat structure on the competitive intensity of territorial white spotted charr
628 Salvelinus leucomaenis. J. Fish Biol. 73(1): 213–219. doi:
629 10.1111/j.1095-8649.2008.02133.x.
630 Hasegawa, K., and Yamamoto S. 2010. The effect of flow regime on the occurrence of
631 interference and exploitative competition in a salmonid species, white-spotted
632 char (salvelinus leucomaenis). Can. J. Fish. Aquat. Sci. 67(11): 1776–1781. doi:
https://mc06.manuscriptcentral.com/cjfas-pubs Page 29 of 47 Canadian Journal of Fisheries and Aquatic Sciences
633 10.1139/F10-100.
634 Hühn, D., Lübke, K., Skov, C., and Arlinghaus, R. 2014. Natural recruitment,
635 density-dependent juvenile survival, and the potential for additive effects of stock
636 enhancement: an experimental evaluation of stocking northern pike (Esox lucius)
637 fry. Can. J. Fish. Aquat. Sci. 71(10): 1508–1519. doi: 10.1139/cjfas-2013-0636.
638 Huntingford, F.A. 2004. Implications of domestication and rearing conditions for the
639 behaviour of cultivated fishes. J. Fish Biol. 65(Suppl. S1): 122–142. doi:
640 10.1111/j.0022-1112.2004.00562.x.
641 Huntingford, F., and Adams, C. 2005. Behavioural syndromes in farmed fish:
642 implications for production and welfare. Behaviour 142(9-10): 1207–1221. doi:
643 10.1163/156853905774539382.Draft
644 Iguchi, K., Watanabe, K., and Nishida, M. 1999. Reduced mitochondrial DNA variation
645 in hatchery populations of ayu (Plecoglossus altivelis) cultured for multiple
646 generations. Aquaculture 178: 235–243. doi: 10.1016/S0044-8486(99)00133-7.
647 Iguchi, K., Ogawa, K., Nagae, M., and Ito, F. 2003. The influence of rearing density on
648 stress response and disease susceptibility of ayu (Plecoglossus altivelis).
649 Aquaculture 220: 515–523. doi:10.1016/S0044-8486(02)00626-9.
650 Ikeda, M, Takagi, S., and Taniguchi, N. 2005. Relationships between genetic and
651 number of successive generations in hatchery populations of ayu Plecoglossus
652 altivelis assessed by microsatellite DNA polymorphism. Nippon Suisan Gakkaishi
653 71(5): 768–774. doi: 10.2331/suisan.71.768.
654 Jonsson, B., Jonsson, N., and Hansen, L.P. 1991. Differences in life-history and
655 migratory behavior between wild and hatchery-reared Atlantic salmon in nature.
656 Aquaculture 98 (1–3): 69–78. doi: 10.1016/0044-8486(91)90372-E.
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 30 of 47
657 Jørgensen, E.H., Christiansen, J.S., and Jobling, M. 1993. Effects of stocking density on
658 food intake, growth performance and oxygen consumption in Arctic charr
659 (Salvelinus alpinus). Aquaculture 110(2): 191–204. doi:
660 10.1016/0044-8486(93)90272-Z.
661 Kaji, K. 2016. Summary of ayu breeding in 2014. Rep. Yamanashi Fish. Tech. Cent. 43:
662 31–37. Available from
663 http://www.pref.yamanashi.jp/suisan-gjt/documents/jiho43_p31-37.pdf [accessed
664 9 November 2018].
665 Katano, O., and Iguchi, K. 1996. Individual differences in territory and growth of ayu,
666 Plecogiossus altivelis (Osmeridae). Can. J. Zool. 74(12): 2170–2177. doi:
667 10.1139/z96-245. Draft
668 Katano, O., and Uchida, K. 2006. Effect of partial fin clipping as a marking technique
669 on the growth of four freshwater fish. Aquac. Sci. 54(4): 577–578. doi:
670 10.11233/aquaculturesci1953.54.577.
671 Katano, O. 2014. Experimental analysis on the relationship between the population
672 density of ayu Plecoglossus altivelis altivelis and fishery catch by ‘‘Tomozuri’’
673 angling. Fish. Sci. 80(5): 897–906. doi: 10.1007/s12562-014-0790-2.
674 Kataoka, Y., and Suzuki T. 2007. Evaluation of hatchery-reared delayed-growth ayu as
675 recreational fishing stock in ayu. Rep. Shiga Fish. Exp. St. 2006: 143–144.
676 Available from
677 http://www.pref.shiga.lg.jp/g/suisan-s/jigyohoukoku/files/p143_2.pdf [accessed 9
678 November 2018].
679 Klefoth, T., Skov, C., Krause, J., and Arlinghaus, R. 2012. The role of ecological
680 context and predation risk-stimuli in revealing the true picture about the genetic
https://mc06.manuscriptcentral.com/cjfas-pubs Page 31 of 47 Canadian Journal of Fisheries and Aquatic Sciences
681 basis of boldness evolution in fish. Behav. Ecol. Sociobiol. 66(4): 547–559. doi:
682 10.1007/s00265-011-1303-2.
683 Klefoth, T., Pieterek, T., and Arlinghaus, R. 2013. Impacts of domestication on angling
684 vulnerability of common carp, Cyprinus carpio: the role of learning, foraging
685 behaviour and food preferences. Fish. Manage. Ecol. 20(2-3) 174–186.
686 doi:10.1111/j.1365-2400.2012.00865.x.
687 Klefoth, T., Skov, C., Kuparinen, A., Arlinghaus, R. 2017. Toward a mechanistic
688 understanding of vulnerability to hook-and-line fishing: Boldness as the basic
689 target of angling-induced selection. Evol. Appl. 10(10) 994–1006. doi:
690 10.1111/eva.12504.
691 Koeck, B., Závorka, L., Aldvén, D.,Draft Näslund, J., Arlinghaus, R., Thörnqvist, P.,
692 Winberg, S., Björnsson, B.T., and Johnsson, J.I. 2018. Angling selects against
693 active and stress-resilient phenotypes in rainbow trout. Can. J. Fish. Aquat. Sci. in
694 press. doi: 10.1139/cjfas-2018-0085.
695 Kostow, K.E. 2004. Differences in juvenile phenotypes and survival between hatchery
696 stocks and a natural population provide evidence for modified selection due to
697 captive breeding. Can. J. Fish. Aquat. Sci. 61(4): 577–589. doi: 10.1139/f04-019.
698 Kostow, K.E. 2009. Factors that contribute to the ecological risks of salmon and
699 steelhead hatchery programs and some mitigating strategies. Rev. Fish Biol. Fish.
700 19(1): 9–31. doi: 10.1007/s11160-008-9087-9.
701 Larsen, M.H., Johnsson, J.I., Näslund, J., Thomassen, S.T., and Aarestrup, K. 2016.
702 Reduced rearing density increases postrelease migration success of Atlantic
703 salmon (Salmo salar) smolts. Can. J. Fish. Aquat. Sci. 73(5): 804–810. doi:
704 10.1139/cjfas-2014-0563.
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 32 of 47
705 Lennox, R.J., Alós, J., Arlinghaus, R., Horodysky, A., Klefoth, T., Monk, C.T., and
706 Cooke, S.J. 2017. What makes fish vulnerable to capture by hooks? A conceptual
707 framework and a review of key determinants. Fish and Fish. 18(5): 986–1010.
708 doi: 10.1111/faf.12219.
709 Lorenzen K. 2000. Allometry of natural mortality as a basis for assessing optimal
710 release size in fish-stocking programs. Can. J. Fish. Aquat. Sci. 57(12): 2374–
711 2381. doi: 10.1139/f00-215.
712 Lorenzen K. 2006. Population management in fisheries enhancement: Gaining key
713 information from release experiments through use of a size-dependent mortality
714 model. Fish. Res. 80(1), 19–27. doi: 10.1016/j.fishres.2006.03.010.
715 Lorenzen, K., Beveridge, M.C.M., andDraft Mangel, M. 2012. Cultured fish: integrative
716 biology and management of domestication and interactions with wild fish. Biol.
717 Rev. 87(3): 639–660. doi: 10.1111/j.1469-185X.2011.00215.x.
718 Louison, M.J., Adhikari, S., Stein, J.A., and Suski, C.D. 2017. Hormonal
719 responsiveness to stress is negatively associated with vulnerability to angling
720 capture in fish. J. Exp. Biol. 220(14): 2529–2535. doi: 10.1242/jeb.150730.
721 Mezzera, M., and Largiadèr, C.R. 2001. Evidence for selective angling of introduced
722 trout and their hybrids in a stocked brown trout population. J. Fish Biol. 59(2):
723 287–301. doi: 10.1111/j.1095-8649.2001.tb00130.x.
724 Miller, L.M., Close, T., and Kapuscinski, A.R. 2004. Lower fitness of hatchery and
725 hybrid rainbow trout compared to naturalized populations in Lake Superior
726 tributaries. Mol. Ecol. 13(11), 3379–3388. doi:
727 10.1111/j.1365-294X.2004.02347.x.
728 Miura, M., Tsuboi, J., Okazaki, T., Oohama, H., and Ashizawa, A. 2012. Assessment of
https://mc06.manuscriptcentral.com/cjfas-pubs Page 33 of 47 Canadian Journal of Fisheries and Aquatic Sciences
729 the characteristic traits of hatchery-reared ayu as a recreational fishing target:
730 comparison of two breeds subcultured under identical conditions. Nippon Suisan
731 Gakkaishi 78(6): 1149–1158. doi: 10.2331/suisan.78.1149.
732 Miwa, S., Sakai, A., and Nakane, M. 2003. Impairment of thymus development in
733 cultured osmerid fish, the ayu, Plecoglossus altivelis. Aquaculture 221: 535–548.
734 doi: 10.1016/S0044-8486(02)00661-0.
735 Moriyama, M. 2013. Evaluation of behavioral quality of ayu Plecoglossus altivelis
736 seedlings via analysis of schooling and fishing trials. J. Fish. Tech. 6(1): 39–43.
737 Available from https://www.fra.affrc.go.jp/bulletin/fish_tech/6-1/05.pdf [accessed
738 9 November 2018].
739 Nagai, T., and Sakamoto, T. 2006. SusceptibilityDraft and immune response to
740 Flavobacterium psychrophilum between different stocks of ayu Plecoglossus
741 altivelis. Fish Pathol. 41(3): 99–104. doi:10.3147/jsfp.41.99.
742 Nakano, S. 1995. Individual differences in resource use, growth and emigration under
743 the influence of a dominance hierarchy in fluvial red-spotted masu salmon in a
744 natural habitat. J. Anim. Ecol. 64(1): 75–84. doi: 10.2307/5828.
745 Nishida, M. 1986. Geographic variation in the molecular, morphological and
746 reproductive characters of the ayu Plecoglossus altivelis (Plecoglossidae) in the
747 Japan-Ryukyu archipelago. Jap. J. Ichthyol. 33(3): 232–248. doi:
748 10.11369/jji1950.33.232.
749 Reinbold, D., Thorgaard, G.H., and Carter P.A. 2009. Reduced swimming performance
750 and increased growth in domesticated rainbowtrout, Oncorhynchus mykiss. Can. J.
751 Fish. Aquat. Sci. 66(7), 1025–1032. doi:10.1139/F09-064.
752 Ruzzante, D. 1994. Domestication effects on aggressive and schooling behavior in fish.
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 34 of 47
753 Aquaculture 120(1-2), 1–24. doi: 10.1016/0044-8486(94)90217-8.
754 Sakae, K., Umino, T., Takahara, Y., Arai, K., and Nakagawa, H. 1996. Biological and
755 biochemical characteristics of yearling ayu in the Ohta River of Hiroshima
756 Prefecture. Nippon Suisan Gakkaishi 62(1): 46–50. doi: 10.2331/suisan.54.1107.
757 Sato, M., and Tsuboi, J. 2018. Stocking effectiveness of ayu Plecoglossus altivelis
758 during the early season, in a tributary of the Yoneshiro River Basin. Aquacult.
759 Sci. 66(3): 227–233.
760 Sutter, D.A.H., Suski, C.D., Philipp, D.P., Klefoth, T., Wahl, D.H., Kersten, P.,
761 Cooke,S.J., and Arlinghaus, R. 2012. Recreational fishing selectively captures
762 individuals with the highest fitness potential. Proc. Natl. Acad. Sci. U.S.A.
763 109(51): 20960–20965. doi:10.1073/pnas.1212536109.Draft
764 Tachihara, K., and Kimura, S. 1988. Some notes on the over-wintered ayu Plecoglossus
765 altivelis in Lake Ikeda, Kagoshima Prefecture. Nippon Suisan Gakkaishi 54(7):
766 1107–1113. doi: 10.2331/suisan.54.1107.
767 Teletchea, F., and Fontaine, P. 2014. Levels of domestication in fish: implications for
768 the sustainable future of aquaculture. Fish Fish. 15(2): 181–195. doi:
769 10.1111/faf.12006.
770 Thériault, V., Moyer, G.R., Jackson, L.S., Blouin, M.S., and Banks, M.A. 2011.
771 Reduced reproductive success of hatchery coho salmon in the wild: insights into
772 most likely mechanisms. Mol. Ecol. 20(9): 1860–1869. doi:
773 10.1111/j.1365-294X.2011.05058.x.
774 Tsuboi, J., Morita, K., Klefoth, T., Endou, S., and Arlinghaus, R. 2016.
775 Behaviour-mediated alteration of positively size-dependent vulnerability to
776 angling in response to historical fishing pressure in a freshwater salmonid. Can. J.
https://mc06.manuscriptcentral.com/cjfas-pubs Page 35 of 47 Canadian Journal of Fisheries and Aquatic Sciences
777 Fish. Aquat. Sci. 73(3): 461–468. doi: 10.1139/cjfas-2014-0571.
778 Tsukamoto, K., Masuda, S., Endo, M., and Ishida, R. 1990a. Influence of fish stocks on
779 the recapture rate of ayu released in the River Tsubusa. Nippon Suisan Gakkaishi
780 56(8): 1169–1176. doi: 10.2331/suisan.56.1169.
781 Tsukamoto, K., Masuda, S., Endo, M., Otake, T. 1990b. Behavioural characteristics of
782 the ayu, Plecoglossus altivelis, as predictive indices for stocking effectiveness in
783 Rivers. Nippon Suisan Gakkaishi 56(8): 1177–1186. doi: 10.2331/suisan.56.1177.
784 Uchida, K., Iguchi, K., and Kiso, K. 1995. Effects of water temperature on aggressive
785 behaviour of the territorial ayu Plecoglossus altivelis in Aquaria. Bull. Natl. Res.
786 Inst. Fish. Sci. 7: 389–401.
787 Uusi-Heikkilä, S., Sävilammi, T., Leder,Draft E., Arlinghaus, R., and Primmer, C.R. 2017.
788 Rapid, broad-scale gene expression evolution in experimentally harvested fish
789 populations. Mol. Ecol. 26(15): 3954–3967. doi: 10.1111/mec.14179.
790 Waples, R.S., and Hendry, A.P. 2008. Special Issue: Evolutionary perspectives on
791 salmonid conservation and management. Evol. Appl. 1: 183–188. doi:
792 10.1111/j.1752-4571.2008.00035.x.
793 Warnock, W.G., and Rasmussen J.B. 2013. Assessing the effects of fish density, habitat
794 complexity, and current velocity on interference competition between bull trout
795 (Salvelinus confluentus) and brook trout (Salvelinus fontinalis) in an artificial
796 stream. Can. J. Zool. 91(9): 619–625. doi: 10.1139/cjz-2013-0044.
797 Wilson, A.D.M., Brownscombe, J.W., Sullivan, B., Jain-Schlaepfer. S., and Cooke, S.J.
798 2015. Does angling technique selectively target fishes based on their behavioural
799 type? Plos One 10(8): e0135848. doi: 10.1371/journal.pone.0135848.
800 Yogo, S. 2010. Production of ayu larvae having wild nature of fish for stock
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 36 of 47
801 enhancement. Nippon Suisan Gakkaishi 76(3): 417–418. doi:
802 10.2331/suisan.76.417.
803 Yoneyama, Y., Hosoya, H., Otsuka, O., Fujita, T., Hoshino, M., and Sato, Y. 1997.
804 Stocking effectiveness of hatchery reared ayu in the Umikawa River with special
805 reference to generation differences. Rep. Niigata Inland Fish. Exp. St. 22: 25–33.
806 Yoshiura, Y., Kamaishi, T., Nakayasu, C., and Ototake, M. 2006. Detection and
807 Genotyping of Flavobacterium psychrophilum by PCR targeted to Peptidyl-prolyl
808 cis-trans isomerase C Gene. Fish Pathol. 41(2): 67–71. doi: 10.3147/jsfp.41.67.
809 Yoshizawa, K. 2003. Inbreeding in hatchery populations of ayu cultured for multiple
810 generations as inferred from allozyme analysis. Rep. Gunma Fish. Exp. St. 9: 67–
811 78. Draft
https://mc06.manuscriptcentral.com/cjfas-pubs Page 37 of 47 Canadian Journal of Fisheries and Aquatic Sciences
812 Table 1 Survival rate for 90 days between hatching to sorting and malformation rate in
813 ayu culturing between 2009 and 2013. Each study year, two cohorts (brood lines) of
814 ayu, which constituted different numbers of generations held in captivity, were
815 produced. A total of three cohorts were produced over four years, A; 5th, 6th, 7th, 8th,
816 B; 9th, C; 1st, 2nd, 3rd generations in captivity.
817
Year 2009-2010 2010-2011 2011-2012 2012-2013
Cohort (brood line) A B C A C A C A
Number of generations 5th 9th 1st 6th 2nd 7th 3rd 8th of ayu in captivity
Estimated number of hatched individuals Draft 2,569,000 2,463,000 1,192,000 4,205,000 1,502,000 4,629,000 1,984,000 2,259,000 (0.9 × estimated N of eyed eggs)
Estimated number of 839,000 701,000 153,000 448,000 670,000 722,000 723,000 700,000 individuals at sorting
Survival rate 32.7 28.5 12.8 10.7 44.6 15.6 36.4 31.0 from hatching to sorting (%)
Malformation rate just before stocking (%) 3.8 5.8 9.5 1.6 2.7 0 0 0 (N of malformed fish / (12 / 313) (15 / 313) (4 / 42) (3 / 185) (3 / 111) (0 / 146) (0 / 170) (0 / 180) N of sampled fish) 818
819 820 821
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 38 of 47
822 Table 2 Summary of stocking and angling experiment between 2010 and 2013. 823 Asterisks show group marking by adipose fin clipping. 824
Year 2010 2011 2012 2013
Cohort (brood line) A B C A C A C A
Number of generations 5th 9th * 1st * 6th 2nd * 7th 3rd 8th * of ayu in captivity
Mean body weight 14.3 14.3 11.3 13.4 13.2 13.5 12.5 12.1 at stocking (g)
Date of marking 2010/6/3 2011/5/30 2012/5/7 2013/5/21
Date of stocking 2010/6/4 2011/6/7 2012/5/28 2013/5/29
Period of 2010/6/26 ‒ 9/22 2011/7/1 ‒ 9/13 2012/6/15 ‒ 9/20 2013/6/16 ‒ 10/2 angling experiment Draft Total angling effort (h) 68.9 30.5 71.3 33.5
Date of 2010/7/27, 8/6 2011/6/29, 8/9 2012/6/14, 8/10 2013/6/11, 8/21 casting-net fishing 825 826 827
https://mc06.manuscriptcentral.com/cjfas-pubs Page 39 of 47 Canadian Journal of Fisheries and Aquatic Sciences
828 Table 3 Summary of G-tests of the number of individuals in single year comparison
829 between two cohorts for the index of survival (the number of individuals stocking and
830 recaptured by casting net), the vulnerability to angling (recaptured by angling and by
831 casting net), and the migration trends (recaptured by casting net at up- and downstream
832 section from stocking point). The p values were Bonferroni-corrected to account for12
833 G-tests.
834
a) Index of suvival Year 2010 2011 2012 2013 Number of generations 5th 9th 1st 6th 2nd 7th 3rd 8th in captivity Draft Stocking 10,000 10,000 10,000 10,000 10,000 10,000 7,515 8,082
Caught by 74 41 76 129 115 130 130 145 casting net p = 0.024 p = 0.003 p = 1.000 p = 1.000 b) Vulnerability to angling Caught by 63 43 13 55 63 98 59 28 angling Caught by 74 41 76 129 115 130 130 145 casting net p = 1.000 p = 0.060 p = 1.000 p = 0.009 c) Migration after stocking
Upstream 8 6 11 12 57 42 82 69
Downstream 66 35 65 117 58 88 48 76
p = 1.000 p = 1.000 p = 0.072 p = 0.120 835
836
837
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 40 of 47
838 Table 4 The best model of a generalized linear mixed model (GLMM) selected by 839 Akaike information criteria (AIC) in a field experiment on the catch per unit effort by 840 angling of ayu reared several generations in captivity. 841 Dependent Error Independent AIC ∆ AIC Coefficient χ 2 p variable distribution variable Number of Poisson 666.9 0.8 Number of 0.530 11.737 < 0.001 individuals generations caught by in captivity angling Square of the -0.052 12.580 < 0.001 number of generations in captivity
Recapture point 0.279 2.852 0.091
Intercept -1.339 842 843 Draft 844 Note: The independent variables were Number of generations in captivity, Square of the 845 number of generations in captivity, Date (number of days from 1st June), and Recapture 846 point (upstream from stocking point = 1, downstream = 0). Angling effort (h) was the 847 offset term in the model. Random effects were cohort (brood lines: A; 5th, 6th, 7th, 8th, 848 B; 9th, C; 1st, 2nd, 3rd generations in captivity, n = 3, variance = 0.000, SD = 0.000) 849 and individual data ID (n = 164, variance = 0.492, SD = 0.702). ∆AIC shows the 850 difference of AIC between the best and full models.
https://mc06.manuscriptcentral.com/cjfas-pubs Page 41 of 47 Canadian Journal of Fisheries and Aquatic Sciences
851 Table 5 The best and full model of a generalized linear mixed model (GLMM) selected 852 by Akaike information criteria (AIC) in a field experiment on the vulnerability to 853 angling of ayu reared several generations in captivity 854
Dependent Error Independent AIC ∆ AIC Coefficient χ 2 p variable distribution variable Caught by Binomial 1279.9 1.1 Number of 0.439 2.991 0.084 angling (1) or generations casting net (0) in captivity Square of the -0.074 11.396 < 0.001 number of generations in captivity
TL 1.133 83.927 < 0.001 Date -0.370 10.399 0.001 DraftRecapture point 1.384 68.652 < 0.001 Intercept -0.710 855 856 857 858 Note: The independent variables were Number of generations in captivity, Square of the 859 number of generations in captivity, TL (total length at recapture), Date (number of days 860 from 1st June), and Recapture point (upstream from stocking point = 1, downstream = 861 0). Random effect was cohort (brood lines: A; 5th, 6th, 7th, 8th, B; 9th, C; 1st, 2nd, 3rd 862 generations in captivity, n = 3, variance = 2.158, SD = 1.469). ∆AIC shows the 863 difference of AIC between the best and second models. 864
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 42 of 47
865 Figure captions 866 867 Fig. 1. Schematic representation of the typical ayu angling, using live ayu as a subject 868 of agonistic behaviour for wild individuals which have foraging territory. 869 870 871 Fig. 2. Map showing locations of study reach. Three weirs prevent upstream migration 872 of ayu in the study area. 873 874 875 Fig. 3. Body weights of ayu until 80 days after hatching in different numbers of 876 generations in captivity. Error bars indicate standard deviation. Growth curve of first 877 generation of ayu are without SD, because there was only one pond caused by few 878 number of mature individuals of wild fish under hatchery condition in 2010. 879 880 Draft 881 Fig. 4. Estimated generation effect (0.530 * Generation – 0.052 * (Generation) 2) on 882 catch per unit effort (N / person hour) by angling using territorial behaviour for a total 883 of 82 days throughout four years stocking and recapture experiments in a natural stream 884 on ayu (see Table 4). Bold line shows median values. Grey buffer indicates 95% 885 confidence interval estimated by non-parametric bootstrap method. 886 887 888 Fig. 5. Estimated generation effect (0.439 * Generation – 0.074 * (Generation) 2) on the 889 probability of being caught by angling throughout four years stocking and recapture 890 experiments in a natural stream on ayu (see Table 5). Bold line shows median values. 891 Grey buffer indicates 95% confidence interval estimated by non-parametric bootstrap 892 method. 893 894 895
https://mc06.manuscriptcentral.com/cjfas-pubs Page 43 of 47 Canadian Journal of Fisheries and Aquatic Sciences
896 Figure 1 897
Draft
898 899
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 44 of 47
900 Figure 2 901
N36°
E139°
Stocking point of hatchery reared ayu Study area
Draft
100m 902
https://mc06.manuscriptcentral.com/cjfas-pubs Page 45 of 47 Canadian Journal of Fisheries and Aquatic Sciences
903 Figure 3 904
350 2009-2010 Fifth 8th 300 2009-2010 Ninth 2010-2011 First 2010-2011 Sixth 250
) 2011-2012 Second 9th g
m 2011-2012 Seventh 3rd (
t 200
h 2012-2013 Third
g 6th i
e 2012-2013 Eighth 7th w 150 5th y d o B 100 1st 2nd 50 Draft
0 0 10 20 30 40 50 60 70 80 Days after hatching
905
906 907 908
https://mc06.manuscriptcentral.com/cjfas-pubs Canadian Journal of Fisheries and Aquatic Sciences Page 46 of 47
909 Figure 4 910 911
Draft
912 913 914 915 916
https://mc06.manuscriptcentral.com/cjfas-pubs Page 47 of 47 Canadian Journal of Fisheries and Aquatic Sciences
917 Figure 5 918
Draft
919 920 921 922
https://mc06.manuscriptcentral.com/cjfas-pubs