Migratory Behaviors in Masu Salmon (Oncorhynchus Masou) and the Influence of Endocrinological Factors

Home , Oncorhynchus masou, Pink salmon, Salmo, Salmonidae, Sockeye salmon

Aqua-BioScience Monographs, Vol. 5, No. 2, pp. 29–65 (2012) www.terrapub.co.jp/onlinemonographs/absm/

Migratory Behaviors in Masu Salmon (Oncorhynchus masou) and the Influence of Endocrinological Factors

Arimune Munakata

Department of Biology Miyagi University of Education Aoba-ku, Sendai, Miyagi 980-0845, Japan e-mail: [email protected]

Abstract Received on April 1, 2011 In the freshwater phase of their lifecycle, masu salmon (Oncorhynchus masou) comprise Accepted on September 22, 2011 two different phenotypes. A portion of the juveniles (migratory form) exhibit downstream Online published on November 20, 2012 migratory behavior after smoltification. However, some masu salmon (non-migratory form) such as precociously mature males live continuously in their natal rivers throughout their Keywords lifetime. The coexistence of migratory and non-migratory forms within the species indi- • cortisol cates that this salmon can be effectively used as a model fish to illuminate both inhibitory • downstream migration and stimulatory physiological control mechanisms of migratory behaviors. In masu salmon, • masu salmon it was found that sex steroid hormones inhibit the occurrence of downstream swimming • Oncorhynchus masou behavior, the initial step in seaward migration. Moreover, after the commencement of • Pacific salmon • sex steroid hormone downstream migration, sex steroid hormones induced the upstream swimming and subse- • spawning quent spawning behaviors. These findings indicate that sex steroid hormones influence • testosterone the occurrence of the downstream and upstream swimming behavior in the resulting rheo- • upstream migration taxis fashion (negative and positive, respectively). In contrast to sex steroid hormones, it was also found that cortisol, which is involved substantially in smoltification, stimulates the downstream swimming behavior. These findings indicate that the occurrence of seaward migration is controlled competitively by sex steroid hormones (sexual maturation) and smolt-inducing factors such as cortisol, in masu salmon and potentially other Pacific salmon.

1. Introduction and Kobayashi 2010). Most of the non-migratory forms will live continuously in their natal rivers throughout Salmonids (family Salmonidae) consist of four gen- their lives (Fig. 1). Regardless of these life history era, Hucho, Salvelinus, Salmo, and Oncorhynchus types, most salmonids will spawn in freshwater envi- (Neave 1958; Norden 1961; Murata et al. 1993). These ronments, mainly in their natal rivers (Fig. 1) (Quinn salmonids originally inhabited tributaries from high- 2005). From these phenomena, salmonids are consid- through mid-latitude areas in the northern hemisphere ered to be of freshwater (fluvial) fish origin and their (Quinn 2005). It is also known that a large part of migratory behaviors by and large start from the rivers salmonids (in quantity: the number of species and (Fig. 1). biomass) are anadromous, and these fish (i.e., migra- Among the four genera of salmonids, two genera tory form) regularly exhibit downstream migratory Hucho and Salvelinus are considered evolutionally behavior from the rivers to the sea (or lakes), after the ancient groups, based on the phylogenic analyses occurrence of parr to smolt transformation (i.e., (Norden 1961; Murata et al. 1993). Genus Hucho in- smoltification) (Fig. 1). However, many species capa- habit only the northern Eurasia continent and genus ble of anadromy also have phenotypes that are full- Salvelinus inhabit northern Eurasia and the American time residents of freshwater habitats (i.e., non- continent (Quin 2005). On the other hand, genus Salmo migratory forms) and display neither smoltification nor (i.e., Atlantic salmon) and Oncorhynchus (i.e., Pacific downstream migratory behavior to the sea (Munakata salmon), which are considered evolutionally new

River Sea

Pink and chum salmon

Oncorhynchus (Pacific salmon) Migrants

Evolution Masu salmon Non-migrants-m

Salmo (Atlantic Salmon) Salvelinus

Hucho

Fig. 1. Schematic drawing that illustrates the diversity of distance covered by non-migratory and migratory forms for four salmonid genera (shown in the order of the evolutional age). In genus Hucho, most fish live continuously in their natal rivers. In genus Salvelinus and Salmo, some fish migrate to the sea after smoltification. In genus Oncorhynchus, most juveniles perform long distance seaward migration for several years. On the other hand, in masu salmon (O. masou), a portion of the fish perform seaward migration for a year after smoltification, while an equivalent portion of them stay in the rivers similar to genera Hucho, Salvelinus, and Salmo.

groups, are widely distributed in the rivers and tributaries around the north Atlantic and Pacific Oceans, respectively (Groot and Margolis 1991). Fig. 2. Photographs of masu salmon (Oncorhynchus masou). In regard to migratory behaviors, the majority of fish (a) precocious male non-migrants, (b) immature parr non- in the genus Hucho and Salvelinus live continuously migrants, (c) pseudo smolt, (d) smolt migrants, and (e) adult in their natal river systems throughout their lifetime, smolt migrants that migrated back from the sea. as non-migratory forms (Fig. 1). If at all existent, the proportions of the migratory forms are much smaller, and their temporal and spatial ranges of migratory movements are shorter and narrower, respectively, than that mainly inhabits Japanese rivers (i.e., western Pa- those in other salmonids such as Atlantic and Pacific cific Ocean), some yearling (1+) fish live continuously salmon. On the other hand, in Pacific salmon such as in their natal rivers similar to the ancient salmonid pink (O. gorbuscha) and chum (O. keta) salmon, which genera including genus Hucho and Salvelinus are considered evolutionally the newest species, most (Machidori and Kato 1984; Kato 1991; Kiso 1995) juveniles undergo long distance seaward migration (Figs. 1, 3). In masu salmon, however, a portion of the (e.g., from Japanese streams to the Bering Sea) which 1+ juveniles exhibit downstream migratory behavior will continue for several years (Groot and Margolis after the occurrence of smoltification, as do other Pa- 1991). Their temporal and spatial ranges of migratory cific salmon such as pink and chum salmon. In masu movements are considerably longer and broader than salmon, such differentiations in lifecycles regularly in other salmonid species. Based on these wide differ- occur within the same population from the same riv- ences in migratory patterns among salmonid genera ers, especially in the northern regions of their habitat from different evolutional time periods, it is inferred (e.g., northern Honshu through Hokkaido) (Machidori that the proportions of migratory forms increased, and and Kato 1984; Kato 1991; Kiso 1995). subsequently the temporal and spatial ranges of mi- Taking the lifecycles of masu and other salmonids gration became longer and broader, respectively, into consideration, the proportion of non-migratory and through the evolutionary processes (Fig. 1). migratory forms seem to vary, not only among differ- In masu salmon (O. masou) (Fig. 2), a Pacific salmon ent salmonid genera, but also within the same genus

River Sea

Feeding Immature SSmolt

Dow Spawning Precocious migration male

Feeding Non-migrants ( year)

Upstream migration Homing migration

Fig. 3. Lifecycles of masu salmon (Oncorhynchus masou). In masu salmon, some immature juveniles (migratory form) display the downstream migratory behavior after they have transformed from parr to smolt (smoltification). However, some juveniles (non-migratory form) such as precociously mature males (precocious males) will live continuously in their natal rivers throughout their lifetime. The lifecycle (migratory behavior, seaward migration) of migrants consists of downstream migration, feeding, homing, upstream migration, and spawning. On the other hand, the lifecycle of non-migrants consists of downstream movement within a river, stream residence, upstream movement, and spawning.

(e.g., Pacific salmon). Since both the non-migratory precocious males) (Machidori and Kato 1984; Kiso and migratory forms appear within the same species, 1995). On the other hand, most of migratory forms are it is hypothesized that the masu salmon possesses both sexually immature male and female smolts, as observed evolutionarily ancestral (i.e., fluvial) and modern (i.e., in other Pacific salmon (Quinn 2005). These phenom- anadromous) characteristics of migratory behaviors. In ena thus indicated that “sexual maturation” is one of Japanese streams, the non-migratory form of masu the key physiological factors that regulate the occur- salmon is called “yamame” meaning mountain girl, and rence of seaward migration. Furthermore, since most the representative migratory form is called “sakura- of the downstream migrants undergo smoltification masu”, meaning cherry blossoms. Why do only a por- before their seaward migration, it was hypothesized tion of masu salmon juveniles exhibit the ocean-bound that some physiological factors which are closely re- migratory behaviors, whereas the rest do not? lated to the smoltification stimulatory regulate the oc- In this monograph, an overview of the migratory currence of downstream migratory behavior (Munakata behaviors, especially the downstream and upstream mi- et al. 2007). Thus, previous studies investigated both gratory behaviors, and subsequent spawning behaviors inhibitory and stimulatory control mechanisms of mi- in masu salmon will be presented. Additionally, a gratory behaviors in relation to sexual maturation (sex theory of hormonal control as a mechanism governing hormones) and smoltification, respectively (Munakata downstream (negative rheotaxis) and upstream (posi- et al. 2000b, 2001a, 2001b, 2007, 2012a, 2012b). New tive rheotaxis) swimming behaviors, major components information will be used to reconsider the physiologi- of downstream and upstream migratory behaviors, and cal control mechanisms, roles, and evolutionary proc- subsequent spawning behaviors in masu and other Pa- esses of the migratory behaviors in masu salmon, and cific and Atlantic salmon will also be presented. Since perhaps in entire salmonids. Furthermore, this analy- both non-migratory and migratory forms appear within sis incorporates not only physiological factors, but also the same population inhabiting the same river, it was environmental factors that influence the migratory hypothesized that individual masu salmon physiologi- behaviors. It is therefore suggested that the findings cally control the inhibitory and stimulatory mechanisms have important implications. Especially, these data can of migratory behaviors (Munakata and Kobayashi serve as new tools for improving our salmon stock- 2010). management, focusing specifically on the conservation In masu salmon, it is generally known that the non- of their migratory behavior. migratory forms are the precociously mature fish (i.e.,

2. Roles of sex steroid hormones in the regulation mentioned above, the representative migratory form of downstream swimming behavior in masu that lives in the rivers for 1.5 years regularly begins salmon and other salmonids the seaward migration following the occurrence of smoltification (Kato 1991). On the other hand, repre- 2-1. Definition of migratory behaviors in sentative non-migratory forms such as 1+ precocious salmonids males live continuously in their natal rivers throughout their lifetime (Utoh 1976, 1977). In this section, Migratory behavior of salmonids regularly consists the lifecycles of masu salmon from hatching to the of downstream migratory behavior (downstream mi- period in which the smolt migrants exhibit downstream gration) from the river to the sea (or lakes), feeding migratory behavior will be summarized, with empha- migratory behavior (feeding migration) in the sea (or sis on environmental factors which induce either the lakes), homing migratory behavior (homing migration) stream residency in non-migrants or smoltification in from the sea (or lakes) to the mouth of their natal riv- migrants. Then, an overview of previous and recent ers, upstream migratory behavior (upstream migration) studies that have investigated the inhibitory roles of from the mouth to the spawning ground in upper sex steroid hormones, such as testosterone (T), 11- β reaches in the natal rivers, and spawning behaviors ketotestosterone (11-KT), and estradiol-17 (E2) in the (Munakata and Kobayashi 2010) (see Fig. 3). The occurrence of smoltification and subsequent down- downstream migratory behavior, the initial step of sea- stream swimming behavior in masu and other ward migration mainly consists of several specific salmonids will be outlined. behaviors, such as schooling behavior, downstream swimming behavior (negative rheotaxis, downstream 2-3. Early growth after emerging movement), and salinity preference (seawater adaptation) (Iwata 1995, 1996; Munakata and Kobayashi After emerging from spawning beds (common name: 2010). Also, the upstream migratory behavior consists redd) which are located in the main stem or tributaries of several types of behaviors, such as upstream swim- in upper rivers, underyearling (0+) masu salmon juve- ming behavior (positive rheotaxis, upstream move- niles (3 cm in standard body length) are typically found ment), and freshwater preference. Among these phe- in shallow areas (e.g., behind large stones, under fallen nomena, downstream and upstream swimming trees, between roots or stems of emergent plants, etc.) behaviors, major components of downstream and up- where the flow rate is moderate (Machidori and Kato stream migratory behaviors, can be observed in an ar- 1984; Kato 1991). These 0+ juveniles then gradually tificial raceway system (see Fig. 5) during their natu- move to deeper areas, such as the edge or center of the ral downstream and upstream migratory periods flow in the pools (Kiso 1995). The 0+ juveniles, after (Munakata et al. 2000b, 2001a, 2001b, 2007, 2012a, emergence, are called “parr”, since these fish display 2012b). In this monograph, therefore, we mainly in- large black round spots (i.e., parr marks) on both sides vestigated the effects of endocrinological (hormonal) of their body. The parr mark is considered to allow factors on the occurrence of downstream and upstream them to be better camouflaged against the background swimming behaviors in the raceway system. of the rivers, that is, stream beds, rocks, and fallen trees. The 0+ masu salmon parr mainly forages on drift 2-2. Lifecycle of masu salmon aquatic insects such as larval Ephemeroptera, Trichoptera, Plecoptera, Chironomidae, and occasion- Masu salmon, a Pacific salmon, is broadly distrib- ally forages on fallen terrestrial insects (Machidori and uted in north western Pacific-rim rivers (Kamchatka Kato 1984; Kato 1991; Kiso 1995). In upper rivers, Peninsula through Kyushu Island) (Machidori and Kato however, distributions of drifting aquatic and terres- 1984). In addition, several sub-species or sub-types of trial insects are generally stratified among the differ- masu salmon are found in this region: amago salmon ent spaces. This suggests that accessibility to prey items (O. masou ishilawae) inhabit streams in southwestern differs considerably among individual 0+ parr. For Japan (e.g., southern Honshu, part of Kyushu, and the these reasons, 0+ parr compete frequently with their Shikoku Islands) (Kato 1991); Biwa salmon (O. masou conspecifics as well as other species (i.e., Japanese rhodurus) inhabit the tributaries around Lake Biwa charr (Salvelinus fontinalis)) with the same dietary (Fujioka et al. 1990); Taiwan salmon (O. masou habits for occupying focal foraging areas (i.e., terri- formosanus), an endangered sub-species, inhabit lim- tory) in which they can access substantial drifting prey ited highland rivers of Taiwan (Oshima 1936); and a items (Nakano et al. 1990; Nakano and Furukawa- hybrid type of the Honmasu salmon (O. masu x Tanaka 1994; Nakano 1995). Moreover, to achieve and rhodurus) inhabit tributaries around Lake Chuzenji maintain their focal foraging areas, 0+ parr frequently (Munakata et al. 1999, 2000a). During the spring, year- exhibit territorial aggressiveness against other indi- ling (1+) masu salmon juveniles can be classified into viduals, and some of the territorial 0+ parr establish two groups, migratory and non-migratory forms. As themselves to be the dominant fish in each focal for-

Fig. 4. Changes in body length (BL), body weight (BW), condition factor (CF), gonad somatic index (GSI), plasma levels of β α α β testosterone (T), 11-ketotestosterone (11-KT), estradiol-17 (E2), progesterone (P), 17 -progesterone (17 -P), 17,20 - dihydroxy-4-pregnene-3-one (DHP), and pituitary hormone luteinizing hormone (LH) in male and female masu salmon. Dif- ferences in mean plasma hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Scheffe’s F-test. * and *** indicates significant difference at P < 0.05 and P < 0.001, respectively. Reprinted from Comp. Biochem. Physiol. Part B, 129, Munakata et al., The involvement of sex steroid hormones in downstream and upstream migratory behavior of masu salmon, 661–669,  2001a, with permission from Elsevier.

aging space during autumn (Machidori and Kato 1984; behavior”, result in the augmentation for distribution Munakata et al. 2000b). On the other hand, juveniles of masu salmon into broader habitats within the riv- that could not occupy the focal foraging areas become ers. the subordinates in the social order. 2-4. Differentiation in non-migratory forms 2-3A. Wintering downstream movement During the same period (autumn through winter), a In juvenile masu salmon, both non-migratory (i.e., significant proportion of 0+ masu salmon, including stream resident) and migratory (i.e., smolt) forms com- non-migrants and migrants, tend to move from their monly originate from dominant and subordinate parr, former habitat to the downstream areas (Machidori and respectively, and both types can be discriminated by Kato 1984; Kato 1991), which is possibly induced by their morphologic characteristics after 0+ summer decreased temperature (Giannico and Hinch 2003). (Machidori and Kato 1984; Kiso and Matsumiya 1992; These movements, which are called “wintering Kiso 1995). In regard to the non-migratory forms, the

0+ dominant precocious males regularly become the ness prior to the smolting period, and these smolts sub- non-migrants that live continuously in their natal riv- sequently exhibit a tendency to gather in open spaces ers. In 0+ precocious males, standard body length (BL), even during the daytime hours. According to Hutchison body weight (BW), depth of body, and testes (gonad and Iwata (1998), such behavioral changes are caused weight (GW)) become larger, and the body color be- by the increase of plasma thyroxine, which is consid- comes darker than that of 0+ immature parr during ered to be one of the smolt-inducing factors. These summer, and these fish subsequently attend to spawn- behavioral changes seemed to convert the “territorial ing activities in the following autumn (Kato 1991; Kiso behavior” into “schooling behavior” in 1+ smolts, as 1995). After spawning, testes in 1+ (former 0+) preco- the peak period of their smoltification approached. cious males regress from winter through spring Subsequently, most 1+ smolts begin downstream swim- (Munakata et al. 2001a; Fig. 4). However, the size of ming behavior throughout the evening (Munakata et the testes, expressed by the gonad somatic index (GSI: al. 2000b), during and after rainfall (Yamauchi et al. 100 x GW/BW), in 1+ precocious males is still larger 1984, 1985), and snow runoff, etc. in favorable peri- than those of immature male parr (Munakata et al. ods during the spring (Iwata 1995, 1996). 2001a; Fig. 4). The 1+ precocious males then begin In summary, two phenotypes (forms) of masu salmon the maturation process again after spring commences, diverge as juveniles at the age of 1+. The precocious while most of 1+ immature migrants exhibit down- males retain many characteristics of parr, despite un- stream migratory behavior following smoltification. dergoing sexual maturation (Utoh 1976, 1977; Machidori and Kato 1984). In contrast, 1+ smolts ex- 2-5. Smoltification in subordinate juveniles perience drastic morphological, physiological, and behavioral changes during smoltification (Iwata 1995, In contrast to the dominant precocious male non- 1996). Based on these observations, one might argue migrants, most of 0+ subordinates live as immature parr that the precocious males are more similar to their flu- from summer through autumn (Machidori and Kato vial ancestors than smolts. In the next section, we will 1984; Kiso 1990, 1995). In the following winter, some discuss how sex hormones are involved in the differ- of the 1+ immature fish initiate smoltification, which entiation between the two forms and identify the likely is completed in the following spring, before the down- factors regulating these processes. stream migration begins. Smoltification regularly consists of a series of mor- 2-6. Inhibitory roles of sex steroid hormones in phological, physiological, and behavioral changes, the smoltification which enable 1+ (former 0+) juveniles to adapt to marine environments (Hoar 1976, 1988; Boeuf 1994; In masu salmon, it was previously found that dissec- Boeuf et al. 1994; Iwata 1995, 1996; McCormick tion of the testes (castration) of 0+ precocious males, 2001). In a short span of time, future smolts start to the non-migrants, during autumn induced display a silvery body color, and black dorsal and dor- smoltification in the following (1+) spring, while sham- sal fin tips, which camouflage them against the colors operated fish remained as precocious males (Aida et of the ocean waters (Quinn 2005), similar to other al. 1984). Plasma androgen (T + 11-KT) levels (0.12 marine migratory fish such as sardines (Engraulis spp.) ng/ml) of 1+ castrated precocious males became lower and saury (Cololabis spp.). The changes in the body than those in sham-operated precocious males (1 ng/ colors are supported fundamentally by the accumula- ml) after surgery. If a portion of the testis was left in tion of granules of pigments such as guanine and the abdominal cavity, those males did not smoltify, just melanophores on the abdominal and dorsal skins, re- as the sham-operated ones did not. Accordingly, the spectively (Hoar 1988). Masu salmon smolts also dis- findings indicate that sexual maturation, more specifi- play lower condition factor (CF: 100 x BW/BL3) when cally, sex steroid hormones released from the gonads compared with those values before starting inhibited the occurrence of smoltification in the pre- smoltification (Aida et al. 1984; Ikuta et al. 1985, 1987; cocious males. Ikuta et al. (1985, 1987) later confirmed Munakata et al. 2000b, 2001a). Their osmoregulatory that treatment (oral administration) with exogenous sex ability is modulated by hormonal factors, such as steroid hormones, such as methyletestosterone (MT), growth hormone (GH) and cortisol (Hirano 1991; T, 11-KT, and E2 in 1+ masu salmon smolts in winter McCormick 2001). through spring impaired some part of the changes as- Physiological changes, such as an increase in gill sociated with smoltification, such as silvery body color, Na+–K+–ATPase activity allow the smolts to adapt to decrease in CF, seawater tolerance capacity, and plasma salt water conditions (Boeuf et al. 1989; Iwata et al. rises in thyroid hormones. On the other hand, synthetic 1990). steroid, 5α-dihydrotestosterone (DHT) which has With regard to behavior, Iwata (1995) suggested that stronger androgenic effects than T did not exhibit sig- 1+ masu salmon smolts cease to exhibit aggressive nificant inhibitory effects (Ikuta et al. 1987). behaviors, which support their territorial aggressive- The inhibitory effects of sex steroid hormones on the

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 35 smoltification are also confirmed in other Pacific Current salmon such as amago salmon (Miwa and Inui 1986). Therefore, it is further hypothesized that inhibitory regulation of smoltification by sex steroid hormones is a common phenomena in a number of Pacific salmon. Upper Net traps pond 2-7. Seasonal changes in plasma sex hormone levels

After the completion of smoltification, most masu and other Pacific salmon smolts subsequently migrate downstream to the sea (Quinn 2005; Munakata and Kobayashi 2010). Therefore, by logical extension, not Fishway only the smoltification, but also the downstream migratory behavior is repressed by some of the sex steroid hormones. To investigate which sex steroid hormones are indeed involved in the occurrence of smoltification and downstream swimming behavior, seasonal changes in the plasma levels of sex steroid hormones and pituitary hormone (luteinizing hormone (LH)) were investigated in masu salmon by radioimmunoassays (RIAs), during the period of smoltification and downstream migration (Munakata Separated et al. 2001a; Fig. 4). Lower Area pond 2-7A. Males In 0+ and 1+ males, precocious males (representative non-migrants) appeared and were distinguishable from the immature males of the same age by their larger BL, BW, CF, and GSI values and plasma sex steroid Current hormone levels (Fig. 4). In 1+ precocious males, values of GSI and plasma levels of T, 11-KT, and 17,20β- dihydroxy-4-pregnene-3-one (DHP) significantly increased from May through September, overlapping the Fig. 5. Schematic drawing of experimental raceway. In or- period of the smoltification and downstream migration, der to study the roles of sex steroid hormones in downstream while 1+ immature males (smolt and parr) did not ex- behavior (negative rheotaxis), the fish were transferred into hibit such phenomena (Fig. 4). On the other hand, the upper pond (2 × 4 × 0.5 m) of a two-step raceway con- plasma levels of progesterone (P), 17α-progesterone nected to the lower pond (2 × 8 × 0.5 m) through a fishway (17α-P), and LH did not show significant increases in (20 cm in diameter by 4 m in length made of a polyvinyl 1+ precocious males. chloride (PVC) half-cut pipe (Munakata et al. 2000b). Spring water was supplied into the upper pond. Flow rate (volume) In 1+ precocious males, it was noteworthy that and velocity of the water in the fishway ranged between plasma levels of T began to increase and attained peak 10–20 l/s and 70–85 cm/s, respectively. Water temperature levels earlier than did 11-KT and DHP (Fig. 4). More- fluctuated between 9–10°C. At the downstream edge of the over, T maintained high plasma levels extensively from fishway in the lower pond, a net trap (2 × 0.7 × 0.7 m) was March through September, overlapping the period of placed to capture fish that moved down from the upper pond. their seaward migration. An individual experimental fish was identified as a downstream migrant if it moved from the upper pond into the net 2-7B. Females trap in the lower pond. In order to investigate the effects of In females, precocious maturation rarely occurred sex steroid hormones on the occurrence of upstream among 0+ and 1+ fish in hatchery-raised strains behavior, the experimental fish and net trap were transferred (Shiribetsu River strain, introduced from Hokkaido in into the lower and upper pond, respectively. The frequency of downstream or upstream migrations is expressed as a 1980) that were used in our studies (Munakata et al. percentage of the initial fish numbers. Reprinted with per- 2000b, 2001a, 2001b, 2007, 2012a). Consequently, a mission from Fish. Sci., 78, Munakata et al., Involvement major part of 1+ immature females exhibited low CF, of sex steroids, luteinizing hormone and thyroid hormones GSI, and plasma sex hormone levels (Fig. 4). in upstream and downstream migratory behaviors in land- The females regularly commence apparent sexual locked sockeye salmon Oncorhynchus nerka, 81–90, Fig. maturation at the age of 2+ during the spring through 1,  2012b, The Japanese Society of Fisheries Science.

Fig. 7. (a) Frequency of migrants and non-migrants, plasma Fig. 6. (a) Frequency of migrants and non-migrants, plasma levels of (b) testosterone (T), (c) pituitary contents of lutei- levels of (b) testosterone (T), (c) pituitary contents of luteinizing hormone (LH), (d) plasma levels of LH, (e) thyrox- µ nizing hormone (LH), (d) plasma levels of LH, (e) thyroxine (T4), and (f) triiodothyronine (T3) in controls, T 5 g, T 50 µg, T 500 µg-treated smolts and precociously mature male ine (T4), and (f) triiodothyronine (T3) in control and T 500 µg-treated 1+ masu salmon smolts. Numbers above columns 1+ masu salmon. Numbers above columns in (a) indicate in (a) indicate the number of migrants and non-migrants. the number of migrants and non-migrants. Differences in Differences in the frequency of downstream behavior from the frequency of downstream behavior from the control group the control group were analyzed by the χ2-test, using were analyzed by the χ2-test, using StatView version 4.5 StatView version 4.5 software (Abacus Concepts, Inc., Cali- software (Abacus Concepts, Inc., California, USA). * indi- fornia, USA). ** indicates a significant difference at P < cates a significant difference at P < 0.05 from the control 0.01 from the control group. Differences in mean plasma group. Differences in mean plasma and pituitary hormone and pituitary hormone levels among experimental groups levels among experimental groups were analyzed by one- were analyzed by one-way analysis of variance (ANOVA) way analysis of variance (ANOVA) followed by Fisher’s followed by Fisher’s PLSD. Differing letters represent sig- PLSD. Differing letters represent significant differences at nificant differences at P < 0.05 among all groups. Reprinted P < 0.05 among all groups. Reprinted with permission from with permission from Zoological Science, 17, Munakata et Zoological Science, 17, Munakata et al., Inhibitory effects al., Inhibitory effects of testosterone on downstream migra- of testosterone on downstream migratory behavior in masu tory behavior in masu salmon, Oncorhynchus masou, 863– salmon, Oncorhynchus masou, 863–870, Fig. 2,  2000b, 870, Fig. 1,  2000b, Zoological Society of Japan. Zoological Society of Japan.

summer period, which coincides with the timing of tained peak levels earlier than did 17α-P, DHP, and their upstream migration (Fig. 4). Their GSI and plasma LH (Fig. 4). Moreover, T maintained high plasma lev- α levels of T, E2, 17 -P, DHP, and LH increased in May els during a broader period than other sex steroid hor- through October. In addition, most females used in the mones. investigation ovulated around October, similar to the Therefore, in males and females, it seems likely that wild populations (Kiso 1995). In 2+ mature females, sex hormones, especially some of the sex steroid hor- plasma levels of T and E2 began to increase and at- mones, are important factors that regulate (inhibit) the

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 37 display of downstream swimming behavior as well as smoltification. Furthermore, it was indicated that the sex steroid hormone T commonly increased earlier and more broadly than did other sex steroid hormones in both sexes (Fig. 4). In general, T is considered a reser- voir substance for impending conversion to other sex steroid hormones, such as 11-KT, and E2 (male), and E2 (female) (Kagawa et al. 1982a, b). In masu salmon, however, it was also hypothesized that T is one of the more important sex steroid hormones that effectively repress the occurrence of smoltification and the following downstream swimming behavior. This hypothesis was tested in the next set of studies which were conducted in an artificial raceway system (Fig. 5). T and other relevant sex steroid hormones were included in these experiments.

2-8. Inhibitory roles of sex steroid hormones in the downstream swimming behavior in masu salmon

I investigated the effects of treatments with exogenous sex steroid hormones, such as T, 11-KT, E2, and DHP, on the occurrence of downstream swimming behavior, in 1+ masu salmon smolts, held in the experimental raceway system (Fig. 5).

2-8A. Roles of T in the downstream swimming behavior In the raceway, one experiment showed that 89.5% (17 of 19 fish) of the control 1+ masu salmon smolts swam from the upper pond to the lower pond through a fishway (Munakata et al. 2000b; Fig. 6). This con- trasts with the behavior of 1+ smolts into which a T 500 µg/fish via a Silastic tube capsule (Dow Corning Corp.: outer diameter 1.95 mm, inner diameter 1.47 Fig. 8. (a) Frequency of migrants and non-migrants, plasma mm, length 20 mm) was inserted. These 1+ smolts dis- β levels of (b) testosterone (T), (c) estradiol-17 (E2), (d) 11- played high plasma T levels (Fig. 6). Under these cir- ketotestosterone (11-KT), (e) 17,20β-dihydroxy-4-pregnene- cumstances, the frequency of downstream swimming 3-one (DHP), (f) thyroxine (T4), and (g) triiodothyronine (T3) behavior in the T 500 µg/fish-treated group was 31.8% µ in controls, T, E2, 11-KT, and DHP 500 g-treated 1+ masu (7 of 22 fish). In the non-migrants of the T-treated fish, salmon smolts. Numbers above columns in (a) indicate the plasma levels of T were higher than those in migrants, number of migrants and non-migrants. Differences in the suggesting that higher plasma T levels are important frequency of downstream behavior from the control group for the suppression of downstream swimming behavior. were analyzed by the χ2-test, using StatView version 4.5 Since Ikuta et al. (1987) demonstrated that T inhibited software (Abacus Concepts, Inc., California, USA). * and natural smoltification, it is further hypothesized that T *** indicate significant differences at P < 0.05 and P < 0.001, impairs not only downstream swimming behavior but respectively, from the control group. Differences in mean plasma and hormone levels among experimental groups were also seawater preference and schooling behavior in the analyzed by one-way analysis of variance (ANOVA) fol- T-treated 1+ smolts. lowed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted 2-8B. Effects of T doses on the occurrence of down- from Comp. Biochem. Physiol. Part B, 129, Munakata et al., stream swimming behavior The involvement of sex steroid hormones in downstream and It was previously demonstrated that T-treatments sig- upstream migratory behavior of masu salmon, 661–669,  nificantly inhibited the occurrence of downstream 2001a, with permission from Elsevier. swimming behavior in 1+ smolts in a dose dependent manner (Munakata et al. 2000b; Fig. 7). The frequency of downstream swimming behavior in control, T5 µg,

et al. 2001a; Fig. 4). On the other hand, plasma DHP levels increased only during the spawning period in autumn in both sexes. To test the potential roles of sex steroid hormones other than T, the effects of treatments µ of T, E2, 11-KT, and DHP (500 g/fish) on the occurrence of downstream swimming behavior were investigated (Munakata et al. 2001a; Fig. 8). It was found that not only T but also E2 and 11-KT 500 µg/fish-treatments resulted in an elevation of each sex steroid hormone and inhibited the occurrence of downstream swimming behavior in 1+ smolts (Fig. 8). Interestingly, DHP 500 µg/fish-treatments did not inhibit the occurrence of downstream swimming behavior. In the raceway system, it was also demonstrated that all of the forty 1+ masu salmon smolts that were transferred into the upper pond performed downstream swimming behavior within a week (Munakata et al. 2000b; Fig. 9, left columns). By comparison, 60% of T500 µg/fish-treated smolts remained in the upper pond during the same period (Fig. 9, right columns). These phenomena suggest that most 1+ smolts spontaneously exhibit negative rheophilic behavior when flowing Fig. 9. (a) Frequency of migrants and non-migrants, plasma water is present, or that some environmental factors levels of (b) testosterone (T), (c) thyroxine (T4), and (d) are involved in the induction of downstream swimming triiodothyronine (T ) in control and T 500 µg-treated 1+ masu 3 behavior. Since 1+ smolts seemed to swim downstream salmon smolts. Numbers above columns in (a) indicate the spontaneously, it is also hypothesized that T directly number of migrants and non-migrants. Differences in mean inhibits the downstream activity, or inhibits the recep- plasma hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) fol- tiveness to some environmental factors which induce lowed by Fisher’s PLSD. * indicates a significant difference the downstream swimming behavior. at P < 0.05 from migrants. Reprinted with permission from Zoological Science, 17, Munakata et al., Inhibitory effects 2-8D. Effects of intra-specific interactions on the of testosterone on downstream migratory behavior in masu downstream swimming behavior salmon, Oncorhynchus masou, 863–870, Fig. 3,  2000b, In one of the previous investigations, 1+ precocious Zoological Society of Japan. male masu salmon, transferred into the upper pond of the raceway, frequently displayed aggressive behaviors towards 1+ smolts (Munakata et al. 2000b, 2012a). Therefore, another acceptable explanation is that the T50 µg, T500 µg/fish-treated 1+ smolts, and 1+ pre- downstream swimming behavior is induced partly by cocious male groups were 21.3, 18.2, 6.9, 4.5, and 0%, some socio-environmental factors, such as intra- respectively. Plasma T levels and pituitary LH contents specific interactions from 1+ precocious males. in the T500 µg/fish-treated smolt group were highest among all experimental groups (Fig. 7). 2-8E. Effects of dosing period of T on the down- In this study, on the other hand, none of the 1+ pre- stream swimming behavior cocious males exhibited the downstream swimming The dosing period was normally approximately 2 behavior in the artificial raceway, whereas their plasma weeks before sex steroid hormone-treated 1+ masu T levels were lower than those in the T500 µg/fish- salmon smolts were transferred into the upper pond of treated smolt (Fig. 7). Such phenomena are consistent the raceway (Munakata et al. 2000b, 2001a). Conse- with the findings that most of the 1+ precocious males quently, some of the fish in the T500 µg/fish-treated stay in their natal rivers even though their plasma T 1+ smolts exhibited downstream swimming behavior levels are not very high. even when their average plasma T level (13.7 ng/ml) was higher than those of 1+ precocious males (3.9 ng/ 2-8C. Roles of sex steroid hormones other than T in ml) (Fig. 7). One possible explanation is that multiple the downstream swimming behavior sex steroid hormones are necessary to impede down- In maturing masu salmon, not only T but also 11-KT stream swimming behavior. Evidence supporting this (males) and E2 (females) levels increased coincident supposition is that in 1+ precocious males, plasma lev- with the period of downstream migration (Munakata els of not only T but also 11-KT significantly increased

(Fig. 4). Moreover, it was demonstrated that treatment androgen levels largely decreased during winter (Aida µ with T, E2, and 11-KT 500 g/fish significantly inhib- et al. 1984; Munakata et al. 2000b, 2001a). Consider- ited the occurrence of downstream swimming behavior ing these factors, it is suggested that chronic release in 1+ smolts (Fig. 8). of sex steroid hormones into the plasma is an impor- However, another acceptable explanation is that a tant factor in the inhibitory regulation of downstream continuous release of a sex steroid hormone into the swimming behavior. plasma can obstruct the downstream swimming behavior completely. As shown previously (see Sub- 2-9. Inhibitory roles of sex steroid hormones in section 2-4 in detail), 1+ precocious males undergo the downstream swimming behavior in other sexual maturation during the summer of 0+ year-old- Pacific and Atlantic salmon life, a half year prior to the downstream migratory period in the river, although their GSI values and plasma Aside from masu salmon, we have discovered that T500 µg/fish-treatments significantly inhibited the occurrence of downstream swimming behavior in 1+ land-locked sockeye salmon (O. nerka) smolts by use Migrants of the same raceway (Munakata et al. 2012b, Fig. 10). 100 (a) Non-migrants Furthermore, inhibitory effects of sex steroid hor- 6 56* mones upon downstream swimming behavior were also 37 50 31 found in Atlantic salmon (Salmo salar) smolts

fish (%) 20 (Berglund et al. 1994). In these smolts, treatment with Frequency of 1 11-ketoandrostendione (11-KA) via implantation of 0 d Silastic tube capsules inhibited the downstream swim- 50 (b) ming behavioral activity occurring along the current cd in the circular round tank. This suggests that inhibitory regulatory mechanisms of downstream swimming 25 behavior by sex steroid hormones are inhered in some

T (ng/ml) T bc ab Pacific and Atlantic salmon. a a 0 50 (c) b 5 3. Roles of sex hormones in the upstream swim- 40 4 ming and spawning behaviors in masu salmon b and other salmonids 30 3 20 ab 2 a Pituitary LH During autumn, most 2+ mature masu salmon Pituitary LH (ng/pituitary) 10 a a 1 ( g/pituitary) 0 0 10 (d) Fig. 10. (a) Frequency of migrants and non-migrants, plasma levels of (b) testosterone (T), (c) pituitary contents of lutei- 5 b ab nizing hormone (LH), (d) plasma levels of LH, (e) thyrox- bab ab LH (ng/ml) a ine (T4), and (f) triiodothyronine (T3) in control and T 500 µg-treated smolts, and precociously mature male 1+ sockeye 0 salmon. In Fig. 10c, unit of Y axis in the control group was 20 (e) ng/pituitary, while those of T-treated and precocious male 15 c groups was µg/pituitary. Numbers above columns in (a) indicate the number of migrants and non-migrants. Differences 10 bc (ng/ml) b in the frequency of downstream behavior from the control 4 T 5 group were analyzed by the χ2-test, using StatView version a a a 4.5 software (Abacus Concepts, Inc., California, USA). 0 * indicates a significant difference at P < 0.05 from the con- 20 (f) trol group. Differences in mean plasma and pituitary hor- 15 mone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fish- 10 (ng/ml) er’s PLSD. Differing letters represent significant differences 3 ab ab ab b T ab 5 a at P < 0.05 among all groups. Reprinted with permission from Fish. Sci., 78, Munakata et al., Involvement of sex ster- 0 Control T-treated Precocious oids, luteinizing hormone and thyroid hormones in upstream male and downstream migratory behaviors in land-locked sockeye salmon Oncorhynchus nerka, 81–90, Fig. 3,  2012b, The Fig. 10. Japanese Society of Fisheries Science.

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. 40 A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 sequentially migrate upstream (i.e., upstream swim- Rivers) in northern Honshu during late winter through ming behavior) from the sea (or lakes), home to their early spring. Also, some of the 2+ smolt migrants, natal rivers and then spawn in the upper reaches of the members of the so-called “late run” migrate into small catchment basin (Machidori and Kato 1984; Kato 1991; rivers (i.e., Kesen, Hirose, and Natori Rivers) along Kiso 1995). Also, during this period, non- the Sanriku coast of northeastern Honshu, just before migratory, precocious males swim upstream from the the spawning period (Machidori and Kato 1984; Kiso mid-to upper-river reaches and consequently spawn 1995; Munakata et al. unpublished data). This illus- with 2+ migrants (see Fig. 3). Since most of these fish trates that there are variations in timing when the 2+ exhibit signs of sexual maturation while engaged in smolts return to their natal rivers. Moreover, it is also these activities, it appears likely that sexual matura- hypothesized that some masu salmon smolts can modu- tion induces or regulates these behaviors. In this sec- late their osmoregulatory ability coincident with their tion, an overview of the lifecycle of migrants and non- entry into the natal rivers. migrants of masu salmon during the periods of hom- On reaching their natal streams, a significant pro- ing and upstream migration through spawning will be portion of masu salmon regularly inhabit deep areas presented with regard to previous investigations of the of the rivers, such as pools and the thalweg (center of stimulatory effects of sex steroid hormones on the oc- the flow), usually located downstream from their au- currence of upstream swimming and spawning tumn spawning sites. Because 2+ migrants are now behaviors (Munakata et al. 2001a, 2001b, 2002, 2012a). larger (50 to 60 cm in BL), it is probable that they stay in deep areas to hide from potential predators. Further- 3-1. Upstream migratory behavior in masu more, water temperatures in deeper areas may be modu- salmon lated and stabilized by spring water upwelling from the river beds. During late summer through autumn, The feeding migration of 1+ masu salmon smolts in most of the 2+ migrants start to move upstream again, the sea occurs between March and May, after which toward their spawning areas (Kato 1991; Kiso 1995; the major part of the run has entered into either the Munakata and Miura, unpublished data). It is thus as- Pacific Ocean or the Sea of Japan (Machidori and Kato sumed that the upstream swimming behavior of 2+ 1984). Most of these smolts are thought to migrate into migrants can be regularly divided into two steps: 1) areas between the Sea of Japan near Hokkaido Island movement from the mouth of the river to the areas in and the Sea of Okhotsk around June. However, it is which 2+ fish spend the summer, and 2) movement also thought that a small number of 1+ masu salmon from the latter areas to their spawning areas. It seems smolts migrate to coastal areas, such as off the Sanriku that the upstream migratory behavior of the “late run”, coast on the Pacific Ocean side of northern Honshu which is generally found in small streams, coincides (Kiso 1995). As mentioned in Section 4, such short with the latter upstream migratory pattern. Most 2+ distance migratory forms are considered to be “coastal migrants that reach their spawning sites exhibit high migrants” (see Subsection 5-2 in detail). The Sea of GSI values (Munakata et al. 2001a, 2012a), and sub- Okhotsk is the summer-late autumn feeding ground for sequently the 2+ males spermiate and 2+ females ovu- most of the 1+ smolts, where they forage on fish, squid, late (Munakata and Kobayashi 2010). Amphipods, Euphausiids, Decapods, Copepods, and a small number of terrestrial insects (Machidori and Kato 3-2. Feeding of 2+ masu salmon migrants during 1984; Kato 1991). During this period of time, most the upstream migration masu salmon reach 50 to 60 cm in BL. During winter through spring when the smolt reach the age of 2+, As is typical for semelparous Pacific salmonids, adult most will head southward towards their spawning masu salmon are thought not to feed after returning to ground in natal rivers (i.e., homing migration, upstream their natal rivers. According to Sano (1947), however, migration) (Machidori and Kato 1984; Kato 1991). some 2+ masu salmon smolts, which are considered Values of GSI and plasma levels of sex steroid hor- the “early run”, in Nishibetsu and Shibetsu Rivers in mones start to increase in both sexes (Munakata et al. Hokkaido, occasionally feed. In 2010, it was also dis- 2001a; Fig. 4). Hence, there is a correlation between covered that a 2+ masu salmon male migrant caught in the initiation of sexual maturation and the occurrence the Hirose River, Miyagi Prefecture, that enters Sendai of homing and upstream migratory behaviors. Bay near the Sanriku coast had consumed a number of In general, most 2+ smolts migrate into their natal larval aquatic insects (Munakata and Miura, unpub- rivers and subsequently show upstream swimming lished data). Interestingly, sport fishermen (i.e., lure, behavior during mid spring through early summer. fly, and bait fishing) consistently catch 2+ “early run” According to information provided by the sports fish- smolts in larger rivers, such as the Kitakami, Mogami, ing industry, however, it seems that some 2+ masu Akagawa, Omono Rivers in northern Honshu in late salmon called the “early run” migrate into their natal winter through spring, while some fishermen also fish rivers (e.g., Kitakami, Mogami, Akagawa, and Omono the 2+ masu salon smolts via bait fishing nearshores

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 41 of Hokkaido coasts. and Kobayashi 2010). Thus there are conflicting reports regarding the feed- Recently, it was discovered that 2+ mature female ing of the 2+ masu salmon migrants in the rivers. Since masu salmon release a sex pheromone, L-kynurenine, most of the 2+ migrants are considered not to feed which attracts sexually mature males (Yambe et al. during summer through autumn, though they seem to 2003, 2006). The urine from 2+ mature females attracts feed during winter through spring in both the rivers and elicits the male spawning behaviors, such as at- and offshore sea, one possible hypothesis is that the tending behaviors (see Figs. 17, 18). The timing of its feeding activity of 2+ migrants is determined by their production in the females clearly indicates that this sex maturity, but not by the entry into the rivers. pheromone could be a signal to non-specific males indicating sexual maturity, location of the redd, and re- 3-3. Lifecycle of non-migratory precocious males ceptiveness of 2+ females to non-specific males. After sexually mature 2+ male migrants arrive at the Precocious male non-migrants generally maintain spawning grounds, these males regularly swim around their territories during 0+ winter through 1+ summer the 2+ digging females and exhibit a series of male in the upper and middle reaches of their natal rivers spawning behaviors, such as attending and quivering, (Machidori and Kato 1984; Kiso 1995). Subsequently, towards the digging female (Munakata and Kobayashi the 1+ precocious males begin to become sexually com- 2010) (see also Fig. 18). Additionally, 2+ males some- petent, resulting in high plasma levels of T and 11-KT times exhibit aggressive behaviors towards other after summer (see Fig. 4). The 1+ precocious males male(s) of the same species, to prevent the antagonis- then exhibit upstream movements during late summer tic males from showing courtship behavior to the dig- through autumn the same as 2+ migrants (Kiso 1995; ging female. After the redd is constructed by the fe- Munakata et al. 2001b). male, both the female and male crouch (i.e., crouching In contrast to the 2+ migratory smolts, most 1+ pre- behavior) on the accomplished redd, and release eggs cocious males continue to feed on insects or small fish, (oviposition) and sperm (ejaculation), respectively prior to and during the spawning period (Munakata et (Munakata and Kobayashi 2010). Thereafter, the fe- al. unpublished data), indicating that the changes in male covers the redd (i.e., covering behavior) with feeding activity do not depend on the maturity of the small stones and pebbles by using its caudal fin in a precocious males. After the occurrence of upstream similar manner to the digging behavior. Most of 2+ movements, these fish attend to spawning together with females and males will repeat such spawning behaviors 2+ male and female migrants. The precocious male several times over a few weeks until most ovulated non-migrants seem to repeat the same phenomena dur- oocytes are released (Machidori and Kato 1984). ing the ages of 0+ through 2+ (Kiso 1995). 3-5. Spawning behaviors in 1+ precocious male 3-4. Spawning behaviors in 2+ migrants non-migrants

Spawning of masu salmon is observed in natal riv- Spawning behaviors of precocious male non- ers approximately from August through October migrants are generally different from those in 2+ male (Machidori and Kato 1984; Kiso 1995). The peak pe- migrants (Munakata and Kobayashi 2010). Since the riod of spawning is generally earlier in northern re- body size of the precocious male non-migrants (10 to gions (i.e., Hokkaido) than in southern ones (i.e., 30 cm in BL) is relatively smaller than that of 2+ male Honshu and Kyushu). migrants (Utoh 1976, 1977), most 0+, 1+, and 2+ pre- During the spawning period, 2+ female migrants start cocious males spawn “as sneakers” (Munakata and to swim above specific river beds where various sizes Kobayashi 2010). Briefly, the sneaker precocious male of stones and pebbles are located and oxygen rich wa- does not display the specific attending and quivering ter indwells, the same as coho salmon (O. kisutch) behaviors towards the nest digging 2+ female. Instead, (Sandercock 1991) (note that pink salmon spawn on these males conceal their bodies behind obstacles, such river beds where spring water upwells, Heard 1991). as large rocks, fallen trees, etc., while larger 2+ fe- Above such a river bed, a 2+ female digs up (i.e., dig- males and males undertake a series of spawning (pair- ging behavior) the pebbles and stones to make a spawning) behaviors (Munakata et al. unpublished data). ing bed (i.e., redd: 170 × 80 cm in length and width, Also, it was observed that some precocious males swim with a depth of 12 to 45 cm) by using her body, espe- posteriorly to the redd, while 2+ females display dig- cially the tail (caudal fin), digging at intervals of 1 to ging behavior. These precocious males are called “ac- 5 min. While digging, the 2+ female frequently checks cessory males”. During the spawning period, dominant the redd’s shape (i.e., depth) and substrates by using 2+ male migrants exhibit aggressive behaviors towards mainly the pectoral fins (i.e., probing behavior). Dur- other fish including these precocious males. About the ing and after digging, most 2+ females ovulate to pre- time when the 2+ females and males release eggs and pare for the oviposition (i.e., egg release) (Munakata sperm, the 0+, 1+, or 2+ precocious male sneaker(s) or

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. 42 A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 accessory males swim onto the redd and immediately spring through autumn, overlapping the period of their release their sperm on the released eggs. homing, upstream migration, and spawning (Fig. 4). Among these sex hormones, plasma T and E2 levels 3-6. Spawning behaviors in female non-migrants began to increase after May and attained peak levels earlier than did other sex hormones, such as 17α-P, Male and female non-migrant masu salmon other DHP, and LH. than the precocious fish have been reported in some rivers (Kiso 1995). The proportions of these types of 3-8. Changes in sex steroid hormone levels before non-migrants generally increases towards the south- and after the occurrence of upstream swim- ern regions (i.e., Honshu and Kyushu), the same as ming behavior in masu salmon precocious males. These fish regularly exhibit low GSI values compared with precocious male non-migrants, 3-8A. Males and their sex steroid hormones remain at low levels. We investigated the changes in the plasma levels of However, their growth rates were relatively higher than sex steroid hormones (T, E2, 11-KT, and DHP) in 2+ those of other immature fish, including the future masu salmon during the following temporal phases: smolts, during the age of 0+ through 1+ (Kiso 1995). before the onset of upstream swimming behavior (ini- These non-migrants regularly mature and spawn up to tial sampling: before transfer into the lower pond of twice, during the 1+ and 2+ autumn. Thus 2+ masu the two-step raceway), during the migratory period salmon spawners are comprised of larger migrants, (sampled from upstream migrants), and after upstream smaller precocious males and other 2+ parr non- activity ceased (sampled from non-migrants remain- migrants in the rivers. In general, the spawning ing in the lower pond of the raceway) (Munakata et al. behaviors of the female and male non-migrants are 2012a; Table 1). It was found that all of the 15 2+ similar to those of 2+ migrants (Munakata et al. un- males continued to spermiate before and during the published data). onset of upstream movements, indicating that the fish spermiate during the upstream migratory phase. Plasma 3-7. Changes in plasma sex hormone levels dur- levels of T, 11-KT, and DHP were considerably higher ing the upstream migratory and spawning during September (Table 1), comparable to the levels periods displayed by hatchery raised 2+ males, as shown in Fig. 4 (Munakata et al. 2001a). Plasma T, E2, and 11- Thus in masu salmon, the appearance of homing, KT levels decreased significantly after the cessation upstream migratory, and spawning behaviors is closely of upstream movement. Even so, plasma T, 11-KT, and related to the progress of sexual maturation and an in- DHP maintained higher levels than those of 1+ imma- crease in sex hormone levels. To understand which ture males (Fig. 4). hormonal factors control the occurrence of upstream swimming behaviors and subsequent male and female 3-8B. Females spawning behaviors, it is required to measure plasma Each of the 10 females at age 2+ that moved upstream levels of sex steroid hormones and LH before, during, in the raceway ovulated while 5 non-migrants that re- and after the behavioral patterns become manifest. This mained in the lower pond did not (Munakata et al. is covered in the following sections. 2012a; Table 1). These phenomena indicate that female masu salmon ovulate during the last phase, or 3-7A. Males after upstream swimming behavior has ceased. Most In 2+ males, values of GSI and plasma levels of T, of the 10 females moved upstream within 1 week after 11-KT, and DHP increased during the periods of hom- the experiment began. Interestingly, prior to the ex- ing migration through spawning (Munakata et al. periment, plasma levels of T and E2 in upstream mi- 2001a; Fig. 4). In 2+ males, moreover, it was noticed grants were significantly higher and lower, respec- that plasma levels of T increased earlier than did 11- tively, than the corresponding levels in non-migrants KT and DHP, and T retained higher plasma levels ex- (Table 1). Furthermore, plasma E2 levels significantly tensively during May through September. decreased after the females completed their upstream Regarding 1+ precocious male non-migrants, such movement. Though the E2 levels decreased, the levels plasma sex steroid hormone elevations coincide with of T, E2, and DHP were considerably higher than those the period when these fish remain in their natal rivers, in the 1+ immature females (Fig. 4). display upstream movement toward spawning areas, In summary, the plasma levels of sex steroid hor- and spawning behaviors (Fig. 4). mones increased during the upstream migratory and spawning periods, for both 2+ males and females. 3-7B. Females These findings are supported by the fact that patterns In 2+ female masu salmon, values of GSI and plasma in plasma elevation of sex steroid hormones during the α levels of T, E2, 17 -P, DHP, and LH increased during upstream migratory and spawning periods are consist-

β Table 1. Frequency of migrants and non-migrants, and plasma levels of testosterone (T), estradiol-17 (E2), 11-ketotestosterone β ± (11-KT), 17,20 -dihydroxy-4-pregnene-3-one (DHP), thyroxine (T4), and (g) triiodothyronine (T3) (mean SEM) in 2+ male and female masu salmon during upstream migratory period. Differences in mean plasma hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. —: no sample. Reprinted from Aquaculture, 362–363, Munakata et al., Involvement of sex steroids and thyroid hormones in upstream and downstream behaviors in masu salmon, Oncorhynchus masou, 158–166,  2012a, with permission from Elsevier.

ent with those detected in chum salmon, sockeye masu salmon do not have high plasma 11-KT levels salmon, chinook salmon (O. tshawytscha), rainbow (Fig. 4, Table 1). Based on these data, it is proposed trout, and Arctic charr (Salvelinus alpinus) engaged in that 11-KT is not involved in the regulation of upstream the same activities (Lou et al. 1984; Ueda et al. 1984; swimming behavior in female masu salmon. Liley et al. 1986; Truscott et al. 1986; Slater et al. 1994; In 1+ precocious males, the control (sham-operated) Frantzen et al. 1997). group after being transferred to the lower pond moved Considering the changing patterns of sex steroid upstream to the upper pond at significantly high fre- hormones in masu salmon, sex steroid hormones espe- quencies (Munakata et al. 2001b; Fig. 12). To the con- cially T, E2, 11-KT, and DHP (males), and T, E2, and trary, castrated 1+ precocious males did not show this DHP (females) seem to be important factors that con- tendency. However, upstream swimming behavior was trol the occurrence of homing, upstream swimming, elicited from castrated 1+ precocious males treated with µ and spawning behaviors in 2+ migrants. T and 11-KT 500 g/fish. In contrast, E2 and DHP 500 In 2+ non-migratory precocious males, moreover, sex µg/fish did not induce significant upstream swimming steroid hormones such as T and 11-KT seem to influ- behavior in castrated fish (Fig. 12). ence the occurrence of stream residency, upstream To summarize these experiments, it appears that T movement from their territory to the spawning ground, and E2 (females), and T and 11-KT (males) play sig- and spawning behaviors. nificant roles in inducing the occurrence of upstream swimming behavior in masu salmon. Furthermore, it 3-9. Stimulatory effects of sex steroid hormones is also suggested that DHP had no significant effect on on the upstream swimming behavior in masu the occurrence of the upstream swimming behaviors salmon in either sex (Figs. 11, 12).

During autumn, 1+ immature parr had low levels of 3-10. Roles of T in the upstream and downstream plasma sex steroid hormones and did not exhibit any swimming behaviors in masu salmon tendency to move upstream in the artificial raceway. µ However, when T, E2, and 11-KT 500 g/fish via Since androgen (male sex steroid hormone) T com- SILASCON tubing (Kaneka, Medics Corp.; outer di- monly increases plasma levels in both sexes, T is re- ameter 1.5 mm, inner diameter 1.0 mm, length 30 mm) garded as one of the most important sex steroid hor- were implanted into the abdominal cavity of 1+ imma- mones in regulating the occurrence of the upstream and ture parr, upstream swimming behavior was induced downstream swimming behaviors in salmonids, such (Munakata et al. 2001a; Fig. 11). Interestingly, it ap- as masu salmon. In teleosts, however, T is also the pre- peared that DHP 500 µg/fish-treatment had little in- cursor converted to other sex steroid hormones such fluence on the occurrence of upstream behavior. as estrogen E2 (males, females) and androgen 11-KT In the interest of full disclosure, however, most of (males) (Kagawa et al. 1982a, b). For these reasons, it the 1+ immature masu salmon used in the research is hypothesized that T itself does not regulate the oc- (Shiribetsu River strain) were females, because most currence of the downstream or upstream swimming males from this stock mature precociously (Munakata behaviors. et al. 2000b). In this strain, the average male to female ratio of 1+ immature parr was approximately 1:9 (for 3-10A. Females example, Munakata et al. 2012a). In general, female In order to investigate the potential role(s) of andro-

Fig. 12. (a) Frequency of migrants and non-migrants, plasma levels of (b) testosterone (T), (c) estradiol-17β Elsevier Sci- ence (USA), with permission from Elsevier. (E ), (d) 11- Fig. 11. (a) Frequency of migrants and non-migrants, plasma 2 β levels of (b) testosterone (T), (c) estradiol-17β (E ), (d) 11- ketotestosterone (11-KT), (e) 17,20 -dihydroxy-4-pregnene- 2 3-one (DHP), (f) thyroxine (T ), and (g) triiodothyronine (T ) ketotestosterone (11-KT), (e) 17,20β-dihydroxy-4-pregnene- 4 3 µ µ µ in castrated, castrated + T 500 g, E2 500 g, 11-KT 500 g, 3-one (DHP), (f) thyroxine (T4), and (g) triiodothyronine (T3) µ µ µ and DHP 500 µg/ fish-treated groups, and sham-operated in controls, T 500 g, E2 500 g, 11-KT 500 g, and DHP 500 µg-treated 1+ immature masu salmon parr. Numbers 1+ precocious male masu salmon. Numbers above columns above columns in (a) indicate the number of migrants and in (a) indicate the number of migrants and non-migrants. non-migrants. Differences in the frequency of upstream Differences in the frequency of upstream behavior from the χ2 behavior from the control group were analyzed by the χ2- control group were analyzed by the -test, using StatView test, using StatView version 4.5 software (Abacus Concepts, version 4.5 software (Abacus Concepts, Inc., California, Inc., California, USA). *, **, and *** indicate significant USA). *, **, and *** indicate a significant difference at P < difference at P < 0.05, P < 0.01, and P < 0.001, respectively, 0.05, P < 0.01, and P < 0.001, respectively from the control from the control group. Differences in mean plasma hormone group. Differences in mean plasma hormone levels among levels among experimental groups were analyzed by one- experimental groups were analyzed by one-way analysis of way analysis of variance (ANOVA) followed by Fisher’s variance (ANOVA) followed by Fisher’s PLSD. Differing PLSD. Differing letters represent significant differences at letters represent significant differences at P < 0.05 among P < 0.05 among all groups. Reprinted from Comp. Biochem. all groups. Reprinted from Comp. Biochem. Physiol. Part Physiol. Part B, 129, Munakata et al., The involvement of B, 129, Munakata et al., The involvement of sex steroid hor- sex steroid hormones in downstream and upstream migra- mones in downstream and upstream migratory behavior of  tory behavior of masu salmon, 661–669,  2001a, with per- masu salmon, 661–669, 2001a, with permission from mission from Elsevier. Elsevier.

β Table 2. Frequency of migrants and non-migrants, and plasma levels of testosterone (T), estradiol-17 (E2), thyroxine (T4), ± and (g) triiodothyronine (T3) (mean SEM) in 1+ immature masu salmon implanted with T, 1,4,6-androstatriene-3,17-dion (ATD) or tamoxifen 500 µg/fish. Differences in the frequency of upstream behavior from the control group were analyzed by the χ2-test, using StatView version 4.5 software (Abacus Concepts, Inc., California, USA). * indicates significant difference at P < 0.05, from the control group. Differences in mean plasma hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. —: no sample. Reprinted from Aquaculture, 362–363, Munakata et al., Involvement of sex steroids and thyroid hormones in upstream and downstream behaviors in masu salmon, Oncorhynchus masou, 158–166,  2012a, with permission from Elsevier.

gen T in the regulation of upstream swimming behavior the T, 11-KT, and E2 are potential factors involved in in females, we examined the effects of implants of T, the regulation of upstream swimming behavior in aromatase inhibitor 1,4,6-androstatrien-3,17-dione males. (ATD), and the estrogen antagonist tamoxifen 500 µg/ In order to investigate the potential role(s) of andro- fish on the occurrence of upstream swimming behavior gens, such as T and 11-KT, in the regulation of up- in 1+ immature parr using the artificial raceway stream swimming behavior, the effects of tamoxifen (Munakata et al. 2012a; Table 2). It was assumed that 500 µg/fish-treatment on the occurrence of upstream the upstream swimming behavior in the female was swimming behavior in 1+ precocious males was in- regulated by aromatized E2 but not T. If this hypoth- vestigated. As a result, it was found tamoxifen did not esis is correct, it is inferred that administration of ATD decrease the stimulatory effects of T on the occurrence and tamoxifen should lower the stimulatory effects of of upstream swimming behavior (Table 3, Exp 4). T treatment on the occurrence of upstream swimming Therefore, it is thought that estrogens such as E2, and behavior. Otherwise, there is a possibility that T itself androgens such as T and 11-KT may regulate the oc- regulates the occurrence of the upstream swimming currence of upstream swimming behavior in males. behavior without being converted to E2. As a result, Thus in masu salmon, we concluded that both estrogens µ ATD and tamoxifen 500 g/fish-treatment did not de- such as E2 (males and females), and androgens such as crease the stimulatory effects of T on the occurrence T (males and females) and 11-KT (males) are involved of upstream swimming behavior in 1+ immature fe- in the regulation of upstream swimming behavior. Fur- male parr (Munakata et al. 2012a; Table 2, Exp 2 and thermore, because of the patterns of changes in the 3). plasma levels (Fig. 4), it is concluded that T, especially, is one of the common sex steroid hormones that regu- 3-10B. Males late the occurrence of downstream and upstream swim- In males, T and 11-KT 500 µg/fish-treatment induced ming behaviors. the occurrence of upstream swimming behavior in 1+ In masu salmon, DHP did not exhibit a significant castrated precocious males (Munakata et al. 2001b; Fig. effect on the occurrence of downstream and upstream µ 12). Moreover, treatment with estrogen E2 500 g/fish swimming behavior (Munakata et al. 2001a). DHP is induced the occurrence of upstream swimming considered to be maturation inducing factor (MIF), behavior in 1+ intact precocious males (Munakata et which mediate final oocyte maturation (ovulation) and al. 2012a; Table 3, Exp 4). Hence, it is speculated that final testicular maturation (spermiation), in most

β Table 3. Frequency of migrants and non-migrants and, plasma levels of testosterone (T), estradiol-17 (E2), thyroxine (T4), ± µ and (g) triiodothyronine (T3) (mean SEM) in castrated, castrated + E2 500 g/fish, sham-operated, sham-operated + E2 500 µg/fish, control, and tamoxifen 500 µg/fish-treated 1+ precocious male masu salmon. Differences in the frequency of upstream behavior from the control group were analyzed by the χ2-test, using StatView version 4.5 software (Abacus Concepts, Inc., California, USA). * indicates significant difference at P < 0.05, from the control group. Differences in mean plasma and pituitary hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. —: no sample. Reprinted from Aquaculture, 362–363, Munakata et al., Involvement of sex steroids and thyroid hormones in upstream and downstream behaviors in masu salmon, Oncorhynchus masou, 158–166,  2012a, with permission from Elsevier.

teleosts (Nagahama 1987a, b). Plasma DHP levels regu- swimming behavior in the later experiments (36.8%) larly increase just prior to the spawning period was lower than in the former experiment (43%) (Munakata et al. 2001a; Fig. 4), while most salmonids (Munakata et al. 2012a; Table 2, Exp 1). Although the have already arrived at their spawning areas. As Mayer inference is not strong, this suggests that the dosing et al. (1994) mentioned, the treatment of DHP via period (duration of plasma sex steroid hormone in- Silastic tube insertion induced the occurrence of male crease) is an important factor in this process. spawning behaviors. These findings indicate that DHP In these experiments, major parts of upstream swim- is more important in regulating spawning behaviors ming behaviors occurred while there were no signifi- than upstream and downstream swimming behaviors. cant changes in plasma levels of T being administered As demonstrated previously, the frequency of down- through Silastic or SILASCON tubing. This suggests stream swimming behavior changes in accordance with that the continuous release of T may play a role as a the treatment dose of T in 1+ immature masu salmon “requirement” for the occurrence of the upstream swim- smolts (Munakata et al. 2000b; Fig. 7). In like fash- ming behavior (Munakata and Kobayashi 2010). ion, the induction of upstream swimming behavior in 1+ immature parr (Fig. 13) and 1+ castrated preco- 3-11. Roles of sex steroid hormones in the up- cious males (Fig. 14) seemed to be a dose-dependent stream swimming behavior in land-locked response (Munakata et al. 2001b). These are also strong sockeye salmon indications that the plasma level of T is an important factor regulating the occurrence of either downstream It was determined that T 500 µg/fish treatment sig- or upstream swimming behaviors. nificantly induced the occurrence of upstream swim- It was also detected that the dosing period may be ming behaviors in 1+ immature land-locked sockeye an important factor in initiating the upstream response. salmon in the raceway (Munakata et al. 2012b; Fig. When 1+ immature masu salmon were implanted with 15). In addition, precocious males of 1+ land-locked T 500 µg/fish for approximately 4 months, the frequen- sockeye salmon with high plasma T levels also mi- cies of upstream migrants was considerably higher grated upstream at significantly high frequencies. From (43%) than for control individuals (19%) (Munakata this, it is inferred that stimulatory regulation of up- et al. 2012a; Table 2, Exp 1). On the other hand, the stream migratory behavior by sex steroid hormones frequency of upstream migrants given a dosage of T might be common to multiple Pacific salmon. 500 µg/fish for approximately 2 months in different trials were 17.1% (Munakata et al. 2001b; Fig. 13), 3-12. Roles of sex hormones other than sex steroid 22% (Munakata et al. 2001b; Fig. 16), 36% (Munakata hormones in the upstream swimming et al. 2012a; Table 2, Exp 2), 52% (Munakata et al. behavior in salmonids 2001b; Fig. 11), and 57% (Munakata et al. 2012a; Ta- ble 2, Exp 3). The average frequency of upstream In 2+ female masu salmon, all of the 10 upstream

Fig. 13. (a) Frequency of migrants and non-migrants, plasma Fig. 14. (a) Frequency of migrants and non-migrants, plasma levels of (b) testosterone (T), (c) thyroxine (T4), and (d) levels of (b) testosterone (T), (c) thyroxine (T4), and (d) µ µ µ triiodothyronine (T3) in controls, T 50 g, T 500 g, and T triiodothyronine (T3) in castrated, cast. + T 50 g, cast. + T 1000 µg-treated 1+ immature masu salmon. Numbers above 500 µg-treated groups, and sham-operated 1+ precocious columns in (a) indicate the number of migrants and non- male masu salmon. Numbers above columns in (a) indicate migrants. Differences in the frequency of upstream behavior the number of migrants and non-migrants. Differences in from the control group were analyzed by the χ2-test, respec- the frequency of upstream behavior from the control group tively, using StatView version 4.5 software (Abacus Con- were analyzed by the χ2-test, using StatView version 4.5 cepts, Inc., California, USA). * indicates significant differ- software (Abacus Concepts, Inc., California, USA). * and ence at P < 0.05, from the control group. Differences in mean *** indicate significant difference at P < 0.05 and P < 0.001, plasma hormone levels among experimental groups were respectively from the control group. Differences in mean analyzed by one-way analysis of variance (ANOVA) fol- plasma hormone levels among experimental groups were lowed by Fisher’s PLSD. Reprinted from General and Com- analyzed by one-way analysis of variance (ANOVA) fol- parative Endocrinology, 122, Munakata et al., The effects lowed by Fisher’s PLSD. Differing letters represent signifi- of testosterone on upstream migratory behavior in masu cant differences at P < 0.05 among all groups. Reprinted salmon, Oncorhynchus masou, 329–340,  2001b, with per- from General and Comparative Endocrinology, 122, mission from Elsevier. Munakata et al., The effects of testosterone on upstream migratory behavior in masu salmon, Oncorhynchus masou, 329–340,  2001b, with permission from Elsevier. migrants ovulated, while 5 non-migrants did not (Munakata et al. 2012a; Table 1). These results sug- indicate that LH levels increase with final ovarian gest that females ovulate during the last phase, or after maturation, especially during ovulation. These studies their upstream swimming behavior ceased. Since sper- support the hypothesis that LH may play a role in the miation and ovulation are controlled physiologically regulation of upstream swimming behavior. by pituitary hormones such as LH (Nagahama 1984; Implants of a gonadotropin-releasing hormone ana- Kobayashi et al. 1986, 1988), it was hypothesized that logue (GnRHa) enhanced the occurrence of homing not only sex steroid hormones but also some other sex behavior, the movement from the center of Lake hormones may be involved in the occurrence of up- Shikotsu (Hokkaido) to the mouth of the natal rivers stream swimming behavior. Investigations into masu in adult land-locked sockeye salmon (Sato et al. 1997). salmon (Amano et al. 1992, 1993), coho salmon Based on this, it is hypothesized that GnRH directly (Swanson 1991), and rainbow trout (Prat et al. 1996) influences the occurrence of upstream swimming

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. 48 A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 behavior. On the other hand, considering that the GnRH movement when compared with corresponding values is one important factor which stimulates the secretion for intact immature fish (Figs. 6, 7, 10, 15, 16). Since of LH from the pituitary gland into the plasma (Amano one of main roles of LH is to stimulate sexual (go- et al. 1995), it is also inferred that GnRH plays a role nadal) maturation (Nagahama 1984), there is the pos- in enhancing the occurrence of homing behavior and sibility that LH influences the upstream swimming perhaps also upstream swimming behavior through the behavior through stimulating the synthesis and/or se- release of LH into the plasma. cretion of sex steroid hormones into the plasma. How- In masu salmon and land-locked sockeye salmon, ever, the critical role of LH in the regulation of migra- however, it was demonstrated that plasma LH levels tory behaviors clearly requires further investigation. in T-treated immature fish did not exhibit clear increases over the period of downstream and upstream 3-13. Roles of sex steroid hormones in the spawning behavior in masu salmon

Since T was identified as one of the common sex 100 (a) 48 42 steroid hormones which stimulated the onset of up- 18 stream swimming behavior, T may also give rise to ** 50 13 spawning behaviors. In order to investigate the stimu- fish (%)

* latory role of T in the male and female spawning

Frequency of 8 behaviors, the following experiments were conducted. 2 0 T 500 µg/fish or T 1000 µg/fish dosages were admin- 50 (b) Migrants istered to 1+ castrated precocious males or immature Non-migrants females placed in an artificial stream chamber (1.5 × c 0.6 m with a water depth of 0.2 m) (Munakata et al. 25 b b 2002). T (ng/ml) T

a a N.S. 3-13A. Roles of T on the spawning behavior in males 0 In general, the differences in spawning behaviors 50 (c) 5 between “sneakers” (precocial males) and anadromous b 40 4 males (2+ males) have been discussed. However, if a 30 3 1+ precocious male and 2+ female are transferred to- 20 2 gether into an artificial chamber, the 1+ precocious Pituitary LH Pituitary LH

a ( g/pituitary) (ng/pituitary) 10 a 1 a a N.S. 0 0 10 (d) Fig. 15. (a) Frequency of migrants and non-migrants, plasma levels of (b) testosterone (T), (c) pituitary contents of luteinizing hormone (LH), (d) plasma levels of LH, (e) thyrox- 5 b b ine (T4), and (f) triiodothyronine (T3) in control and T 500 ab LH (ng/ml) aab µg/fish-treated 1+ immature fish, and 1+ precociously ma- N.S. ture male sockeye salmon. In (c), unit of Y axis in the con- 0 trol group was ng/pituitary, while those of T-treated and pre- 20 µ (e) b cocious male groups was g/pituitary. Numbers above col- 15 umns in (a) indicate the number of migrants and non-migrants. N.S. represents no sample. Differences in the fre- 10 quency of upstream behavior from the control group were (ng/ml) a 4

T a analyzed by the Fisher’s exact probability test, using 5 a a StatView version 4.5 software (Abacus Concepts, Inc., Cali- N.S. 0 fornia, USA). * and ** indicate a significant difference at P 20 (f) < 0.05 and P < 0.01 from the control group, respectively. b Differences in mean plasma and pituitary hormone levels 15 among experimental groups were analyzed by one-way 10 analysis of variance (ANOVA) followed by Fisher’s PLSD. (ng/ml) a 3 a Differing letters represent significant differences at P < 0.05 T 5 among all groups. Reprinted with permission from Fish. Sci., a a N.S. 78 et al 0 , Munakata ., Involvement of sex steroids, luteiniz- Control T-treated Precocious ing hormone and thyroid hormones in upstream and down- male stream migratory behaviors in land-locked sockeye salmon Oncorhynchus nerka, 81–90, Fig. 6,  2012b, The Japanese Fig. 15. Society of Fisheries Science.

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 49 male exhibits male spawning behaviors typically associated with 2+ mature males (see Fig. 18). 1+ precocious males (sham-operated fish) frequently showed attending and quivering behaviors towards the 2+ females (Munakata et al. 2002; Figs. 17, 18). 1+ castrated males did not display such behaviors (Fig. 17). However, the full behavioral array of spawning males (e.g., attending and quivering) can be restored by the administration of T 500 µg/fish.

3-13B. Roles of T on the spawning behavior in females In the artificial stream chamber, 2+ ovulated isolated females frequently exhibited digging behaviors on the gravel substrates (Figs. 19, 20). This behavior con- forms field observations. Female Pacific salmon generally arrive at spawning sites earlier than do males (Groot and Margolis 1991), and females, but not males, release a pheromone L-kynurenine to attract non- specific mature males (Yambe et al. 2003, 2006). On the other hand, 1+ immature females did not perform digging behaviors (Fig. 19). The GSI values and plasma sex steroid hormone levels of 1+ immature females were low (Fig. 4), and their ovaries were in the primary growth stage (Kiso 1995). 1+ immature females, however, significantly exhibited digging behaviors when treated with a dose of T 1000 µg/fish (Munakata et al. 2002; Fig. 19).

3-14. Stimulatory effects of sex steroid hormones on the spawning behavior in male rainbow trout

It was demonstrated that treatment of DHP induced the occurrence of quivering and attending behaviors in castrated male rainbow trout towards sexually re- Fig. 16. (a) Frequency of migrants and non-migrants, plasma ceptive females (Mayer et al. 1994). Interestingly, ad- levels of (b) testosterone (T), (c) pituitary contents of lutei- ministration of 11-ketoandrostendione (11-KA) did not nizing hormone (LH), (d) plasma levels of LH, (e) thyrox- induce male spawning behaviors to any significant ine (T4), and (f) triiodothyronine (T3) in control and T 500 degree. This evidence supports the notion that male µg/fish-treated 1+ immature masu salmon. Numbers above spawning behaviors in some Pacific salmon are con- columns in (a) indicate the number of migrants and non- trolled by some sex steroid hormones, potentially T migrants. N.S. represents no sample. Differences in the fre- and/or DHP, released into the plasma. quency of upstream swimming behavior from the control group were analyzed by the Fisher’s exact probability test, using StatView version 4.5 software (Abacus Concepts, Inc., 4. Roles of thyroid hormones, cortisol, growth California, USA). ** indicates a significant difference at P hormone, and environmental factors in the < 0.01 from the control group. Differences in mean plasma regulation of downstream and upstream swim- and pituitary hormone levels among experimental groups ming behaviors in salmonids were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent sig- During the period of smoltification and associated nificant differences at P < 0.05 among all groups. Reprinted downstream migration, salmonid smolts exhibit in- from General and Comparative Endocrinology, 122, creases in various types of hormones, such as thyrox- Munakata et al., The effects of testosterone on upstream ine (T ), triiodothyronine (T ), cortisol, GH, and prol- migratory behavior in masu salmon, Oncorhynchus masou, 4 3  actin (Ikuta et al. 1985; Young et al. 1989; Prunet et 329–340, 2001b, with permission from Elsevier. al. 1989; Hirano 1991; Nagae et al. 1994; Dickhoff et al. 1997; McCormick 2001). Since these hormones are involved in both physiological and morphological

Fig. 17. Frequency of quivering and attending behaviors in the sham-operated, castrated, and castrated + testosterone (T) 500 µg/fish-treated 1+ precocious male masu salmon. Differences in mean frequencies of quivering and attending behaviors among experimental groups were analyzed by one- way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted with permission from Fish. Sci., 68, Munakata et al., Sex steroids control migration of masu salmon, 49–52, Fig. 3,  2002, The Japanese Society of Fisheries Science. Fig. 18. Male masu salmon showing (a) attending and (b) quivering behaviors against 2+ mature females performing a series of female spawning behaviors. changes during smoltification, it is likely that these hormones also play some roles in the regulation of downstream swimming behavior, and perhaps in up- ment of some environmental factors, such as intra- stream swimming behavior. In previous studies, dras- specific interaction, day-night cycle, etc., in the regu- tic increases in plasma levels of T4 (i.e., T4 surge) were lation of downstream swimming, upstream swimming, discovered during the peak periods of smoltification and spawning behaviors will be presented. in several salmonids, such as coho and masu salmon (Grau et al. 1981; Yamauchi et al. 1984, 1985). Our 4-1. Roles of thyroid hormones in the downstream studies on masu and land-locked sockeye salmon also and upstream swimming behavior in masu demonstrated that plasma T4 levels in 1+ smolts were and land-locked sockeye salmon higher than those of T-treated 1+ smolts (Munakata et al. 2000b, 2001a, 2012a, 2012b) (Figs. 6–10, Table Grau et al. (1981) and Yamauchi et al. (1985) dis- 4). Furthermore, some 1+ mature and immature masu covered in coho and masu salmon smolts that plasma salmon parr exhibited considerably high plasma T4 lev- T4 levels increased rapidly and significantly (T4 surge) els during the upstream migratory period (Munakata during the peak periods of their smoltification. In our et al. 2001a, 2001b, 2012a, 2012b) (Figs. 11–16, Ta- studies, most 1+ masu and land-locked sockeye salmon bles 1–3). Considering these facts, thyroid hormone smolts exhibited high plasma thyroid hormone levels T4 has been recognized as an important factor involved during the downstream migratory period. In masu and in the downstream and upstream swimming behaviors. land-locked sockeye salmon, and rainbow trout, how- Plasma levels of cortisol and GH continually increase ever, plasma T4 and T3 levels in downstream migrants during the period of downstream migration (Prunet et tended to be lower than those of non-migrants that re- al. 1989; Nagae et al. 1994; McCormick 2001; Mizuno mained in the upper pond of the raceway (Ewing et al. et al. 2001; Zydlewski et al. 2005). The involvement 1994; Munakata et al. 2000b, 2001a, 2012a, 2012b) of cortisol and ovine GH (oGH) treatment on the oc- (see Figs. 6–10 and Table 4). Based on these observa- currence of downstream swimming behavior in 1+ and tions, it was hypothesized that plasma T4 and T3 levels 0+ masu salmon juveniles will be summarized in the decrease during or after the initiation of downstream following sections. swimming behavior. Environmental factors (i.e., inorganic and organic In our previous studies, however, the downstream (e.g., biological) factors) also appear to play an indis- migratory smolts were usually sampled every morn- pensable role in the occurrence of downstream swim- ing at 09:00, over several weeks or months (Munakata ming, upstream swimming, and spawning behaviors. et al. 2000b, 2001a, 2012b) (Figs. 6–10). Therefore, In the following section, an overview of the involve- there is a possibility that the sampling delay (i.e., sam-

Fig. 19. Frequency of female digging behavior in the control, testosterone (T) 500 µg/fish treated, and T 1000 µg/ fish treated groups. Differences in mean frequency of digging behavior among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fish- er’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted with permission from Fish. Sci., 68, Munakata et al., Sex steroids control migration of masu salmon, 49–52, Fig. 4,  2002, The Japa- nese Society of Fisheries Science. pling conducted considerably after the downstream swimming behavior) may have influenced the decrease in plasma T4 and T3 levels in downstream migrants. Fig. 20. Female masu salmon performing digging behavior. In the previous study, therefore, the downstream (a) 2+ mature females and (b) 1+ immature females treated migrants from the lower pond of the three-step raceway with testosterone (T) 500 µg/fish. (Fig. 21) were sampled within 60 minutes of the occurrence of migration (Table 4). To avoid the dispar- ity in the sampling dates between migrants and non- duced by the net trap (Yada et al. 2007) caused a de- µ migrants, six of the 1+ control smolts, T500 g/fish- crease in plasma T4 levels in downstream migrants. treated smolts, and precocious males were each sam- Thus in a separate study, so as to avoid captive stress pled from the middle pond on May 3 during the April from the net trap, we used a hand dip net to capture the 29–May 11 observation period (Munakata et al. 2012a). land-locked sockeye salmon downstream migrants di- As a result, there were no clear differences in plasma rectly from a separated area (2 × 4 × 0.5 m) in the lower thyroid hormone levels between downstream migrants pond (Fig. 5), a few hours after the onset of down- and non-migrants (Table 4). Ikuta (1994) reported that stream swimming behavior. After the downstream plasma T4 levels in 1+ land-locked sockeye salmon swimming behavior ceased, however, plasma T4 lev- smolts of downstream migrants differed: those that els of downstream migrants became lower than those were sampled immediately after the onset of down- of non-migrants in the control, T 500 µg/fish-treated, stream swimming behavior tended to have higher and the precocious male groups (Fig. 10). These re- plasma concentrations than those of non-migrants, sults indicate that plasma T4 levels in downstream mi- which were sampled simultaneously in the upper pond grants become lower than those of non- of the raceway. Therefore, one possible hypothesis is migrants, independent of the use of a net trap. that decreases in plasma thyroid hormone levels occur innately after the start of the downstream swimming 4-1B. Role of T4 and T3 in the upstream swimming behavior. behavior In ayu, Plecoglossus altivelis, an amphidromous fish, 4-1A. Effects of net trap on plasma thyroid changes it was found that plasma T4 levels of immature down- during the downstream migratory period in land- stream migrants became lower than in the initial lev- locked sockeye salmon els, likewise masu and land-locked sockeye salmon In previous studies (Munakata et al. 2000b, 2001a, smolt migrants, and that the levels of upstream migrants 2012a) (Figs. 6–9, Table 4), we regularly used a net became higher than in the initial plasma levels trap (2 × 0.7 × 0.7 m) to catch the downstream migrat- (Tsukamoto et al. 1988). In adult chum salmon, on the ing 1+ masu and land-locked sockeye salmon smolts. other hand, plasma levels of the thyroid hormones in Therefore, there is a possibility that captive stress in- upstream migrants were lower than those of migrants

Table 4. Frequency of upstream and downstream swimming behaviors and, plasma levels of testosterone (T), thyroxine (T4), ± µ and (g) triiodothyronine (T3) (mean SEM) in controls and T 500 g/fish-treated 1+ smolts and 1+ precocious male masu salmon. Differences in the frequency of upstream and downstream swimming behavior from the control group were analyzed by the χ2-test, using StatView version 4.5 software (Abacus Concepts, Inc., California, USA). *** indicates significant difference at P < 0.001, from the control group. Differences in mean plasma hormone levels among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted from Aquaculture, 362–363, Munakata et al., Involvement of sex steroids and thyroid hormones in upstream and downstream behaviors in masu salmon, Oncorhynchus masou, 158–166,  2012a, with permission from Elsevier.

which swim in the coastal sea during their upstream was demonstrated that T4-treated 0+ coho salmon parr migratory period (Ueda et al. 1984). Based on these tended to exhibit a higher frequency of downstream findings, it was suggested in some Pacific salmon that swimming behavior than the control parr (Munakata the plasma T4 and T3 levels decreased in association and Schreck, unpublished data). However, it was also with the progression of final gonadal maturation. In a previously reported that T4 itself does not induce the previous study, on the other hand, we used 1+ imma- occurrence of downstream swimming behavior in ture masu and land-locked sockeye salmon as a surro- salmonids (see a review by Iwata 1995). The role of gate for 2+ maturing fish in the upstream migratory the thyroid hormones in mediating the downstream and period (Munakata et al. 2001a, b). Consequently, upstream swimming behaviors clearly requires further plasma T4 and T3 levels in upstream migrants tended investigation. to be lower than those of non-migrants in 1+ masu and land-locked sockeye salmon (Munakata et al. 2001a, 4-2. Roles of cortisol and growth hormone in the 2001b, 2012a, 2012b) (Figs. 11–16, Tables 2, 3). downstream swimming behavior in masu Hence, it is suggested that the decrease in thyroid hor- salmon mones coincided with the occurrence of the upstream swimming behavior. However, the upstream migrants 4-2A. Cortisol were sampled every morning at 09:00 over several In anadromous salmonids, cortisol and GH have been weeks or months in our previous studies (Munakata et known to regulate the hypo-osmoregulatory ability, al. 2001a, b). Therefore, the sampling delay may have during the smoltification period (Hirano 1991; influenced the decrease in plasma thyroid hormone McCormick 2001). In both 1+ smolts and 0+ parr of levels in upstream migrants. In a separate experiment masu salmon, on the other hand, it was found that treat- (Munakata et al. 2012a; Table 4), the upstream mi- ment of cortisol, but not GH (oGH), caused the occur- grating T 500 µg/fish-treated 1+ masu salmon were rence of downstream swimming behavior in the sampled within 60 minutes of the occurrence of up- raceway (Munakata et al. 2007; Figs. 22, 23). The fre- stream swimming behavior from the upper pond of the quency of downstream swimming behavior in the cor- three-step raceway. It was found that there were no dif- tisol 2 mg/fish-treated group (72%) and oGH 250 µg/ ferences in plasma T4 and T3 levels between upstream fish + cortisol 2 mg/fish-treated group (82%) were sig- migrants and non-migrants. As a result, decreases in nificantly higher than those in the control (23%) and µ plasma T4 and T3 levels may initiate a few hours after oGH 250 g/fish-treated group (18%) (Fig. 22). The the occurrence of the upstream swimming behavior. plasma cortisol levels of migrants in the cortisol 2 mg/ To summarize these investigations, thyroid hor- fish-treated 1+ smolts (Fig. 22) were similar to those mones, such as T4 and T3, appear to play some roles in levels of naturally occurring 1+ smolts (Nagae et al. the regulation of downstream and upstream swimming 1994; Mizuno et al. 2001). behavior in some anadromous salmonids. Recently, it The results indicate that the downstream swimming

Fig. 22. Frequency of migrants and non-migrants, plasma levels of oGH and cortisol in the control, oGH 250 µg, cortisol 2 mg, and oGH 250 µg + cortisol 2 mg-treated 1+ masu salmon smolts. Figures beside columns indicate the number of migrants and non-migrants. Differences in the frequency of downstream behavior from the control group were analyzed by the χ2-test, using StatView version 4.5 software (Abacus Concepts, Inc., California, USA). *** indicates a significant difference at P < 0.001 from the control group. Differences in mean plasma hormone levels among experi- Fig. 21. Schematic drawing of three-step raceway that con- mental groups were analyzed by one-way analysis of vari- sists of upper, middle, and lower ponds. This raceway ena- ance (ANOVA) followed by Fisher’s PLSD. Differing let- bles us to quantify upstream and downstream swimming ters represent significant differences at P < 0.05 among all behaviors at the same period. After the experimental fish groups. Reprinted from General and Comparative had been reared in tanks, they were transferred to the mid- Endocrinology, 150, Munakata et al., Effects of growth hor- dle pond (4 × 2 × 0.5 m) of the raceway. The middle pond mone and cortisol on the downstream migratory behavior in was connected to the upper (4 × 2 × 0.5 m) and lower (4 × 2 masu salmon, Oncorhynchus masou, 12–17,  2007, with × 0.5 m) ponds through square holes (50 × 25 cm, thickness permission from Elsevier. 3 cm) which were made on wooden walls. Flow rate (volume) and velocity of the water in the raceway were 10 l/s and 75-85 cm/s. The upstream and downstream migrants are identified when the fish swam from the middle pond to the behavior in the cortisol 2 mg/fish-treated group (82%) µ net traps (2 × 0.7 × 0.7 m) that were located in the upper and and oGH 250 g/fish + cortisol 2 mg/fish-treated group lower ponds, respectively. Reprinted from Aquaculture (in (90%) were higher than those in the control (18%) and press), Munakata et al., Involvement of sex steroids and thy- oGH 250 µg/fish-treated group (0%) (Fig. 23). Based roid hormones in upstream and downstream behaviors in on these findings, it can also be hypothesized that cor- masu salmon, Oncorhynchus masou,  2012a, with permis- tisol induces the occurrence of downstream swimming sion from Elsevier. behavior not only in 1+ smolts but also in 0+ parr in masu salmon. A causal mechanism for how the increases of plasma cortisol levels induce the downstream swimming behavior (negative rheotaxis) is controlled competi- behavior in 1+ smolts and 0+ parr remains unclear. tively by both sex steroid hormones and smolt induc- Since only portions of 1+ masu salmon undergo the ing factors, such as cortisol. smoltification, one possible explanation is that the In 0+ parr, frequency of downstream swimming plasma cortisol levels increase innately as the

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. 54 A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 smoltification process advances. Alternatively, since various types of acute and chronic environmental stimuli, such as inter-and intra-specific interactions (Iwata 1996; Kagawa and Mugiya 2000; Kelsey et al. 2002), fasting (Varnavsky et al. 1995), exposure to low water (Pankhurst and Kraak 2000), fright responses from being chased by a netter (Nichols and Weisbart 1984; Yada et al. 2007), and water quality (Barton et al. 1987; Redding et al. 1987), increased circulating plasma cortisol levels within a few hours in teleosts, another explanation is that some environmental cues directly or indirectly enhance the additional secretion of cortisol into the plasma, and subsequently downstream swimming behavior is induced. Some studies indicated that inter-and intra-specific interactions, among larger and smaller individuals, may be an important factor that not only increased plasma cortisol levels in small subordinates (Kelsey et al. 2002), but also induced movement of smaller salmonid juveniles out of their focal foraging areas (Nakano et al. 1990; Nakano 1995). Alternatively, as mentioned in Subsection 2-6, it was found that all of the 40 1+ masu salmon smolts transferred into the upper pond of the raceway migrated down within a week (Munakata et al. 2000b; Fig. 9). Fig. 23. Frequency of migrants and non-migrants, plasma µ Based on this, it was hypothesized some other envi- levels of oGH and cortisol in the control, oGH 250 g, cortisol 2 mg, and oGH 250 µg + cortisol 2 mg-treated 0+ masu ronmental factors induce the plasma cortisol elevations salmon parr. Figures beside columns indicate the number of and subsequent downstream swimming behavior. Ei- migrants and non-migrants. Differences in the frequency of ther downstream or upstream swimming behavior oc- downstream swimming behavior from the control group were curs mainly during the evening, through the night, and analyzed by the χ2-test, using StatView version 4.5 software when it rains (Yamauchi et al. 1985; Munakata et al. (Abacus Concepts, Inc., California, USA). *** indicates a 2000b; Munakata et al. unpublished data; Fig. 24). Ac- significant difference at P < 0.001 from the control group. cordingly, environmental factors, such as photoperiod, Differences in mean plasma hormone levels among experi- temperature, flow rate, and water quality (e.g., turbid- mental groups were analyzed by one-way analysis of vari- ity), which exhibit diurnal fluctuations, affect the oc- ance (ANOVA) followed by Fisher’s PLSD. Differing let- currence of downstream swimming behavior through ters represent significant differences at P < 0.05 among all increases in plasma cortisol. groups. Reprinted from General and Comparative Endocrinology, 150, Munakata et al., Effects of growth hormone and cortisol on the downstream migratory behavior in 4-2B. GH masu salmon, Oncorhynchus masou, 12–17,  2007, with In anadromous salmonids, GH is considered to be permission from Elsevier. an important factor which regulates the hypo- osmoregulatory ability during smoltification (Hirano 1991; McCormick 2001). Previous studies also demonstrated that treatment of oGH as well as native GH downstream swimming behavior in masu salmon. influenced the physiological processes of smoltification Recently, Ojima et al. (2009) reported that and several types of behaviors such as salinity prefer- hypothalamic hormone growth hormone-releasing hor- ence, foraging, and anti-predator behaviors (Iwata et mone (GHRH) caused the downstream swimming al. 1990; Boeuf et al. 1994; Johnsson et al. 1996; behavior in 0+ chum salmon fry. The findings thus in- Jönsson et al. 1996; Yada et al. 1999). In our previous dicate that GHRH directory modulates the occurrence study, however, most oGH-treated 1+ smolt and 0+ parr of downstream movements in chum salmon fry. Fur- did not exhibit downstream swimming behavior dur- thermore, this also indicates that GH which is stimu- ing the downstream migratory period (Figs. 22, 23). lated by GHRH plays a role in the downstream swim- One possible explanation is that the treatment dose of ming behavior. Therefore, it is necessary to investi- oGH may have been insufficient, or the treatment pe- gate the effects of treatment dose and period of both riod may have been too short to affect the downstream oGH and native GH on downstream swimming behavior. On the other hand, it is also possible that GH, behavior. including oGH, is not involved in the occurrence of

including phenotypes that will become smolts (Nakano (a) 1995; Kiso 1995; Hutchison and Iwata 1998; Munakata et al. 2000b). In 1+ masu salmon, moreover, the BL and BW of non-migrants, especially the 1+ precocious males, tended to be larger than those of downstream migrants such as 1+ smolts (Munakata et al. 2000b; Table 5). Similarly, BL, BW, and CF in downstream migrants were smaller than those in non-migrants among T500 µg/fish-treated immature smolts (b) (Munakata et al. 2000b; Table 6). Although further investigation is needed, an acceptable rationale is that intra-specific interactions, which depend partly on their body size (growth), play roles in stimulating the downstream behavior in smaller fish.

4-3B. Roles of external factors in the upstream swimming behavior Fig. 24. (a) Number of fish that exhibited the downstream During the upstream migratory period, a significant migratory behavior in the 1+ masu salmon smolts (white portion of the sex steroid hormone-treated 1+ imma- column) and the T 500 µg/fish-treated smolts (dark column). ture masu salmon, and 2+ masu and land-locked (b) date of rainfall during experimental period in May. sockeye salmon moved upstream during and after dusk (Munakata et al. 2001a, 2001b, 2012a, 2012b) (Fig. 25). Thus, we must consider that upstream swimming behavior is triggered or regulated by some environ- 4-3. Roles of external factors in the downstream mental factors such as photoperiodicity. swimming, upstream swimming, and spawn- Because plasma sex steroid hormone levels of sex ing behaviors steroid-treated immature fish did not show apparent changes during the upstream migratory period 4-3A. Roles of external factors in downstream swim- (Munakata et al. 2001a, b; Figs. 11–16, Tables 2–4), ming behavior sex steroid hormones may play roles in the regulation A significant portion of 1+ masu salmon smolts ex- of upstream behavior as a “requirement” (Munakata hibit the downstream movement during favorable pe- and Kobayashi 2010) and sex steroid hormones may riods in the spring. Furthermore, it is believed that the modulate receptivity of fish from some environmental downstream migratory period in the southern regions stimulations. However, additional research will be re- tends to be earlier than that in the northern regions quired to validate this linkage between the physiologi- (Machidori and Kato 1984; Kato 1991; Kiso 1995). cal response and the environmental change. This trend clearly indicates that some external (environmental) factors such as photoperiod and tempera- 4-3C. Roles of external factors in the spawning ture, which show seasonal fluctuations and regional behavior differences, may be involved in the occurrence of When 2+ ovulated female masu salmon are reared downstream swimming behavior. As mentioned previ- continuously in artificial environments such as in ously, it is further indicated that some environmental hatchery fiber-reinforced plastic (FRP) tanks, these fish stimulations such as photoperiod and temperature, seldom display spawning behavior, including the dig- which exhibit diurnal patterns, may play roles in stimu- ging of redds. On the other hand, we discovered that lating this behavior. the 2+ ovulated and T 1000 µg/fish-treated 1+ imma- According to Yamauchi et al. (1985), 1+ masu salmon ture females that were transferred into the previously smolts exhibited downstream swimming behavior af- mentioned artificial stream chamber frequently dis- ter precipitation occurred. Our previous study also sug- played digging behaviors (Fig. 18). The chamber con- gested that the number of downstream migrants in the tained both gravel and flowing water, indicating that 1+ masu salmon smolts increased when it rained (Fig. physical environmental cues may be necessary before 24). Accordingly, it is indicated that rain, snow, and a particular spawning behavior is elicited. Satou et al. the concomitant increase in flow may trigger the oc- (1984) demonstrated that spawning behavior in male currence of downstream swimming behavior land-locked sockeye salmon adults was elicited by (Munakata et al. unpublished data). decoy fish which exhibit movements and oscillations As mentioned repeatedly, dominant precocious male resembling a mature female in spawning condition. parr frequently initiate aggressive behaviors (e.g., at- This example shows that visual stimulation or oscilla- tacking, nipping, chasing) towards subordinate fish tion from females can induce spawning behavior in

Table 5. Body length (BL), body weight (BW), and condition factor (CF) in the control, T5 µg/fish-, T 50 µg/fish-, and T 500 µg/fish-treated 1+ smolts and 1+ precocious male masu salmon. Differences in mean BL, BW, and CF among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted with permission from Zoological Science, 17, Munakata et al., Inhibitory effects of testosterone on downstream migratory behavior in masu salmon, Oncorhynchus masou, 863–870, Table 2,  2000b, Zoological Society of Japan.

males in the same manner as pheromones, such as L- in some juveniles, and these fish regularly exhibit kynurenine (Yambe et al. 2003, 2006). downstream swimming behavior along with the river currents, in association with the smoltification (Figs. 5. Conclusion and discussion 3, 26). In comparison, masu salmon that begin to sexually mature and consequently have high plasma sex 5-1. Roles of sex steroid hormones in the occur- steroid hormone levels in the sea (or lakes) tend to rence of migratory behaviors in masu salmon move against the river current (positive rheotaxis), and home upstream to their natal reach. This behavior is In masu salmon, it was demonstrated that sex ster- homologous with the upstream movement of preco- oid hormones, such as T inhibited the occurrence of cious non-migrants that live continuously in their na- smoltification and downstream swimming behavior tal rivers (Figs. 3, 26). Finally, masu salmon with high (negative rheotaxis), the initial phase in seaward mi- levels of plasma sex steroid hormones initiate the dis- gration (Aida et al. 1984; Ikuta et al. 1985, 1987; play of spawning behavior after they arrive at a poten- Munakata et al. 2000b, 2001a). On the other hand, tial spawning habitat. treatment with T induced the occurrence of upstream Because sex steroid hormones commonly regulate swimming behavior (positive rheotaxis), a component the occurrence of downstream swimming, upstream of upstream migration in 1+ immature parr and cas- swimming, and spawning behaviors, can it be consid- trated precocious males (Munakata et al. 2001a, b). ered that downstream (negative rheotaxis) and up- Furthermore, T commonly induced spawning behavior stream migratory behavior (positive rheotaxis) are in- in both sexes (Munakata et al. 2002). Therefore, it is volved as an obligatory phase before the spawning concluded that sex steroid hormones, such as T, regu- takes place? More specifically, these migratory late the occurrence of downstream and upstream swim- behaviors are involved as part of the spawning activi- ming behavior, in negative and positive rheotaxis fash- ties, which take place in the upper river. Because masu ions, and that sex steroid hormones negatively control salmon use different habitats (rivers and sea) during the occurrence of the seaward migration. their lifecycle, it may be that the inhibition of down- Based on these phenomena, we now understand why stream swimming behavior and the stimulation of up- mature masu salmon that have high levels of plasma stream swimming behavior by the sex steroid hormones sex steroid hormones live continuously in their natal are “biological endorsements” that orient sexually rivers as non-migratory forms (Figs. 3, 26). On the mature masu salmon to natal spawning areas, before other hand, delay of sexual maturation in the rivers the initiation of spawning behaviors. results from low levels of plasma sex steroid hormones While masu salmon non-migrants such as 1+ preco-

Table 6. Body length (BL), body weight (BW), and condition factor (CF) in the control (Raceway 1) and T500 µg/fish-treated smolts (Raceway 2) in masu salmon (Oncorhynchus masou). Differences in mean BL, BW, and CF among experimental groups were analyzed by one-way analysis of variance (ANOVA) followed by Fisher’s PLSD. Differing letters represent significant differences at P < 0.05 among all groups. Reprinted with permission from Zoological Science, 17, Munakata et al., Inhibitory effects of testosterone on downstream migratory behavior in masu salmon, Oncorhynchus masou, 863–870, Table 3,  2000b, Zoological Society of Japan.

cious males exhibit high plasma sex steroid hormone ferentiation from parr (non-migrants) to smolts (mi- levels and stay in their natal rivers, representative grants) is influenced by the percipiency of environ- downstream migratory forms, such as 1+ smolts regu- mental factors and concomitant cortisol elevations. larly exhibit increases in plasma T4, T3, cortisol, GH, According to Machidori and Kato (1984) and Kiso and prolactin levels during the period of smoltification (1995), it is clear that 0+ precocious males grow faster and through downstream migration (e.g., Dickhoff et than immature parr (e.g., smolt migrants) of the same al. 1997). Among these hormones, cortisol significantly age during 0+ summer, a half year prior to the induced the occurrence of downstream swimming smoltification period. Because growth and the follow- behavior in 1+ smolts (Munakata et al. 2007; Fig. 22). ing sexual maturation of non-migrants appear to be Based on these results, the initiation of seaward (down- supported partly by their active foraging behaviors and stream) migration in masu salmon seems to be con- territorial aggressiveness (e.g., Nakano 1995), it is trolled competitively by sex steroid hormones (sexual hypothesized that intra-specific interactions, such as maturation in rivers) and cortisol (metamorphosis of territorial aggressiveness and other concomitant phe- smoltification: preparation of marine life). nomena (e.g., hunger, delay of growth, etc.), play key Since cortisol is an smoltification-inducing factor roles in regulating the transformation from non- (Hirano 1991; McCormick 2001), it has been suggested migrants (parr) to migrants (smolts). As a result, non- that the levels innately increase with the progression migratory forms, such as 1+ precocious males, and of smoltification. On the other hand, it was inferred migrants, such as 1+ smolts, are eventually recipro- that socio-environmental factors, such as intra-, or cally balanced in some rivers (see Fig. 27). inter-specific interactions acting as stressors, cause acute and/or chronic plasma cortisol elevations in some 5-2. Sub-types of non-migratory and migratory teleosts (e.g., salmonids) (Schreck 2000). Previous in- forms vestigations also discovered that plasma cortisol levels in 1+ masu salmon smolts exhibited considerable In masu salmon, it has been noted that non-migrants fluctuations within one day and among different sam- such as precocious male parr and migrants such as 1+ pling dates in the Kesen River, in northern Honshu smolts appear in most rivers, and the two forms can be (Munakata et al. unpublished data). These facts sug- distinguished by their diagnostic characteristics, such gest that some environmental factors, such as tempera- as appearance and increasing plasma sex steroid hor- ture, cause or regulate the occurrence of smoltification mone levels (Fig. 4). In masu salmon, however, it has and subsequent downstream swimming behavior via also been recognized that there are phenotypes, which elevations of plasma cortisol levels. exhibit intermediate migratory patterns between rep- If smoltification and downstream swimming behavior resentative non-migrants and migrants (Kiso 1995) in 1+ masu salmon migrants are caused or regulated (Fig. 26, Table 7). For example, in a significant num- partly by some environmental factors via plasma cor- bers of rivers, there are so-called “immature parr non- tisol elevations, it is further hypothesized that the dif- migrants”—some females and a small number of males

Fig. 25. Eight-day running average of the number of upstream migrants in 2+ male and female sockeye salmon in the raceway. Dark and white bars indicate light dark-periods. This graph shows that the majority of 2+ land-locked sockeye salmon exhibit upstream migratory behavior before, during, and after dusk between 16:00 and 22:00. There was no clear difference between sexes. Reprinted with permission from Fish. Sci., 78, Munakata et al., Involvement of sex steroids, luteinizing hormone and thyroid hormones in upstream and downstream migratory behaviors in land-locked sockeye salmon Oncorhynchus nerka, 81– 90, Fig. 4,  2012b, The Japanese Society of Fisheries Science.

(Kiso 1995), that live continuously in their natal riv- more likely to reside in the middle to upper reaches of ers as do 1+ precocious males (Fig. 26, Table 7). Most rivers as do the precocious males and the immature of these 1+ non-migrants are considered to be imma- parr non-migrants, whereas “regressive smolts” display ture parr, based on their appearance (Kiso 1995) and more distinct downstream migration and will reach the plasma sex steroid hormone levels (Munakata et al. lower part of their natal rivers (Kiso 1995). unpublished data). Besides such sub-types of migratory forms, there are In masu salmon, moreover, there are some 1+ down- also other migratory forms considered to be “coastal stream migrants, which are identified as smolts by their smolt migrants” that will migrate near the coastal seas appearance, but do not travel along migratory routes between their natal rivers and the Sea of Okhotsk to the offshore seas, likewise the representative smolt (Machidori and Kato 1984; Kiso 1995). There is also migrants (Machidori and Kato 1984; Kiso 1995) (Fig. no clear difference in the appearance between the “re- 26). For instance, some of these migrants migrate fur- gressive smolts” and “coastal smolt migrants” (Kiso ther downstream than the 1+ precocious males and non- 1995). Regularly, however, body size (25 to 40 cm in migratory 1+ parr. However, the majority of these fish BL) of the “coastal migrants” during the upstream mi- do not enter the sea and instead stay in the mid through gratory period is larger and smaller than those of “re- lower part of their natal rivers from spring onward, gressive smolts” and “representative smolt migrants”, then move upward in the rivers through summer and respectively, mainly because of the short migration autumn (Fig. 26). These fish exhibit a silvery body (less than a year) period in the sea (Kiso 1995). color and low plasma sex steroid hormone levels as do In masu salmon, there is an additional type of non- representative smolt migrants, but their body size and migrants in which BL and BW are considerably lower CF values tend to be high (Kiso 1995). Therefore, these than those of other non-migrants and migrants. Accord- sub-types of migratory forms are referred to as “pseudo ing to Kiso (1995), such small sized non-migratory smolts” and “regressive smolts” (Kiso 1995). Although masu salmon can be considered to be a “poor growth there is no clear difference in the appearance between fish”, which do not differentiate into either precocious the “pseudo smolts” and “regressive smolts” during the parr or smolt migrants during the age of 1+, indicating downstream migratory period, “pseudo smolts” are that not only non-migrants but also smolts migrants

Hatching 1+ year 2+ year 3+ year 4+ year

Growth Maturation Upper river (1) Precocious parr (1) Growth Maturation (2) Immature parr (2) Parr non-migrants Growth Maturation (3) Upstream (7) Poor growth fish Less growth (7) migration Less growth type II Smoltificationf at Mid–lower river (3) Pseudo smolts

Downstreame Diversity (4) Regressive smolts migration Growth Maturationat (4)4) Rare?

Growth Maturationtu (5)5) Rare? Sea (5) Coastal smolt migrants Rare? Upstream Feeding migration migration Growth Maturation type I (6) Smolt migrants (6) Homing

Fig. 26. Diagrammatic representation of the lifecycle of masu salmon (Oncorhynchus masou) (modified the figure by Kiso 1995). There seems to be a sequential diversity in migratory patterns among representative non-migratory (precocious parr) and migratory (smolt) forms. In masu salmon, 1+ parr that occupy focal foraging areas (territory), exhibit precociously sexual maturation, and have high plasma sex steroid levels become representative non-migrants, precocious parr. In contrast, some of the 1+ immature fish that could not gain focal territories become representative migrants, smolt migrants. Their migratory patterns seem to be modulated in an inhibitory fashion by their maturity and/or growth performance (see text for details). Note that the “poor growth fish” which does not differentiate into either precocious parr or smolt migrants during the age of 1+ appears to differentiate into non-migratory parr or migratory smolts in another year, mainly during the age of 2+.

need to grow before they initiate the smoltification. tuitary hormones such as GnRH, follicle stimulating Generally, the “poor growth fish” seems to differenti- hormone (FSH), and LH, prior to the elevation in ate into non-migratory parr or migratory smolts in an- plasma sex steroid hormones to stimulate gonadal other year, mainly at age 2+ (Kiso 1995) (Fig. 26). maturation after the spring (e.g., Amano et al. 1998; Such variations, especially among migratory Munakata et al. 2000b). In masu salmon “pseudo behaviors, may be influenced heavily by environmen- smolts” and “regressive smolts”, it is also important to tal and physiological factors. In Japanese streams, the note that the GSI values and/or ovarian development proportions of “1+ immature parr non-migrants” tend stages were slightly higher than those in the 1+ smolt to increase in the southern regions when compared to migrants (Table 7). To account for these phenomena, the northern regions, which is also a trend observed it is hypothesized that some sex hormones other than for 1+ precocious males (Fig. 27). In Japan, therefore, the sex steroid hormones also play some roles in in- these two types of masu salmon are commonly called hibiting the occurrence of downstream swimming “yamame” in Japanese. Although the GSI values and behavior in immature non-migratory forms. plasma sex steroid hormone levels in 1+ immature parr It is generally thought that salmonids initiate gonadal non-migrants are considerably lower than those of the maturation after they have attained sufficient growth precocious males during spring (Kiso 1995; Munakata (Nordeng 1983; Kiso 1995). Therefore, it appears likely et al. unpublished data), the ovarian development stage that not only sex hormones but also other hormones in 1+ immature female non-migrants is the “yolk vesi- such as GH modulate the occurrence of downstream cle stage”, while most 1+ immature female smolt mi- swimming behavior and smoltification, depending on grants are in the “early peri-nucleolus stage” (Kiso their growth phase. Until now, however, such a hypoth- 1995) (Table 7). It is thus indicated that some of the esis is far from being established, and which physi- immature non-migrants progress gonadal maturation ological factors are actually involved in the occurrence in rivers, though their sex steroid hormone levels are of downstream swimming behaviors is not fully un- low. derstood. This topic clearly needs further investigation. The existence of such non-migrants indicates that stream residency of non-migrants is regulated not only 5-3. Variation of migratory behavior in salmonids by high plasma sex steroid hormone levels, but also by other physiological factors. In general, most In masu salmon, it becomes apparent that some salmonids exhibit increases in hypothalamic and pi- strains exhibit varieties of lifecycle between the rep-

(a) (b) River Sea River Sea

Anadromy Hokkaido

Amphidromy Honshu

Catadromy Kyushu Low Latitude High Late Spawning period Early Early Downstream migratory period Late

Fig. 27. Diagrammatic representations of (a) life histories of anadromous, amphidromous, and catadromous fish (modified from Gross 1987) and (b) differences in lifecycle of masu salmon (Oncorhynchus masou) among different regions. In (a) (e.g., northern hemisphere), anadromy and catadromy exceed in northern temperate and southern tropic areas, respectively, while amphidromy are frequent between these areas. In (b), proportion of non-migratory and migratory (e.g., smolts) forms increased in southern (e.g., Kyushu) and northern (e.g., Hokkaido) Japanese streams, respectively, maybe by the differences in productivity among different areas. Also, spawning and downstream migratory period in masu salmon become later in southern and northern Japan, respectively (Machidori and Kato 1984). resentative non-migratory and migratory forms. Also, 5-4. Driving force of migration from rivers to the it is speculated that these varieties of non-migratory sea—Why do salmonids migrate?— and migratory patterns are sequentially regulated by some physiological factors depending on the gonadal According to Gross (1987), it is thought that the maturation and/or growth stages, prior to and during migration (anadromy) of salmonids evolved in rela- the downstream migratory period (Table 7). tion to the availability of food (or productivity) be- Among the four salmonid genera, there also are va- tween the rivers and the sea (Gross 1987) (Fig. 27). rieties of non-migratory and migratory forms that re- Concisely, it is thought that productivity in the sea is semble those lifecycles observed in masu salmon higher than that in the rivers in the northern hemisphere (Groot and Margolis 1991; Thorpe 1994; Quinn 2005) regions, whereas contrasting patterns are found in the (Fig. 1). Considering the hypothesis that salmonids are southern tropic regions. This also indicates that there of a freshwater origin, it is acceptable to think that the are gradual variations in the productivities of the riv- evolutionally-ancient genera such as the genus Hucho ers and the sea along the latitude gradient within each and Salvelinus remain to show tendency depending on of the hemispheres. freshwater life, and spawn in the rivers as do the masu Actually, in masu salmon, proportions of “precocious salmon non-migrants, whereas evolutionally new gen- male non-migrants” and “immature non-migratory era, such as the genus Salmo and Oncorhynchus parr” are typically higher in the southern streams (i.e., evolved to rely more heavily on ocean life (seaward Kyushu) than those in northern regions (i.e., Hokkaido) migration), as do the masu salmon migrants (Figs. 3, (Machidori and Kato 1984) (Fig. 27). On the other 26). If the migratory patterns of masu salmon can be hand, the proportion of representative downstream considered as an “epitomization” of the variations of migratory smolts is higher in northern streams than in salmon migration, it may be hypothesized that the oc- the southern ones (Machidori and Kato 1984) (Fig. 27). currence of migratory behavior for a major part of the Considering a research hypothesis that the occurrence salmonids is also controlled by their gonadal matura- of non-migrants and migrants (smolts) are modulated tion and/or growing stages in the rivers, which will by their gonadal maturation and/or growth stages, require further investigations for validation. which are influenced by the productivity of the rivers,

Table 7. Appearance of frequency and migratory pattern (area) of non-migratory and migratory forms of masu salmon (Oncorhynchus masou) that inhabit streams along Sanriku coast in northern Honshu and its plasma sex steroid hormone levels, gonad somatic index (GSI), and gonadal development and growth stages prior to and during the period of downstream migration, based on Kiso (1995). Shades indicate the potential physiological factors that inhibit or modulate the occurrence of downstream migratory behavior.

Male Group Migratory status Appearance freq. Migratory areas Sex steroid hormone GSI Gonadal development Growth Precociou parr Non-migratory High Upper- middle river High High Sperm-formation stage Very good Immature parr non-migrants — Rare — Low Moderate Late multiplication stage — Pseudo smolts Migratory? —— —Low — Good Regressive smolts Migratory — Middle- lower river —— — — Coastal smolt migrants ——Coastal sea ——Early multiplication stage Moderate Representative smolt migrants — Moderate Sea of Okhotsk —— — — Poor growth fish Non-migratory Rare Upper- middle river —— — Poor

Female Group Migratory status Appearance freq. Migratory areas Sex steroid hormone GSI Gonadal development Growth Precociou parr Non-migratory Rare Upper- middle river Moderate High Oil drop stage Very good Immature parr non-migrants — Moderate — Low Moderate Yolk vesicle stage — Pseudo smolts Migratory? Rare ————Good Regressive smolts Migratory — Middle- lower river —— — — Coastal smolt migrants ——Coastal sea — Low Late peri- nucleolus stage Moderate Representative smolt migrants — High Sea of Okhotsk —— — — Poor growth fish Non-migratory Rare Upper- middle river ——Early peri- nucleolus stage Poor

such differences in the proportions of non-migrants and high productivity of the sea is considered to be less of migrants seem to be shaped by the gradual changes of a causal factor of the migration than a “refuge” for the productiveness in the rivers throughout different re- immature salmonids that perform the downstream mi- gions. gratory behavior. Again, Gross (1987) has indicated that the seaward migration of salmonids is induced evolutionally by the 5-5. Evolution of migratory behavior in salmonids differences in productivity between the rivers and the sea in high latitude areas of the northern hemisphere. Most of the juveniles of pink and chum salmon, At the mouth of the rivers, however, there is a clear which are considered evolutionarily new species, ex- boundary between the fresh and salt (sea) waters, and hibit long distance migration, about six months after it is still unclear as to why the ancestral form of their hatching (see Fig. 1). Therefore, it is suggested salmonids (freshwater origin) crossed over the osmotic that evolutionarily new species are more likely to de- boundary and discovered the favorable feeding envi- pend on marine life when compared to the older ronments (higher productivity) in the sea. salmonid species (Gross 1987). Based on the variations In this monograph, it was shown that the dominant of migratory behavior in salmonids, and the physiologi- non-migratory forms of masu salmon regularly occupy cal control mechanisms of masu salmon and some other focal foraging areas (see Subsection 2-3). Accordingly, species, however, these patterns in migratory behavior it is suggested that most of the downstream migratory are an extension of the variations of the salmon migra- behavior in migrants is caused by a reduction in food tions. Considering these phenomena, it is acceptable availability in the natal rivers, as shown by Gross to think that the majority of the maturing salmon con- (1987), but more specifically, one of the important and tinue to aspire to spawn in their natal rivers, at the end direct factors that cause the downstream migration from of their lifecycle. That is, the seaward migration may the favorable habitat is the intra-specific interactions be undertaken as a subsidiary choice in the lifecycle between the dominant non-migrants and subordinate of most salmonids. Previously, non-migratory forms migrants, depending on their stock density, or other of masu salmon had been considered to be a “land- environmental stressors, such as changes in tempera- locked population”, and worse still the precocious male ture, flow, water quality, and photoperiodicity. parr within the entire salmonid species were generally If that is the case, it is easier to understand why some considered as a “biological mistake”, according to of the migratory forms that could not stay in their fo- Gross (1987). Based on the migratory patterns and cal territories ultimately crossed the boundary between physiological control mechanism of migratory the rivers and the sea. Based on these phenomena, the behaviors in masu salmon, however, it is apparent that

doi:10.5047/absm.2012.00502.0029 © 2012 TERRAPUB, Tokyo. All rights reserved. 62 A. Munakata / Aqua-BioSci. Monogr. 5: 29–65, 2012 these types of fish are so-called “reversions” and they teaching and helping with the masu salmon sampling in the represent the ancestral forms of salmonids that matured Kesen River from 2004 through 2009. and spawned in their natal rivers. References 5-6. Conservation implications for masu salmon Aida K, Kato T, Awaji M. Effects of castration on the and their habitats smoltification of precocious male masu salmon Oncorhynchus masou. Nippon Suisan Gakkaishi 1984; 50: 565–571. 1) Numbers (biomass) of wild masu salmon—both Amano M, Aida K, Okumoto N, Hasegawa Y. Changes in “Yamame” and “Sakura masu” are continuously de- salmon GnRH and chicken GnRH-II contents in the brain creasing in a number of rivers (e.g., Kato 1991). and pituitary, and GTH II contents in the pituitary in fe- 2) Because the non-migrants and migrants diverge male masu salmon, Oncorhynchus masou, from hatching depending on their sexual maturity and growth stages, through ovulation. Zool. Sci. 1992; 9: 375–386. which are supported by the productivity in their natal Amano M, Aida K, Okumoto N, Hasegawa Y. Changes in rivers, both non-migrant (yamame) and migratory levels of GnRH in the brain and pituitary and GTH in the (sakura masu) forms can be increased by improving pituitary in female masu salmon, Oncorhynchus masou, the productivity of the rivers. In northern streams, car- from hatching to maturation. Fish. Physiol. Biochem. 1993; casses of salmonids which migrate back from the sea 11: 233–240. Amano M, Hyodo S, Kitamura S, Ikuta K, Suzuki Y, Urano are considered to be an important resource which plays A, Aida K. Salmon GnRH synthesis in the preoptic area a primary role in the increase of productivity in rivers and the ventral telencephalon is activated during gonadal (JoAnna and Richard 2006). Without consideration of maturation in female masu salmon. Gen. Comp. such implications, the stock management program of Endocrinol. 1995; 99: 13–21. masu salmon would become an insufficient exercise. Barton BA, Schreck CB. Influence of acclimation tempera- 3) Migratory behaviors seem to be regulated by both ture on internal and carbohydrate stress responses in ju- physiological and environmental factors. As a result, venile chinook salmon (Oncorhynchus tshawytscha). it is important to prevent artificial physiological and Aquaculture 1987; 62: 299–310. environmental disruptions that potentially become Berglund I, Lundqvist H, Fangstan H. Downstream migra- stressors and influence the migratory behaviors. For tion of immature salmon (Salmo salar) smolts blocked by example, we need to focus our attentions not only on implantation of the androgen 11-ketoandrostendione. Aquaculture 1994; 121: 269–276. the direct impacts of dam constructions, but also on Boeuf G. Salmonid smolting: a pre-adaptation to the oce- the concomitant modifications in flow, temperature, anic environment. In: Rankin GC and Jenson GB (eds.). and turbidity, which all have a natural cycle critical to Fish Ecophysiology. Chapman and Hall. 1994. fish. Boeuf G, LeBail PY, Prunet P. Growth hormone and thyroid 4) Since most of the 2+ masu salmon migrant smolts hormones during Atlantic salmon, Salmo salar L., that migrate from the sea will stay in the deeper areas smolting, and after transfer to seawater. Aquaculture 1989; of the mid reaches in their natal rivers throughout the 82: 257–268. summer, where they can avoid predation and sudden Boeuf G, Marc AM, Prunet P, Bail PYL, Smal J. Stimula- changes in water flow and temperature, the conservation of parr-smolt transformation by hormonal treatment tion of habitat diversity not only in upper reaches in Atlantic salmon (Salmo salar L.). Aquaculture 1994; (spawning ground), but also in entire rivers is essen- 121: 195–208. Dickhoff WW, Beckman BR, Larsen DA, Duan C, Moriyama tial. S. The role of growth in endocrine regulation of salmon smoltification. Fish Physiol. Biochem. 1997; 17: 231–236. Acknowledgments Ewing RD, Barratt D, Garlock D. Physiological changes re- This study was supported partly by research fellowships lated to migration tendency in rainbow trout from the Japan Society for the Promotion of Science and the (Oncorhynchus mykiss). Aquaculture 1994; 121: 277–287. Saito Houonkai Research Grant. I am grateful to Prof. Frantzen M, Johnsen HK, Mayer I. Gonadal development Katsumi Aida, for providing the opportunity to write this and sex steroids in a female Arctic charr brood stock. J. monograph. I thank Dr. Shoji Kitamura, Dr. Kazumasa Ikuta, Fish Biol. 1997; 51: 697–709. Dr. Masafumi Amano, Dr. Makito Kobayashi, Dr. Takashi Fujioka Y, Fushiki S, Tagawa M, Ogasawara T, Hirano T. Yada, Dr. Hidenobu Yambe, Dr. Carl Schreck, Dr. Hiram Li, Downstream migratory behavior and plasma thyroxine and Dr. David Noakes for their input and for their open dis- levels of Biwa salmon, Oncorhynchus rhodurus. Nippon cussion of many of these investigations. Mr. Toshio Shikama Suisan Gakkaishi 1990; 56: 1773–1779. and Mr. Hidefumi Nakamura helped in part of experiments. Giannico RG, Hinch SG. The effect of wood and tempera- I am also grateful to Dr. Hiram Li, Mr. Ralph Lampman, ture on juvenile coho salmon winter movement, growth, and Mr. Aalon Brock for reading the manuscript and provid- density and survival in side-channels. River Res. Applic. ing useful advice, and Dr. Tsukasa Fukushi, Dr. Nobuharu 2003; 19: 219–231. Goto, Dr. Kimiharu Ishizawa, Dr. Ryusaku Deguchi and Mr. Grau EG, Dickhoff WW, Nishioka RS, Bern HA, Folmar LC. Hiroki Suzuki for their conceptual support during the re- Lunar phasing of the thyroxine surge preparatory to sea- search. I would also like to thank Mr. Akira Shishido for ward migration of salmonid Fish. Sci. 1981; 211: 607–

609. auratus) to bluegills (Lepomis macrochirus) enhances Groot C, Margolis L (eds.). Pacific Salmon Life Histories. expression of stress protein 70 mRNA in the brains and UBC Press, British Columbia, Vancouver. 1991. increases plasma cortisol levels. Zool. Sci. 2000; 17: 1061– Gross M. Evolution of diadromy in fishes. Am. Fish. Soc. 1066. Symp. 1987; 1: 14–25. Kato F. Life histories of masu and amago salmon Heard RH. Life history of pink salmon (Oncorhynchus (Oncorhynchus masou and Oncorhynchus rhodurus). In: Gorbuscha). In: Groot C, Margolis L (eds.). Pacific Groot C, Margolis L (eds.). Pacific Salmon Life Histo- Salmon Life Histories. UBC Press, British Columbia, Van- ries. UBC Press, British Columbia, Vancouver. 1991; 448– couver. 1991; 119–230. 520. Hirano T. Endocrine control of osmoregulation in migratory Kelsey DA, Schreck CB, Congleton JL, Davis LE. Effects fishes. In: Mauchline J, Nemoto T (eds.). Marine Biol- of juvenile steelhead on juvenile chinook salmon behavior ogy: Its Accomplishment and Future Prospect. Hokusen- and physiology. Transactions of the American Fisheries sha, Japan. 1991, 3–14. Society 2002; 131: 676–689. Hoar WS. Smolt transformation: evolution, behavior, and Kiso K. Polymorphism of life form in masu salmon physiology. J. Fish. Res. Bd. Canada. 1976; 33: 1233– (Oncorhynchus masou) in the rivers of southern Sanriku 1252. District, Honshu, Japan. Bull. Inst. Zool. Academia Sinica Hoar WS. The physiology of smolting salmonids. In: Hoar 1990; 29(3): 27–39. WS, Randall DJ (eds.). Fish Physiology, Vol. II B. Aca- Kiso K. The life history of masu salmon Oncorhynchus demic Press, New York. 1988; 275–343. masou originated from rivers of the Pacific coast of north- Hutchison MJ, Iwata M. Effect of thyroxine on the decrease ern Honshu, Japan. Bull. Natl. Res. Inst. Fish. Sci. 1995; of aggressive behaviour of four salmonids during the parr- 7: 1–188 (in Japanese with English abstract). smolt transformation. Aquaculture 1998; 168: 169–175. Kiso K, Matsumiya Y. Growth of the fluviatile form masu Ikuta K. Effects of steroid hormones on migration of salmon Oncorhynchus masou in rivers of southern Sanriku salmonid fishes. Bull. Natl. Inst. Aquacult. Suppl. 1994; District, Honshu, Japan. Nippon Suisan Gakkaishi 1992; 2: 23–27. 58: 9–13. Ikuta K, Aida K, Okumoto N, Hanyu I. Effects of thyroxine Kobayashi M, Aida K, Hanyu I. Gonadotropin surge during and methyletestosterone on smoltification of masu salmon spawning in male goldfish. Gen. Comp. Endocrinol. 1986; (Oncorhynchus masou). Aquaculture 1985; 45: 289–303. 62: 70–79. Ikuta K, Aida K, Okumoto N, Hanyu I. Effects of sex ster- Kobayashi M, Aida K, Hanyu I. Hormone changes during oids on the smoltification of masu salmon, Oncorhynchus the ovulatory cycle in goldfish. Gen. Comp. Endocrinol. masou. Gen. Comp. Endocrinol. 1987; 65: 99–110. 1988; 69: 301–307. Iwata M. Downstream migratory behavior of salmonids and Liley NR, Fostier A, Breton B, Tan ES. Endocrine changes its relationship with cortisol and thyroid hormones: A re- associated with spawning behavior and social stimuli in a view. Aquaculture 1995; 135: 131–139. wild population of rainbow trout (Salmo gaidneri). Gen. Iwata M. Downstream migratory behaviors and endocrine Comp. Endocrinol. 1986; 62: 157–167. control of salmonid fishes. Bull. Natl. Res. Inst. Aquacult. Lou SW, Aida K, Hanyu I, Sakai K, Nomura M, Tanaka M, 1996; Suppl. 2: 17–21. Tazaki S. Endocrine profiles in the female of a twice- Iwata M, Yamauchi K, Nishioka RS, Lin R, Bern HA. Ef- annually spawning strain of rainbow trout. Aquaculture fects of thyroxine, growth hormone and cortisol on salin- 1984; 43: 13–22. ity preference of juvenile coho salmon (Oncorhynchus Machidori S, Kato F. Spawning populations and marine life kisutch). Mar. Behav. Physiol. 1990; 17: 191–201. of masu salmon Oncorhynchus masou. Int. North Pacific JoAnna LL, Richard WM. Influence of marine-derived nu- Fisheries Commission 1984; 43: 1–138. trients from spawning salmon on aquatic insect commu- Mayer I, Liley N, Borg B. Stimulation of spawning behavior nities in southeast Alaskan streams. Oikos 2006; 113(2): in castrated rainbow trout (Oncorhynchus mykiss) by 334–343. 17α,20β-dihydroxy-4-pregnene-3-one, but not by 11- Johnsson JI, Petersson E, Jönsson E, Björnsson BTh, Järvi ketoandrostendione. Hormones and Behavior 1994; 28: T. Domestication and growth hormone alter anti-predator 181–190. behavior and growth patterns in juvenile brown trout, McCormick SD. Endocrine control of osmoregulation in Salmo trutta. Can. J. Fish. Aquat. Sci. 1996; 53(7): 1546– teleost fish. American Zool. 2001; 41: 781–794. 1554. Miwa S, Inui Y. Inhibitory effects of of 17α-methyle testo- Jönsson E, Johnsson JI, Björnsson BTh. Growth hormone sterone and estradiol-17β on smoltification of sterilized increases predation exposure of rainbow trout. Proc. Roy. amago salmon (Oncorhynchus rhodurus). Aquaculture Soc. Lond. Ser. B. 1996; 263: 647–651. 1986; 53: 21–39. Kagawa H, Young G, Nagahama Y. Estradio-17β produc- Mizuno S, Ura K, Onodera Y, Fukada H, Misaka N, Hara A, tion in isolated amago salmon (Oncorhynchus rhodurus) Adachi S, Yamauchi K. Changes in transcript levels of ovarian follicles and its stimulation by gonadotropins. Gen. gill cortisol receptor during smoltification in wild masu Comp. Endocrinol. 1982a; 47: 361–365. salmon, Oncorhynchus masou. Zool. Sci. 2001; 18: 853– Kagawa H, Young G, Adachi S, Nagahama Y. Estradiol-17β 860. production in amago salmon (Oncorhynchus rhodurus) Munakata A, Kobayashi M. Endocrine control of sexual ovarian follicles: role of the thecal and granulosa cells. behavior in teleost fish. Gen. Com. Endocrinol. 2010; 165: Gen. Comp. Endocrinol. 1982b; 47: 440–448. 456–468. Kagawa N, Mugiya Y. Exposure of goldfish (Carassius Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. Growth

of wild honmasu salmon parr in a tributary of Lake species of stream-dwelling chars. Ecol. Res. 1994, 9: 9– ChuzenjiGrowth masu salmon, Oncorhynchus masou. 20. Fish. Sci. 1999; 65(6): 965–966. Nakano S, Kachi T, Nagoshi M. Restricted movement of the Munakata A, Björnsson BTh, Jönsson E, Amano M, Ikuta fluvial form of red-spotted masu salmon, Oncorhynchus K, Kitamura S, Kurokawa T, Aida K. Post-release adapta- masou rhodurus, in a mountain stream, central Japan. tion processes of hatchery-reared honmasu salmon parr. Japanese J. Ichthyol. 1990; 37(2): 158–163. J. Fish Biol. 2000a; 56: 163–172. Neave F. The origin and speciation of Oncorhynchus. Trans. Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. In- Roy. Soc. Canada 1958; 551: 25–39. hibitory effects of testosterone on downstream migratory Nichols DJ, Weisbart M. Plasma cortisol concentrations in behavior in masu salmon, Oncorhynchus masou. Zool. Sci. Atlantic salmon, Salmo salar: Episodic variations, diur- 2000b; 17: 863–870. nal changes, and short term response to adrenocortico- Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. The trophic hormone. Gen. Comp. Endocrinol. 1984; 56: 169– involvement of sex steroid hormones in downstream and 176. upstream migratory behavior of masu salmon. Comp. Norden CR. Comparative osteology of representative Biochem. Physiol. Part B 2001a; 129: 661–669. salmonid fishes, with particular reference to the grayling Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. The (Thymallusus arcticus) and its phylogeny. J. Fish. Res. effects of testosterone on upstream migratory behavior in Bd. Can. 1961, 18: 679–791. masu salmon, Oncorhynchus masou. Gen. Comp. Nordeng H. Solution to the “char problem” based on Arctic Endocrinol. 2001b; 122: 329–340. char (Salvelinus alpinus) in Norway. Can. J. Aquat. Sci. Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. Sex 1983; 40: 1372–1387. steroids control migration of masu salmon. Fish. Sci. 2002; Ojima D, Iwata M. Central administration of growth 68 Suppl. 1: 49–52. hormone-releasing hormone triggers downstream move- Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. Ef- ment and schooling behavior of chum salmon fects of growth hormone and cortisol on the downstream (Oncorhynchus keta) fry in an artificial stream. Comp. migratory behavior in masu salmon, Oncorhynchus masou. Biochem. Physiol. Part A. 2009; 152: 293–298. Gen. Comp. Endocrinol. 2007; 150: 12–17. Oshima M. Ecological study on the masu of the Taiko River. Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. In- Botany and Zoology 1936; 4: 1–13. volvement of sex steroids and thyroid hormones in up- Pankhurst NW, Kraak GVD. Evidence that acute stress in- stream and downstream behaviors in masu salmon, hibits ovarian steroidogenesis in rainbow trout in vivo, Oncorhynchus masou. Aquaculture 2012a; 362–363: 158– through the action of cortisol. Gen. Comp. Endocrinol. 166. 2000; 117: 225–237. Munakata A, Amano M, Ikuta K, Kitamura S, Aida K. In- Prat F, Sumpter JP, Tyler CR. Validation of radioimmu- volvement of sex steroids, luteinizing hormone and thy- noassay for two salmon gonadotropins (GTH I and GTH roid hormones in upstream and downstream migratory II) and their plasma concentrations throughout the repro- behaviors in land-locked sockeye salmon Oncorhynchus ductive cycle in male and female rainbow trout nerka. Fish. Sci. 2012b; 78: 81–90. (Oncorhynchus mykiss). Biol. Reproduction. 1996; 54: Murata S, Takasaki N, Saitoh M, Okada N. Determination 1375–1382. of the phylogenic relationships among Pacific salmonids Prunet P, Boeuf G, Bolton JP, Young G. Smoltification and by using short interspersed elements (SINEs) as temporal seawater adaptation in Atlantic salmon (Salmo salar): landmarks of evolution. Proc. Nat. Acad. Sci. USA. 1993; Plasma prolactin, growth hormone, and thyroid hormones. 90: 6995–6999. Gen. Comp. Endocrinol. 1989; 74: 355–364. Nagae M, Fuda H, Hara A, Saneyoshi M, Yamauchi K. Quinn, TP. The Behavior and Ecology of Pacific Salmon and Changes in serum concentrations of immunoglobulin M Trout. University of Washington Press, Seattle. 2005. 378

(IgM), cortisol and thyroxine (T4) during smoltification pp. in the masu salmon Oncorhynchus masou. Fish. Sci. 1994; Redding JM, Schreck CB, Everest FH. Physiological effects 60(2): 241–242. on coho salmon and steelhead of exposure to suspended Nagahama Y. Mechanism of gonadotropin control of ster- solids. Transactions American Fisheries Society 1987; 116: oidogenesis in teleost gonads. In: Gunma Symp. 737–744. Endocrinol. Center for Academic Publication, Tokyo, Ja- Sandercock FK. Life history of coho salmon (Oncorhynchus pan. 1984; 21: 167–182. kisutch). In: Groot C, Margolis L (eds.). Pacific Salmon Nagahama Y. Gonadotropin action on gametogenesis and Life Histories. UBC Press, British Columbia, Vancouver. steroidogenesis in teleost gonads. Zool. Sci. 1987a; 4(2): 1991; 396–445. 209–222. Sano S. Changes in masu salmon during the no-feeding sea- Nagahama Y. 17α,20β-Dihydroxy-4-pregnen-3-on a teleost son. Salmon J. 1947; 44: 9–14. maturation-inducing hormone. Develop. Growth and Dif- Sato A, Ueda H, Fukaya M, Kaeriyama M, Zohar Y, Urano fer. 1987b; 29(1): 1–12. A, Yamauchi K. Sexual differences in homing profiles and Nakano S. Individual differences in resource use, growth shortening of homing duration by gonadotropin-releasing and emigration under the influence of a dominance hier- hormone analog implantation in lacustrine sockeye salmon archy in fluvial red-spotted masu salmon in a natural habi- (Oncorhynchus nerka) in Lake Shikotsu. Zool. Sci. 1997; tat. J. Animal Ecol. 1995; 64: 75–84. 14: 1009–1014. Nakano S, Furukawa-Tanaka T. Intra- and interspecific domi- Satou M, Oka Y, Kusunoki M, Matsushima T, Kato M, Fujita nance hyerarchy and variation in foraging tactics of two I, Ueda K. Telencephalic and preoptic areas integrate

sexual behavior in hime salmon (landlocked red salmon, seawater transfer and fasting on plasma growth hormone Oncorhynchus nerka): results of electrical brain stimula- levels of yearling coho salmon (Oncorhynchus kisutch) tion experiments. Physiol. Behav. 1984; 33: 441–447. parr. Aquaculture 1995; 135: 141–145. Schreck CB. Accumulation and long-term effects of stress Yada T, Nagae M, Moriyama S, Azuma T. Effects of prolac- in fish. In: Moberg G, Mench J (eds.). The Biology of Ani- tin and growth hormone on plasma immunoglobulin M mal Stress. C.A.B. International Press, Wallingford, U.K. levels of hypophysectomized rainbow trout, Oncorhynchus 2000; 147–158. mykiss. Gen. Comp. Endocrinol. 1999; 115: 46–52. Slater CH, Schreck CB, Swanson P. Plasma profiles of the Yada T, Azuma T, Hyodo S, Hirano T, Grau EG, Schreck sex steroids and gonadotropins in maturing female spring CB. Differential expression of corticosteroid receptor chinook salmon (Oncorhynchus tshawytscha). Comp. genes in rainbow trout (Oncorhynchus mykiss) immune Biochem. Physiol. 1994; 109A: 167–175. system in response to acute stress. Can. J. Fish Aqua Sci. Swanson P. Salmon gonadotropins: Reconciling old and new 2007; 64(10): 1382–1389. ideas. In: Scott et al. (eds.). Proc. Fourth Int. Symp Re- Yamauchi K, Koide N, Adachi S, Nagahama Y. Changes in productive Physiology of Fish. University of East Anglia, seawater adaptability and blood thyroxine concentrations Norwich. 1991; 2–7. during smoltification of the masu salmon, Oncorhynchus Thorpe JE. An alternative view of smolting in salmonid. masou, and the amago salmon, Oncorhynchus rhodurus. Aquaculture 1994; 121: 105–113. Aquaculture 1984; 42: 247–256. Truscott B, Idler DR, So YP, Walsh JM. Maturation steroids Yamauchi K, Ban M, Kasahara N, Izumi T, Kojima H, Harako and gonadotropin in upstream migratory sockeye salmon. T. Physiological and behavioral changes occurring during Gen. Comp. Endocrinol. 1986; 62: 99–110. smoltification in the masu salmon, Oncorhynchus masou. Tsukamoto K, Aida K, Otake T. Plasma thyroxine concen- Aquaculture 1985; 45: 227–235. tration and upstream migratory behavior of juvenile ayu. Yambe H, Munakata A, Kitamura S, Aida K, Fusetani N. Nippon Suisan Gakkaishi 1988; 54(10): 1687–1693. Methyltestosterone induces male sensitivity to both primer Ueda H, Hiroi O, Hara A, Yamauchi K, Nagahama Y. and releaser pheromones in the urine of ovulated female Changes in serum concentrations of steroid hormones, masu salmon. Fish Physiol. Biochem. 2003; 28: 279–280. thyroxine, and vitellogenine during spawning migration Yambe H, Kitamura S, Kamio M, Yamada M, Matsunaga S, of the chum salmon, Oncorhynchus keta. Gen. Comp. Fusetani N. L-Kynurenine, and amino acid identified as a Endocrinol. 1984; 53: 203–211. sex pheromone in the urine of ovulated female masu Utoh H. Study of the mechanism of differentiation between salmon. Proc. Natl. Acad. Sci. USA 2006; 103: 15370– the stream resident form and the seaward migratory form 15374. in masu salmon, Oncorhynchus masou Brevoort, I. Growth Young G, Björnsson BTh, Prunet, P, Lin, RJ, Bern, HA. and gonadal maturity of precocious masu salmon parr. Bull. Smoltification and seawater adaptation in coho salmon Fac. Fish. Hokkaido Univ. 1976; 26: 321–326 (in Japa- (Oncorhynchus kisutch): Plasma prolactin, growth hor- nese). mone, thyroid hormones, and cortisol. Gen. Comp. Utoh H. Study of the mechanism of differentiation between Endocrinol. 1989; 74: 335–345. the stream resident form and the seaward migratory form Zydlewski GB, Haro A, McCormick D. Evidence for cumu- in masu salmon, Oncorhynchus masou Brevoort, II. lative temperature as an initiation and terminating factor Growth and sexual maturity of precocious masu salmon in downstream migratory behavior of Atlantic salmon parr (2). Bull. Fac. Fish. Hokkaido Univ. 1977; 28: 66–73 (Salmo salar) smolts. Can. J. Fish. Aquat. Sci. 2005; 62: (in Japanese). 68–78. Varnavsky VS, Sakamoto T, Hirano T. Effects of premature