INTRODUCTION
Grey mullet(Mugil cephalus Linnaeus, 1758) are dis-tributed circumglobally between 42° N and 42° S, and are an economically important species for both com-mercial fisheries and aquaculture because their roes have a high culinary value (Thomson 1966, Nash & Shehadeh 1980). The adult mullet migrates annually from the coastal waters of mainland China to the off-shore waters of SW Taiwan to spawn during the winter (December through January) when 3 to 4 yr old (Tung
1981, Chen & Su 1986, Huang & Su 1989). Their eggs and larvae drift with the coastal current to estuaries along the SW to NE coasts of Taiwan where they become juveniles at the age of 1 to 2 mo post-hatching (Tung 1981, Chang & Tzeng 2000, Chang et al. 2000). Current knowledge of the migratory history of the mullet has been interpreted fragmentally from the spa-tio-temporal distribution of the migrating spawners and estuarine-recruited juveniles. The lifetime migra-tory hismigra-tory, especially between estuarine arrival and spawning migration, is not completely understood.
© Inter-Research 2004 · www.int-res.com *Corresponding author. Email: wnt@ccms.ntu.edu.tw
Migratory environmental history of the grey mullet
Mugil cephalus as revealed by otolith Sr:Ca ratios
C. W. Chang
1, Y. Iizuka
2, W. N. Tzeng
1, 3,*
1Institute of Zoology, and 3Institute of Fisheries Sciences, College of Life Science, National Taiwan University, Taipei,
Taiwan 106, ROC
2Institute of Earth Sciences, Academia Sinica, Nankang, Taipei, Taiwan 115, ROC
ABSTRACT: We used an electron probe microanalyzer (EPMA) to determine the migratory environ-mental history of the catadromous grey mullet Mugil cephalus from the Sr:Ca ratios in otoliths of 10 newly recruited juveniles collected from estuaries and 30 adults collected from estuaries, nearshore (coastal waters and bay) and offshore, in the adjacent waters off Taiwan. Mean (± SD) Sr:Ca ratios at the edges of adult otoliths increased significantly from 6.5 ± 0.9 × 10– 3in estuaries and
nearshore waters to 8.9 ± 1.4 × 10– 3in offshore waters (p < 0.01), corresponding to increasing
ambi-ent salinity from estuaries and nearshore to offshore waters. The mean Sr:Ca ratios decreased sig-nificantly from the core (11.2 ± 1.2 × 10– 3) to the otolith edge (6.2 ± 1.4 × 10– 3) in juvenile otoliths (p <
0.001). The mullet generally spawned offshore and recruited to the estuary at the juvenile stage; therefore, these data support the use of Sr:Ca ratios in otoliths to reconstruct the past salinity history of the mullet. A life-history scan of the otolith Sr:Ca ratios indicated that the migratory environmen-tal history of the mullet beyond the juvenile stage consists of 2 types. In Type 1 mullet, Sr:Ca ratios range between 4.0 × 10– 3and 13.9 × 10– 3, indicating that they migrated between estuary and offshore
waters but rarely entered the freshwater habitat. In Type 2 mullet, the Sr:Ca ratios decreased to a minimum value of 0.4 × 10– 3, indicating that the mullet migrated to a freshwater habitat. Most mullet
beyond the juvenile stage migrated from estuary to offshore waters, but a few mullet less than 2 yr old may have migrated into a freshwater habitat. Most mullet collected nearshore and offshore were of Type 1, while those collected from the estuaries were a mixture of Types 1 and 2. The mullet spawning stock consisted mainly of Type 1 fish. The growth rates of the mullet were similar for Types 1 and 2. The migratory patterns of the mullet were more divergent than indicated by previous reports of their catadromous behavior.
KEY WORDS: Mugil cephalus · Otolith · Sr:Ca ratio · Migration · Environmental history
Tung (1981) speculated that all the juveniles and adults after spawning might migrate northward to the coastal waters of mainland China to feed until the spawning migration. Liu (1986) postulated that a few juveniles did not migrate northward but resided in the coastal waters around Taiwan. Chen et al. (1989) and Lee (1992) also found immature mullet of various sizes in estuaries and bays of Taiwan. The mullet was also believed to be catadromous, migrating from estuarine brackishwaters to high salinities offshore for breeding (Breder & Rosen 1966, De Silva 1980, Torricelli et al. 1982). However, the definition of catadromy for the mullet remains controversial because its use of fresh-water habitats is not clear (Blaber 1987, McDowall 1988). Thus, the determination of whether the mullet enters freshwater before spawning is critical to clarify the extent of its catadromy.
The fish otolith is composed of calcium carbonate deposited rhythmically as aragonite crystals in a pro-tein matrix (Degens et al. 1969, Pannella 1971). Except for their physiological functions in balance and hear-ing, fish otoliths also function as a chronological and environmental recorder of the fishes’ past life history (Campana & Neilson 1985, Campana 1999). The ratios of strontium (Sr) to calcium (Ca) in otoliths are posi-tively correlated to ambient salinities (Secor et al. 1995, Tzeng 1996, Kawakami et al. 1998). The temporal
changes of Sr:Ca ratios in otoliths have been widely applied to determine the migratory environmental histories of diadromous fishes between freshwater and marine environments, including anadromous salmonids (Kalish 1990, Howland et al. 2001, Limburg et al. 2001), striped bass (Secor 1992, Secor & Piccoli 1996) and catadromous freshwater eels (Tzeng 1994, 1995, 1996, Tzeng & Tsai 1994, Tzeng et al. 1997, 1999, 2000, 2002a, 2003, Tsukamoto & Arai 2001, Jessop et al. 2002). Accordingly, the analysis of Sr:Ca ratios in otoliths, in combination with age data, makes possible the reconstruction of the migratory environmental his-tory of the mullet and the clarification of its freshwater habitat use.
The aims of the present study were: (1) to validate whether otolith Sr:Ca ratios could be used to determine retrospectively the migratory environmental history of the mullet, (2) to elucidate its lifetime migratory envi-ronmental history, particularly freshwater habitat use, and (3) to reconsider the definition of catadromy as applied to the mullet.
MATERIALS AND METHODS
Sampling design. We collected 10 newly recruited juvenile mullet from the estuaries of Shuang Creek (S), Ta-an Creek (TA) and Szu-chung Creek (SC) on the northern and western coasts of Taiwan from December 1998 through February 1999 (Fig. 1, Table 1).
In addition, 30 adults were collected from the estuar-ies, nearshore (coastal waters and bay) and offshore across the dispersal range of the fish (Fig. 1). Of these, 15 were collected from the estuaries of I-lan River (IL), Gong-shy-tyan Creek (GST), Ta-tu River (TT) and Hsin-hu-wei Creek (HHW); 3 were taken from
Site, Fork length Weight
Sampling date (mm) (g)
Shuang Creek (S)
Dec 23, 1998 25.32 0.13
Jan 24, 1999 30.31 0.27
Feb 22, 1999 32.98 0.33
Ta-an Creek (TA)
Dec 25, 1998 31.40 0.25 Jan 25, 1999 31.87 0.25 Feb 25, 1999 35.79 0.47 Feb 25, 1999 33.76 0.36 Szu-chung Creek (SC) Dec 11, 1998 24.83 0.14 Jan 17, 1999 28.66 0.20 Feb 9, 1999 24.09 0.12
Table 1. Mugil cephalus. Biological characteristics of
juve-niles used for otolith Sr:Ca ratio analysis
Fig. 1. Mugil cephalus. Sampling sites of grey mullet.
Estu-ary = IL: I-lan River; S: Shuang Creek; GST: Gong-shy-tyan Creek; TA: Ta-an Creek; TT: Ta-tu River; HHW: Hsin-hu-wei Creek; SC: Szu-chung Creek. Nearshore = MT: Ma-tzu; TPW: Ta-peng-wan Bay. Offshore = TS: Tan-shui; WC: Wu-chi; KS:
Ta-peng-wan bay (TPW) on the NE and western coasts of Taiwan; 3 were collected from coastal waters influ-enced by freshwater runoff off Ma-tzu (MT) on the SE coast of mainland China; and 9 were collected during their spawning migration offshore of Tan-shui (TS), Wu-chi (WC) and Kan-shan (KS) on the western coast of Taiwan. Sampling sites and dates, sex, length, weight, gonadosomatic index and age of the mullet are given in Table 2.
Salinities at the sampling sites were obtained from Su & Jane (1974), Anonymous (1998), Cheng et al. (1990) and Tzeng et al. (2002b).
Otolith preparation and microprobe analysis. Sagit-tal otoliths of the mullet were used for the measure-ment of Sr and Ca concentrations. The otoliths were extracted, cleaned, air-dried and embedded in epoxy resin (Epofix, Struers). The lateral view of the mullet otolith is oval in shape in the juvenile stage but becomes oblong in the adult stage. The otolith grows fastest along the antero-posterior axis, moderately
along the dorso-ventral axis and slowest along the disto-proximal axis. The otoliths were polished along the sagittal plane for juveniles, but along the frontal plane for adults because the adult otolith curves and elongates dorsally and is concave distally and convex proximally (Fig. 2a). Each otolith was polished using a Metaserv grinder-polisher (Buehler) equipped with various grit sandpapers and 0.05 µm alumina slurry and polishing cloth. After polishing, the otoliths were rinsed with deionized water, air-dried and coated with carbon.
The concentrations of Sr and Ca in the otoliths were measured with an electron probe microanalyzer, EPMA (JXA-8900R, JEOL) from the primordium to the otolith edge along the posterior maximum growth axis. The beam condition of the EPMA was similar to that used by Tzeng et al. (2002a), but the beam currents, beam sizes, intervals and peak counting times were differed for juveniles (3 nA, 5 µm, 10 µm and 120 s, respectively, for Sr) and adults (5 nA, 10 µm, 20 µm and
Code Site Sampling date Sex FL (mm) Wt (g) GSI Age (yr)
Estuary E1 IL May 24, 2000 – 207 109 < 0.01 1+ E2 IL May 24, 2000 – 292 324 < 0.01 2+ E3 IL Nov 23, 2000 m 297 296 2.80 1+ E4 IL Nov 23, 2000 f 393 549 15.16 4+ E5 GST Dec 29, 1997 – 235 168 < 0.01 1+ E6 GST Jun 23, 1998 – 237 149 < 0.01 1+ E7 GST Oct 30, 1997 – 376 730 0.22 3+ E8 GST Oct 8, 1998 f 423 940 0.34 4+ E9 TT Feb 27, 1998 m 268 270 < 0.01 1+ E10 TT Feb 27, 1998 m 292 364 0.03 2+ E11 TT Feb 27, 1998 m 333 485 < 0.01 2+ E12 TT Feb 27, 1998 f 376 626 0.28 3+ E13 HHW May 15, 2001 – 72 4 < 0.01 0+ E14 HHW May 15, 2001 – 91 8 < 0.01 0+ E15 HHW May 15, 2001 – 269 216 < 0.01 1+ Nearshore N1 MT Apr 19, 1998 f 424 934 0.48 3+ N2 MT Apr 19, 1998 f 468 1147 1.29 4+ N3 MT Apr 19, 1998 f 471 1061 0.96 5+ N4 TPW Dec 15, 1997 m 260 215 0.02 1+ N5 TPW Dec 15, 1997 m 280 308 0.02 1+ N6 TPW Dec 15, 1997 m 440 1120 0.03 5+ Offshore O1 TS Dec 16, 1997 m 421 1000 0.32 4+ O2 TS Dec 16, 1997 f 422 945 11.57 4+ O3 TS Dec 16, 1997 f 446 1200 0.93 4+ O4 WC Dec 25, 1997 f 453 1198 11.82 4+ O5 WC Dec 25, 1997 f 488 1321 14.09 4+ O6 WC Dec 25, 1997 f 491 1466 12.58 5+ O7 KS Jan 12, 1997 f 448 1254 20.08 4+ O8 KS Jan 12, 1997 m 481 1309 8.70 5+ O9 KS Jan 12, 1997 f 500 1525 2.45 5+
Table 2. Mugil cephalus. Biological characteristics of adults used for otolith Sr:Ca ratio analysis. Sampling sites abbreviated as in
90 s, respectively, for Sr). Strontianite (SrCO3; National
Museum of Natural History [NMNH] R10065) and cal-cite (CaCO3; NMNH 136321) were used as standards
for the calibration of Sr and Ca concentrations in otoliths (Jarosewich & White 1987). The quantitative data were calculated by the ZAF method (Z: atomic number; A: absorption; F: fluorescence correction; Philibert & Tixier 1968) and Ca was used to normalize the concentration of Sr. The temporal changes in Sr:Ca ratios in the otoliths were similar with the same beam condition (e.g. see Mullet Nos. E11, N2, N6 and O6 in Fig. 5) and similar with different beam conditions for the same otolith (e.g. see E15 in Fig. 5 and E13 and 14
in Fig. 6). This indicates that the measurements of Sr:Ca ratios in the mullet otoliths were stable and repeatable.
After electron probe analysis, the otoliths were repolished to remove the carbon coating and etched with 5% ethylenediaminetetraacetate (EDTA) to reveal annulus marks for age determination. The otolith Sr:Ca ratios were correlated with life history events includ-ing estuarine arrival and annulus deposition.
Data analysis. The migratory environmental histo-ries of the adult mullet were classified into 2 types according to age and the temporal changes in the Sr:Ca ratios in the otoliths. Differences in frequency
Fig. 2. Mugil cephalus. Frontal section of otolith photographed with reflected light. (a) Whole otolith; (b) estuarine mark
distributions among habitats and among migration patterns for the different mullet environmental history types were examined by a G-test (Siegel 1956). Differ-ences in otolith mean radius and Sr:Ca ratios among life stages, age groups and salinity habitations were tested by analysis of variance (ANOVA) and a Tukey multiple-comparison test at a significance level of α = 0.05 (Winer 1971).
In addition, the growth curve of the mullet was fitted with the von Bertalanffy growth equation (Ricker 1958). The mean lengths at annulus formation were compared between migratory types by a t-test.
RESULTS
Otolith microstructure
Estuarine checks and annuli in otoliths from adult mullet were discernible with reflected light after EDTA etching (Fig. 2b,c). The mean (± SD) radius of the estu-arine check in adult mullet otoliths was 435.4 ± 76.8 µm (range 318.6–595.0 µm). The radius of the estuarine check in adult otoliths was close to the mean (± SD) otolith radius of newly recruited juveniles (387.0 ± 73.9 µm, range 280.0–510.0 µm, Table 3). Thus, the estuarine check in otoliths of the adult mullet are deposited during their estuarine arrival at the juvenile stage and can be used as a marker to identify the estuarine arrival of the adult.
Otolith Sr:Ca ratios in relation to habitat salinity
The Sr:Ca transect pattern indicated that Sr:Ca ratios in the otoliths of juvenile mullet significantly de-creased from the core to the otolith edge (r = –0.73, p < 0.001, Fig. 3). The mean (± SD) Sr:Ca ratios in the
core of the otolith (11.2 × 10– 3± 1.2 × 10– 3) were
signif-icantly greater than that in otolith edge (6.2 × 10– 3 ±
1.4× 10– 3, p < 0.001, Table 3). The decrease in Sr:Ca
ratios from core to edge of the otolith was consistent with larval dispersal from the high-salinity offshore spawning ground to the low-salinity estuarine nursery ground.
The Sr:Ca ratios at the estuarine check of the adult mullet otoliths averaged 5.1 × 10– 3± 1.0 × 10– 3, did not
differ significantly among sampling sites, and were similar to those at the juvenile otolith edge (6.2 × 10– 3±
1.4 × 10– 3; Table 3). This indicates that the adult mullet
probably migrated to a low-salinity estuary at the juve-nile stage.
The mean salinity in the sampling sites changed from 23.5 ‰ (range from 1.0 to 34.9 ‰) in the estuaries and 23.9 ‰ (14.3–33.5 ‰) in the nearshore coastal zone to 34.4 ‰ (34.0–34.8 ‰) in the offshore (Fig. 4a). Maxi-mum salinities were similar among sampling sites, but their mean values increased from estuary and nearshore to offshore. Similarly, the mean (± SD) Sr:Ca
Otolith n Otolith radius (µm) Sr:Ca ratio × 10– 3
area Range Mean ± SD CV HG Range Mean ± SD CV HG
Juveniles Core 10 11.3–16.3 14.5 ± 1.6 10.7 a 9.4–13.5 11.2 ± 1.2 10.5 d Edge 10 280.0–510.0 387.0 ± 73.9 19.1 b 3.9–8.00 6.2 ± 1.4 22.9 a Adults Check 30 318.6–595.0 435.4 ± 76.8 17.6 b 2.9–7.00 5.1 ± 1.0 20.3 a A1 28 2840.0–4030.0 3421.1 ± 349.5 10.2 c 3.6–10.7 7.9 ± 1.8 23.2 b A2 20 3680.0–4760.0 4283.0 ± 283.6 6.6 d 5.3–11.2 8.6 ± 1.6 18.9 bc A3 17 4260.0–5220.0 4927.1 ± 266.8 5.4 e 7.7–13.9 10.0 ± 1.5 15.2 cd A4 14 4600.0–5900.0 5383.6 ± 331.5 6.2 f 5.7–12.4 9.7 ± 1.7 17.8 cd A5 5 5600.0–6280.0 5864.0 ± 259.8 4.4 g 8.1–11.7 9.8 ± 1.4 14.0 bcd
Table 3. Mugil cephalus. Mean (± SD) radius and Sr:Ca ratios at otolith core and edge for juveniles and of estuarine check and
an-nulus (A1– 5) of otoliths for adults. HG: homogeneous group; same letters indicate that mean values were homogeneous among areas
Fig. 3. Mugil cephalus. Changes in mean otolith Sr:Ca ratios
ratios at the otolith edge of the adult mullet of ≥3 yr old significantly increased from 6.5 × 10– 3 ± 0.8 × 10– 3
(sample size, n = 4) in the estuaries (Sites IL, GST, TT, HHW) and 6.5 × 10– 3± 1.2 × 10– 3(n = 4) nearshore (Bay
Site TPW and Coastal Site MT) to 8.9 × 10– 3± 1.4 × 10– 3
(n = 9) offshore (Sites TS, WC, KS) (p < 0.01, Fig. 4b). The correspondence between the Sr:Ca ratios at the edge of adult mullet otoliths and the ambient salinity between estuary and offshore sites indicates that the use of Sr:Ca ratios in mullet otoliths to reconstruct their migratory salinity history is reliable.
The upper and lower limits of the 95% confidence interval (CI) of the mean (± SD) Sr:Ca ratios of 5.1 × 10– 3 ± 1.0 × 10– 3 at the estuarine check of the adult
mullet otoliths were 3 × 10– 3and 7 × 10– 3, respectively,
and were used as criteria to define the migratory envi-ronmental history of the mullet between freshwater and seawater. In other words, the fish was expected to have migrated to the high-salinity offshore sites if the otolith Sr:Ca ratios were greater than the upper limit of the CI, to have migrated to freshwater sites if the ratios were less than the lower limit of the CI, and to have migrated to brackish estuaries if the ratios fluctuated between the upper and lower limits of the CI.
Temporal changes in Sr:Ca ratios in adult otoliths
The migratory environmental history of the mullet was divided into 2 types based on temporal changes in otolith Sr:Ca ratios:
Type 1. Most Sr:Ca ratios beyond the estuarine check in the otoliths of Type 1 mullet adults fluctuated between 4 × 10– 3and 13.9 × 10– 3and were seldom less
than 3 × 10– 3(Fig. 5), indicating that after the juvenile
stage the mullet had migrated between estuary and offshore waters and had rarely entered freshwaters. The ages of Type 1 mullet ranged from 1 to 5 yr old, but most were aged 4 to 5 yr (Fig. 5, Table 2).
Type 2. Sr:Ca ratios beyond the estuarine check in the otoliths of Type 2 adult mullet fluctuated between 0.4 × 10– 3and 12.5 × 10– 3(Fig. 6), indicating that they
had migrated between freshwater and offshore waters. The recruitment time of the mullet to freshwater dif-fered among individuals, lasting from 1 to 2 yr. In some cases, the mullet all directly entered freshwater after estuarine arrival (e.g. Mullet Nos. E2, 5, 13, 14, and N1 and O4 in Fig. 6). Among these, Mullet E2 and O4 remained in freshwater for approximately 1 yr and then migrated to higher-salinity waters, whereas Mullet N1 remained in freshwater for a short time only (ca. 1–2 mo) and then migrated to higher salinities. Mullet E5 migrated between freshwater and brackish-water in the first year. Mullet E13 and E14 remained in freshwater after estuarine arrival until they were captured. In other cases, the mullet delayed entering freshwater after estuarine arrival (e.g. Mullet Nos. E1, 6, 9, 10 in Fig. 6). Among these, Mullet E1 and E10 resided in brackish or highly saline waters for approx-imately 6 mo and then migrated between freshwater and brackishwaters. Mullet E6 and E9 had migrated between brackish and high-salinity waters for 1 yr and then entered freshwater. Thus, the Type 2 mullet migrates to freshwater after the juvenile stage, but the duration of freshwater residence differs among individuals.
Composition of migratory patterns
The composition of Types 1 and 2 mullet differed sig-nificantly among habitats (p < 0.001, Table 4). The
pro-Fig. 4. Mugil cephalus. Mean (± SD) salinity and the mean
(± SD) Sr:Ca ratios at otolith edge of adults collected in estuar-ies (i.e. Sites IL, GST, TT, HHW), nearshore coastal zone (Sites TPW, MT) and offshore (Sites TS, WC, KS) (site abbreviations as in Fig. 1). Same letters above data points indicate that mean Sr:Ca ratios at otolith edge were homogeneous among
designated sites
Season, n Composition (%)
Sites Type 1 Type 2
Non-spawning season Estuary (IL, GST, TT, HHW) 15 46.7 53.3 Nearshore (MT, TPW) 6 83.3 16.7 Spawning season Offshore (TS, WC, KS) 9 88.9 11.1 Total 30 66.7 33.3
Table 4. Mugil cephalus. Percent composition of Types 1 and
2 as a function of migratory status and sampling sites. Site abbreviations as in Fig. 1
Fig. 5.
Mugil cephalus
. Changes in Sr:Ca ratios fr
om cor
e to otolith edge in otoliths of T
ype 1 adults. (
Z
) estuarine check; (
y
) annuli. Sr and Ca concentrations in Mullet
Nos.
E11, N2, 6 and O6 wer
e measur
ed for fish fr
om 2 transects with the same electr
on pr
obe micr
oanalyzer (EPMA) beam conditions
and those in Mullet E15 wer
e with
dif
fer
ent beam conditions. Gr
ey bands between the Sr:Ca ratios 3 to 7
×
10
–3
portion of Type 1 mullet was significantly higher in the nearshore (83.3%) and offshore (88.9%) sites irrespec-tive of season (spawning or non-spawning), but the proportion of Type 2 mullet was slightly higher in the estuaries (53.3%) during non-spawning season. In other words, the proportion of mullet entering fresh-waters (Type 2) was lower for the nearshore and off-shore populations than the estuarine population.
Conversely, the proportion of Type 1 mullet (88.9%) was higher than Type 2 during the spawning season (11.1%, Table 4) suggesting that the spawning stock of the mullet was mostly composed of Type 1 fish that did not enter freshwater.
Age-related habitat use
The habitat use of the mullet changed with age (Fig. 7). At Age 1 yr, about 26.7% of mullet inhabited freshwater; this decreased to 10.5% at Age 2 yr, and to none at Age 3 yr and older. In contrast, the proportion of mullet inhabiting offshore waters increased from 13.3% in the first year to 41.7% in the fifth year. Thus, a small portion of mullet inhabited freshwater and the number of mullet migrating offshore from freshwaters and the estuary increased with age.
Beyond the estuarine check in adult otoliths, the peak Sr:Ca ratios corresponded to annuli (Figs. 5 & 6). Mean (± SD) Sr:Ca ratios at the annuli ranged from 7.9 ± 1.8 × 10– 3to 10.0 ±
1.5 × 10– 3 and were significantly
greater than at the estuarine check in otoliths of both adult (5.1 ± 1.0 × 10– 3)
and juvenile mullet (6.2 ± 1.4 × 10– 3,
Table 3). The frequency distribution of peak Sr:Ca ratios could be roughly divided into 2 groups, one comprising the first and second annuli and the other the third to fifth annuli. The mean peak Sr:Ca ratios were larger in the latter than in the former group (Fig. 8). The mullet presumably tended to migrate to higher-salinity offshore waters after their third year.
Growth
The maximum otolith radius and fork length of the mullet were linearly correlated (Fig. 9), indicating that otolith growth reflects the somatic growth of the fish. The mean radius of the estuarine check and the succes-sive annuli in otoliths of the adult dif-fered significantly (Table 3). Annual increments within the otoliths were largest (ca. 3500 µm) within the first year and decreased dramatically to ca. 850 µm within the second year and ca. 650 µm to < 500 µm within the third to fifth years, indicating that mullet growth was fastest in the first year and declined with age. The coefficient of variation (CV) values of the otolith radius at both the estuar-ine check and first annulus were larger than for succeeding annuli for adults, indicating greater variability of growth rate in the early life stage.
Fig. 6. Mugil cephalus. Changes in Sr:Ca ratios from core to otolith edge in the
otoliths of Type 2 adults. (Z) estuarine check; (y) annuli. Sr and Ca
concentra-tions in Mullet Nos. E13 and E14 were measured for fish from 2 transects with different electron probe microanalyzer (EPMA) beam conditions. Grey bands
The mean lengths of the mullet at annulus formation were not significantly different between migratory his-tory types (Fig. 10). This indicated that migrahis-tory envi-ronments did not affect growth rates of the mullet. The age–length data of different migratory types were combined and a theoretical growth curve was fitted with the von Bertalanffy growth equation as follows:
Lt= 571.3 (1 – e(– 0.27[t + 1.29]))
whereLtwas the theoretical length at age t, the maxi-mum estimated length was 571.3 mm and the growth coefficient was 0.27 yr–1.
DISCUSSION
Sr:Ca otolith ratios as indicator of environmental salinity
The Sr:Ca ratios in otoliths of both juvenile and adult mullet were positively correlated with the salinities of the sampling sites (Fig. 4). Thus, the Sr:Ca ratios in otoliths can be used to reconstruct the migratory envi-ronmental history of the mullet, as for other fishes (Tzeng & Tsai 1994, Secor et al. 1995, Secor & Rooker 2000, Limburg et al. 2001).
Based on our findings, the Sr:Ca ratios in otoliths of the mullet decreased to less than 3 × 10– 3when the fish
migrated to freshwater, fluctuated between 3 and 7 × 10– 3 in estuarine brackishwaters, and were greater
Fig. 7. Mugil cephalus. Frequency distribution of age-related
habitat use. n = sample size
Fig. 8. Mugil cephalus. Frequency distribution of Sr:Ca ratios
at annulus (A1– 5) in otoliths. n = sample size
Fig. 9. Mugil cephalus. Relationship between fork length and
otolith radius
Fig. 10. Mugil cephalus. Relationship between fork length
than 7 × 10– 3in high-salinity offshore waters. Secor &
Rooker (2000) found a similar relationship between otolith Sr:Ca ratios and environmental salinity. They divided estuarine fishes into freshwater (salinity range of 0 to 5 ‰; mean Sr:Ca ratios of 1.97 × 10– 3), estuarine
(5 to 25 ‰; 5.03 × 10– 3) and marine (> 25 ‰; 7.43 × 10– 3)
habitats. Tsukamoto & Arai (2001) divided the Japan-ese eel Anguilla japonica into river, estuarine and sea eels when otolith Sr:Ca ratios were < 2.5 × 10– 3,
between 2.5 and 6 × 10– 3and > 6 × 10– 3, respectively.
Jessop et al. (2002) divided the American eel A. ros-trata into freshwater and estuarine residents when otolith Sr:Ca ratios were < 4 × 10– 3 and > 5 × 10– 3,
respectively. Tzeng et al. (2003) used the Sr:Ca ratio of 4 × 10– 3in otoliths of the Japanese eel as a criterion to
distinguish freshwater and seawater contingents of the eel. Otolith Sr:Ca ratios relative to ambient salinity seem similar among different species suggesting that their use to reconstruct the past salinity history of a fish is reliable.
Sr:Ca otolith ratios and their implication in migratory history of the mullet
The migratory environmental histories of juvenile mullet collected in the estuaries and those of the adult mullet collected in the estuaries, nearshore and off-shore, were reconstructed from the temporal changes of the Sr:Ca ratios in their otoliths. The Sr:Ca ratios in juvenile otoliths decreased from the core to the otolith edge (Fig. 3), indicating that the mullet larvae dis-persed away from the high-salinity offshore spawning ground and grew to juveniles in the estuaries, as expected (Tung 1981, Chang et al. 2000). The Sr:Ca ratios beyond the estuarine check in the otoliths of Type 2 mullet (e.g. Mullet Nos. E2, 5, 13, 14, N1 and O4 in Fig. 6) indicated that the fish resided in freshwater in Taiwan during the first year. The Sr:Ca ratios at the edge of adult mullet otoliths collected in Taiwanese estuaries were lower (e.g. E1, 6, 9, 10 in Fig. 6), indi-cating that the mullet resided in freshwater after 1 yr. The temporal change in Sr:Ca ratios in otoliths of Type 2 mullet supports the hypothesis that some mullet did not migrate to the coastal waters of mainland China but resided in the coastal waters of Taiwan (Liu 1986). The mullet migrated to the spawning ground at approximately 3 to 5 yr of age (Table 2). The Sr:Ca ratios in otoliths of mullet older than 3 yr were found to have oscillated annually, with a peak Sr:Ca ratio at the annulus check (Figs. 5 & 6). The annuli in otoliths of adult mullet are formed in the winter during spawning (Ibáñez-Aguirre & Gallardo-Cabello 1996). Thus, the peak Sr:Ca ratio and annulus in adult mullet otoliths were deposited simultaneously in the winter, as in the
European eel (Tzeng et al. 1999). In other words, the fluctuations in the Sr:Ca ratios in adult mullet otoliths corresponded to their migration between feeding grounds in the low-salinity coastal waters of mainland China and the high-salinity offshore spawning ground off the SW coast of Taiwan, which is influenced by the Kuroshio Current in winter (Tung 1981, Huang & Su 1989). Also, a few adult mullet also showed a peak Sr:Ca ratio 1 and/or 2 yr of age (e.g. Mullet Nos. O5, 6, 8 in Fig. 5 and O4 in Fig. 6). Some immature mullet might migrate with the spawners to the spawning ground, as reported by Hwang (1986) and Huang & Su (1989).
On the other hand, the feeding activity of mullet may decrease and their growth rates become lower in the winter. The deposition rate of Sr in the otolith may be negatively correlated to both water temperature (Radtke 1989, Townsend et al. 1992, Tzeng 1994) and growth rate of the fish (Sadovy & Severin 1992). There-fore, the temperature and fish growth rate may interact with salinity to regulate the Sr:Ca ratios in mullet otoliths. Interactive effects of temperature and salinity on otoliths were also found for black bream Acantho-pagrus butcheri (Elsdon & Gillanders 2002).
Divergent migratory behaviour of the mullet
The mullet was shown to comprise 2 coexisting migratory types in coastal waters of Taiwan (Table 4). Consequently, we do not completely agree with the hypothesis that the mullet is a catadromous fish spending most of its life in freshwater before migrating offshore for spawning (De Silva 1980, Torricelli et al. 1982).
Type 1 mullet is not a typical catadromous fish, because it does not migrate to freshwater but moves between estuarine and offshore waters (Fig. 5). One of the Type 2 mullet may have been catadromous, as are mullet in Australia (Thomson 1966) and the USA (Shireman 1975), which reside in freshwater before maturation at 3 to 4 yr of age (e.g. Mullet No. E10 in Fig. 6). Conversely, 2 of the Type 2 mullet emigrated from the freshwater habitat before 1 yr of age and did not remain in freshwater until maturation (i.e. N1 and O4 in Fig. 6) as do mullet in freshwaters of South Africa (Bok 1979) and western Australia (Chubb et al. 1981). Strictly, these 2 mullet could be only categorized as marine amphidromous because they emigrated before maturation (McDowall 1988). The remaining 7 mullet of the Type 2 mullet also could not be categorized as catadromous or amphidromous because they resided in freshwater only at an early stage (≤1 yr old). Thus, the migratory pattern of mullet is more plastic and complicated than previously thought. The term
‘catadromy’ is insufficient to represent the diverse migratory patterns of mullet, which seem to be more a diadromous continuum (McDowall 1988) in a sequen-tial succession of 4 ecophenotypes, namely marine, euryhaline, amphidromous and catadromous, as in the directional evolution of diadromous fishes as proposed by Gross (1987).
Mullet can inhabit freshwater, estuarine and off-shore waters, but tend to inhabit high-salinity offoff-shore waters after maturation (Figs. 7 & 8). This phenomenon was also found in the Mediterranean grey mullet, in which the mature mullet prefer polyhaline areas and strongly avoid freshwater sites (Cardona 2000). Young mullet may have greater phenotypic plasticity and more divergent migratory patterns than older mullet. The migratory patterns of mullet collected from estuar-ies were more divergent than those from nearshore and offshore coastal areas (Table 4). The salinity was more varied and productivity was higher in the estuary than offshore (Haedrich & Hall 1976). The mullet is a euryhaline fish and can tolerate a wide range of salin-ity. Consequently, the divergent migratory behaviour of the mullet might be simply an expression of its eury-haline nature and/or an attempt to seek the maximum evolutionary fitness for foraging (Gross 1987).
Acknowledgements. The study was financially supported by
the National Science Council, Republic of China (Grant num-bers NSC-89-2611-M002-039, NSC-90-2611-M002-019 and NSC-91-2611-M002-023). The authors are grateful to Mr. C. S. Huang of the Department of Fisheries Sciences, National Taiwan Ocean University, Dr. S. C. Lee of the Institute of Zoology, Academia Sinica, Mr. C. C. Liu and Dr. Y. S. Lin of the Institute of Zoology, National Taiwan University, Mr. C. C. Han of the National Museum of Marine Biology and Aquar-ium for providing specimens; to Mr. C. S. Chen for the field-work; to Mr. B. M. Jessop of the Department of Fisheries and Oceans, Bedford Institute of Oceanography, Canada, and the anonymous referees for helpful comments on an earlier draft of the manuscript.
LITERATURE CITED
Anonymous (1988) A comprehensive oceanographic survey of the central and northern part of the Taiwan Strait. Science Publication, Beijing
Blaber SJM (1987) Factors affecting recruitment and survival of mugilids in estuaries and coastal waters of southeastern Africa. Am Fish Soc Symp 1:507–518
Bok AH (1979) The distribution and ecology of two mullet species in some fresh water rivers in the Eastern Cape, South Africa. J Limnol Soc South Africa 5:97–102 Breder CM, Rosen DE (1966) Modes of reproduction in fishes.
TFH Publishers, Nepture City, NJ
Campana SE (1999) Chemistry and composition of fish oto-liths: pathways, mechanisms and applications. Mar Ecol Prog Ser 188:263–297
Campana SE, Neilson JD (1985) Microstructure of fish otoliths. Can J Fish Aquat Sci 42:1014–1032
Cardona L (2000) Effects of salinity on the habitat selection and growth performance of Mediterranean flathead grey
mullet Mugil cephalus (Osteichthyes, Mugilidae). Estuar
Coast Shelf Sci 50:727–737
Chang CW, Tzeng WN (2000) Species composition and seasonal occurrence of mullet (Pisces, Mugilidae) in the Tanshui estuary, northwest Taiwan. J Fish Soc Taiwan 27: 253–262
Chang CW, Tzeng WN, Lee YC (2000) Recruitment and
hatching dates of grey mullet (Mugil cephalus L.)
juve-niles in the Tanshui estuary of northwest Taiwan. Zool Stud 39:99–106
Chen WY, Su WC (1986) Reproductive biology of the grey
mullet, Mugil cephalus L. of Taiwan. In: Su WC (ed) Study
on the resource of grey mullet in Taiwan, 1983–1985. Kaohsiung Branch TFRI, Kaohsiung, p 73–80
Chen WY, Su WC, Shao KT, Lin CP (1989) Morphometric
studies of the grey mullet (Mugil cephalus) from the
waters around Taiwan. J Fish Soc Taiwan 16(3):153–163 Cheng HH, Hsieh CS, Su MS (1990) Ecological and en-vironmental survey of the Tapen Bay. II. Distribution of plankton. TungKang Branch TFRI, TungKang
Chubb CF, Potter IC, Grant CJ, Lenanton RCJ, Wallace J (1981) Age structure, growth rates and movements of sea
mullet, Mugil cephalus L., and yellow-eyed mullet,
Aldri-chetta forsteri (Valenciennes), in the Swan-Avon River
system, Western Australia. Aus J Mar Freshw Res 32: 605–628
Degens ET, Deuser WG, Haedrich RL (1969) Molecular struc-ture and composition of fish otoliths. Mar Biol 2:105–113 De Silva SS (1980) Biology of juvenile grey mullet: a short
review. Aquaculture 19:21–36
Elsdon TS, Gillanders BM (2002) Interactive effects of tem-perature and salinity on otolith chemistry: challenges for determining environmental histories of fish. Can J Fish Aquat Sci 59:1796–1808
Gross M (1987) The evolution of diadromy in fishes. Am Fish Soc Symp 1:14–25
Haedrich RL, Hall CAS (1976) Fishes and estuaries. Oceanus 19:55–63
Howland KL, Tonn WM, Babaluk JA, Tallman RF (2001) Iden-tification of freshwater and anadromous inconnu in the Mackenzie River system by analysis of otolith strontium. Trans Am Fish Soc 130:725–741
Huang CS, Su WC (1989) Studies on the fluctuations of
fish-ing conditions for grey mullet (Mugil cephalus Linnaeus)
from the western coast of Taiwan. J Fish Soc Taiwan 16: 47–83
Hwang SY (1986) Age composition, size distribution and
length-weight relationship of grey mullet (Mugil cephalus)
in 1984. In: Su WC (ed) Study on the resource of grey mullet in Taiwan, 1983–1985. Kaohsiung Branch TFRI, Kaohsiung, p 57–62
Ibáñez-Aguirre AL, Gallardo-Cabello M (1996) Age
determi-nation of the grey mullet Mugil cephalus L. and the white
mullet Mugil curema V. (Pisces: Mugilidae) in Tamiahua
lagoon, Veracruz. Cienc Mar 22:329–345
Jarosewich E, White JS (1987) Strontianite reference sample for electron microprobe and SEM analyses. J Sediment Petrol 57:762–763
Jessop BM, Shiao JC, Iizuka Y, Tzeng WN (2002) Migratory
behaviour and habitat use by American eels Anguilla
ros-trata as revealed by otolith microchemistry. Mar Ecol Prog
Ser 233:217–229
Kalish JM (1990) Use of otolith microchemistry to distinguish the progeny of sympatric anadromous and non-anadro-mous salmonids. Fish Bull US 88:657–666
Kawakami Y, Mochioka N, Morishita K, Tajima T, Nakagawa H, Toh H, Nakazono A (1998) Factors influencing otolith
strontium/calcium ratios in Anguilla japonica elvers.
Environ Biol Fish 52:299–303
Lee SC (1992) Fish fauna and abundance of some dominant species in the estuary of Tanshui, northwestern Taiwan. J Fish Soc Taiwan 19:263–271
Limburg KE, Landergren P, Westin L, Elfman M, Kristiansson P (2001) Flexible modes of anadromy in Baltic sea trout: making the most of marginal spawning streams. J Fish Biol 59:682–695
Liu CH (1986) Survey of the spawning grounds of grey mullet. In: Su WC (ed) Study on the resource of grey mullet in Taiwan, 1983–1985. Kaohsiung Branch TFRI, Kaohsiung, p 63–72
McDowall RM (1988) Diadromy in fishes: migrations between freshwater and marine environments. Cambridge Univer-sity Press, Cambridge
Nash EC, Shehadeh ZH (1980) Review of breeding and
prop-agation techniques for grey mullet, Mugil cephalus L.
ICLARM Stud Rev 3:1–40
Pannella G (1971) Fish otolith: daily growth layers and peri-odical patterns. Science 173:1124–1127
Philibert J, Tixier R (1968) Electron penetration and the atomic number correction in electron probe microanalysis. Br J Appl Phys 2:685–694
Radtke RL (1989) Strontium–calcium concentration ratios in fish otoliths as environmental indicators. Comp Biochem Physiol A 92:189–193
Ricker WE (1958) Handbook of computations for biological statistics of fish populations. Bull Fish Res Board Can 119: 1–300
Sadovy Y, Severin KP (1992) Trace elements in biogenic arag-onite: correlation of body growth and strontium levels in
the otoliths of the white grunt, Haemulon plumieri (Pisces.
Haemulidae). Bull Mar Sci 50:237–257
Secor DH (1992) Application of otolith microchemistry analy-sis to investigate anadromy in Chesapeake Bay striped
bass Morone saxatilis. Fish Bull US 90:798–806
Secor DH, Piccoli PM (1996) Age- and sex-dependent migrations of striped bass in the Hudson River as deter-mined by chemical microanalysis of otoliths. Estuaries 19: 778–793
Secor DH, Rooker JR (2000) Is otolith strontium a useful scalar of life cycles in estuarine fishes? Fish Res 46:359–371 Secor DH, Henderson-Arzapalo A, Piccoli PM (1995) Can
otolith microchemistry chart patterns of migration and habitat utilization in anadromous fishes? J Exp Mar Biol Ecol 192:15–33
Shireman JV (1975) Gonadal development of striped mullet (Mugil cephalus) in fresh water. Prog Fish-Cult 37:205–205
Siegel S (1956) Nonparametric statistics for the behavioral sciences. McGraw-Hill, New York
Su WC, Jane KF (1974) Oceanographic investigation in the
Taiwan Strait during the grey mullet season of 1973. Bull Taiwan Fish Res Inst 24:55–94
Thomson JM (1966) The grey mullet. In: Barnes H (ed) Oceanography and marine biology — an annual review, Vol 4. Allen & Unwin, London, p 301–355
Torricelli P, Tongiorgi P, Almansi P (1982) Migration of grey mullet fry into the Arno River: seasonal appearance, daily activity and feeding rhythms. Fish Res 1:219–234 Townsend DW, Radtke RL, Corwin S, Libby DA (1992)
Stron-tium:calcium ratios in juvenile Atlantic herring (Clupea
harengus L.) otoliths as a function of temperature. J Exp
Mar Biol Ecol 160:131–140
Tsukamoto K, Arai T (2001) Facultative catadromy of the eel Anguilla japonica between freshwater and seawater
habitats. Mar Ecol Prog Ser 220:265–276
Tung IH (1981) On the fishery biology of gray mullet, Mugil
cephalus L., in Taiwan. Rep Inst Fish Biol Minist Econ Aff
Natl Taiwan Univ 3:38–102
Tzeng WN (1994) Temperature effects on the incorporation
of strontium in otolith of Japanese eel Anguilla japonica.
J Fish Biol 45:1055–1066
Tzeng WN (1995) Migratory history recorded in otoliths of the
Japanese eel, Anguilla japonica, elvers as revealed from
SEM and WDS analyses. Zool Stud 34(Suppl 1):234–236 Tzeng WN (1996) Effects of salinity and ontogenetic move-ments on strontium:calcium ratios in the otoliths of the
Japanese eel, Anguilla japonica Temminck and Schlegel.
J Exp Mar Biol Ecol 199:111–122
Tzeng WN, Tsai YC (1994) Changes in otolith microchemistry
of the Japanese eel, Anguilla japonica, during its
migra-tion from the ocean to the rivers of Taiwan. J Fish Biol 45: 671–683
Tzeng WN, Severin KP, Wickström H (1997) Use of otolith microchemistry to investigate the environmental history of
European eel Anguilla anguilla. Mar Ecol Prog Ser 149:
73–81
Tzeng WN, Severin KP, Wickström H, Wang CH (1999) Stron-tium bands in relation to age marks in otoliths of European eel Anguilla anguilla. Zool Stud 38:452–457
Tzeng WN, Wang CH, Wickström H, Reizenstein M (2000) Occurrence of the semi-catadromous European eel
Anguilla anguilla (L.) in Baltic Sea. Mar Biol 137:93–98
Tzeng WN, Shiao JC, Iizuka Y (2002a) Use of otolith Sr/Ca ratios to study the riverine migratory behaviors of
Japan-ese eel Anguilla japonica. Mar Ecol Prog Ser 245:213–221
Tzeng WN, Wang YT, Chang CW (2002b) Spatial and tem-poral variations of the estuarine larval fish community on the west coast of Taiwan. Mar Freshw Res 53:419–430 Tzeng WN, Iizuka Y, Shiao JC, Yamada Y, Oka HP (2003)
Identification and growth rates comparison of divergent
migratory contingents of Japanese eel (Anguilla japonica).
Aquaculture 216:77–86
Winer B (1971) Statistical principles in experimental design. McGraw-Hill, New York
Editorial responsibility: Otto Kinne (Editor), Oldendorf/Luhe, Germany
Submitted: October 14, 2002; Accepted: November 21, 2003 Proofs received from author(s): March 15, 2004
