Hydrobiologia DOI 10.1007/s10750-010-0300-1 PRIMARY RESEARCH PAPER Variation of life traits of glass eels of Anguilla anguilla (L.) during the colonization of Rı́os Nalón and Minho estuaries (northwestern Iberian Peninsula) Tania Iglesias • Javier Lobón-Cerviá Sérgia Costa Dias • Carlos Antunes • Received: 10 September 2009 / Revised: 7 May 2010 / Accepted: 10 May 2010 Ó Springer Science+Business Media B.V. 2010 Abstract Analyses of the pigmentation stages and the size of glass eels during the arrival period to two northwestern Iberian estuaries (Rı́os Nalón and Minho) and comparison of the patterns elucidated in this study with those reported for other geographical locations of the Atlantic and Mediterranean ranges showed nine pigmentation stages ranging from practically transparent to fully pigmented in the twofirst locations. Overall, stage VB was predominant but within a broad variance across rivers and seasons. Consistent with previous studies, pigmentation stages tended to increase when the season progressed whilst, Handling editor: M. Power T. Iglesias (&) Facultad de Biologı́a. Dept. B. O. S. (Zoologı́a), Universidad de Oviedo, C/Catedrático Rodrigo Urı́a, s/n, 33071 Oviedo, Spain e-mail: iglesiastania@uniovi.es T. Iglesias  J. Lobón-Cerviá  S. C. Dias Museo Nacional de Ciencias Naturales (CSIC), C/José Gutiérrez Abascal 2, 28006 Madrid, Spain S. C. Dias  C. Antunes CIMAR/CIIMAR—Centro Interdisciplinar de Investigação Marinha e Ambiental. Universidade do Porto, Rua dos Bragas, 289, 4050-123 Porto, Portugal S. C. Dias ICBAS—Instituto de Ciências Biomédicas de Abel Salazar/Universidadedo Porto, Lg. Prof. Abel Salazar 2, 4099-003 Porto, Portugal for a given individual length, mass tended to be higher during the early months of arrival. Moreover, both mass and condition factor were lower in individuals of advanced stages. Whilst we found no significant differences on glass eels length among pigmentation stages, both mass and condition factor were lower in more pigmented individuals. Given that larger glass eels with better physical condition may reveal greater energy content implying improved upstream migration earlier individuals may be better migrants than latter individuals. Keywords Glass eel  Body size  Pigmentation stages  Condition factor  Northwestern Iberian Peninsula Introduction The abundance of riverine stocks of European eel Anguilla anguilla (L.) depends entirely on the annual recruitment of the youngest juveniles (i.e., glass eels) reaching continental waters to commence a new yearclass. Recent studies have emphasized the variability of the external factors during the oceanic migration of early life stages of the European eel (Kettle & Haines, 2006; Friedland et al., 2007; Bonhommeau et al., 2008) so recruitment to rivers would, to a large extent, reflect larvae physiological conditions. Bureau du Colombier et al. (2007) have further suggested that estuarine colonization may depend on the energetic 123 Hydrobiologia conditions of glass eels at their arrival that may also influence the dynamics of river stocks. Moreover, even if the ocean and estuary conditions may be indirectly influenced by human interventions, it is from recruitment onwards that both population assessment and management are best achieved. Therefore, understanding the arrival to the estuarine waters and the later recruitment to the stock is critical to evaluate the dynamics underlying riverine stocks. The relative abundance of recruiting glass eels (Dekker, 2003) and their sizes and biometrics (Desaunay & Guerault, 1997) have been widely reported (Tesch, 2003). In contrast, the pigmentation stages and size have been scarcely documented and only one previous study has assessed levels of variability in the pigmentation stages during the arrival period (Boëtius & Boëtius, 1989) to several locations of northern Europe. Therefore, the objective of this study was to assess the pigmentation stages and the size of glass eels during the arrival period to two northwestern Iberian estuaries which are proximal to the eastward migratory route from the spawning grounds in the Sargasso Sea. Subsequently, we compared the patterns elucidated in this study with those reported for other geographical locations along the west-east colonization pathways of the Atlantic and Mediterranean ranges. Study area and glass eels collections Rı́o Nalón is located in Asturias (Fig. 1; drainage = 3,692 km2, length = 145 km; mean discharge = 100 m3 s-1) and flows north across the northern slope of the Cantabrian corridor of northwestern Spain to discharge into the Cantabrian Sea at San Juan de La Arena. The sources of Rı́o Minho are also in Cantabrian corridor of northwestern Spain. However, Fig. 1 Geographical location of the study sites in Rı́o Nalón (Asturias, Cantabrian Sea, northwestern Spain) and Rı́o Minho (Atlantic Ocean, northern border between Spain and Portugal) 123 Hydrobiologia Rı́o Minho flows south-southwest to discharge into the Atlantic Ocean at the northern border between Spain and Portugal (Fig. 1; drainage = 17,080 km2, length = 340 km; mean discharge = 300 m3 s-1). In the estuaries of the two rivers, monthly samples of glass eels were collected from November to April of the years 2004–2005 and 2005–2006 on new moon nights, when captures are more numerous. The fishing operations differed between the two estuaries. The professional fishing in Rı́o Nalón has traditionally been based on fishing vessels operating along the estuary. Fishing vessels with two lateral 4–6 m long nets moving upstream, filter the 3 km long estuary waters. In contrast, Rı́o Minho’ glass eels were collected by experimental fishing procedures with stow nets. These nets had a maximum of 14–15 m width at the mouth, 10 m long (1–2 mm mesh size) and 8 m height at the mouth; maintained by fishing weights at the bottom and 10–20 l buoys at the surface along the sides. Net sides converge to a central posterior end fastened to a boat where fishermen scoop glass eels with a hand net. These stow nets were anchored facing the upstream current on rising tide for &2 h and checked for glass eel captures every 15 min. Further details on these fishing techniques can be seen in Lara (1994) for Rı́o Nalón and Weber (1986) and Antunes (2002) for Rı́o Minho. Thus, Rı́o Nalón’ glass eels were sub-sampled from those captured by fishermen whereas those from Rı́o Minho were collected directly from the experimental fishing procedures. Since different workers collected and examined glass eels independently on each river, a methodological inter-calibration was carried out to prevent a potential mismatch in the assignment of pigmentation stages. Fresh glass eels just after capture, were measured (total length, mm), weighed (g) and their pigmentation stages were assigned following the classification established by Elie et al. (1982). Unfortunately, the individual assignment of pigmentation stages and size failed in Rı́o Minho’ glass eels in the year 2004–2005. In these samples the length and pigmentation stages were quantified separately for each individual glass eel. Comparison among studies was hindered by the variety of methods used by different authors. For example, several authors described these stages based on preserved material (which causes a marked reduction in length and mass of glass eels, Iglesias, T. unpubl.) whereas other authors described pigmentation stages on different scales and/or used different classification systems as to those described by Strubberg (1913) and Boëtius (1976). To smooth the progress of our comparison we applied the equivalence of pigmentation stages defined by Briand et al. (2005) that made compatible our stages V to VI—as described by Elie et al. (1982)—with stages A to E as proposed by Boëtius (1976). This equivalence among pigmentation stages is as follows: VB = A; VIA0 ? VIA1 = B; VIA2 ? VIA3 = C and VIA4 = D; stage E has no correspondence. Data analysis To infer the effects of river (R), year (Y), month (M) and pigmentation stage (PS) on glass eel length and mass we used Analysis of Variance (ANOVA, Zar, 1999). In these ANOVAs, the explanatory variables were treated as categorical variables. Since the Rı́o Minho data set did not include the individual assignment of size and pigmentation stage during the first sampling year (2004–2005) and also missed samples from December 2004 and March 2006, the interactions between size, pigmentation and river were disregarded and comparisons between rivers were focused on the 4 months common to the two rivers. Length-to-mass relationships were explored in the log-linear form Log (mass, g) = a ? b Log (length, mm) where a, and b are constants. The physical conditions of glass eels were examined through the condition coefficient (K) calculated as K = 106 * mass/lengthb, where the exponent b was obtained from the length-to-mass relationships. Both length-to-mass relationships and condition coefficient were explored for the effects of the month, river, year, and pigmentation stage. Among-month and between-river comparisons were made through Analysis of Variance (ANOVA) and Analysis of Covariance (ANCOVA, Zar 1998). The null hypothesis that pigmentation stages are independent of the arrival month was tested through contingency tables as 9 pigmentation stages * 6 months for each single year in Rı́o Nalón and as 9 pigmentation stages * 4 months for each single year when Rı́o Minho was included in the analyses. For these comparisons, pigmentation stages with less than \ 5 individuals were omitted. 123 Hydrobiologia Results Pigmentation stages Monthly samples averaged 233 (range 55–365) and 68 (range 39–202) individuals within a total of 2,798 and 683 glass eels examined for Rı́os Nalón and Minho, respectively. Individuals in nine pigmentation stages and sub-stages IV, VA, VB, VIA0, VIA1, VIA2, VIA3, VIA4 and VIB were identified in the two rivers. These stages ranged from the least pigmented or completely transparent (i.e., stage IV) to the fully pigmented individuals (i.e., stage VIB). Overall, there was a predominance of individuals in stage VB that contributed a minimum monthly percentage of 55.8% and 33.3% for Rı́os Nalón and Minho, respectively. Nonetheless, the relative contribution of each stage and sub-stage varied widely across months (Fig. 2). Contingency tables were conclusive to reject the null hypothesis that pigmentation stages were independent of the arrival month. Rejection of this hypothesis was consistent for each single year and for the two rivers. Contingency tables for each single year for Rı́o Nalón, at least v2 [ 100, df = 30, 0.001 \ P \ 0.005 and for Rı́o Minho at least v2 [ 62, df = 18, P \ 0.001. In Rı́o Nalón, the monthly differences in the relative contribution of the pigmentation stages were expressed in the contribution of stage VB. The abundance of individuals in this stage declined over successive months to become replaced by individuals in more advanced sub-stages. By the end of the winter and early spring the number of individuals in advanced stages increased and very pigmented individuals in sub-stages VIA0 to VIA4 encompassed the bulk of the individuals whereas those in transparent or near transparent stages practically disappeared (Fig. 2). Moreover, slight differences between years were reflected in a greater relative abundance of stages IV, VA, and VIA0 and a lower relative abundance of individuals in stage VIA2 in 2004–2005 relative to a total absence of stage IV; a decrease of VA and VIA0 and an increase of individuals in stage VIA2 in 2005– 2006. Thus, individuals of the second year tended to show a greater proportion of advanced stages. Overall the temporal patterns of pigmentation stages in Rı́o Minho glass eels were similar (Fig. 2) with the only exception in a less predominant stage VB in the winter of 2004–2005 and a markedly higher diversity of stages in December of 2005. Glass eels size Visual inspections of the temporal variations in length and mass of glass eel from Rı́o Nalón suggested concurrent declines over time in the two study years (Fig. 3a). However, an ANOVA for the effects of year, month and pigmentation stage revealed a remarkable amount of unexplained variance (74.7% for length and 69.9% for mass, respectively). Nonetheless, the month had a significant effect on mass (F5,2721 = 24.717; P \ 0.01, 26.1% of the variance explained) but not on length. Fig. 2 Monthly occurrence of successive pigmentation stages (%) in the glass eels of the Rı́o Nalón (grey) and Rı́o Minho (black)’ estuaries over 2004–2005 and 2005–2006. NN and NM are the number of individuals examined in each estuary 123 Hydrobiologia Fig. 3 Mean monthly length (mm) and mass (g) (mean ± SE) of glass eels examined in a Rı́o Nalón and b Rı́o Minho over the years 2004–2005 and 2005–2006 Since length and mass are closely related, we explored whether length-induced effects were responsible for the temporal variations in mass. For any given month of the 2 years, log-transformed length and mass proved to be strongly related and the variations in logtransformed length explained between 52.5 and 82.1% of the variations in the log-transformed mass (Table 1). An ANCOVA for these 12 length-to-mass regressions with month and year as factors revealed no significant difference among slopes but highly significant differences among intercepts (Table 2a) and a Tukey-test revealed significant differences among practically all these 12 regressions. As a consequence, the monthly length-to-mass regressions were assumed to be parallel to each other with a common slope equal to 2.795 (Table 1). Consistent with these results, an ANOVA for the condition factor (K) re-calculated for a common slope (b = 2.795) revealed significant effects of the year (F1,2786 = 75.5; P \ 0.001), month (F5,2786 = 73.8; P \ 0.001) and interaction year * month (F5,2786 = 16.0; P \ 0.001). Accordingly, the mass of individuals of a given length tended to be higher during the early months of recruitment than in later months. In Rı́o Nalón, an independent analysis of the effects of pigmentation stage and year (for stages VA to VIA4) on length indicated no significant effects of the Table 1 Monthly (November to April of 2004–2005 and 2005–2006) length-to-mass relationships in the form log mass (g) = a ? b log length (mm) for the Rı́o Nalón’ glass eels Year 2004/2005 2005/2006 Month N a b r2 Nov 201 -5.85 2.90 0.81 Dec 301 -5.44 2.66 0.69 Jan 365 -5.51 2.69 0.69 Feb 209 -5.52 2.69 0.70 Mar 301 -5.58 2.73 0.70 Apr Nov 248 234 -5.87 -5.37 2.87 2.62 0.67 0.52 Dec 169 -5.62 2.75 0.82 Jan 293 -6.42 3.18 0.80 Feb 258 -5.79 2.83 0.72 Mar 166 -6.13 3.02 0.76 Apr 55 -5.38 2.58 0.59 2 All instances are significant in, at least, P \ 0.001. r is the variance explained. An ANCOVA detected no significant differences with a common slope b = 2.795 pigmentation stage (F6,2766 = 0.49, P = 0.82) but significant effects of the year (F1,2766 = 6.03, P = 0.01) and of the interaction year * pigmentation (F6,2766 = 2.67, P = 0.014). In contrast, significant effects of the pigmentation stage (F6,2766 = 2.78, P = 0.01) and of the interaction year*pigmentation 123 Hydrobiologia Table 2 Results of ANCOVA for the length-to-mass relationships in the form log mass (g) = a ? b log length (mm) for the glass eels from (a) Rı́o Nalón, two successive years (2004– 2005 and 2005–2006) and six months (November, December, January, February, March and April) and (b) Rı́os Nalón and Minho combined, two successive years (2004–2005 and 2005– 2006) and 4 months (November, January, February, and April) DF P Table 3 Results from ANOVAs for the effects of river (Nalón and Minho), year (2004–2005 and 2005–2006), and month (November, January, February and April) on the length and mass of glass eels DF Length %V Mass P %V P River 1 3.2 <0.001 0.1 [0.05 1 3 0.1 11.4 [0.05 <0.001 0.1 22.3 [0.05 <0.001 Year 1 [0.05 Year Month Month 5 [0.05 R*Y 1 0.4 <0.01 0.0 [0.05 <0.001 R*M 3 0.1 [0.05 0.2 [0.05 (a) Length 1 Y*M 5 <0.05 Y*M 3 0.4 <0.05 0.5 <0.001 Y*L 1 [0.05 R*Y*M 3 0.4 <0.05 0.2 [0.05 M*L 5 [0.05 Residual 2,429 Y*M*L 5 <0.05 Significant results in bold Error (b) 2,774 76.7 % V is the amount of variance explained in percent (%) River 1 <0.05 Year 1 [0.05 Month 3 [0.05 Length 1 <0.001 R*Y 1 [0.05 R*M 3 [0.05 [0.05 Y*M 3 R*L 1 <0.01 Y*L 1 [0.05 M*L 3 [0.05 R*Y*M 3 [0.05 R*Y*L 1 [0.05 R*M*L 3 [0.05 Y*M*L 3 [0.05 R*Y*M*L Error 3 2,413 [0.05 Significant results in bold stage (F6,2766 = 4.02, P \ 0.001) were detected on mass. Concurrently, an ANOVA for the condition coefficients revealed significant effects of the pigmentation stage (F6,2766 = 7.94, P \ 0.001) and a weak but yet significant effect of the interaction pigmentation stage * year (F6,2766 = 2.21, P = 0.04). Similar to Rı́o Nalón, the length and mass of glass eels from Rı́o Minho declined over time (Fig. 3b). Moreover, an ANOVA for the 4 months common to the two rivers (November, January, February and April) also revealed a large proportion of unexplained variance, 84.1% for length and 76.7% for mass, 123 84.1 respectively (Table 3) and highlighted differences between the two rivers in length but not in mass and also emphasized the effects of the month on both, length and mass (Table 3). Interestingly, in Rı́o Minho the month had a significant effect on both, length and mass (F3,573 = 29.01, P \ 0.001 for length and F3,573 = 61.59, P \ 0.001 for mass) with smaller individuals predominating during the latest months of arrival but, like in Rı́o Nalón, an ANCOVA for the length-to-mass relationships revealed no significant differences neither among months nor between years (Table 2 b). Hence, we compared the two sets of length-to-mass regressions for the two rivers and an ANCOVA highlighted no significant difference (F1,2440 = 0.57, P = 0.45). As a consequence, we re-calculated the condition factors for all individuals across months, years and rivers based on a common slope equal to 2.915 and these coefficients differed significantly (at least P \ 0001) among months, rivers and years (Fig. 4). Correspondingly, an analysis of the effects of the pigmentation stage on length, mass and condition of glass eels for the only year available in Rı́o Minho provided evidence of a similar pattern. An ANOVA highlighted no significant effect of the pigmentation stage on length (F6,376 = 1.19, P = 0.31) but a significant effect on mass (F6,376 = 2.76, P = 0.01) and on the condition factor (F6,376 = 7.06, P \ 0.001). Thus, whilst glass eel length did not vary significantly among pigmentation stages both, mass and condition factor were lower in individuals of Hydrobiologia Fig. 4 Temporal decline in the physical condition of glass eels as inferred from the condition factor calculated for a common slope b = 2.915 for Rı́o Nalón and Rı́o Minho combined increasingly advanced stages to an extent that individuals in the most advanced pigmentation stage VIA4 weighted, on the average, some 18% less than those in stage VB. Comparison with other locations Comparative data collected from the literature and the corresponding geographical locations (Table 4) include 15 studies on western Atlantic glass eels, two from the northern Atlantic area, one from the North Sea and six from the Mediterranean. Apparently, glass eels from Atlantic locations appeared larger in size (range 68.6–78.4 mm) than those from the Mediterranean (60.7–73.0 mm, Table 4). However, western Atlantic glass eels show no marked differences in size or in pigmentation stages among locations in Portugal, Spain, France and Ireland although Boëtius & Boëtius (1989) noted that pigmentation stages varied widely with no obvious temporal pattern among several northern Atlantic locations. Studies available for the Mediterranean [La Camargue, southern France (Maes et al., 2009) and Rı́o Arno, northeastern Italy (Gandolfi et al., 1984)] also suggested temporal replacement of pigmentation stages with an increased predominance of more pigmented individuals as the season progressed and interestingly, an overall predominance of more pigmented stages than in the Atlantic glass eels (Table 4). Nonetheless, relative to the whole-season predominance of stage VB in the Iberian Rı́os Nalón and Minho, in the Mediterranean Rı́o Arno where pigmentation stages are well documented (Fig. 5) such predominance only occurred in December– February and declined sharply from March to May to become replaced by the most pigmented individuals in stages VIA4 and VIB. Discussion This comparative study of the pigmentation stages and size of glass eels during the arrival period to two northwestern Iberian estuaries located some 400 km apart highlighted concurrent patterns of temporal variation and in turn, differences in these life history traits. In the two estuaries, the size and pigmentation stages of glass eels varied substantially from November to April. Although most of the temporal variation remained unexplained, several patterns were detected. Consistent with previous studies (summarized by Tesch, 2003), in the two estuaries, glass eels decreased in size and condition but augmented their pigmentation as the season progressed from November to April. Larger individuals in pigmentation stage VB predominated at the onset of the arrival period to 123 Hydrobiologia Table 4 Comparison of pigmentation stages and size of glass eels across Atlantic and Mediterranean localities Location Date Length Mass Pigm. stage* Preservation Author Atlantic locations 1. Minho (Portugal) Jan–Jul 1982 72.0–65.0 0.36–027 2. Minho (Portugal) Nov 2004–Apr 2006 73.5–65.3 0.40–0.25 VB Early Formol Weber (1986) Fresh This study 3. Nalón (Spain) Nov 1989–Mar 1990 72.4–68.2 0.33–0.25 Early Fresh Lara (1994) 4. Nalón (Spain) Nov 2004–Apr 2006 75.2–66.8 0.38–0.23 VB Fresh This study 5. Santander (Spain) Nov 1917–May 1918 76.5–67.5 0.53–0.36 VIAII Formol Gandolfi-Hornyold (1918) 6. Aguinaga (Spain) Dec 1935–Jun 1936 78.4–68.4 0.58–0.30 – Formol Gandolfi-Hornyold (1936) 7. Oria (Spain) Oct 1928–Apr 1929 77.6–70.0 0.53–0.38 – Formol Gandolfi-Hornyold (1929) 8. Oria (Spain) Oct 1929–Apr 1930 76.1–69.7 0.51–0.30 – Formol Gandolfi -Hornyold (1930b) 9. Adour (France) Nov 1997–Mar 1998 73.5–67.4 0.35–0.26 – Fresh de Casamajor et al. (2000) 10. Adour (France) Dec 1999–Feb 2000 75.0–68.3 0.41–0.30 VB Fresh de Casamajor et al. (2003) 11. Adour (France) Jan–Mar 2000 68.6–68.0 0.30–0.30 VB Fresh Bardonnet et al. (2003) 12. Adour (France) Oct 1979–Aug 1980 76.0–70.3 0.45–0.29 Variable Fresh Charlon & Blanc (1982) 13. Gironde (France) Dec 1997–Apr 1998 70.0–66.0 0.32–0.23 Only VB Fresh Lambert et al. (2003) 14. Vilaine (France) 15. Seudre (France) Feb ? Mar 1992 Dec 1984–Jan 1987 69.7 – VB 70.0–68.1 0.31–0.29 A Fresh Frozen Lecomte-Finiger et al. (1996) Boëtius & Boëtius (1989) 16. Burrishoole (Ireland) Feb–May 1988 72.8–70.6 0.36–0.29 VB Fresh Poole (1994) 17. Severn (England) 18. Vidå´ (Denmark) Apr 1985–Apr 1987 72.5–68.1 0.32–0.18 Variable Frozen Boëtius & Boëtius (1989) Apr 1984–May 1987 72.2–69.9 0.22–0.12 C or D Boëtius & Boëtius (1989) Frozen Mediterranean locations 19. Sebou (Morocco) Nov 1980–Jun 1981 68.8–63.5 0.32–0.24 – Formol Yahyaoui et al. (1983) 20. Mallorca (Spain) Jun–Mar 1918 66.9–62.4 0.35–0.24 VIAIII Formol Gandolfi-Hornyold (1920) 21. Valencia (Spain) Dec 1984 Frozen Boëtius & Boëtius (1989) 22. Camarge (France) Jan 2004–Jan 2006 70.3–59.1 0.34–0.17 Variable Fresh Maes et al. (2009) 23. Arno (Italy) Dec 1978–Apr 1979 73.0–65.5 0.47–0.21 C Formol Gandolfi et al. (1984) 24. Alexandrie (Egypt) Mar 1928 60.7 Formol Gandolfi-Hornyold (1930a) 68.15 0.32 0.14 A VIAIII Length (mm), mass (g) and date represent the ranges and time periods of the collections.* Pigmentation Stage refers to the most abundant stage. In locations 1 and 3 ‘‘early’’ refers to combinations of stages VA, VB, and VIA0 in the scale of Elie et al. (1982). Locations 5, 16 and 24 according to Strubberg (1913). Locations 15, 17, 18, 21 and 23 according to Boëtius (1976). All other locations according to Elie et al. (1982) Fig. 5 Month-to-month variation in the pigmentation stages of Rı́o Arno (Italy)’ glass eels during 1978–1979 as reported by Gandolfi et al. (1984) and re-calculated according to the pigmentation stages defined by Briand et al. (2005) that makes compatible the stages V to VI (Elie et al. 1982) to the stages A to E (Boëtius, 1976) 123 Hydrobiologia contribute 50–75% of all individuals in November. The relative contribution of this pigmentation stage declined as the season progressed to become replaced by smaller, more pigmented individuals to the extent that in April, markedly-to-fully pigmented individuals (stage VI) contributed similarly to the bulk of the arriving glass eels. Although both rivers followed the same tendency, Rı́o Minho shows a markedly higher diversity of stages in some early months, such as December of 2005. The reason for this phenomenon remains unclear. The existence of any structure that may block down glass eel migration leading to a mix of different migration waves and pigmented stages can be discarded. However, when analyzing the pigmentation stages of samples with very different number of individuals (for example, January 2005— Fig. 2) accuracy of results may be affected. Moreover, even if the analytical method was inter-calibrated, such phenomenon may be caused by different potential perceptions for visual assignments by the two different workers who assigned pigmentation stages on Rı́os Nalón and Minho. The pigmentation stage had a strong effect on mass and condition coefficient but not on length and because of the tendency of the advanced pigmentation stages to occur later in the season, the two factors, month and pigmentation stage, obscure one another. Whilst glass eel length did not vary significantly among pigmentation stages, both mass and condition factor were lower in individuals of increasingly advanced stages to an extent that individuals in the most advanced pigmentation stage VIA4 weighted, on the average, some 18% less than those in stage VB. In turn, glass eels in the western Rı́o Minho were smaller in size and showed higher condition than those in Rı́o Nalón. Although we found no obvious explanation for such differences, larger glass eels with better physical condition may reflect greater energy content implying an enhanced upstream migration (Bureau du Colombier et al., 2007) but such differences may also be caused by the occurrence of different migration waves or by selective behaviour of tidal river transports. A possible explanation may be related to feeding. It is admitted that most glass eels do not feed during estuarine migration. However, Bardonnet & Riera (2005), using stable isotope analysis, reported that a small amount of individuals may feed. It could be hypothesized that fish in Rı́o Minho were older and had a better condition because some individuals may stop migration to feed. The analyses of pigmentation results for Rı́o Minho should be treated cautiously because only 4 months from one single year were considered. Current results, however, were confirmed by the conclusions drawn for the Rı́o Nalón data set. A comparison among studies highlighted consistent patterns in size and pigmentation stages among Atlantic locations from Portugal to Denmark with slightly larger sizes at the western localities and more advanced pigmentation stages in the northern locations. Moreover, Atlantic glass eels tend to be larger in size (range 68.6–78.4 mm) than those from the Mediterranean (59.1–73.0 mm). This pattern is consistent with those elucidated by Gandolfi-Hornyold (1920, 1930a, 1936)’s pioneering studies on Spanish Atlantic and Mediterranean locations. This author reported smaller glass eels in the Balearic Islands and remarkably smaller individuals in Alexandria (Egypt) located in the Far East Mediterranean (GandolfiHornyold, 1930b). Such counter-intuitive difference in the Mediterranean with smaller sizes at the eastern locations appears unrealistic and may have plausibly been caused by sampling bias involving at least sampling dates and different manipulation, measures and preservation techniques. Nevertheless, more advanced pigmentation stages in the Mediterranean Camargue and Rı́o Arno may result from temperaturedependent effects (Briand et al., 2005). Higher temperatures in the Mediterranean may accelerate the pigmentation process resulting in a predominance of more pigmented individuals relative to a predominance of the early stage VB in the Atlantic Rı́os Nalón and Minho. For several Atlantic locations, Boëtius & Boëtius (1989) reported a replacement of weak pigmented glass eels by more pigmented stages than in the Mediterranean Rı́o Arno (Gandolfi et al., 1984; Fig. 5) but here the pigmentation stages appeared in more advance stages since the beginning of the season. Several causes have been suggested to explain the temporal reduction of length and mass. First, larger leptocephalus may swim faster and reduce the time needed to reach the European continent relative to the smaller, slower individuals. Second, fasting during the metamorphosis when leptocephalus larvae lose their teeth (Elie, 1979; Desaunay & Guerault, 1997) and reduce the length of the gastrointestinal tract, so glass eels may spend prolonged time periods without feeding and living at the expense of their lipid 123 Hydrobiologia reserves (Tesch, 2003). On the other hand, the larger size of Atlantic glass eels compared with the smaller Mediterranean individuals may be caused by a greater osmotic exchange cost in the markedly higher-salinity Mediterranean Sea. Nevertheless, a remarkable heterogeneity and a large proportion of the variance unexplained (*75%) may result from genetic heterogeneity (Maes et al., 2009) but still suggests that environmental factors such as water temperature and salinity may affect the pigmentations stages (Briand et al., 2005) and should be taken into account in future investigations on glass eel recruitment. Acknowledgments Thanks are due to the fishermen of the two rivers and very especially to Pepı́n ‘‘El Ruleru’’ and Ester Diaz (staff of the Rı́o Nalón Fishermen Association) and to Alfredo Oliveira and Eduardo Martins (members of the Rı́o Minho fishery crew) for their continuous help during glass eels collection. Brian Knights gave valuable comments on an earlier draft and Maite Lavandeira made her best to improve the English. During this study T. Iglesias was funded by the INDICANG Project of the INTERREG IIB—EU Program whereas S. Costa Dias was funded by the Portuguese FCT through the grant Ref. No. SFRH/BD/16922/2004. References Antunes, J. C., 2002. Monitoring of glass eels recruitment in Portugal. In Dekker, W. (ed.), Monitoring of Glass Eels Recruitment. Netherlands Institute of Fisheries Research, Ijmuiden, Report C007/02-WD: 219–226. Bardonnet, A. & P. Riera, 2005. Feeding of glass eels (Anguilla anguilla) in the course of their estuarine migration: new insights from stable isotope analysis. Estuarine, Coastal and Shelf Science 63: 201–209. Bardonnet, A., S. Dasse, M. Parade & M. Heland, 2003. Influence del0 alternance jour/nuit sur les desplacements de civelles en fluvarium. Bulletin Français de la Pêche et de la Pisciciculture 368: 9–20. Boëtius, J., 1976. Anguilla anguilla and Anguilla rostrata from two Danish localities. Size, body mass, developmental stage and number of vertebrae related to time of ascent. Meddelelser fra Danmarks Fiskeri-Og Havundersogelser 7: 199–220. Boëtius, I. & J. Boëtius, 1989. Ascending elvers, Anguilla anguilla, from five European localities. Analyses of pigmentation stages, condition, chemical composition and energy reserves. Dana 7: 1–12. Bonhommeau, S., E. Chassot & E. Rivot, 2008. Fluctuations in European eel (Anguilla anguilla) recruitment resulting from environmental changes in the Sargasso Sea. Fisheries Oceanography 17: 32–44. Briand, C., D. Fatin, E. Cicotti & P. Lambert, 2005. A stagestructure model to predict the effect of temperature and salinity on glass eel Anguilla anguilla pigmentation development. Journal of Fish Biology 67: 993–1009. 123 Bureau du Colombier, S., V. Bolliet, P. Lambert & A. Bardonnet, 2007. Energy and migratory behavior in glass eels (Anguilla anguilla). Physiology & Behavior 92: 684–690. Charlon, N. & J. M. Blanc, 1982. Etude des civelles d’Anguilla anguilla L. dans la région du bassin de l’Adour. Archiv für Hydrobiologie 93: 238–255. de Casamajor, M. N., P. Prouzet & P. Lazure, 2000. Identification des flux de civelles (Anguilla angilla) à partir des relations d0 allométrie en fonction des conditions hydrodynamiques de l0 estuarie de l0 Adour. Aquatic Living Resources 13: 411–420. de Casamajor, M. N., R. Lecomte-Finiger & P. Prouzet, 2003. Caracteristiques biologiques des civelles (Anguilla anguilla) lors de la transition en estuaire. Bulletin Français de la Pêche et de la Pisciciculture 368: 109–124. Dekker, W., 2003. Status of the European eel stock and fisheries. In Aida, K., K. Tsukamoto & K. Yanauchi (eds), Eel Biology. Springer-Verlag, Tokyo: 237–254. Desaunay, Y. & D. Guerault, 1997. Seasonal and long-term changes in biometrics of eel larvae: a possible relationship between recruitment variation and North Atlantic ecosystem productivity. Journal of Fish Biology 51: 317–339. Elie, P., 1979. Contribution à l’étude des montées de civelles d’Anguilla anguilla L. dans l’estuaire de la Loire : Pêche, écophysiologie et èlevage. Thèse de 3eme Cycle Universitè Rennes I, 383 pp. Elie, P., R. Lecomte-Finiger, I. Cantrelle & N. Charlon, 1982. Définition des limites des différentes stades pigmentaires durant la phase civelle d’Anguilla anguilla L (Poisson, Téléostéen, Anguilliformes). Vie et Mileu 32: 149–157. Friedland, K. D., M. J. Miller & B. Knights, 2007. Oceanic changes in the Sargasso Sea and declines in recruitment of the European eel. ICES Journal of Marine Science 64: 519–530. Gandolfi, G., M. Pesaro & P. Tongiori, 1984. Environmental factors affecting the ascent of elvers, Anguilla anguilla (L.). Oebalia 10: 17–35. Gandolfi-Hornyold, A., 1918. Algunas medidas de las angulas de Santander. Boletı́n de Pescas 21: 153–192. Gandolfi-Hornyold, A., 1920. Algunas observaciones sobre la angula en Palma de Mallorca. Boletı́n de Pescas 41-42-43: 1–128. Gandolfi-Hornyold, A., 1929. Mensurations de civelles de l’Oria pendant la saison de pêche 1928–29. Bulletin de la Société d’Océanographie de France 49: 897–908. Gandolfi-Hornyold, A., 1930a. Mensurations de civelles de Mex pêchées en mars 1928, près d’Alexandrie. Bulletin de la Société d’Océanographie de France 51: 936–940. Gandolfi-Hornyold, A., 1930b. Mensurations de civelles de l’Oria pendant les saisons de pêches 1928–1929 et 1930– 1931. Bulletin de la Société d’Océanographie de France 56: 1012–1023. Gandolfi-Hornyold, A., 1936. La civelle d’Aguinaga de 1935 à 1936. Bulletin de la Société d’Océanographie de France 92: 1597–1599. Kettle, A. J. & K. Haines, 2006. How does the European eel (Anguilla anguilla) retain its population structure during its larval migration across the Northern Atlantic Ocean? Canadian Journal of Fisheries and Aquatic Sciences 63: 90–106. Hydrobiologia Lambert, P., M. Sabih, E. Rochard, J. Marchelidon, S. Dufour & P. Elie, 2003. Variabilite morphlogique et du taux d’hormone de croissance des civelles d’anguilles europeennes (Anguilla anguilla) dans l’estuaire de la Gironde au cours de la saison 1997–1998. Bulletin Français de la Pêche et de la Pisciciculture 368: 69–84. Lara, M. J., 1994. Catch statistics, capture methods, size, and development stages of glass eels in Asturias (northwestern Spain). Bulletin of the Sea Fisheries Institute 1: 31–39. Lecomte-Finiger, R., Y. Desaunay, D. Guerault & P. Grellier, 1996. The inmigration of Anguilla anguilla (L.) glass eels in coastal waters: questions about the determinism of the otolith structures. Archives of Polish Fisheries 4: 235– 243. Maes, G. E., B. Van Vo, A. J. Crivelli & F. A. M. Volckaert, 2009. Morphological and genetic seasonal dynamics of European eel Anguilla anguilla recruitment in southern France. Journal of Fish Biology 74: 2047–2068. Poole, W.R., 1994. A population study of the European Eel (Anguilla anguilla (L.)) in the Burrishoole System, Ireland, with special reference to growth and movement. PhD Thesis, Dublin University, Ireland: 416 pp. Strubberg, A., 1913. The metamorphosis of elvers as influenced by outward conditions. Meddelelser fra Kommissionen for Havundersogelser, Serie Fiskeri 4: 1–11. Tesch, F. W., 2003. The Eel: Biology and Management of Anguillid Eels, 5th ed. Blackwell Science, Oxford. Weber, M., 1986. Fishing method and seasonal occurrence of glass eels (Anguilla anguilla L.) in the river Minho, west coast of the Iberian Peninsula. Vie et Milieu 36: 243–250. Yahyaoui, A., P. Aguesse & P. Beaubrun, 1983. Etude du recrutement en civelles (Anguilla anguilla L., 1758) dans l’estuaire du Sebou au de la saison de peche 1980–1981. Boletin de l’Institute Scientifique, Rabat 7: 149–164. Zar, J. H., 1999. Biostatistical Analysis, 4th ed. Prentice-Hall, Inc., Upper Saddle River, NJ. 123 View publication stats