Seasonal and interannual variability of fat content of juvenile albacore (Thunnus alalunga) and bluefin (Thunnus thynnus) tunas during their feeding migration to the Bay of Biscay

Progress in Oceanography 86 (2010) 115–123

Contents lists available at ScienceDirect

Progress in Oceanography

journal homepage: www.elsevier .com/ locate /pocean

Seasonal and interannual variability of fat content of juvenile albacore(Thunnus alalunga) and bluefin (Thunnus thynnus) tunas during their feedingmigration to the Bay of Biscay

Nicolas Goñi *, Haritz ArrizabalagaAZTI-Tecnalia, Marine Research Division, Herrera kaia portualdea z/g, 20110 Pasaia, Gipuzkoa, Spain

a r t i c l e i n f o a b s t r a c t

Article history:Received 14 March 2008Received in revised form 16 June 2009Accepted 10 April 2010Available online 24 April 2010

0079-6611/$ - see front matter � 2010 Published bydoi:10.1016/j.pocean.2010.04.016

* Corresponding author. Tel.: +34 943 00 48 00; faxE-mail address: [email protected] (N. Goñi).

The fat content of 2945 juvenile albacore and 618 juvenile bluefin tunas caught in the Bay of Biscay wasmeasured. Individuals were caught in 2004, 2005, 2006 and 2007 from June to early November by pelagictrawling, trolling and baitboat gears. The results for the two species show different seasonal trends. Thefat content of albacore tuna increased linearly throughout the fishing season, which reflects their feedingmigration. The seasonal trend of bluefin tuna showed a minimum in early August, which may be relatedto a different behaviour, physiology or feeding strategy. An interannual increase of fat content wasobserved in albacore tuna and in age-2 to age-5+ bluefin tuna, which is possibly related to a density-dependence phenomenon. The seasonal increase of fat content was strongest and appeared in the fouryears studied for age-3 and age-4 albacore tuna, which can be related to a different vertical habitat ora more efficient use of their ecological niche by the individuals of these age-groups, relatively to theyounger age-groups. Condition factor and girth/length ratio do not appear to be relevant indicators offat content.

� 2010 Published by Elsevier Ltd.

1. Introduction

Albacore (Thunnus alalunga) and bluefin (Thunnus thynnus) arethe main tuna species encountered in the Northeast Atlantic; juve-niles of both species perform large scale feeding migrations duringthe summer months to the Bay of Biscay and surrounding waters(Bard, 2001; Fromentin and Powers, 2005). The feeding nature oftheir migration is reflected in the significantly higher growth ratesobserved for both species during this season (Cort, 1991; Santiagoand Arrizabalaga, 2005). They are mainly exploited in this zonefrom June to October by surface passive gears such as baitboatsand trolling lines, and by pelagic trawling (ICCAT, 2006, 2008).Baitboat and trolling fleets work during daytime, pelagic trawlerswork at night, generally by pair trawling.

Tunas have high metabolic rates due to their obligate continu-ous activity (Graham and Laurs, 1982), and high standard meta-bolic rates compared to strictly poikilotherm fish species (Brill,1979; Stevens and Dizon, 1982; Korsmeyer and Dewar, 2001). Thismetabolic rate is particularly high in juvenile individuals (i.e. thosein rapid-growth phase) and in individuals that perform long-dis-tance seasonal migrations (Harden Jones, 1984). In this context,the amount of stored energy of juvenile albacore and bluefin tunas

Elsevier Ltd.

: +34 943 00 48 01.

is a crucial parameter of their biology, as it can impact their growthand rate of survival.

The amount of energy that a tuna is able to store, integratesinfluences of prey availability, abiotic variables that limit distribu-tion (such as water temperature), and of their physiological abili-ties (such as thermoregulation, buoyancy and diving capacity,and potential swimming speed) which impacts feeding efficiency.Observations of stored energy can therefore highlight ecosystemchanges affecting population dynamics.

In contrast to mammals that use carbohydrates as an energysource, fish get their main source of energy from lipids (McKeown,1984; Shulman and Love, 1999), and lipid content is actually anaccurate measure of their energy reserves (Adams, 1999; Shulmanand Love, 1999). Migrating fish species such as tunas have a higherand more variable fat content compared to most non-migratingfish species, indicating that lipids are their main source of energy,particularly during these long-range migrations (Stansby, 1976).Furthermore, migration distance in pelagic fish has been linkednot only to body size but also to available fat stores (Nøttestadet al., 1999).

Because fat content is considered an indicator of somatic condi-tion, this study focuses on the fat content of juvenile albacore andbluefin tunas and examines whether its trends reflect tuna feedingperformance during their seasonal feeding migrations to the Bay ofBiscay. Physiological state of fish has often been described by bio-metrics-based indices, such as condition factor (Beckman, 1948) or

116 N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123

girth/length ratio, so we also compared these indices to tuna fatcontent.

The objective of this work was to identify seasonal, interannualand size-related patterns in the variability of the fat content ofjuvenile bluefin and albacore tunas in the Bay of Biscay. We com-pared and interpreted our findings considering the different lifecycles, physiological and ecological characteristics of both species.

2. Material and methods

2.1. Samples

From June to early November in 2004, 2005, 2006 and 2007, wesampled 2945 albacore and 618 bluefin tunas in the Bay of Biscayand adjacent waters, in the ports of Ziburu and Hondarribia, bothlocated at the southeast corner of the Bay of Biscay. The Hondarri-bia fleet comprises 19 baitboat and seven trolling gears. The Ziburufleet comprises 18 pelagic trawlers that occasionally use trollingline during daytime. Both fleets operate in the whole Bay of Biscay(with exception of the continental shelf) and adjacent waters up to12�W. Each boat, or pair of boats in the case of pair trawling, searchfish individually. We assumed the sampled individuals are repre-sentative of the individuals available in the fishing area.

The individuals sampled were fresh fish, generally landed 1–6 days after catch, sorted by weight-classes and presented in portson pounded ice before selling. Individual catch dates and locationswere unknown. Landing date was considered an indicator of catchdate, assuming the lag time between catch and landing was thesame for all individuals.

The fork length, girth and weight, as well as the fat content ofeach individual, were measured. The girth was measured at theinsertion of the pectoral fins. Age was attributed to each fish bytransforming fork-length measurements using the growth equa-tion by Cort (1991) for bluefin tuna, and the one by Santiago andArrizabalaga (2005) for albacore tuna.

2.2. Fat content measurements

The muscular fat content of each individual was estimatedusing a Distell� Fish Fatmeter with four measurements taken oneach side of the fish between the insertion of the pectoral fin andthe dorsal fin (Fig. 1), using the same areas as those used by Jensen(2003). The principle of the Fish Fatmeter is based on the inverserelationship between moisture and lipid content in fish, followingKent (1990) who stated that the measurement of one variableserves to determine the other. In the case of albacore tuna, thisrelationship has been confirmed by evidence from Garcia Ariaset al. (1994) and Wheeler and Morrissey (2003). The Fatmeter isfitted with a microstrip sensor that is sensitive to the water con-tent of the sample and that, when in contact with the fish skin,measures the percent water composition and calculates the per-cent lipid content. The precision of the lipid content is 0.5–1% for

Fig. 1. Location of the fat co

a lipid content between 2% and 15%; 1–2% for a lipid content be-tween 16% and 30%; and 2–4% for a lipid content of more than 30%.

In order to confirm the accuracy of the measurements given bythe Fatmeter, Bligh and Dyer (1959) tests were performed on mus-cles samples of 50 albacore tuna. Bligh and Dyer method is a lipidextraction procedure designed for aqueous samples, using chloro-form and methanol to dissolve and extract total lipids. These 50samples were taken in the zone sampled by the Fatmeter, andthe lipid content estimates were correlated to the fat content mea-sured by the Fatmeter. In the case of bluefin tuna, measurements oflipid content by the Fatmeter and by Fosslet method are reportedto be significantly correlated (Distell.com, 2003).

2.3. Statistical analyses

Interannual, seasonal and size-related variability of fat content,condition factor and girth/length ratio were analyzed for both spe-cies using generalized additive models (Hastie and Tibshirani,1990). Year and gear were used as factors, fork length and Julianday as variables. The type of gear was used as an explanatoryparameter to assess possible variations in the fat content betweenfish caught by baitboat, trolling line and pelagic trawling. A gauss-ian distribution was assumed for all random variables. A stepwisemodel selection procedure was adopted where the explanatoryvariables of fork length and Julian day, could either be modelledas absent, as a linear function or as a spline function. Rows withmissing values were omitted before the stepwise model selection,so that all models were based on the same observations. The signif-icance of nonparametric effects was assessed with an F test, andthe model with lowest Akaike Information Criteria (AIC) was se-lected, following Hastie (1992). In the case of albacore tuna, sea-sonal and size-related variability of fat content, of conditionfactor and of girth/length ratio were also analyzed for each age-group and each year studied, using generalized linear models.The relationships between the fat content of both species and (1)their condition factor, and (2) their girth/length ratio, were ana-lyzed using Pearson’s correlation tests. Shapiro–Wilk tests wereused to assess the normality of the residuals. R statistical software(version 2.7.2., R Development Core Team, 2007) was used for allstatistical analyses. The generalized additive models were builtstepwise using the gam 1.0 package; their corresponding analysesof variance were performed using the mgcv 1.4–1 package.

3. Results

3.1. Correlation between measurements by the Fatmeter and the Bligh& Dyer method

The measurements performed by the Fatmeter and the lipidcontent of the corresponding individuals as estimated by the Blighand Dyer method show a significant (p < 2.2 � 10�16, R2 = 0.90) po-sitive relationship (Fig. 2). The residuals of the linear regressionshow a normal distribution (p = 0.1472 for Shapiro–Wilk test).

ntent sampling zones.

Fig. 2. Correlation between the measures of lipid content (%) using the Fatmeter(y-axis) and the Bligh and Dyer (1959) method (x-axis).

Table 1Analysis of variance of the generalized additive models of albacore and bluefin tunasfat content as a function of fork length, julian day and year. Estimated degrees offreedom and nonparametric F are given in the case of smooth terms (indicated initalic).

Subsample Terms Degree offreedom

F p-value

Albacore s (fork length) 7.342 68.72 <2 � 10�16

Julian day 1 278.94 <2 � 10�16

Year 3 27.94 <2 � 10�16

Age-1 bluefin Fork length 1 39.06 1.96 � 10�9

s (julian day) 5.271 3.269 9.13 � 10�4

Year 2 15.25 6.04 � 10�7

Age-2 to age-5+ bluefin Fork length 1 69.99 2.32 � 10�15

s (julian day) 7.259 6.00 9.84 � 10�8

Year 3 19.93 8.4 � 10�12

N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123 117

3.2. Gear-related variability of fat content

No significant effect of gear type (trolling line, baitboat, pelagictrawling) on fat content was found, either for albacore or bluefintunas. In other words, no fat-related selectivity of the fishing gearsappears in our data.

3.3. Size-related variability of fat content in albacore and bluefin tunas

For juvenile albacore tuna, a significant nonlinear relationshipwas observed between fat content and fork length (Table 1) withfat content reaching its maximal value for 87-cm long individuals,and being minimal for both the smallest and largest individuals(Fig. 3). Within age-groups, an increase of fat content with fish sizewas found in 2005 for age-group 1, and in 2004, 2005 and 2006 forage-group 2 (Table 2). For age-groups 3 to 5+, we did not observeany recurrent significant effect of size on fat content.

A significant linear relationship between fat content and forklength was observed for juvenile bluefin tuna in age-class 1, andfor those aged 2 to 5+ (Fig. 4).

3.4. Interannual variability of fat content in albacore and bluefin tunas

We found significant interannual variation in the fat content ofall juvenile albacore tuna and bluefin tuna in age-groups 2–5+,with this fat content globally increasing over the 4 years of study(Fig. 4, Table 1). The fat content of age-1 bluefin tuna showed noclear interannual pattern, although its interannual variability wassignificant (Table 1). In albacore tuna, the interannual increase infat content was not significant between 2005 and 2006; therewas an apparent decrease observed for the 5+ age-group (Fig. 5),although this decrease was also not significant.

3.5. Seasonal patterns of fat content in albacore tuna

The fat content of juvenile albacore tuna shows a significantglobal increase during the fishing season (Fig. 3 and Table 1). Forage-3 and age-4 albacore tuna, the seasonal increase of the fat con-tent was highly significant for each of the four years studied (Table2). In both age-groups the slope of this seasonal increase has nosignificant interannual variation.

The seasonal increase in fat content for the age-groups 1 and 2had globally a lower slope and a lower significance level, and didnot occur every year. No seasonal increase was shown in 2004,2006, 2007 for age-1 albacore tuna, and in 2004 for age-2 albacoretuna. For the age-group 5+, no recurrent seasonal pattern of fatcontent was observed over the four years studied.

The residuals of the generalized linear models of fat content,girth/length ratio and condition factor of albacore tuna in functionof fork length and Julian day were normally distributed in all mod-els (p-values of Shapiro–Wilk tests ranged from 0.090 to 0.7356).

3.6. Seasonal patterns of fat content in bluefin tuna

In the case of bluefin tuna, not all age-groups were representedduring the whole fishing season in our samples. Age-groups 2 to 5+appeared mainly from July to mid-August (Julian days 180–220),and the age-group 1 appeared mainly from the end of August tothe end of September (Julian days 240–270). Consequently, twoseparated analyses of bluefin tuna fat content were performed.The fat content of age-groups 2 to 5+ showed a significant decreasein the first part of the fishing season (Fig. 4, Table 1); the minimumvalues corresponding to the end of July and beginning of August. Aslower, non-significant increase was observed in the second part ofthe season. For age-group 1, we observed a significant increase infat content during the month of September, followed by a decreaseat the end of the season (Fig. 4). The decrease of fat content for age-groups 2 to 5+, and the increase for age-group 1, showed a similarpattern to that observed for monthly bluefin tuna catches by thebaitboat fleet in the Bay of Biscay (Fig. 6).

3.7. Correlation between fat content and biometrics-based indices

For albacore tuna, the seasonal and size-related variability ofgirth/length ratio and condition factor did not reflect the seasonaland size-related variability of fat content (Table 3). The girth/length ratio and condition factor do not show any clear seasonaltrend; the relationship between these variables and date of capturebeing negative in some cases (Table 3).

In both albacore and bluefin tunas, fat content was not signifi-cantly correlated with girth/length ratio or condition factor withinage-groups and years (Table 4). This absence of correlation wasespecially notable for juvenile bluefin tuna where sample sizes werealso relatively limited. In albacore tuna, fat content and girth/lengthratio were significantly correlated in age-groups 1–4, for the years2004 and 2005. No significant correlation was observed betweenthese variables in 2006. Fat content and condition factor showedno recurrent correlation among age-groups and years.

4. Discussion

4.1. Seasonal patterns in fat content

The different seasonal patterns of fat content between bluefinand albacore tunas may indicate different migration paths or feed-ing strategies, or may reflect differences in their physiology, meta-bolic rates or use of stored energy. Metabolic rates of albacore andbluefin tunas have been studied separately (Graham and Laurs,

Fig. 3. Generalized additive model of the fat content of juvenile albacore tuna as a function of: (a) fork length, (b) date, and (c) year. Intervals represent standard errors.

Table 2Generalized linear models of albacore tuna fat content as a function of fork length and julian day, for age-groups 1–4 and years 2004–2007.

Parameter Year Age 1 Age 2 Age 3 Age 4

Adjusted R2 2004 – 0.1069 0.2094 0.26852005 0.3175 0.1384 0.3123 0.32472006 – 0.1361 0.2181 0.11332007 – 0.0829 0.2252 0.1587

p-value for fork length 2004 – 2.60 � 10�6 – –2005 2.72 � 10�2 3.34 � 10�3 – –2006 – 6.17 � 10�5 – –2007 – – 6.71 � 10�3 –

p-value for julian day 2004 – – 8.08 � 10�11 8.3 � 10�7

2005 5.36 � 10�3 1.81 � 10�6 1.11 � 10�11 4.51 � 10�10

2006 – 1.24 � 10�12 <2 � 10�16 9.56 � 10�7

2007 – 9.17 � 10�4 9.0 � 10�9 2.19 � 10�5

Slope for julian day 2004 – – 1.02 � 10�2 1.36 � 10�2

2005 6.06 � 10�3 7.90 � 10�3 1.51 � 10�2 1.71 � 10�2

2006 – 6.97 � 10�3 8.14 � 10�3 5.63 � 10�3

2007 – 6.94 � 10�3 9.52 � 10�3 7.81 � 10�3

118 N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123

Fig. 4. Generalized additive model of the fat content of age-1 and aged 2 to 5+ juvenile bluefin tuna as a function of: (a) fork length, (b) date and (c) year. Intervals representstandard errors.

Fig. 5. Mean fat content (standard deviation) of albacore tuna for each age-group during the 4 years sampled.

N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123 119

1982; Blank et al., 2007), but the respective methodologies do notallow for a direct comparison between species. Nevertheless, dueto the respective growth patterns of albacore (Bard, 1981; Santiagoand Arrizabalaga, 2005) and bluefin tunas (Cort, 1991; Rodriguez-

Cabello et al., 2007), we can suppose higher energetic needs forjuvenile bluefin tuna than for albacore tuna.

Therefore, the different seasonal patterns in fat content for alba-core and bluefin tunas can be interpreted in two ways. The

Fig. 6. Monthly distribution of bluefin tuna catches by the Basque baitboat fleet operating in the Bay of Biscay during the years 2004–2006.

Table 3P-values of the analysis of variance of the generalized linear models of condition factor and ratio of girth-length of albacore tuna, as a function of fork length and julian day forage-groups 1–4 and years 2004–2007. P-values of negative correlations in italics.

Term Variable tested Year Age 1 Age 2 Age 3 Age 4

Fork length Condition factor 2004 – – – –2005 – 4.35 � 10�2 2.15 � 10�2 –2006 – – 1.48 � 10�6 –2007 – – – –

Girth/length ratio 2004 – 4.91 � 10�2 – –2005 – 8.60 � 10�3 – –2006 – – 5.28 � 10�4 –2007 – – – –

Julian day Condition factor 2004 4.54 � 10�2 – 3.01 � 10�9 –2005 1.93 � 10�2 1.07 � 10�2 – –2006 6.25 � 10�3 <2 � 10�16 4.04 � 10�5 2.66 � 10�2

2007 4.65 � 10�2 – 2.42 � 10�2 –

Girth/length ratio 2004 – – 5.84 � 10�13 5.96 � 10�4

2005 – – 1.07 � 10�2 –2006 3.76 � 10�4 6.52 � 10�15 4.03 � 10�2 –2007 2.56 � 10�2 – – –

Table 4P-values of correlation tests between: (1) fat content and girth/length ratio and (2) fat content and condition factor, for albacore and bluefin tunas over the 4 years and 5 age-groups sampled. P-values of significant correlations in bold; negative correlations in italics. A dash represents the absence of a sample for the particular age or year.

Species Test Age-group 2004 2005 2006 2007

Albacore Fat content and girth-length ratio Age 1 8.12 � 10�5 5.85 � 10�5 0.197 0.115Age 2 6.37 � 10�3 9.25 � 10�6 0.150 4.07 � 10�2

Age 3 5.05 � 10�8 5.04 � 10�4 0.939 1.26 � 10�2

Age 4 4.57 � 10�2 2.88 � 10�2 0.610 6.42 � 10�2

Age 5+ 0.145 8.25 � 10�2 0.717 0.481

Fat content and condition factor Age 1 2.80 � 10�3 0.181 0.104 0.224Age 2 0.118 2.28 � 10�2 0.827 0.332Age 3 3.44 � 10�6 0.357 0.116 0.430Age 4 0.222 0.683 0.862 0.117Age 5+ 0.489 0.125 0.489 0.286

Bluefin Fat content and girth/length ratio Age 1 0.012 (n = 87) 0.003 (n = 128) 0.148 (n = 39) –Age 2 0.587 (n = 54) 0.004 (n = 64) 0.076 (n = 28) 0.107 (n = 13)Age 3 0.849 (n = 5) 0.507 (n = 34) 0.014 (n = 62) 0.379 (n = 4)Age 4 0.122 (n = 4) 0.008 (n = 15) 0.787 (n = 12) 0.171 (n = 6)Age 5+ 0.066 (n = 10) – – –

Fat content and condition factor Age 1 0.002 (n = 87) 0.139 (n = 128) 0.836 (n = 39) –Age 2 0.130 (n = 54) 1 � 10�4 (n = 64) 0.005 (n = 28) 0.105 (n = 13)Age 3 0.484 (n = 5) 0.280 (n = 34) 0.599 (n = 62) 0.292 (n = 4)Age 4 0.496 (n = 4) 0.002 (n = 15) 0.971 (n = 12) 0.153 (n = 6)Age 5+ 0.014 (n = 10) – – –

120 N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123

N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123 121

decrease in fat content for age-2 to age-5+ bluefin tuna may be theconsequence of a greater use of stored energy, due to the high met-abolic demand related to their growth in early summer. Themonths of June and July are actually a period of important growthfor juvenile bluefin tuna in the Bay of Biscay (Cort, 1991). On thecontrary, lower growth rates in albacore tuna may demand less en-ergy, allowing for a continuous increase in fat content during thefeeding migration.

An alternative or complementary hypothesis would be that thetrophic resources within the Bay of Biscay in August do not satisfythe high energetic needs of age-2 to age-5+ bluefin tuna. In fact,monthly bluefin tuna catches between 2004 and 2007 showed asimilar pattern to that of fat content, with highest values at thestart and end of the fishing season and mimimum values in August(Fig. 6). This suggests that because of the poor trophic resources inthe Bay of Biscay, bluefin tuna may have migrated to alternativefeeding zones, reducing availability to fishing gears operatingwithin the Bay. Considering that most tuna aged 2–5+ are caughtat the beginning of the fishing season and age-1 fish are caughtmostly at the end of the fishing season, we could further hypothe-size that the Bay of Biscay is probably not a suitable feeding zoneduring June to August for aged 2–5+ juvenile bluefin tuna, whereasit is suitable for age-1 juvenile albacore tuna, allowing them tostore fat. This interpretation would weaken the hypothesis thatthe Bay of Biscay is an important feeding zone for juvenile bluefintuna in the Atlantic (Fromentin and Powers, 2005), at least for age-groups 2–5+. However, given the relatively small and variablebluefin tuna sample size during the four years studied, this appar-ent decrease of fat content should be interpreted with care and cor-roborated over longer time periods. Moreover, potential migrationto alternative trophic areas should be analyzed with electronic tag-ging techniques.

A linear increase in albacore tuna fat content throughout thefishing season has also been observed by Morrissey et al. (2004);this increase being much higher than in our study. This differenceis probably related to a more caloric diet for albacore tuna in theCalifornia Current than in the Bay of Biscay. According to Glaser(2007), more than 60% of the prey weight of troll-caught albacoretuna in the California Current is composed of fish with high lipidcontent (i.e. Engraulis mordax, Sardinops sagax, Scomber japonicus).In troll-caught albacore tuna in the Bay of Biscay, the weight pro-portion of fish with high lipid content in their prey (i.e. Engraulisencrasicolus, Scomber scombrus and Trachurus trachurus) variesfrom 0% to 26% during this time period (Goñi, 2008).

The strong and recurrent seasonal increase in the fat content ofage-3 and age-4 albacore tuna, compared to the weaker and non-recurrent increase in the fat content of younger individuals is prob-ably related to differences in vertical habitat or a more efficient useof their ecological niche by age-3 and age-4 fish. Bard (1981) re-ported that individuals of these age-groups have developed physio-logical abilities that allow them to dive below the seasonalthermocline, whereas younger individuals usually stay above thisthermocline. Consequently, age-3 and age-4 albacore tuna mayhave access to more important food resources from being able to ac-cess prey below the thermocline. This could explain the strongerand recurrent seasonal increase of their fat content, reflecting theirhigher feeding efficiency. The apparent variability in the seasonalpatterns of fat content for age-1 and age-2 albacore tuna is possiblyrelated to variations in the depth of the seasonal thermocline. How-ever, the period considered is short (4 years), and further samplingwould be required to confirm this hypothesis. Regarding age-5+albacore tuna, that appeared late in the fishing season, the absenceof significant seasonal variation of fat content from sampled indi-viduals does not rule out the possibility of a seasonal trend in fatcontent for this age-group. However, additional sampling duringsubsequent months would be required to test this hypothesis.

The difference in feeding performance, as it relates to the phys-iological transition between age-groups 2 and 3, could influencethe respective survival rates of these age-groups. This hypothesisfavours the use of age-dependent natural mortality vectors in fu-ture modelling approaches of albacore tuna populations—with adecreasing natural mortality for ages 1 to 3—rather than assump-tions of constant natural mortality for all ages. Hampton (2000) re-vealed important size-related variations in the natural mortality ofskipjack, yellowfin and bigeye tunas, with the natural mortalitydecreasing with body size for juvenile individuals.

4.2. Interannual variability of fat content

The interannual increase of fat content of juvenile albacoretuna, and of age-2 to age-5+ bluefin tunas, may be related to therelative densities of tunas and their prey. Albacore and bluefinprey, in the period considered, included an important number ofsmall pelagic fish and crustacean species (Goñi, 2008), but the onlyestimates of local abundance published to date are for anchovy(Engraulis encrasicolus) juveniles (Boyra et al., 2008). This limitedinformation does not allow us to identify any relationship betweenprey abundance and tuna fat content.

This interannual increase could also be related to food competi-tion. In particular, the bluefin tuna population decline (Fromentinand Powers, 2005; ICCAT, 2006) could lead to significant variationsin the biomass of small pelagic fish species (Tiews, 1978), or tolower intraspecific and interspecific competition for food. Thiswould allow better feeding performance, reflected by a higher fatcontent. This contrasts with findings in the West Atlantic bluefintuna population, as during a declining population period (Fromen-tin and Powers, 2005), Golet et al. (2007) and Neilson et al. (2007)observed an interannual decline in the condition of bluefin tuna inthe Gulf of Maine and Gulf of St Lawrence. However, only adult fishwere observed in this work, which limits the possible comparisonswith our observations.

In the case of albacore tuna, no significant interannual increaseappears in the seasonal slope of fat content for age-groups 3 and 4.This suggests that interannual fat content variation is probably re-lated to the relative food availability in the wintering zones moreso than in the Bay of Biscay and adjacent waters.

Moreover, our results are based on samples from only four years– three years in the case of age-1 bluefin tunas – which representsa very short time-series. Consequently, several other years of sam-pling would be required to confirm observed results and allow forthis interpretation.

4.3. Size-related variability of fat content in albacore tuna

Unlike Dotson’s (1978) findings for albacore tuna of the North-east Pacific, the relationship between albacore tuna size and fatcontent was significant in our samples. The nonlinear relationshipbetween fat content and size of albacore tuna, with a maximal fatcontent for 87-cm long individuals, can be compared to Bard’s(1981) observations of albacore tuna growth and biometrics. Bardreported a change in the growth pattern when juvenile albacoretuna reach a fork length of 85 cm, which corresponds to the sexualmaturity and full development of the swim bladder. According toour observations, and to those of Bard (1981, 2001) and Bardet al. (1998), adult albacore tuna usually appear in the fisheries be-tween mid-August and late August. It is hypothesized that they mi-grate to the Northeast Atlantic after spawning, consequentlydepleting their fat reserves. More generally, sexual maturity im-plies a different metabolism and use of stored energy; a lower partof this energy being dedicated to growth and a higher part to thedevelopment and functioning of the gonads. This could explainthe overall decrease in fat with increasing size in adult individuals.

122 N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123

The increasing fat content for juvenile albacore tuna up to87 cm, and the strong and recurrent seasonal increase of fat con-tent of age-3 and age-4 individuals, compared to the weaker andnon-recurrent increase of fat content of younger individuals, couldbe related to the decreasing growth rate of larger individuals. Thiswould allow them to store, in the form of fat, a higher part of theenergy ingested (e.g. to be used in migrations) than smaller indi-viduals with a higher growth rate.

4.4. Representation of sampled individuals

Different gears operating in different manners, at differentdepths and during different times of the day could lead, in princi-ple, to individuals with different fat contents. However, our resultsdo not show any fat-related selectivity of the three fishing gears.This confirms the random representation of individuals sampled.

4.5. Discrepancies between measured fat content and biometric-basedindices

The absence of relationships and similar seasonal patterns be-tween fat content and our biometric-based indices can be relatedto variability of body density. Variations in fat content can actuallyinfluence the body density of fish, so that two individuals can havethe same length and weight with one having a higher fat contentthan the other one. In such a case, the condition factor is not a rel-evant descriptor of fat content. Morrissey et al. (2004) actually re-ported a very weak correlation between fat content and conditionfactor in North Pacific juvenile albacore tuna. According to Bard(1981), the linear growth of albacore tuna is more rapid, and theb coefficient of the length-weight relationship is inferior to 3 forfish under 85 cm, whereas linear growth is slower and the b coef-ficient of the length-weight relationship is superior to 3 for fishover 85 cm. Because of this allometry, the shape of albacore tuna,which determines their girth/length ratio and their condition fac-tor, probably does not adequately reflect their fat content. Thiswas also observed in species other than tunas (e.g. cod), for whichMarshall et al. (2004) reported that girth is not necessarily indica-tive of the magnitude of stored energy reserves.

5. Conclusion

Overall, we found evidence of different seasonal patterns in fatcontent of juvenile albacore and juvenile bluefin tunas in the Bay ofBiscay, and a difference between age-groups in the seasonal in-crease of albacore tuna fat content. These results may questionthe status of the Bay of Biscay as the main feeding zone for Atlanticjuvenile bluefin tuna, but this should be considered with greatcare. Regarding albacore tuna, there is concordance with apotential shift in vertical habitat between age-groups 2 and 3, asmentioned by Bard (1981) and highlighted by previous findings(Goñi and Arrizabalaga, 2005). Archival tagging experimentsshould be helpful to characterize the habitat and feeding behaviourof albacore of different age-groups.

Acknowledgements

Josep Lloret contributed to the original idea of this work.Jean-Pierre Esain gave us a precious help at the beginning of thisstudy, and we are grateful to Peio Bilbao for allowing us to samplein Ziburu. We also thank Luis Alberto Martín (AZTI-Tecnalia) andLuis Naval, who provided us a great help for sampling in Hondarri-bia. We finally thank Eduardo Saitua (AZTI-Tecnalia) for havingperformed the Bligh & Dyer tests, Jean-Marc Fromentin andPhilippe Gaudin for their respective suggestions about the analysis

of the bluefin dataset and the interpretation of interannual trends.Nicolas Goñi was supported by a research grant from the Funda-ción Centros Tecnológicos, Iñaki Goenaga. This paper is contribu-tion number 497 of the Marine Research Division of AZTI-Tecnalia.

References

Adams, S.M., 1999. Ecological role of lipids in the health and success of fishpopulations. In: Arts, M.T., Wainman, B.C. (Eds.), Lipids in FreshwaterEcosystems. Springer-Verlag, New York, pp. 132–160.

Bard, F.X., 1981. Le thon germon (Thunnus alalunga Bonnaterre 1788) de l’OcéanAtlantique. De la dynamique des populations à la stratégie démographique.Thèse de Doctorat d’État. Université Paris VI, 334p (in French).

Bard, F.X., 2001. Extension of the Geographical and Vertical Habitat of Albacore (T.alalunga) in the North Atlantic. Possible Consequences on True Rate ofExploitation of this Stock. SCRS/00/119. Collective Volume of Scientific PapersICCAT 52, pp. 1447–1456.

Bard, F.X., Bach, P., Josse, E., 1998. Habitat et écophysiologie des thons: Quoi de neufdepuis 15 ans? Collective Volume of Scientific Papers ICCAT 50, pp. 319–342 (inFrench).

Beckman, W.C., 1948. The length–weight relationship factors for conversionsbetween standard and total length and coefficients of condition for sevenMichigan fishes. Transactions of the American Fisheries Society 75, 237–256.

Blank, J.M., Farwell, C.J., Morrissette, J.M., Schallert, R.J., Block, B.A., 2007. Influenceof swimming speed on metabolic rates of juvenile Pacific bluefin tuna andyellowfin tuna. Physiological and Biochemical Zoology 80, 167–177.

Bligh, E.G., Dyer, W.J., 1959. A rapid method for total lipid extraction andpurification. Canadian Journal of Biochemistry and Physiology 37, 911–917.

Boyra, G., Cotano, U., Martinez, U., Peña, M., Uriarte, A., 2008. JUVENA series reviewof the spatial distribution of anchovy juveniles in the Bay of Biscay. In: XIInternational Symposium on Oceanography of the Bay of Biscay, San Sebastián,Spain, April 2008.

Brill, R.W., 1979. The effect of body size on the standard metabolic rate of skipjacktuna, Katsuwonus pelamis. Fisheries Bulletin 77, 494–498.

Cort, J., 1991. Age and growth of the bluefin tuna, Thunnus thynnus(L.) of theNortheast Atlantic. Collective Volume of Scientific Papers ICCAT 35, pp. 213–230.

Distell.com, 2003. User & Technical Manual, Distell Fish Fatmeter. Model 692, April2003 (Version 2.4).

Dotson, R.C., 1978. Fat deposition and utilization in albacore. In: Sharp, G.J., Dizon,A.E. (Eds.), The Physiological Ecology of Tunas. Academic Press, New York, pp.343–355.

Fromentin, J.M., Powers, J.E., 2005. Atlantic bluefin tuna: population dynamics,ecology, fisheries and management. Fish and Fisheries 6, 281–306.

Garcia Arias, M.T., Sanchez-Muniz, F.J., Castrillon, A.M., Navarro, M.P., 1994. Whitetuna canning total fat and fatty acid changes during processing and storage.Journal of Food Composition and Analysis 7, 119–130.

Glaser, S.M., <http://web.pml.ac.uk/globec/structure/regional/cliotop/symposium/WG3/WG3_OP10.pdf>, 2007. Quantifying Predator-prey Interactions for NorthPacific albacore in the California Current. 1st GLOBEC-CLIOTOP Symposium,December 2007, La Paz, Mexico.

Golet, W.J., Cooper, A.B., Campbell, R., Lutcavage, M., 2007. Decline in condition ofnorthern bluefin tuna (Thunnus thynnus) in the Gulf of Maine. Fisheries Bulletin105, 390–395.

Goñi, N., 2008. Habitat et écologie trophique du germon (Thunnus alalunga) dansl’Atlantique Nord-Est: variabilité, implications sur la dynamique de lapopulation. Thèse de doctorat, Université de Pau et des Pays de l’Adour,France, 194p (in French).

Goñi, N., Arrizabalaga, H., 2005. Analysis of juvenile North Atlantic albacore(Thunnus alalunga) catch per unit effort by surface gears in relation toenvironmental variables. ICES Journal of Marine Science 62, 1475–1482.

Graham, J.B., Laurs, R.M., 1982. Metabolic rate of the albacore tuna Thunnusalalunga. Marine Biology 72, 1–6.

Hampton, J., 2000. Natural mortality rates in tropical tunas: size really does matter.Canadian Journal of Fisheries and Aquatic Sciences 57, 1002–1010.

Harden Jones, F.R., 1984. A view from the ocean. In: McCleave, J.D., Arnold, G.P.,Dodson, J.J., Neill, W.H. (Eds.), Mechanisms of Migration in Fishes. Plenum Press,New York, pp. 1–26.

Hastie, T.J., 1992. Generalized additive models. In: Chambers, J.M., Hastie, T.J. (Eds.),Statistical Models. Chapman and Hall, London, UK, pp. 249–307.

Hastie, T.J., Tibshirani, R., 1990. Generalized Additive Models. Chapman and Hall,London, UK.

ICCAT, 2006. Report of the 2006 Atlantic Bluefin Tuna Stock Assessment Session(Madrid, Spain, June 12–18, 2006). Collective Volume of Scientific Papers ICCAT60, pp. 652–880.

ICCAT, 2008. Report of the 2007 ICCAT albacore stock assessment session (Madrid,Spain, July 5–12, 2007). Collective Volume of Scientific Papers ICCAT 62, pp.697–815.

Jensen, J.M., <http://afrf.org/research/lipidreport2002.pdf>, 2003. Report to theAmerican Fishermen’s Research Foundation (AFRF) on pilot project todetermine albacore tuna lipid content during the 2002 North Pacific albacore(Thunnus alalunga) fishing season. American Fishermen’s Research Foundation.

Kent, M., 1990. Hand-held instrument for fat/water determination in whole fish.Food Control 1, 47–53.

N. Goñi, H. Arrizabalaga / Progress in Oceanography 86 (2010) 115–123 123

Korsmeyer, K.E., Dewar, H., 2001. Tuna metabolism and energetics. In: Block, B.A.,Stevens, E.D. (Eds.), Tunas: Physiology, Ecology and Evolution. Academic Press,San Diego, CA, pp. 35–78.

Marshall, C.T., Needle, C.L., Yaragina, N.A., Ajiad, A.M., Gusev, E., 2004. Derivingcondition indices from standard fisheries databases and evaluating theirsensitivity to variation in stored energy reserves. Canadian Journal ofFisheries and Aquatic Sciences 61, 1900–1917.

McKeown, B.A., 1984. Fish Migration. Cromm Helm, London, UK.Morrissey, M.T., Rasmussen, R., Carrol, S., <http://afrf.org/research/

lipidreport2003.pdf>, 2004. Lipid Content in Troll-caught Albacore Tuna andCorrelations with Geographical Location, Physical Measurements andSeasonality. American Fishermen’s Research Foundation, Lipid Content Project2003 Final Report.

Neilson, J.D., Paul, S.D., Ortiz, M., 2007. Indices of stock status obtained from theCanadian bluefin tuna fishery. ICCAT Standing Committee on Research andStatistics (SCRS)/2006/076, pp. 976–1000.

Nøttestad, L., Giske, J., Holst, J.C., Huse, G., 1999. A length-based hypothesis forfeeding migration in pelagic fish. Canadian Journal of Fisheries and AquaticSciences 56, 26–34.

R Development Core Team, <http://www.r-project.org/>, 2007. R: A Language andEnvironment for Statistical Computing. R Foundation for Statistical Computing,Vienna, Austria.

Rodriguez-Cabello, C., Restrepo, V.R., Rodríguez-Marín, E., Cort, J.L., de la Serna, M.,2007. Estimation of Northeast Atlantic Bluefin Tuna (Thunnus thynnus) GrowthParameters from Tagging Data. Collective Volume of Scientific Papers ICCAT 60,pp. 1258–1264.

Santiago, J., Arrizabalaga, H., 2005. An integrated growth study for North Atlanticalbacore (Thunnus alalunga Bonn 1788). ICES Journal of Marine Science 62, 740–749.

Shulman, G.E., Love, R.M., 1999. The biochemical ecology of marine fishes. In:Southward, A.J., Tayler, P.A., Young, C.M. (Eds.), Advances in Marine Ecology, vol.36. Academic Press, London, UK.

Stansby, M.E., 1976. Chemical Characteristics of Fish Caught in the Northeast PacificOcean. MFR Paper 1198. Marine Fisheries Research Review 38, pp. 1–16.

Stevens, E.D., Dizon, A.E., 1982. Energetics of locomotion in warm-bodied fish.Annual Review of Physiology 44, 121–131.

Tiews, K., 1978. On the disappearance of bluefin tuna in the North Sea and itsecological implications for herring and mackerel. Rapports et procès-verbauxdes réunions Conseil International pour l’Exploration de la Mer 172, 301–309.

Wheeler, S., Morrissey, M.T., 2003. Quantification and distribution of lipid, moisture,and fatty acids of West Coast albacore tuna (Thunnus alalunga). Journal ofAquatic Food Product Technology 12, 3–16.