408 



Fishery Bulletin 102(3) 



Materials and methods 



Statistical analysis 



Samples and condition indices 



During the eastward migration, 62 adult male bluefin 

 tuna weighing between 71 and 273 kg (mean 195.17 kg) 

 were obtained from the trap fishery in the area of the 

 Strait of Gibraltar (Barbate. Cadiz, southwestern Spain) 

 from late April to early June 1999. 2000, and 2001. 

 Thirty-four mature males, weighing between 19 and 

 349 kg (mean 115.11 kg), were sampled in June-July 

 1999-2001 from the purse-seine fleet operating in the 

 Mediterranean spawning grounds of bluefin tuna off 

 the Balearic Islands. Whenever possible, the total body 

 weight (W) was recorded to the nearest kg. When indi- 

 vidual body weights were not available, W was estimated 

 from the fork length (L F ) measurements (recorded to the 

 nearest cm), according to the formula: W = 0.000019 x 

 L F 3 (Table VIII in Rodriguez-Roda, 1964). Following 

 dissection, the liver, testes, and the fat bodies associ- 

 ated with the gonads were removed and weighed to the 

 nearest g. The condition of the fish was assessed by three 

 different indices. The gonad index (gonadosomatic index) 

 (I G ) is indicative of the maturation state and was calcu- 

 lated as: I G = (W G I W) x 100, where W G = gonad weight. 

 The liver index (hepatosomatic index) (I L ) and fat-body 

 index (I F ) were calculated as I L = (W L I W) x 100, and 

 Ip- = (W F I W) x 100 (where W L and W F represent liver 

 and fat-body weights), respectively, and are considered 

 as good indicators of the metabolic condition and energy 

 reserves of the fish. All measurements are expressed as 

 means +SD. 



The bluefin tuna specimens used in this study showed 

 considerable variability in size, especially those caught 

 by purse seine in Balearic waters, where weight ranged 

 between 12 and 349 kg. The purse-seine fishery is, in 

 fact, much less size-selective than are traps, which 

 seldom catch small bluefin tuna (Rodriguez-Roda, 1964: 

 Mather et al., 1995). Analysis of covariance (ANCOVA), 

 with body weight as covariate, was used as the most 

 suitable method (Garcia-Berthou, 2001) to test interan- 

 nual differences in the weight of the organs within the 

 two sampling sites. ANCOVA was likewise applied to 

 compare the weights of the three organs between both 

 areas. All data were previously log-transformed to meet 

 the prerequisites of normality and homoscedasticity 

 (Zar, 1996). Linear least-squares regression analyses 

 were performed to test possible correlations between 

 I G and the two other indices (I L and I F ) by using the 

 pooled data of Barbate and the Balearic Islands. In 

 the regression between I G and I F , the Balearic samples 

 corresponding to year 2001 were excluded because the 

 reduced fat-body size (adipose tissue was almost non- 

 existent in the mesentery) of these small bluefin tuna 

 did not permit an accurate weight measurements on 

 board. The values of the indices were arcsine-trans- 

 formed prior to the statistical analysis (Zar, 1996). A 

 P-value <0.05 was considered statistically significant 

 for all tests. 



Results 



Histology 



For light microscopy, tissue samples from the central 

 part of the testes were fixed for 48-96 hours in 10% 

 formalin in phosphate buffer, 0.1 M, pH 7.2. After dehy- 

 dration in ascending concentrations of ethanol, a part 

 of each sample was embedded in paraffin wax and the 

 remainder was embedded in plastic medium (2-hydroxy- 

 ethyl-methacrylate). Paraffin sections (6 /im thick) were 

 stained with haematoxylin-eosin, and plastic sections (3 

 /jm thick) were stained with toluidine blue. These were 

 examined and photographed on a Leitz DMR BE light 

 microscope. 



For electron microscopy, small fragments of testis were 

 fixed for 3-4 hours in 2.5% glutaraldehyde buffered with 

 0.1 M sodium cacodylate buffer (pH 7.2). Following two 

 30-min washes in cacodylate buffer, they were postfixed 

 for 1 hour at 4C in cacodylate-buffered 1% osmium 

 tetroxide, rinsed several times in buffer, dehydrated in 

 ascending concentrations of acetone, and embedded in 

 epoxy resin (either Epon 812 or Spurr). Thin sections 

 (-80 nm thick) were picked up on copper grids, stained 

 with uranyl acetate and lead citrate, and examined in 

 a Jeol 1200 EX transmission electron microscope. Ap- 

 proximate dimensions provided for germ cells are mea- 

 surements (means ±SD) of the largest cell diameters on 

 electron micrographs. 



Condition indices 



ANCOVA did not reveal significant interannual differ- 

 ences in gonad, liver, and fat-body weight in the samples 

 of Barbate as well as in those of the Balearic Islands. In 

 contrast, a strongly significant difference in testicular 

 size (P<0.0001) was found in comparing data of matur- 

 ing bluefin tuna from Barbate (migrant tuna) with fully 

 mature fish from the Balearic Islands (spawning fish). 

 In fact, as shown in Figure 1, the average I was more 

 than fourfold higher in the Balearic Islands than it was 

 in Barbate (4.81 ±1.77 vs. 1.12 ±0.57). This finding may 

 indicate a noticeable increase in sperm production during 

 reproductive migration to the Mediterranean spawning 

 grounds. Significant differences between maturing and 

 spawning tuna were also found in fat-body weight, the 

 volume of which dropped to about half by spawning 

 time. Thus, I F fell from 0.36 ±0.24 in migrating fish to 

 0.16 ±0.12 in spawning fish isee Fig. 1). The liver mass, 

 however, did not differ significantly (P=0.31> between 

 the two samples. 



Figure 2 illustrates linear regression analysis be- 

 tween I G and I, , and between I G and I F . A significant 

 negative correlation (;-'- = 0.34; P<0.0001) was found be- 

 tween Iq and Ip, indicating that the amount of mesenteric 

 fat tissue decreases as the gonad matures. In contrast, 

 there was a positive, though somewhat weak, correlation 



